diff --git a/README.md b/README.md
index 3be9deffb..298faccf7 100644
--- a/README.md
+++ b/README.md
@@ -880,6 +880,7 @@ will return much faster than the first query and we'll be certain the authors ma
 | `answer.evidence_retrieval`                  | `True`                                 | Use retrieval vs processing all docs.                                                                   |
 | `answer.evidence_summary_length`             | `"about 100 words"`                    | Length of evidence summary.                                                                             |
 | `answer.evidence_skip_summary`               | `False`                                | Whether to skip summarization.                                                                          |
+| `answer.evidence_text_only_fallback`         | `False`                                | Whether to allow context creation to retry without media present.                                       |
 | `answer.answer_max_sources`                  | `5`                                    | Max number of sources for an answer.                                                                    |
 | `answer.max_answer_attempts`                 | `None`                                 | Max attempts to generate an answer.                                                                     |
 | `answer.answer_length`                       | `"about 200 words, but can be longer"` | Length of final answer.                                                                                 |
diff --git a/packages/paper-qa-pypdf/tests/test_paperqa_pypdf.py b/packages/paper-qa-pypdf/tests/test_paperqa_pypdf.py
index 6f5a298ce..f5d61ffc2 100644
--- a/packages/paper-qa-pypdf/tests/test_paperqa_pypdf.py
+++ b/packages/paper-qa-pypdf/tests/test_paperqa_pypdf.py
@@ -19,6 +19,7 @@ def test_parse_pdf_to_pages() -> None:
     parsed_text = parse_pdf_to_pages(filepath)
     assert isinstance(parsed_text.content, dict)
     assert "1" in parsed_text.content, "Parsed text should contain page 1"
+    assert isinstance(parsed_text.content["1"], str)
     matches = re.findall(
         r"Abstract\nWe introduce PaSa, an advanced Paper ?Search"
         r"\nagent powered by large language models.",
diff --git a/pyproject.toml b/pyproject.toml
index 727339c8a..d5ea54801 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -61,7 +61,7 @@ dev = [
     "ipython>=8",  # Pin to keep recent
     "litellm>=1.68,<1.71",  # Lower pin for PydanticDeprecatedSince20 fixes, upper pin for VCR cassette breaks (https://github.com/BerriAI/litellm/issues/11724)
     "mypy>=1.8",  # Pin for mutable-override
-    "paper-qa[ldp,pypdf,pymupdf,typing,zotero,local,qdrant]",
+    "paper-qa[image,ldp,pypdf,pymupdf,typing,zotero,local,qdrant]",
     "pre-commit>=3.4",  # Pin to keep recent
     "pydantic~=2.11",  # Pin for start of model_fields deprecation
     "pylint-pydantic",
@@ -78,6 +78,9 @@ dev = [
     "typeguard",
     "vcrpy>=6",  # Pin for https://github.com/kevin1024/vcrpy/issues/884
 ]
+image = [
+    "pillow>=10.3.0",  # Pin for py.typed
+]
 ldp = [
     "ldp>=0.25.0",  # For new LLM client interface
 ]
diff --git a/src/paperqa/core.py b/src/paperqa/core.py
index c654fb371..450d739da 100644
--- a/src/paperqa/core.py
+++ b/src/paperqa/core.py
@@ -2,16 +2,21 @@
 
 import contextlib
 import json
+import logging
 import re
 from collections.abc import Callable, Sequence
 from typing import Any, cast
 
+import litellm
 from aviary.core import Message
 from lmi import LLMModel
 
+from paperqa.prompts import text_with_tables_prompt_template
 from paperqa.types import Context, LLMResult, Text
 from paperqa.utils import extract_score, strip_citations
 
+logger = logging.getLogger(__name__)
+
 
 def llm_parse_json(text: str) -> dict:
     """Read LLM output and extract JSON data from it."""
@@ -136,6 +141,7 @@ async def map_fxn_summary(
     parser: Callable[[str], dict[str, Any]] | None = None,
     callbacks: Sequence[Callable[[str], None]] | None = None,
     skip_citation_strip: bool = False,
+    evidence_text_only_fallback: bool = False,
 ) -> tuple[Context, LLMResult]:
     """Parses the given text and returns a context object with the parser and prompt runner.
 
@@ -154,6 +160,8 @@ async def map_fxn_summary(
             Should return dict with at least 'summary' field.
         callbacks: Optional sequence of callback functions to execute during LLM calls.
         skip_citation_strip: Optional skipping of citation stripping, if you want to keep in the context.
+        evidence_text_only_fallback: Opt-in flag to allow retrying context creation
+            without media in the completion.
 
     Returns:
         The context object and LLMResult to get info about the LLM execution.
@@ -163,25 +171,61 @@ async def map_fxn_summary(
     extras: dict[str, Any] = {}
     citation = text.name + ": " + text.doc.formatted_citation
     success = False
+    used_text_only_fallback = False
 
+    # Strip newlines in case chunking led to blank lines,
+    # but not spaces, to preserve text alignment
+    cleaned_text = text.text.strip("\n")
     if summary_llm_model and prompt_templates:
+        media_text: list[str] = [m.text for m in text.media if m.text]
         data = {
             "question": question,
             "citation": citation,
-            # Strip newlines in case chunking led to blank lines,
-            # but not spaces, to preserve text alignment
-            "text": text.text.strip("\n"),
+            "text": (
+                text_with_tables_prompt_template.format(
+                    text=cleaned_text,
+                    citation=citation,
+                    tables="\n\n----\n\n".join(media_text),
+                )
+                if media_text
+                else cleaned_text
+            ),
         } | (extra_prompt_data or {})
-        message_prompt, system_prompt = prompt_templates
-        messages = [
-            Message(role="system", content=system_prompt.format(**data)),
-            Message(role="user", content=message_prompt.format(**data)),
-        ]
-        llm_result = await summary_llm_model.call_single(
-            messages=messages,
-            callbacks=callbacks,
-            name="evidence:" + text.name,
-        )
+        message_prompt, system_prompt = (pt.format(**data) for pt in prompt_templates)
+        try:
+            llm_result = await summary_llm_model.call_single(
+                messages=[
+                    Message(role="system", content=system_prompt),
+                    Message.create_message(
+                        text=message_prompt,
+                        images=(
+                            [i.to_image_url() for i in text.media]
+                            if text.media
+                            else None
+                        ),
+                    ),
+                ],
+                callbacks=callbacks,
+                name="evidence:" + text.name,
+            )
+        except litellm.BadRequestError as exc:
+            if not evidence_text_only_fallback:
+                raise
+            logger.warning(
+                f"LLM call to create a context failed with exception {exc!r}"
+                f" on text named {text.name!r}"
+                f" with doc name {text.doc.docname!r} and doc key {text.doc.dockey!r}."
+                f" Retrying without media."
+            )
+            llm_result = await summary_llm_model.call_single(
+                messages=[
+                    Message(role="system", content=system_prompt),
+                    Message(content=message_prompt),
+                ],
+                callbacks=callbacks,
+                name="evidence:" + text.name,
+            )
+            used_text_only_fallback = True
         context = cast("str", llm_result.text)
         result_data = parser(context) if parser else {}
         success = bool(result_data)
@@ -199,9 +243,7 @@ async def map_fxn_summary(
             except KeyError:
                 success = False
     else:
-        # Strip newlines in case chunking led to blank lines,
-        # but not spaces, to preserve text alignment
-        context = text.text.strip("\n")
+        context = cleaned_text
         # If we don't assign scores, just default to 5.
         # why 5? Because we filter out 0s in another place
         # and 5/10 is the other default I could come up with
@@ -213,6 +255,8 @@ async def map_fxn_summary(
 
     if not success:
         score = extract_score(context)
+    if used_text_only_fallback:
+        extras["used_images"] = False
 
     return (
         Context(
diff --git a/src/paperqa/docs.py b/src/paperqa/docs.py
index bc62fada5..be8b6b0c4 100644
--- a/src/paperqa/docs.py
+++ b/src/paperqa/docs.py
@@ -380,7 +380,7 @@ async def aadd(  # noqa: PLR0912
                 doc, **(query_kwargs | kwargs)
             )
 
-        texts = await read_doc(
+        texts, metadata = await read_doc(
             path,
             doc,
             chunk_chars=parse_config.chunk_size,
@@ -388,9 +388,10 @@ async def aadd(  # noqa: PLR0912
             page_size_limit=parse_config.page_size_limit,
             use_block_parsing=parse_config.pdfs_use_block_parsing,
             parse_pdf=parse_config.parse_pdf,
+            include_metadata=True,
         )
         # loose check to see if document was loaded
-        if (
+        if metadata.parse_type != "image" and (
             not texts
             or len(texts[0].text) < 10  # noqa: PLR2004
             or (
@@ -669,6 +670,7 @@ async def aget_evidence(
                         parser=llm_parse_json if prompt_config.use_json else None,
                         callbacks=callbacks,
                         skip_citation_strip=answer_config.skip_evidence_citation_strip,
+                        evidence_text_only_fallback=answer_config.evidence_text_only_fallback,
                     )
                     for m in matches
                 ],
diff --git a/src/paperqa/prompts.py b/src/paperqa/prompts.py
index 5786d04c8..b9028b963 100644
--- a/src/paperqa/prompts.py
+++ b/src/paperqa/prompts.py
@@ -1,10 +1,9 @@
 from datetime import datetime
 
-# ruff: noqa: E501
-
 summary_prompt = (
     "Summarize the excerpt below to help answer a question.\n\nExcerpt from"
-    " {citation}\n\n----\n\n{text}\n\n----\n\nQuestion: {question}\n\nDo not directly"
+    " {citation}\n\n------------\n\n{text}\n\n------------"
+    "\n\nQuestion: {question}\n\nDo not directly"
     " answer the question, instead summarize to give evidence to help answer the"
     " question. Stay detailed; report specific numbers, equations, or direct quotes"
     ' (marked with quotation marks). Reply "Not applicable" if the excerpt is'
@@ -12,9 +11,16 @@
     " newline indicating relevance to question. Do not explain your score.\n\nRelevant"
     " Information Summary ({summary_length}):"
 )
+# This prompt template integrates with `text` variable of the above `summary_prompt`
+text_with_tables_prompt_template = (
+    "{text}\n\n------------\n\nMarkdown tables from {citation}."
+    " If the markdown is poorly formatted, defer to the images"
+    "\n\n------------\n\n{tables}"
+)
 
 summary_json_prompt = (
-    "Excerpt from {citation}\n\n----\n\n{text}\n\n----\n\nQuestion: {question}\n\n"
+    "Excerpt from {citation}\n\n------------\n\n{text}\n\n------------"
+    "\n\nQuestion: {question}\n\n"
 )
 
 # The below "cannot answer" sentinel phrase should:
@@ -45,7 +51,7 @@
 
 qa_prompt = (
     "Answer the question below with the context.\n\n"
-    "Context:\n\n{context}\n\n----\n\n"
+    "Context:\n\n{context}\n\n------------\n\n"
     "Question: {question}\n\n"
     "Write an answer based on the context. "
     "If the context provides insufficient information reply "
@@ -99,15 +105,19 @@
 )
 
 # NOTE: we use double curly braces here so it's not considered an f-string template
-summary_json_system_prompt = """\
-Provide a summary of the relevant information that could help answer the question based on the excerpt. Respond with the following JSON format:
-
-{{
-  "summary": "...",
-  "relevance_score": "..."
-}}
-
-where `summary` is relevant information from the text - {summary_length} words. `relevance_score` is an integer 1-10 for the relevance of `summary` to the question."""
+summary_json_system_prompt = (
+    "Provide a summary of the relevant information"
+    " that could help answer the question based on the excerpt."
+    " Your summary, combined with many others,"
+    " will be given to the model to generate an answer."
+    " Respond with the following JSON format:"
+    '\n\n{{\n  "summary": "...",\n  "relevance_score": "..."\n  "used_images"\n}}'
+    "\n\nwhere `summary` is relevant information from the text - {summary_length} words."
+    " `relevance_score` is an integer 1-10 for the relevance of `summary` to the question."
+    " `used_images` is a boolean flag indicating"
+    " if any images present in a multimodal message were used,"
+    " and if no images were present it should be false."
+)
 
 env_system_prompt = (
     # Matching https://github.com/langchain-ai/langchain/blob/langchain%3D%3D0.2.3/libs/langchain/langchain/agents/openai_functions_agent/base.py#L213-L215
diff --git a/src/paperqa/readers.py b/src/paperqa/readers.py
index 1f7b787e3..e0c060902 100644
--- a/src/paperqa/readers.py
+++ b/src/paperqa/readers.py
@@ -2,10 +2,12 @@
 
 import asyncio
 import os
+from collections.abc import Awaitable, Callable
 from math import ceil
 from pathlib import Path
 from typing import Literal, Protocol, cast, overload, runtime_checkable
 
+import anyio
 import tiktoken
 from html2text import __version__ as html2text_version
 from html2text import html2text
@@ -13,6 +15,7 @@
 from paperqa.types import (
     ChunkMetadata,
     Doc,
+    ParsedMedia,
     ParsedMetadata,
     ParsedText,
     Text,
@@ -30,6 +33,42 @@ def __call__(
     ) -> ParsedText: ...
 
 
+async def parse_image(
+    path: str | os.PathLike, validator: Callable[[bytes], Awaitable] | None = None, **_
+) -> ParsedText:
+    apath = anyio.Path(path)
+    image_data = await anyio.Path(path).read_bytes()
+    if validator:
+        try:
+            await validator(image_data)
+        except Exception as exc:
+            raise ImpossibleParsingError(
+                f"Image validation failed for the image at path {path}."
+            ) from exc
+    parsed_media = ParsedMedia(index=0, data=image_data, info={"suffix": apath.suffix})
+    metadata = ParsedMetadata(
+        parsing_libraries=[],
+        paperqa_version=pqa_version,
+        total_parsed_text_length=0,  # No text, just an image
+        count_parsed_media=1,
+        parse_type="image",
+    )
+    return ParsedText(content={"1": ("", [parsed_media])}, metadata=metadata)
+
+
+def _make_chunk(
+    parsed_text: ParsedText, doc: Doc, text: str, lower_page: str, upper_page: str
+) -> Text:
+    media: list[ParsedMedia] = []
+    for pg_num in range(int(lower_page), int(upper_page) + 1):
+        pg_contents = cast(dict, parsed_text.content)[str(pg_num)]
+        if isinstance(pg_contents, tuple):
+            media.extend(pg_contents[1])
+    # pretty formatting of pages (e.g. 1-3, 4, 5-7)
+    name = "-".join([lower_page, upper_page])
+    return Text(text=text, name=f"{doc.docname} pages {name}", media=media, doc=doc)
+
+
 def chunk_pdf(
     parsed_text: ParsedText, doc: Doc, chunk_chars: int, overlap: int
 ) -> list[Text]:
@@ -48,27 +87,25 @@ def chunk_pdf(
             f" {doc.dockey}, either empty or corrupted."
         )
 
-    for page_num, page_text in parsed_text.content.items():
+    for page_num, page_contents in parsed_text.content.items():
+        page_text = (
+            page_contents if isinstance(page_contents, str) else page_contents[0]
+        )
         split += page_text
         pages.append(page_num)
         # split could be so long it needs to be split
         # into multiple chunks. Or it could be so short
         # that it needs to be combined with the next chunk.
         while len(split) > chunk_chars:
-            # pretty formatting of pages (e.g. 1-3, 4, 5-7)
-            pg = "-".join([pages[0], pages[-1]])
             texts.append(
-                Text(
-                    text=split[:chunk_chars], name=f"{doc.docname} pages {pg}", doc=doc
-                )
+                _make_chunk(parsed_text, doc, split[:chunk_chars], pages[0], pages[-1])
             )
             split = split[chunk_chars - overlap :]
             pages = [page_num]
 
     if len(split) > overlap or not texts:
-        pg = "-".join([pages[0], pages[-1]])
         texts.append(
-            Text(text=split[:chunk_chars], name=f"{doc.docname} pages {pg}", doc=doc)
+            _make_chunk(parsed_text, doc, split[:chunk_chars], pages[0], pages[-1])
         )
     return texts
 
@@ -232,6 +269,9 @@ def chunk_code_text(
     return texts
 
 
+IMAGE_EXTENSIONS = tuple({".png", ".jpg", ".jpeg"})
+
+
 @overload
 async def read_doc(
     path: str | os.PathLike,
@@ -287,7 +327,7 @@ async def read_doc(
     parse_pdf: PDFParserFn | None = ...,
     **parser_kwargs,
 ) -> tuple[list[Text], ParsedMetadata]: ...
-async def read_doc(
+async def read_doc(  # noqa: PLR0912
     path: str | os.PathLike,
     doc: Doc,
     parsed_text_only: bool = False,
@@ -326,6 +366,8 @@ async def read_doc(
         parsed_text = await asyncio.to_thread(
             parse_text, path, html=True, **parser_kwargs
         )
+    elif str_path.endswith(IMAGE_EXTENSIONS):
+        parsed_text = await parse_image(path, **parser_kwargs)
     else:
         parsed_text = await asyncio.to_thread(
             parse_text, path, split_lines=True, use_tiktoken=False, **parser_kwargs
@@ -351,6 +393,11 @@ async def read_doc(
             overlap=overlap,
             chunk_type="overlap_pdf_by_page",
         )
+    elif str_path.endswith(IMAGE_EXTENSIONS):
+        chunked_text = chunk_pdf(
+            parsed_text, doc, chunk_chars=chunk_chars, overlap=overlap
+        )
+        chunk_metadata = ChunkMetadata(chunk_chars=0, overlap=0, chunk_type="no_chunk")
     elif str_path.endswith((".txt", ".html")):
         chunked_text = chunk_text(
             parsed_text, doc, chunk_chars=chunk_chars, overlap=overlap
diff --git a/src/paperqa/settings.py b/src/paperqa/settings.py
index 5e42ae330..0cadbeb58 100644
--- a/src/paperqa/settings.py
+++ b/src/paperqa/settings.py
@@ -116,6 +116,13 @@ class AnswerSettings(BaseModel):
     evidence_skip_summary: bool = Field(
         default=False, description="Whether to summarization."
     )
+    evidence_text_only_fallback: bool = Field(
+        default=False,
+        description=(
+            "Opt-in flag to allow creating contexts without media (just text),"
+            " if the media is problematic for the LLM provider or network."
+        ),
+    )
     answer_max_sources: int = Field(
         default=5, description="Max number of sources to use for an answer."
     )
@@ -523,7 +530,11 @@ class IndexSettings(BaseModel):
         ),
     )
     files_filter: Callable[[anyio.Path | pathlib.Path], bool] = Field(
-        default=lambda f: f.suffix in {".txt", ".pdf", ".html", ".md"},
+        default=lambda f: (
+            f.suffix
+            # TODO: add images after embeddings are supported
+            in {".txt", ".pdf", ".html", ".md"}
+        ),
         exclude=True,
         description=(
             "Filter function to apply to files in the paper directory."
diff --git a/src/paperqa/types.py b/src/paperqa/types.py
index 3cd876ae2..3e9c3a23c 100644
--- a/src/paperqa/types.py
+++ b/src/paperqa/types.py
@@ -2,6 +2,8 @@
 
 import ast
 import csv
+import hashlib
+import json
 import logging
 import os
 import re
@@ -10,6 +12,8 @@
 from copy import deepcopy
 from datetime import datetime
 from enum import StrEnum
+from pathlib import Path
+from random import Random
 from typing import Annotated, Any, ClassVar, Self, cast
 from uuid import UUID, uuid4
 
@@ -21,8 +25,10 @@
 from pybtex.scanner import PybtexSyntaxError
 from pydantic import (
     BaseModel,
+    BeforeValidator,
     ConfigDict,
     Field,
+    JsonValue,
     PlainSerializer,
     computed_field,
     field_validator,
@@ -30,11 +36,13 @@
 )
 
 from paperqa.utils import (
+    bytes_to_string,
     create_bibtex_key,
     encode_id,
     format_bibtex,
     get_citation_ids,
     maybe_get_date,
+    string_to_bytes,
 )
 from paperqa.version import __version__ as pqa_version
 
@@ -139,6 +147,9 @@ class Text(Embeddable):
             " (e.g., 'Wiki2023 chunk 1', 'sentence1')."
         )
     )
+    media: list[ParsedMedia] = Field(
+        default_factory=list, description="Optional list of associated media."
+    )
     doc: Doc | DocDetails = Field(
         union_mode="left_to_right",
         description="Source document this text chunk originates from.",
@@ -167,6 +178,9 @@ def __hash__(self) -> int:
 class Context(BaseModel):
     """A class to hold the context of a question."""
 
+    # We allow extras so one can extend the summary JSON prompt
+    # to have the LLM answer with more conclusions such as alternate scores
+    # or useful excerpts of text
     model_config = ConfigDict(extra="allow")
 
     id: str = Field(
@@ -344,7 +358,8 @@ def filter_content_for_user(self) -> None:
                     # on why we drop embeddings, drop embeddings here too because
                     # embeddings aren't displayed to front end users
                     doc=c.text.doc.model_dump(exclude={"embedding"}),
-                    **c.text.model_dump(exclude={"text", "embedding", "doc"}),
+                    # We drop media since images can be quite large
+                    **c.text.model_dump(exclude={"text", "embedding", "doc", "media"}),
                 ),
             )
             for c in self.contexts
@@ -421,8 +436,8 @@ def __init__(self, *args, **kwargs):
 class ChunkMetadata(BaseModel):
     """Metadata for chunking algorithm."""
 
-    chunk_chars: int
-    overlap: int
+    chunk_chars: int = Field(description="Chunk size (chars), or 0 for no chunking.")
+    overlap: int = Field(description="Chunk overlap (chars), or 0 for no overlap.")
     chunk_type: str
 
 
@@ -431,15 +446,86 @@ class ParsedMetadata(BaseModel):
 
     parsing_libraries: list[str]
     total_parsed_text_length: int
+    count_parsed_media: int = Field(default=0, ge=0)
     paperqa_version: str = pqa_version
     parse_type: str | None = None
     chunk_metadata: ChunkMetadata | None = None
 
 
+class ParsedMedia(BaseModel):
+    """Raw image or table parsed from a document's page."""
+
+    index: int = Field(
+        description="Index of the image in a given page, or 0 if solely an image."
+    )
+    data: Annotated[
+        bytes,
+        PlainSerializer(bytes_to_string),
+        BeforeValidator(lambda x: x if isinstance(x, bytes) else string_to_bytes(x)),
+    ] = Field(description="Raw image, ideally directly savable to a PNG image.")
+    text: str | None = Field(
+        default=None,
+        description="Optional associated text content (e.g. markdown export of a table).",
+    )
+    info: dict[str, JsonValue | tuple[float, ...] | bytes] = Field(
+        default_factory=dict,
+        description=(
+            "Optional image metadata. This may come from image definitions sourced from"
+            " the PDF, or attributes of custom pixel maps."
+        ),
+    )
+
+    def __hash__(self) -> int:
+        return hash(
+            (self.index, self.data, self.text, json.dumps(self.info, sort_keys=True))
+        )
+
+    def to_id(self) -> UUID:
+        """Convert this media to a UUID4 suitable for a database ID."""
+        # We only hash the image and text content, since we don't want
+        # minor parsing details (e.g. inconsequentially-small decimal values
+        # in bounding boxes) to change the resultant ID
+        to_hash: bytes = (
+            self.data if not self.text else self.data + self.text.encode("utf-8")
+        )
+        seed_hash = hashlib.sha256(to_hash).hexdigest()
+        seed_uint32 = int(seed_hash, 16) % (2**32)
+
+        # Convert uint32 to UUID4
+        uuid_int = Random(seed_uint32).getrandbits(128)
+        uuid_int &= ~(0xF << 76)  # Clear version bits
+        uuid_int |= 0x4 << 76  # Then set version to 4
+        uuid_int &= ~(0x3 << 62)  # Clear variant bits
+        uuid_int |= 0x2 << 62  # Then set variant to 10 for RFC 4122
+        return UUID(int=uuid_int)
+
+    def __eq__(self, other) -> bool:
+        if not isinstance(other, ParsedMedia):
+            return NotImplemented
+        return (
+            self.index == other.index
+            and self.data == other.data
+            and self.text == other.text
+            and json.dumps(self.info, sort_keys=True)
+            == json.dumps(other.info, sort_keys=True)
+        )
+
+    def to_image_url(self, image_type: str = "png") -> str:
+        """Convert the image data to an RFC 2397 data URL format."""
+        return f"data:image/{image_type};base64,{bytes_to_string(self.data)}"
+
+    def save(self, path: str | os.PathLike) -> None:
+        """Save the image to the input file path."""
+        with Path(path).open("wb") as f:
+            f.write(self.data)
+
+
 class ParsedText(BaseModel):
     """All text from a document read, before chunking."""
 
-    content: dict[str, str] | str | list[str] = Field(
+    content: (
+        dict[str, str] | str | list[str] | dict[str, tuple[str, list[ParsedMedia]]]
+    ) = Field(
         description=(
             "All parsed but not further processed (e.g. not chunked) contents from a"
             " document. It may be structured, depending on the parser's implementation."
@@ -448,6 +534,8 @@ class ParsedText(BaseModel):
             "\n- `dict[str, str]` (e.g. page number -> page text) for PDFs."
             "\n- `str` for text files."
             "\n- `list[str]` for line-by-line parsings."
+            "\n- `dict[str, tuple[str, list[ParsedMedia]]]` (e.g. page number"
+            " -> (page text, page images)) for PDFs."
         )
     )
     metadata: ParsedMetadata = Field(
@@ -473,7 +561,9 @@ def reduce_content(self) -> str:
             return self.content
         if isinstance(self.content, list):
             return "\n\n".join(self.content)
-        return "\n\n".join(self.content.values())
+        return "\n\n".join(
+            x[0] if not isinstance(x, str) else x for x in self.content.values()
+        )
 
 
 class BibTeXSource(StrEnum):
diff --git a/src/paperqa/utils.py b/src/paperqa/utils.py
index 82c70cfe7..25397828d 100644
--- a/src/paperqa/utils.py
+++ b/src/paperqa/utils.py
@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 import asyncio
+import base64
 import contextlib
 import hashlib
 import logging
@@ -15,7 +16,7 @@
 from functools import reduce
 from http import HTTPStatus
 from pathlib import Path
-from typing import Any, BinaryIO, ClassVar, TypeVar
+from typing import TYPE_CHECKING, Any, BinaryIO, ClassVar, TypeVar
 from uuid import UUID
 
 import aiohttp
@@ -33,6 +34,12 @@
     wait_incrementing,
 )
 
+if TYPE_CHECKING:
+    from typing import IO
+
+    from PIL._typing import StrOrBytesPath
+
+
 logger = logging.getLogger(__name__)
 
 T = TypeVar("T")
@@ -616,3 +623,37 @@ def clean_possessives(text: str) -> str:
     # Remove standalone apostrophes
     text = re.sub(r"\s+'\s+", " ", text)
     return re.sub(r"(?<!\w)'\s*", "", text)
+
+
+def bytes_to_string(value: bytes) -> str:
+    """Convert bytes to a base64-encoded string."""
+    # 1. Convert bytes to base64 bytes
+    # 2. Convert base64 bytes to base64 string,
+    #    using UTF-8 since base64 produces ASCII characters
+    return base64.b64encode(value).decode("utf-8")
+
+
+def string_to_bytes(value: str) -> bytes:
+    """Convert a base64-encoded string to bytes."""
+    # 1. Convert base64 string to base64 bytes
+    # 2. Convert base64 bytes to original bytes
+    return base64.b64decode(value.encode("utf-8"))
+
+
+def validate_image(path: StrOrBytesPath | IO[bytes]) -> None:
+    """
+    Validate that the file at the given path is a valid image.
+
+    Raises:
+        OSError: If the image file is truncated.
+    """  # noqa: DOC502
+    try:
+        from PIL import Image
+    except ImportError as exc:
+        raise ImportError(
+            "Image validation requires the 'image' extra for 'pillow'. Please:"
+            " `pip install paper-qa[image]`."
+        ) from exc
+
+    with Image.open(path) as img:
+        img.load()
diff --git a/tests/cassettes/test_get_reasoning[deepseek-reasoner].yaml b/tests/cassettes/test_get_reasoning[deepseek-reasoner].yaml
index a0b6bb85d..a8c27f9ec 100644
--- a/tests/cassettes/test_get_reasoning[deepseek-reasoner].yaml
+++ b/tests/cassettes/test_get_reasoning[deepseek-reasoner].yaml
@@ -19,7 +19,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -29,7 +29,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -45,18 +45,18 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jJJNb9swDIbv/hUCz8kQJ6kT5NYCPexQYB8Y2q4ObEWiHWWyJEh0uyLI
-          fx9kJ3G6dUAvOvDhS/EluU8YAyVhxUBsOYnG6fHN6131w3zfyMd69qXd3WaP5la6h/ufX9PlHYyi
-          wm52KOik+iRs4zSSsqbHwiMnjFXTxdVsMc2u0qwDjZWoo6x2NJ7b8XQynY/TdDydHIVbqwQGWLGn
-          hDHG9t0bWzQSf8OKTUanSIMh8BphdU5iDLzVMQI8BBWIG4LRAIU1hKbruizLXbAmN/vcMJYDKdKY
-          w4rl8HD9mX3DZ4UvOYx6ylvaWh8if8rhHrXmL5wIGRLjOof1MU9aFXNMq3VuDrkpy/Lyf49VG7g+
-          ZlwAbowlHsfXOV8fyeHsVdvaebsJf0mhUkaFbeGRB2uir0DWQUcPCWPrbqbtmzGB87ZxVJD9hd13
-          i2lfDoYlDnB2gmSJ6yGeTuajd8oVEokrHS6WAoKLLcpBOmyQt1LZC5BcmP63m/dq98aVqT9SfgBC
-          oCOUhfMolXjreEjzGG/8f2nnIXcNQ0D/rAQWpNDHRUiseKv784PwGgibolKmRu+86m+wcoXIFots
-          WWXLCSSH5A8AAAD//wMAmWDqAYwDAAA=
+          H4sIAAAAAAAAAwAAAP//jFJLi9swEL77V4g5J8VOvBvIrbQUlpal5NAH62DL0tjWVpaENN60hPz3
+          IjsbZ9st9KLDfA/NNzPHhDFQErYMRMdJ9E4v32Uf779nu7prxX394ZOXu4Pc3X0OX/z7Q4BFVNj6
+          EQU9q94I2zuNpKyZYOGRE0bXbHOT5+vVJk1HoLcSdZS1jpa5Xa7SVb7MsuUqPQs7qwQG2LKHhDHG
+          juMbWzQSf8KWjTZjpccQeIuwvZAYA291rAAPQQXihmAxg8IaQjN2XVXVY7CmMMfCMFYAKdJYwJYV
+          8O3tHdvhk8JDAYsJ5QN11oeIPxTwFbXmB06EDIlxXcD+zJNWRY4ZtC7MqTBVVV3/77EZAtdnxhXA
+          jbHE4/jG5Pszcrpk1bZ13tbhDyk0yqjQlR55sCbmCmQdjOgpYWw/znR4MSZw3vaOSrI/cPxus5rs
+          YF7iDK6fQbLE9VzP0nzxil0pkbjS4WopILjoUM7SeYN8kMpeAclV6L+7ec17Cq5M+z/2MyAEOkJZ
+          Oo9SiZeJZ5rHeOP/ol2GPDYMAf2TEliSQh8XIbHhg57OD8KvQNiXjTIteufVdIONK29ryZvbdX6D
+          kJyS3wAAAP//AwAjefZ9jAMAAA==
       headers:
         CF-RAY:
-          - 96665c683b2d9e5c-SJC
+          - 96a9b5391cb9ce98-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -64,14 +64,14 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:17 GMT
+          - Tue, 05 Aug 2025 22:25:00 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=fbASGENIWM6WUGwxO83qItf.23NH2kgQXfA1uog0cXQ-1753726517-1.0.1.1-eDkiJARyv_9PqM1Pl2o3eQhT1GkFHGV0jsXiE66uNNNZBmkJZ_snWvlCtJI46rb_sRKoyGSgNxWfDzLOwdVP6PHoVtA1veHO0kOHW1XSed4;
-            path=/; expires=Mon, 28-Jul-25 18:45:17 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=TucZD8dgmdbo4g5zsCtPh52bMXjJbFdEcIdhIVw8eUI-1754432700-1.0.1.1-lVxjkojC57mKjrusxVqZyfevH2b9sHaPs9so5aEESBrg2nuRraoNFREeW8V7fuea0XPy3h5NbxAK41i4G.svVNAV8aYZWP8HVxOrm2Na9ik;
+            path=/; expires=Tue, 05-Aug-25 22:55:00 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=PFYGlZsoULP.UNmLeJtCVev9sdnHPwvrB2YHWJTmvcM-1753726517464-0.0.1.1-604800000;
+          - _cfuvid=6iT0ZRuuhfAWhcV4MeYyMhYWIjDiHOGoXB409bgss8c-1754432700783-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Transfer-Encoding:
           - chunked
@@ -86,7 +86,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "825"
+          - "370"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -94,7 +94,7 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "831"
+          - "378"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -102,13 +102,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29999934"
+          - "29999935"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 0s
         x-request-id:
-          - req_cc7bf4e1090cda31bc9f3e39c4b115de
+          - req_997c197bfb5c6d259915734d933ea7e8
       status:
         code: 200
         message: OK
@@ -130,14 +130,14 @@ interactions:
           null}, "publicationTypes": ["JournalArticle", "Conference", "Review"], "publicationDate":
           "2024-11-20", "journal": {"name": "2024 8th International Conference on System
           Reliability and Safety (ICSRS)", "pages": "511-517"}, "citationStyles": {"bibtex":
-          "@Article{Nguyen2024EnhancingPW,\n author = {Duc An Nguyen and Khanh T. P.
-          Nguyen and Kamal Medjaher},\n booktitle = {International Conference on System
-          Reliability and Safety},\n journal = {2024 8th International Conference on
-          System Reliability and Safety (ICSRS)},\n pages = {511-517},\n title = {Enhancing
-          PHM with XAI: Review and Dataset for System Health Indicator Construction},\n
-          year = {2024}\n}\n"}, "authors": [{"authorId": "2167438752", "name": "Duc
-          An Nguyen"}, {"authorId": "2184162840", "name": "Khanh T. P. Nguyen"}, {"authorId":
-          "2267984723", "name": "Kamal Medjaher"}], "matchScore": 57.629684}]}
+          "@Article{Nguyen2024EnhancingPW,\n author = {D. Nguyen and Khanh T. P. Nguyen
+          and Kamal Medjaher},\n booktitle = {International Conference on System Reliability
+          and Safety},\n journal = {2024 8th International Conference on System Reliability
+          and Safety (ICSRS)},\n pages = {511-517},\n title = {Enhancing PHM with XAI:
+          Review and Dataset for System Health Indicator Construction},\n year = {2024}\n}\n"},
+          "authors": [{"authorId": "2167438752", "name": "D. Nguyen"}, {"authorId":
+          "2184162840", "name": "Khanh T. P. Nguyen"}, {"authorId": "2267984723", "name":
+          "Kamal Medjaher"}], "matchScore": 59.218933}]}
 
           '
       headers:
@@ -146,31 +146,31 @@ interactions:
         Connection:
           - keep-alive
         Content-Length:
-          - "1448"
+          - "1440"
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:18 GMT
+          - Tue, 05 Aug 2025 22:25:01 GMT
         Via:
-          - 1.1 fb1699c4cb8ff04b39762e99ca06e3d2.cloudfront.net (CloudFront)
+          - 1.1 b735cef950dccc7c59398fa8df6c4f7e.cloudfront.net (CloudFront)
         X-Amz-Cf-Id:
-          - dSWAvGeb0-zvFa4P0Cai61BW4J375nDq7uscBWhrJICNqC3O9bo5Fw==
+          - u0hPbL-0-C8MdJgMvbFkWCkOK_-Br_CN9v-lRaeeNaXV_xlIJGlSFQ==
         X-Amz-Cf-Pop:
           - SFO53-P7
         X-Cache:
           - Miss from cloudfront
         x-amz-apigw-id:
-          - ObqocGZiPHcEOrQ=
+          - O2mtmFwdPHcEPpw=
         x-amzn-Remapped-Connection:
           - keep-alive
         x-amzn-Remapped-Content-Length:
-          - "1448"
+          - "1440"
         x-amzn-Remapped-Date:
-          - Mon, 28 Jul 2025 18:15:18 GMT
+          - Tue, 05 Aug 2025 22:25:01 GMT
         x-amzn-Remapped-Server:
           - gunicorn
         x-amzn-RequestId:
-          - 99bca4a5-a373-4270-9eb2-e361ea2a53f6
+          - 0ba6b8d9-f059-4a79-a9cd-a4409223da0b
       status:
         code: 200
         message: OK
@@ -182,11 +182,11 @@ interactions:
     response:
       body:
         string:
-          '{"status":"ok","message-type":"work-list","message-version":"1.0.0","message":{"facets":{},"total-results":22730,"items":[{"indexed":{"date-parts":[[2024,8,7]],"date-time":"2024-08-07T06:47:19Z","timestamp":1723013239657},"reference-count":0,"publisher":"Transstellar
-          Journal Publications and Research Consultancy Private Limited","issue":"3","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IJMPERD"],"published-print":{"date-parts":[[2020]]},"DOI":"10.24247\/ijmperdjun2020913","type":"journal-article","created":{"date-parts":[[2020,8,27]],"date-time":"2020-08-27T00:33:19Z","timestamp":1598488399000},"page":"9575-9582","source":"Crossref","is-referenced-by-count":0,"title":["A
-          Review Paper on Food Security"],"prefix":"10.24247","volume":"10","author":[{"given":"Ansumansamal
-          et al.,","family":"Ansumansamal et al.,","sequence":"first","affiliation":[]},{"name":"TJPRC","sequence":"additional","affiliation":[]}],"member":"10346","container-title":["International
-          Journal of Mechanical and Production Engineering Research and Development"],"language":"en","deposited":{"date-parts":[[2020,8,27]],"date-time":"2020-08-27T00:33:19Z","timestamp":1598488399000},"score":24.880903,"resource":{"primary":{"URL":"http:\/\/tjprc.org\/publishpapers\/2-67-1598502792-913IJMPERDJUN2020913.pdf"}},"issued":{"date-parts":[[2020]]},"references-count":0,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2020]]}},"alternative-id":["Arch-13770"],"URL":"https:\/\/doi.org\/10.24247\/ijmperdjun2020913","ISSN":["2249-6890"],"issn-type":[{"type":"print","value":"2249-6890"}],"published":{"date-parts":[[2020]]}}],"items-per-page":1,"query":{"start-index":0,"search-terms":null}}}'
+          '{"status":"ok","message-type":"work-list","message-version":"1.0.0","message":{"facets":{},"total-results":22876,"items":[{"indexed":{"date-parts":[[2024,8,7]],"date-time":"2024-08-07T06:45:10Z","timestamp":1723013110666},"reference-count":0,"publisher":"Transstellar
+          Journal Publications and Research Consultancy Private Limited","issue":"6","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IJASR"],"published-print":{"date-parts":[[2017]]},"DOI":"10.24247\/ijasrdec201724","type":"journal-article","created":{"date-parts":[[2017,11,9]],"date-time":"2017-11-09T01:18:32Z","timestamp":1510190312000},"page":"173-180","source":"Crossref","is-referenced-by-count":0,"title":["Bovine
+          Mastitis, Global Quandary \\\"A Review\\\""],"prefix":"10.24247","volume":"7","author":[{"given":"Kamaldeep
+          et al.,","family":"Kamaldeep et al.,","sequence":"first","affiliation":[]},{"name":"TJPRC","sequence":"additional","affiliation":[]}],"member":"10346","container-title":["International
+          Journal of Agricultural Science and Research"],"language":"en","deposited":{"date-parts":[[2017,11,9]],"date-time":"2017-11-09T01:24:49Z","timestamp":1510190689000},"score":24.764158,"resource":{"primary":{"URL":"http:\/\/tjprc.org\/publishpapers\/2-50-1510208310-24.IJASRDEC201724.pdf"}},"issued":{"date-parts":[[2017]]},"references-count":0,"journal-issue":{"issue":"6","published-print":{"date-parts":[[2017]]}},"alternative-id":["Arch-9275"],"URL":"https:\/\/doi.org\/10.24247\/ijasrdec201724","ISSN":["2250-0057"],"issn-type":[{"type":"print","value":"2250-0057"}],"published":{"date-parts":[[2017]]}}],"items-per-page":1,"query":{"start-index":0,"search-terms":null}}}'
       headers:
         Access-Control-Allow-Headers:
           - X-Requested-With, Accept, Accept-Encoding, Accept-Charset, Accept-Language,
@@ -204,7 +204,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:20 GMT
+          - Tue, 05 Aug 2025 22:25:02 GMT
         Server:
           - Jetty(9.4.40.v20210413)
         Vary:
@@ -213,9 +213,9 @@ interactions:
           - interest-cohort=()
         x-api-pool:
           - plus
-        x-ratelimit-interval:
+        x-rate-limit-interval:
           - 1s
-        x-ratelimit-limit:
+        x-rate-limit-limit:
           - "150"
       status:
         code: 200
@@ -1283,7 +1283,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -1293,7 +1293,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -1309,1692 +1309,1691 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA1RWSa+zOhLdv19xdbdpKUzB9rcjTCEYMAESSK8gDAEyMdnA+/Ote5/Urd5Ycvks
-          rKo6w99/fX19v7OmuI3ff76+H/Uwfv/rp5anY/r95+vff319fX39/Xv+H7J4ZkWe16/qF/77WL/y
-          Yv7+88X9t/I/0J+vb+7dKN4amQ+wus3JBNNFuBN1c7RCgb+4NXxvhxNNDcMN190WpzD2xh1xbcvX
-          Zgkda4TCqKDxKVqyJczqAEIR7al+Ns7aYi5yC9/CvCFu/MIaF2pzAovPPiO6zj21zh2cAR7aLJ42
-          ZbjvxbWEOjxuDudplS7NSn1rH4DHqeCIFugRXu5sI8NeGezf/2SzQHoGgfCRPPa0P3g48E4B9U93
-          I8TZxD3z8eTB9K4AajaPnTaDmZiw/kBCzvi99nMoMhkNYDlPrYIEjY2vRUZ4JDox27zGi4qgA8Lx
-          FtDkU/R4fJZhDRVv8/CAxE39st/vTXS4iofpaapqxreGvwDnMEckPh+2mLWMW1BDPhM1/dN2ZbzX
-          CVDR1CNVIHA0zhh7Gfq7VScHk9uEM89XG3Te7m0aUOHQjMXVUtAurRlVjnHUU2lRVcgjsyT4Lkjh
-          3PJqjfYbPaLXKHbxxB0dW64SbBPLr81e8N6ijPCQidTUuUvGroFcg587Meu31ohEJgqE6v5MXD2c
-          15mtHwe6W/VInDgpMPd+rDY6NSwiZ2VLANtLYwUThfMJ9miA1/5atVAOh3XqwyZu1oHfXuA1eRu0
-          jF9ZJvJp5sPWUDKaJE6TiWDXPVHCQ0qcR78Bq6qqCirnd0FPRf3E/Yl7OnAtd2ha8FUIWS7WHSq4
-          RCKZLFSYr8rXAqfJSom+OUrN3EaTBcpnuKd6Tx49ZxnBhHjsXGgqICEcZb22oJSfbJqeLzBbiqGx
-          0LWXM+oIrp7N2/PWh7/9LtEO4xFujz60QQSICsITXnZZoCPRbTMa9yfST3v/UshVZLzoUSGeNsiV
-          tCBuaBtS7qd7KD46rMBN/GFU12Cuzf3J0NEgDgvV2+vUL0leq+hzUDDxy7uBxf4YpAizMqOHZ25j
-          bqfdPYQctZ2gEfhgntF52c3ECqii0WM4tkbCgPmCG5qPrdX/s6+fvBaploySNl9N2QQs94801fJa
-          W97uEqDNcnNpQiYFDLHlWuCn38S0hQNgB+nko0BNDuS0M/mVOYe0gPXDKKmmsrQR1SBxwFoCRA5X
-          rGbLLagStJSdR5zt+6atyen+RKckFam57IJGkPvKR0JnSNTCog84MztO6EyF2GN8ofYcep8itPDn
-          idg3remFYjfUUPNanyofRQm5T5hAaDw/PCnrSO0Zo9kN6oW/JZcuUDLxvJMXtNMccRJXaGji2OQ2
-          RCSQiadWLl7Cnt3QD/8noc1VzHG+ICDbrc9kD49VJhx82wPzyAyCXTA0S88JKWTDDRLF9fSGK1Xf
-          RE9yupKoet7XNZUeA3SInE7ztcszntQXC9DXJSe6gi4aN8bDBPFwFYkSSfts5d2TDPM8ITQ01gdm
-          lurK0KoLmzqW9daYc3kJKGO0o64LbqC9WyIHm917M70OvKjNyae7ATgmM7GL/QyWQZU6FDmmQYJQ
-          W/F8w4YAH7YtT5L0mDPxxbEUwXPYe0K5OCFXREuA4jEeiJVuYm2IlOEJYtKYxN/pELOUPCaYShFH
-          biwxsYAUNUDAh1dqixuChZ7bJKB33k8ao6wKx/DgXaAvbUVvQ3DdL6U5QrjNlRsJSebjdevpT8ja
-          ViMpw2ozt7xdgR/9pI5S4GxYrsNNrqaPQM0wsvFivxhEDt539NCfSDOP2VL8+geJfvmhS2GKYNVb
-          9PAq1JUny3ABKXc6UX0dX2BRC79DP/pL971ihYxt7x3SobTQC0zcdeGvQQSiEC9E+/GPYTonEzyI
-          /eg9z8d3syItdmBtloOHbkEF2CpwKlLleqT2o5+a1Sr0Gpp+b5C9NpaAx2O8wG4+2TTi7m7DZGZa
-          ELgl/7O/Wr9c5V0Ll4Lfk7N0iTRBs/kOQG/ZTOx+afp5tMoC+oqiElvLVU28PbUa+QcZUFNgak/X
-          RpHkU7NEVNvgWhse4wtCH8OCxAtwsvmQIA7wSC+nVoM5Zg13bOHbVQBRn/Yuo6SbO9kN3yMNfH3A
-          bDLUGrWvvJ1k9jOvD/AVlLfTkXp3S9GEJiMS7J/Zx4MlmfDqko0JB8DO9JAhQVtevucDe5sW9Mhv
-          XngdtV0LZ2IH1IVSjcWbcJ1g80zf5IAyJRMnX4bQ2i2ahyz7vjIBXxQUzfODuM7mkPFddJJR6sWt
-          Nwv5DOiP3wA1ebfk+ql0wCT+CaFWFSklt2DW1iOIdMSpSkTjNpOaefaZgLhr5HocyT9goR/IoFny
-          Iolh+mq4+9VU0K8fqnpTZosf7Sf4YPqLnuTovg4zjDh0uxkmyfv6AjiuvQrw3oMbTagpYHa6vhNk
-          1TebFPoZ96Ia+B4Y6KOn0SmwQl6PavMfvlqvj9OLcYha4PppTH/5znfNK/hH3/OfPLSsxWqjyvdP
-          9ArWR8i6Krgg28gIcd7TJ3xbwU2HzztkpITSkq3P+C385iVyMwIGmNvWCjKsgBArvLvNfM8zJrvT
-          DL0uV67aGt0aDyljmxCckydeDTNfgCBHBvXAPsWL8dg5cOQ2CY229hyyHn1kxJbthx5ztW7WT7lP
-          kGQrHHVu2rqyUdrryA+DhDripgznML7VMFbXfJq1sVxZuRw5+KNHE0sKee1PZ7uDAYihx513adM9
-          HnyAilzCJGPHK2aP6b3AeFdfiCUtbdN/ymOy00QWUseBbsNPb8BBbA4JsUArZWwb1BH60W9qM1w3
-          a4kWH12B4Xl3eFTC8VzvJbjbWQ21d880Ww2BVHDibxFR8VXIFjHcVfBi3EWijpygDY1WKoC54nF6
-          GCsfzoeb2sFKFqqJvfB+5Wele0NFU47EunNT9psPweNcxTRGMcIrx2wfat7TJx4+bMNBr+oKAXdS
-          qbskHuYSY7YQ8SaJqh+F9Wyn3R142C4rOWaujsfj5cNBwnUPYrWd0c9Beo/AHoUHaul6m63dYDPY
-          KlAiBRLncK5TbwOnThcn6CIzE6vDJEP9fr8Q7wGNXshO9Ak2vR1O3H4zgZVl7wt8LcVAvY712sop
-          SQ1+9Hna5uFxpek7rnb8nWrUlBuajULAWih0B4l6lpVk891PU6gEDqRYFhRN3LdyCk+i2k1cEN4x
-          e3h6CnMjPHq7iqfZ6JN1s/tAq6OnrfUK58DObDjP6o3oP/NnZfEU5OrmjdQtfKKta8npMMGopAVL
-          LhkX8FtbWv8DAAD//6R7ybKyQJvmvfTWjpBJMnPJPEsiIOIOFBEQkSmBjKh7r+B8f+1q1b005Hjg
-          nZ7hTXY+B+farKm5zCNcmitHdr6b93y0lmgSvHlmnjWfk1lqFhjeR4n89UMvdvkD3vDbmA9mvNX0
-          8gk4aJ0WldjSQ87po3Ur2IjQ9Fm71/OVXywH/vWbRjgejPdamiH3enc4ieLeXcdDz/zpG2Id8y7a
-          CrgmSFNONU45ZKjjzh/BE3gsuRxA7FJ70AUoOJw8sxmc3P6eXT2YJ/CLHX8SonXuKAOLj68Tz4h+
-          w4CiTwbj2/lKtMbLoxWH+ABGW/Dw9fOQXfa6pRpExWvyefFjq9OnKmPIH4Yb0V9LOKx9UxqQZdCd
-          ONrFpLydHUd4IePiM3T60k0cuhAQ6dNgN6wOdFZEkgJDjzridi+Y//Fv8JdvuTskYJG7Wwx/1/uK
-          ncVVhjWX7wE0aJzj22MFYOcDI/zjD8xVZ9V1nOMOTqqGyStmvzV1T5YBF1mciDYJds2qNV/A7qyA
-          meXe7LDa7aKgYjIO/mZsmsv95WvHZ4xj9jsshtmO4JFfTj7aXBF0gs56p7HMTnt8QL6ExjmDH3Dm
-          /NND/oGOWXMfItb/krjNJZcLZ+8BEuF8noUoFiNyUoIQMYoS++ItncCCfSqJk38u8EvQPMDuehQW
-          11Wff7PJu11l1Azq0bkl3lesh5V/nSCs1tDGnmsSdS3VuwL/+EOhXYd6PfHdjI58MxLj+/ipNF3P
-          DozzUCAKXzbR5jJiAlU6Lz7zqA4D2fFTLMlTxTZowz98HKGlMjeCMSqGZTgvjhiofEqs84hzaiT1
-          AbAL4Ig+qzSiD+4yQoaHEo7W36Oejsmiwe0QyDjzJyFf3ndDgXXOpcToAp9ui8q3UCQaJjfQvumk
-          muUIwUNgseUN3B9fLgDvUR9LtXkYZkczFzhHjfuf+Zr/0hjeMjMhiv8o6V4fCawuvP0X33xlP9kG
-          d72AvTlhogb7VAF85ww+fAuhuu38EN5l/YiVkS+HNb5KFaDR8T0Ty9nUMcPTDG6V/yTWZksD/cvX
-          nZCVeGZc5dNlUqs/vkqM97cH1FzaEar6gIhqtxcwvk9VCp/AZ7Fn2yWgz+BRoXYCHpGx79WsRFYB
-          rknhYVlZ+nwG+mCI+/zEt6D06Hq2ogzt+fAXrSYuPSlpAKVfyeJCmcpoTY73Ap4+4pmYmuIMi70t
-          D1E8XBSsXXXWpUgLFPSXPwvPJdhqdyuh2kTLv/vvDjEXw12fYVv1r+rKVucDHGo38ikR3vnanYoA
-          6Jo3kkzTmJrY4KEBexYrrJyuCqDgbvhitIw/HMTCO9/EVA/QzezOM98HHKUR0TMUSIpCvDNqIzrR
-          OoFbthbEd5wfWITETeGuR/HjLYTuOp8UBb4VJ8Om+Pm522EqJbTc/Rabs9nS9X0+t7DOVZeYoH2D
-          lbDIg4IMZmLkAEbrzwwq9HbUCFsLiAf+iA4JoKXlEkmrsbraqs/BITqU2E1zqeZ3vIFrmP2wNrNH
-          ddmSVw9f7ChjaWI+dCtOYwkH8eURu32K0bKMfg+oOGdY41453fXyAqui2YiMRBBtuK0FtFV6iq3M
-          Wevt5Acz9L/oSNTbaYto44EH+PMb4r/51MStBfAUDzgiwjvaruxHgayqB0Q+Qg8wOXIfcLRFj/zp
-          Vb5tOR9UYZfgyxch+k8fCkBExH7IP7pYK9/B1kxL7FW6ntM//f7HJ+UvQmDVi6SDq6l6WN169s8/
-          KUFtNco82jqqF+jKC0S3TJ6HWrDB7qco8DStBlHVQzuMxRt0onocPP/gmse8H3St/+eH6UQbVarf
-          hQ3oFfPy09X60vm79B3880f89U7cFStiA970nGCTJxe6vIqZg8OhJNjY9c4aLGkMevXR7P7NhS4k
-          VAWEoqSYvxsYo6X7AAv64FGRwjVfEVVflwZl/rXBNvwOKulhcEDKYE7YR/wl2vlGA/b68MU50dzN
-          lekGRenxxG4/v91d31XgddN8LJsNAbOY6iH8jFyAw5GXhhXOQIPcrSMzqUwSbcJz1mD9MmaMmYvk
-          jrxAFwgdkmNdkbph5b0EormYspk/NUG0TaXkQeEIS5zXmpYv71/swbyKNaJHYRBtl/USgD3evqiH
-          Us2d8tAQAf/Sib068sDk58sI+oKR5hGDeNjYX1lByzQmrM68VXO3xfOheNsS/3OJK3VrUir+zafp
-          wJ2Zej0aPwE90/qFlXUwXPKNYwXOZv7Gnsgkw3aTsgL2xZfO/LeoANfW7xCej5KNL+QrDrQoqxmO
-          ZXryASu/6aIty4ggMw/7/FXqLY9hCoacvfvL2W/q7cOpjz+9R7QU6uoW1DcDVB9xxFYUZ3l3PQQl
-          2vUYdvf6Wm5pnsDHwzSI/lOyehMmJkDWwi//vmcqe+mR9useRLme+nx6iMcGxujgY+MSCvU6nxwF
-          ohVe8GPnY2zfdBq8xgWYm+fS1+vp5sV//TDzOz8gWDQluGzoR2Lkqe6y3w9M15NHnJfS18ufPzi8
-          BIhvJntTp7NwtuBdd6t5edtBxFn3wwZvBTTIq979j/VxCkF4zF++6IjD8PlUZYKURpJxLm6QEkVx
-          FKjfXgox/LIZ1vczWtB+PVGMTVNpclNS9PWq4wxq4UcJrzUhioRv5Yvf3wOM5SPjgNB8NeLb9sVd
-          e4oesK6tEKu/4xGsNy8U/vTfP/28vbjbDNGPd4g9GUW+4qqwgBVi9OffuRu4ngK4+18+/Ip1var1
-          sYAn5BvEGuMq3y4BFP76e/dzIJ35YBzhjpdYG5Twn38Jr2lV+Ict+9K/+jpFBrn7jLHF0QhFUIFW
-          eD2w+mg6d/sYlxGWG8mI0st2xN8PfgU/5e/7z78db2kUQ3XyVXxWCpqvoegJEE3iRPRl5KPRMOcZ
-          atjwiFxUhsvK46SBr9YZ2DXDruZfvlTB9zn67Xp2rOcP5z5EO8+Pf/N94JXLLYN/ftD9ljDqqJV9
-          BW+zdcIh/THR/OJeI0RJyxDjEqYDO7yjEDTeKSZWHwbDpnW6B3Pz/SVKoI3qMq/fAuhWgMnryN+j
-          WV5EDiw2bbH/it7Drn8ZeLxoEr7kow82qBTBiR993++/k6QyjaxKUGKEkGjbSNQtoRcJ9mwyEdm+
-          +znlmUSC96OzziBYazrs+kj8q3fb1HWV/cLBAHs/+4yPEVj/9JBzTAuy67t8Gt55AHf9Mu/8Ll9+
-          qrMBhcDn/Md3/vQ+nL3ggm1yVV3mfOu6f3528YGfese/DvJ2KRCLQ4ZLe7dN4eMERuJvmUmXD+Kc
-          v78nZpOn9YJMCOHO3/01ULZoOcc5B39FrOH7Nf3QrR4KQ5yS5DPP5lsHvDi+fbjXB5EEsx0mNLQh
-          TCUu8OEZtTm9RaiFdvQ2sEFFd/e3Bwv+4UHhOD+6HSDKhN21J26w1oBUYRxCizlFONv1xb/5GCXO
-          3WeUozlsuz8KkSc1WP2eBzofzmqPOCnmyEsZODB0F9ODel59sG4wszuanauJ7y972vFBVsd8RAH8
-          KKGM/ZM6AOqkwww3s7b25xeG31Lnyl/+iWn316Ha/Rpka0mEzX3ejAp3KeEtE2oi9SuIRrl7JZDv
-          rAGfH2FJl2962kABLyv+42eTanYj/C2eT5K9P9ZQ1AR4fTQyxo9XEy2nWttgIx0Ecp4a65+f9G8f
-          JOU9q05//l7y7BH+y/8WqLREo7waRKnTTzTbzi8Fezzmefz8hvXiEQf88TNjwCz42yfAnc/7s1Tk
-          w+bcqgV0Z8bCN3ZV601Srwba91VYVtfaHcGzPsA/P0RtfA2w1wsdkTA+EVYI9x0WbRFmGBnT/R+f
-          W1j59wBRKKjYcHmpZs0UcZDjggonwbV2/+23Ps882PchaU3xcwnBPe114l8k2WWga2/gZAYZ0W+2
-          lNOfDzwwpJKKi9qYoz/+It7Ygs4HQYnUFUfSgna+8Y//rpIkWXB/HvKEg08X6NoLfJ3eNyJFagUW
-          r+Uk0HVp6q9/evlc3zVR5bcIO47o1vPO5wB0pnwWd/+m9w6tAM2n/px/u//HZsHcgyeUFKJLRwJm
-          /+RrYJSp4YtqMNM1OV4e4Pn4Knt95HQV4yWGf/HyBEWhq3uxJbBePg+sljeGjkneWeBvX6LfdOQu
-          HDAX+KdPjq+L4K7pr3qgU/I6z2D3/2lU8D6UhKnDCbq39eIlhINj/cPY2E7bQA6XSwZ3/YptOTDA
-          qEY/B77H+TczB1sYVuU0F1AcpJEYk8ZEy+jUMfzDA/czddHwt88s7lznw2628+1PH3NcWO3Xx9G4
-          +yfwstki8X5EcZmP76Wwe5xnrDKvCBCu/z3gjkfY42Wb8pu+JrC4Ut1vQdOpJFS6Au5+ry+EJUPX
-          bZYyJB+MmNizeXO7vswSWCgVwfbffKnmSkBbe7Cw+ZKkmtn1OPpKTT0DObip272vU+SDosKG3QHa
-          6WjT4IudZfzHJ2gzvhRAD/Xi55YxR5Nt5B1c7LXFEbSXmt5UksIrNWZiVV1EN7MMUiS+BrznF4Nu
-          989gCK7Q73c9Ou31D/NQ5zFmfbam4iOSxHRIX1hCsAG7/yJC+/WqiGbSSeXm+9zCw2vj5qP+ntU1
-          X10JYmG2iKM87IiOxrkB+z4UX07Rz52mXHyAT7dNxArLGKx28maQeYE+vt3dKiImpRU82ouBzat5
-          VH/6euqhIscDUXw8DVNtNgsS4Oxg1e3UfLGunQBf2anB+nPTotXAcgghPRVEs7uhHqWhceD/3L8G
-          2NtweAD/WOlYqVM9575lXMCOowf/FdIr7XKkPtCuz3Ekxpu7vS9RgsCxccmrsd/uiiNrgRfb1/7t
-          29bTTYuh/D4NPhA3CNbz93CAu1+FDW6phlmWbUNMo8sde3Hr5Lyy0gUZpaYT63fh6XR5oAQ80/eL
-          2BRe8y3jxwX+n79TAf/1f/8fThSw//uJgt4aFSxVgkpZSrwQXrljga3nSVP5TSQFeHnfJ3GEHoLl
-          ddgWGGzZYW78U0VXoP5mxM9pTLLOknJef0oteg/PmDhPcK75L9wYiE9fOld5IKsM4e8cGD6KT1Sq
-          a/n6jWOIJhlMRL9tN8DdxzmAOjeHfmRrNCKb+H1Al/kFOA/XedjM4qKcuNt7JJqeD2ATr2IA6Tlm
-          ibpsq7smDOBAZxWQ4F/+cOeL1XSwE7QTCU3DB9TOywzNq/vFpijKgEZ9x0GhkC4kk06DOgpfJ4Nn
-          KbXxNea/OXHEJQSXKHNm1h7VYaZnd4Md9X9EWiQzpxf5GMLX1PxIcX5qLtNdpxTY27PDV+Yauuxy
-          ayt4UHxC/FesUL7MjQzOZROTkD0GEZ+jpIMz/prYBNVR3ebYEiE7eiO5h9JvWA/Cd0HBtG+8juMw
-          EDk8iXCKAws/2qSgNBR1A1U/hxBF6ltALu1owF4M3gS77jVnnHIVUbTxMZFeHEMXeFI9GPAXH3sm
-          UiPeJnIKfdIpvjioCxj8U6ChzOcJdplqBOvSbQnyBvOIlRSd6bZmIYOUzi/8LUUToO3vwsDh7GGf
-          e14obcM4qlCoKibxcqXN+Ve6euj3+6Uk3O+PMavMQJeOn/zvBGi+eh56wLZKA/90ecsqI32yBzwW
-          pYjNZ5jW9CLzIfIeTx+H74QdFu1hOdBoY+zTlS8jKsutBu9+/yJazFUD00WBjxZ5W4j5+gzgr75g
-          W2UBSc/BL6dnWQlgzgc8ccTCrCdOPlvwEqUOyfbPSxDfGOj6fomVG/kNq1fIAVpvoU38X3bL5zUs
-          x1M0qFdyjoRvvjFmyyE2MFds/rwb5YbrW0PnN+xILngndcZKOKIi9jwid0fkjoUxHmBymVx8u6T3
-          nP+rx+JQH4nT0idgn/qjgj/4McgzvZB8i32uheKZL7ATwKO6+fPaQY52HbnZ8+SS2w1y8J0JMrm2
-          s6Ey1ehvUFYmfe+vJurqJu6RI70Nch2HD92Gn6PBLzYKIhV2l//YN+uhvMq+2A6lX02nB7/Bx9OK
-          yI0XQsq6ThnAsx7esTIz40CjN7YgJ74yrMe8GfHue0vQ11nfRH8GWGVCDTRw/hQNMeS4HHj5mHJo
-          FPWZnLHtAt46fDqYLw8dB185HTjzNFrg+5wwxrIuq7y0K7rh7GN8V4w+3/tdgv5jjPElPNo5U34g
-          h36GdcMx95VqVj93PbTR54a1MxMNPDDDDTqb+cHqIbBczmhVB523N5i3768DnNCDBzTfP0zs6tdE
-          yzC5Dny0AoMdS5spDY6VgkAdC9iYxLlehkm1YPTTKFbG9z3nU94bxZP2kn0Y8VXNGL3mIyEUqhkK
-          ZwS227nI4F++i3MrABpUJBGTC3GJOwplvby9gwgnONznXD2r6rq+JQ8mJuxnTNK83uJSyU6h9BuJ
-          wTnyQPn05iNFSkt8u9s1XTfLCf+ux1L0nFwu8n4ZtBV2wOrKSxFzbBQFFYf3cRbi/OxSeuczmMt5
-          joMUaBFnj08Dskcnw76RX91FJ28HPfX3x2+ecwh4L5M8dM6cHzHlQlTJInkFPNZ3lxipiunEx7MF
-          BduRZ8hxGpiKi5qBkv0K2PfSmHZnVxhh5rOEuEEBom8Psh7Oct3PSz5JLvO1DyX8vJgGm7erQDfp
-          xs8oNIWb/wXVy11unBaizyQo5MlGF8ooUPDg5WV65AyvEtjKN93QgJ2VyCteVDqXTIX+7u/1XjGd
-          35ncwdgkDT77vZqvj6SLoTPh2OcZVEXbamMJfodtmrkyqsDKx60Dq1k5Erc+XcDKZq4BDLuTcKjF
-          kcv4RrTB+tBZJKK6FjE/MfPgMXg+sf5QXHU718qCfs1q4dvwclX2Y6QCNOxemo+35umuufzaIAkt
-          ij3n7tabwpstAm55IjfEhgN7GlUBSo/gSkw3vw/TVJUpEq2jj61klnKW+5wgXPJ3QM6XT+wSm9ip
-          +IdPRtR/waKkaoqC03khryjv8i18nHswvY4ONm5x7G46S1KIUFyQqxiuw/R6bAE6L5KB46/3A0t/
-          ryywvJgzvjCLOSy/n58Cw20qon9KxWU1zYf/8EYaLsPAHhe2R4JBW+JwPlNPZ3R3QJdnPbY/+ljT
-          ebNi1Gn2ERtfP8pZU7q0SAzo2W8z1h02+yIf4PFYGn/xcrm8EwTIG20yA4/BKm/hU4nUui6I5vKU
-          rs5hTWCfYx5rQQDAmkzyiNrcbnD28N8q/yteENI5lnHUfop84fOoAnv+iXXU8nzTaOghM5KKv/rP
-          Of6WFdAA/ovcO5LScXoxGkqfhoQjo3Pd/fjfBvKDuTsunFrzkukqUBOPd+ImYeuuzKuToFGUAbn/
-          GMtdzbyZ0Xw8a+TaWsdoeOmBj4wMR8Quk4u7qLUYQjbQ1xmGG40+pQpjmKhahsMqb8By2k4hvOvP
-          EF+P8tn9h78rLhzitkCimyL9Yth8DAF7F72lVDlR8S+/s5Cl47B1fR+CMH31MxjDO9iOPrKgbhpX
-          UpilFnGK9EuQtQUpif7wsr63HuLXWCX3Q3urV0m9Q3juaEb8W53mNPuWBVLB7eHPUVbXy9IaEjpe
-          rxM2r50XUa/BFTig94yly1LWo6BcW1Q09IUVZ8ndv/wClpEGomuoc5c9ntDOtIQY5Jur/FW6aZDQ
-          KcTqvi1Y1OASIN5oEpL24UBX/hRCZHg2iz3NEHLa8VcRmjYfkLOyNO56PZ8YOByW2f8In4+7ZPcU
-          QvstXLE29m+VKkW5QOva2v7dUnt1dT80g09/3rDGxOeIhm4K4edCzfl4DEWwMYy+obOnu+Tpa3Cg
-          dHB9gAkbYgnD60BugTKi5cWdCSbYAPzhPI7Qu3YP4lbDGWyfY/0An2Zmfep9HLDw0S+FL6cH/nU9
-          2XRbaLBA2ssa0ZoTV4/jd/bgnl8sXZw7WLe8Dv/xBWsoMWCz7DnD9e4N2GfpyyXtqkKor4xC9Htb
-          gbVvpBadwurlC3xG1ClV+AMEbnXyV+CRYa4Iz0C+Dj/zSZdzdfMdjYGpHigkMrrBJRZeKzgZXu/P
-          ZWWqfP9de/iabeiX3SHNlyv4+ZC9h4d5pNdzTnn2LogmuBMSJVUf0UukGsjeXt3MPsmt3k5cH4N8
-          qhK/vS8w+sMrMUyfPQlu5infwC8P4HUI993fb8h/30+5QFFxHvOq84b6K7ufBf/4spoetoH8ihuE
-          4angZmgiNd8UHjdg5+cz3IJCXdiubwHLKAPBJAV1d3oFIcqztsP6evqB0fBPBoQUyjhZkgQQ/Wm1
-          kD1aGb7M1AT88Z63UMwFSGzls1GKKXBgZkkIv1bPqLckpAk8d2tGQiFPB1qAwwNGvHMh+qYcht/l
-          plkwPD04fJ4tX+VaN0wR5SzpX/0uzPIYTwcIc2y+Pi6dovktQD+7nvAff98a3XpAXSPiDG7HrW4y
-          fAyAkZ0jIjVsvdef6sMTbpZ//JRPflYHbdPqsLvhUz72994CW/i1sZcGfT0j6WTAINrOxL2eEzB+
-          uWdx+sl8RpRCPOeU3o8pVN8txMaAfLBV8TjCIlcO2OscGzCLIwriGF4WIs/UpDsfsJDO3Hjidc6P
-          0neLQsjE20r0P37ZPPoKOKc42Pm1C9hkkmfQCcaJyGqzDQv88imEX17HssvowwI+lQ+jVP/+my/j
-          +G19uAj49Ze/YYkG1P7hqc+k5QxokLE+FIy1JfHxtQD6bK0AqpKyb6iHaNgAhCEMyXwlO18d6A3+
-          OthCa8G3j6q6/C8OevTm28BnwOEejTHfPNCDgSG5Ou6orrL+tUBbHWvsCbSptwg8K6g/aoGYZ9wM
-          G9XvPUr5O/Ovf3Z9UkA5F0MiaeumTmpzVGA76yGWo2IF284vAdRiDV+OD2Vgr++Qge/XciGy9Kvc
-          RTyUFtR0E82Vacy0s2YLwi48P33hcmzoKh4bAf5UCrH0138KTSpx1xP/vv9d3yEHxeMdzcdhxtHW
-          tlMFvZP7JEYlkWi6SdUCi0XMsXSMX8Oa80EG71919bdLJA/sb9jGv/nrH3a+N/8+egx2/CfJclZU
-          Lpn9Appoc7A0vjCgYygUEH2qEhtH5UIX8U4TqI7zgDWebeoprKiE/MoV/FMbBQNNXu0BNqXfEeUS
-          vYfldwYhuJ9Sj8iPpzSsQDQP8Jd/RaK/j0693od3hc6Z9SNRd7fzdeMOKbyRo0CcB1fk+3w+AFqy
-          PDa3fgYU0NOMVBQd5v4UqDXbacIMR/iw8Qt95pzu8xlE6vmGXwIXgnnQriFCx4vkb1shgI29SCUS
-          P8MT20/RAvOnCRJ0u9wsgnu5AqvaHCXwM5ybL37vzUCLsxOD6UcG4hy+j3z7XZpE3PWwP32E3F2J
-          LCQidB0O28ehpfQjcAbskjKbmXKRAUUyYkAFtSfJBJfJl218HQDiTIrdLM/oIq9yAgWWC/7xqUUL
-          2we6tpn7D0/59MBxyNviwQctKCmdrnr/xw+xbOqOu5zdZYT9Pbpi9YPe6vyZbz3c1rOILYd57yeG
-          jwHc9bp/NJVyWFB5DuER2QNRj4WnkmC0JcjNy42kdKnoH97B09sNsJR1C9j1iQD++KX8Sle1n6xM
-          +dMb/jIssbrERIyBSGeAla9TRsvvTEOkn4GJXcG7q0sx3Dfg3xiNOLVwVpltvB3g5Goi9opxq+eX
-          eBohd17dv36uF/lZxmg6yw//+/N4unztQwXEM1uQcw3yaGncZIRF4sxYH2Cvro8ie8Cw53TspYSq
-          lB5GH76HV0yMcnNz9sPqEmD7wCFeLSbDelTkCkSNX+PCwqW6vTO7A6M8P/1VbbZ63fm1WMcv4jPG
-          D9PVzMcRfk+Mi6/kKQ/8InkPqPkXB59fWzv806NxNy5YupmnaNvkWICscQ0IjnIr50ZGOEBa8jz+
-          q5+hOP12/meuM8qaatic6OUA+rEYfLHfWz5/RqVAh05zifr09Xq906iDd9bE2M4LaVhQGjnQWVz1
-          j//t9cVyaEhPoS+29EkndH+J4Pgufli/JJiuw4lbENuHDvY/l6pe+JkpISN2Z3IVPh91GwPNQl5z
-          OfszTPR6sc+p+C//ODxrEbOujAjUELozPdwhmOQ+P5wQSgosM8t3WJgHnSEK+sAvkmucr7PdP0DY
-          vi2cWzqt93qHkNxUdWakcYl2vG7AFIeWz6UqBpQ9mSHY/Sb8er4hWLr25KFJPk1EC8yv+2MzVYOf
-          SVSwpHPvaGrEHwRTJZ7+8IaSR3ZPAQkdip2XvdbEJnIG72kuzXfFcHJ+adYQWig19+dp8gVpmwMZ
-          VwmwVV69iHknQQtTxwr9g1SwQ3d6TwV8/ooXsdqViabEigWUJIqFXd1RAPta8vDPf9nnSQq2ujvP
-          cLvnYD6+Vwz4MVwKFG+vC1YeX9FdxUiH8KZ3OrYeTEy/OO0keDPbrz9WZ8f985PgM31GxNG7tl78
-          QWqQqSEJaz2XuFQOVwFIt0zf51VH11NaCYjmWeSPcr+4K1isHrq+V5JoshpKrkIjIFH4uDg9TdWw
-          0a4QIQ3CZj78Mj6ancfCIAaLJfFnO6SU4TYDgtsZYgvaDuUe8CJATigF7LMxV89oCiS48zu/2p9n
-          CopDAP3lnJB41MScrEIWwJf3efpH9Jl3PW0qf/MWp78zrLefLXqQmevLLA5qQJnmUVVI4PKayH4w
-          0bH8MAy4f+V1RukFR9vpgipYWkmCNels0nlr7RmmT00it/XRAzpdz92/+Hc7/lNVrxi0JaJB/GeZ
-          uovDZCKEBnvCStucBvLHV5P0jfxjB6NhbadMAL69tFijn3WgrfZRoNanxt/9DvPnOBQQaXVMtCDI
-          ASX6L4UfemvJHx5TR+kfol0cWqwe7pCuSycmsHz6LdH8CLrbd4szIL2tluCjINfsuSwe4FYl8R8f
-          pUvDsz08vh8/onHfsqaVWISQco40q/jWDtMtZ0XotsoLS6XRuetBIAvAjRWQyA/5aGliIYMcJhMx
-          AmF0+1tR+TD7KbZ/Wo4DWMMKSHD3K8hf/Ec1uPxn3vTH0R0Y+Fy0f/NMdeyqZu/DrwSWWQ5EK2BC
-          V+w+Q9A6zQXHq6C7/HbrfTFTrYnE1XOO5sIplH/+VxqIbD02sZBCFzFPkv39fteuHuRu9ejfpCBx
-          yb0wFrjj2dzKz2u0Tnd/hEoN3sSmqzl8+2TiQDubIZYfz3Kg39SQ4DdUzliLOaXenpbEQalVuJnZ
-          9dKW2W8FCK5R++TWIHVdmbcHwfmZEzM4LWDiPSjBd9NVez8SOubW2QGXzIMYs1ewv6GX+HDvH+zZ
-          ouou/AxLIKNiJeolNqJOeaH9vZIrnrmjcgF0kpv+7359aIuqylhyJcFcvud49yMALQD3AMz6lEgx
-          PtqIdtZvgTuf/Kcnh/zHdoANHx72wXKo//Q5lE1FwZ7umjUd0JeBxvWmYmPNW3UUD6WDHtPUYiV5
-          6zVv7BaEfaV3/zZOOGdOv1sLF2OQiH6UJ3duzGsHeza94Ce2BzDdhKVF/Jqo2Inwlk+WCBsQR0yJ
-          w15f1YUcXQjKp9cSTUsew2pTpYDe5frCweq7ERUTXoHNRxN88HKSqJu07gB3/8OH4+mTTwUr+PDP
-          v5Sf+pNS99c8AL2ZCfa++ewO5K5k4M+P0SXfG+ifHzM/+w82E5MBdDDlFsG3FJF7JtzBuvNZ0Kjl
-          nTxvFR1GB+oKSjIhxY9d71HjawrA1I4SsZ4RiugMcQ+rBc5YEoRbvb0txIFEIkd/vax6Th+23sAF
-          lo/dn2vyzWl6CH+LoJEdP2sWXT4bXAsNkLv/ClXix3XwT79j3WfpZNQFJ455gbBeukcwyMeAQe9n
-          zvlcGSn7fuWsgffhrsx05aWcTRUeQnQe38S82zUY9/n2zw/l5Fiq1w4UENwvjY9fXNbXc9t+KrQe
-          NROHTKjS7TfyBRTpCGaBNXPAzd+fAM3PwcN/+7VVqd4c7GuV+/NzczZ3TgXc/Vhf/LEe3RTyLOCM
-          PybOdacCmz50BvzJofnPj99E5RLC20FOZoaJp5zK81kCL/7mEBPc+3xhrnYDVXQ54L95xbNP6MCC
-          J+fdv2XocnqlAdz3MzMvim+wvX6TyKdgXEnBf22wwJPrAUmys/0EkQuW8sNw4lZ2X2y3292dgIgP
-          YPgJBXa26enSwHz4IIeu6dNBLIfRvqqS+DcvlX2fQ8sj8GF46QX/K3zner4Oa4hA8fj6+zmxaNcP
-          EHZlCPztOdoRcZSqgIm/Pvbfk4YtLp0MKJMx+7u/EK2ubDQwXb4KMYcDHdadD8JtHjUsn161u3gz
-          grBsj+a8pKDJF0G5NuJ1sUYcBwGg9NQ5HLj6jYe9XDFy/k9f3lkdY5NBVb46h1MCs9PnMQs2WcHe
-          PxIUTWX82yfklWAYHVy+nUeKte+jfV8nAHz60JkZgJ6PwcnrYLhEv/kkwiGagoILIbHQh/j+YLn8
-          3z4r9Ik9g3D165HPoxL1Tt/s/R1G3EH1Q3DkkpMvpEPpzp2OCsji6oldKf9G9KdqHLzIMpkVcR7A
-          vIbd/gYY4+Aoxry61k8SgP3/+5STWHWd+rD884/nW5x10SIaYYy6AaGZ2fF+5T1GgnPZxj4QDwzY
-          MvfgAzYsvPkSjIlKpsdx+ff7r7J2wfL0XxqsXplPdj1JCboWDCRQeJGrZ7lgsl7vQvzT02bkNeoP
-          u9dQ+L1EHp8/pI2oIFscVPvFxm4uy8OyJr8NyrkQEvvHdCo/nbYC6ueT6Yvd8qxXYe0YWOpHQAwx
-          SnY9AgLoSYuHs3T6DVN48J1//uP1VVf5YiS1hwrLDLDfHojaSdKnP106dppP8WOlvUYzH9bDnBFl
-          VALw2fsdJYlk4Zvw9et535+cqFGp2JkSLWL2/RGwFX4g1jgz6pD7ffXnT+Lz7ieuh7krgIeKDBvX
-          rznwS35r4b30jth/xRVdzuLI/NuH3G/IdvlglCUhS4ycnG9VNGyb/BDgEuRvbAdAjLbAl1uYTcJ9
-          5ne/Y6s+SwW/KiQkVo71sKL7TQQtfryx8k6uA1+c3gFKO6r6kzXVYKR2OMPdr5uh+8X1cn7dx/+f
-          EwXc/36i4DDb3/1M2Dda6VXzYN21xD9dRaFerdor4JK6E1Hax9cdS785wOQT+8TKqqu6kQn3UOpi
-          cz6Ozw1sAyYFvIpRQ3BT+TXbgXQD4HHNCVZAoPIXxjNgRMsjsc7fSV3MYmsQ4tYG28bRAlxm/zh4
-          CB3RB/d8HUb7Wfeguntn/LpwR7D9jlUPpPgYkDO3icP4PJQOlK1hmQUScfn0ZFwOskHG+SiTe3dK
-          bC2Duf/OyAWnbzAywcGBJ/ZDMUYRAWuz3BR0HkBM3FtuU6Ya7QRGpCQ+d+NtMMkG8EGL1Q/xF0mq
-          ucJUD+LjCaD/0zQeLB/73aJXranEr7AzMMNN62DEexUO4XnMN1h7M3zCwSBadKnURY6QAqOHKxBj
-          nhBdCA01UFbrBT9cwaOUeVw92KhQJxlHmnxR5SxFwgETYnftN/9J7RogLi8eGJuA7tc/feg92ILY
-          gRiDtQipgryuLonsVjfA+sfrjOa3eZ9P9P2t176XBJiFzAM75DTTxcttCENrNYk5S1z+9z2a2DQj
-          WqOoA9sPUY/u6tLikIaaOgF/FVBnP1ZyVq78sK32PQFKPrFYuuuwXkI+9+GYHKG/0Rt2mUnrJcha
-          6Y0kh/vbZawZVIg/lG98d8X3QMNtjRHrZe1+wmCIPsnJTSEYRhXb8KZRDr7XAzqezj5WircC+HX6
-          BXD0j0+igl6sx/f00IC0xC1JdPmTc1FRJehyMqmvvNrLMLmvQEK5473I5SRa0extgQZM9SCTM5py
-          dwsPXQqBNQjErS+qylcWY8HGiiSsLj3JqQqXB3JtSybJqaY5OVl1DyPqR/PBruOcgafUQKqyBDh8
-          vY2ax8l3hM33eCFyFl3qzY1KiEKekYgcPoWcpmWRQO3HNDjZug2w9KM7SFOsjSQ3/ge49RQkSBmM
-          J7nwR3egny8xYA9CCSef7UpZW+pE4BjaieSjWeaz4p4NOC/LRBLdCKJFoQcRXJXQxv6q42gZdWaD
-          UtBfiW/XTLRMV3ZE1jZaxI1Lga7nV8Sg8mVO2Fa0RV0/aCtQ3DYmSZs0UTmxtgT4Tp7rv3rY6D3R
-          4M2+Zf5qfI/qhpbSQfJD+pInDo711vuBgFzbkQleU3fgiFOG6PBhnj64lQygf/HRPz6P5SSqcuby
-          4xdg3SRjnkJDp6wQzjF01gfGd3hr6GC4fQVQtnHYz+I2ny3DnlEPAgmn5+9ZZQzJLWARK2fsA8kF
-          fC0cINzrFz+fMjssg74ckD1PG1HureMy12Zp4PvHdPgMJhot59/ioJ5+XZ++Tme6uQ3sYaa+mJkr
-          eDaa8y6AqNKr3n9TLQRrHVwduNzKq4+YLVIZ+zn0ohlHPXGt/jWsYPEryN/cnEgXSclXGq0PUHeh
-          io1hQ/kYt24MVrHJiH+Q+Hzc+x/G6hvOB6PuAPWS0oB3s7Bn1v7JgPVwKEKcDTn2Z8cHU3JSM7Qe
-          wHVmt2syMO+HlAIte0EfbZMUcd0CC/C0Du0Mb3PrEtSsBto/40vrs3R5vNZ93jXdjFr+606qcZTA
-          +8d1PpIC3eUgPXdgU5oTyeMypcs+f1G2+fKM0ATUtZQHSwytIp9jUejUZWR+CgTNTcTKmQVgWer4
-          ACdEGmJ6uHKJf707sEivFnHi82Pf0MQMLJUlJoVUy/l6escVYgSi+keT6+jPnx0FkuZ19vmX3tb8
-          Or0D+Fx+CbH3+bX3r4G80k9IASSXcvReGNCF3JvgP3wIsJDAoY4e2DysVr69bp4Anet2JrpCveiH
-          5zMDaq+Vsb3Pn5V8awXd7spMfKPu6MDnRghnObfmOjx8AL9NHwM9dFCQ87jVYB3TkwRrGIQ4FxSt
-          ZkbTLBGXPx44lceOrtef8oDu9ZXPJ1+c3OlZCY7I5dTCOudKw/jAi4G+xv1HbrHT5mvg4wW0nCGR
-          cwn7mqre+kD2wciI9Y1Md9X0MYDzORWw7xbyMD/ubSOGjfX0ef9+HChOyAjx/OvJXT5+3KUa5QTx
-          2IFYYUJLpZX5TtHrq4XklY+XaGuiOEBoLu5+10Z1vr2uogPtIrNxPppSzimurqG6C1Ry7uAJ0Hjt
-          U2hIwslvYV0A9i9fxxP2/TKR2h2/xQaWVfHBuKpaSutbFSMLxxFOf9EMeNBKChrnmzIj19Vz/hgD
-          CPf5i+/MQ/rDlw5lPuNh8+44KlOiu4fS45KQ3PmMlLzcJwc3v/0RfFljsG2nhwXBcQ5xvCpJzb/c
-          K4O26sthbTSfgLYfuUFWfpVJtp6SYZ1vvfbXrziae3NgL+AXQpTrFcl+hzDa6yVGrasmf/Mgp8UQ
-          CeiduDrxz8UvYoqi1+BjChvi9p468NrBlKC7cCtxX/ndZR+araCFlJQE67oBOsqZCK9vmJGrNws5
-          jSrogQlNjc89t7betG4/c7yNFn7QBoHl2wUz+irajKOlJxG9E638+0xctfci7vgKApQX3oxV9mO5
-          Cy1BCEuuYPZ5cxzWSdhEZMTNAzsrMwwjj6wR1vpZIK/mWdfLobN89GAjlfhfp44GCUcWmi/ag0Se
-          daSrKx8LcPhwT//ASZ1LHrDekC8ih3gqsmoi4dz6yweO+/0ETdWMFtz5mi/UT+SSn24mCMTSBQd3
-          /THQxsgK6OfhibjTRXSJ+A036F4tQly9j1XaskqJ+J47ECPlz4Cd+NJCcLhfiP55nWp6MgPhD5/+
-          4pGv15/zL17kfhZ/+cZ0WggH85vNy7gSd7ncLgnojyEmrlPKO184QRi8U5OYDOeCzfiGJVSzbfaZ
-          yrvQrWzKGDwWaJFr/VrBdjkCB3XndSSv7FsN3EXWD/B8DNydHwB3BnFRwuCUfsht5cp8i6eghH/x
-          1IOmyjcugSmA0fTy+Riv+XIeHBF8L4uCY0aQBvpp4wS8IpLPqIAV2By1k2DjfcFfvbubyVkzwtvV
-          x/IjOuV09dEBvtws8OHzVwHifekCkb9Z86la2GGv7xYW/Dz5TOYpA32ObQ+TK2qJPb1LdZF0n4Fq
-          pJ+IdxaYenuakvMvHva3VaM1MdcezudMIObDYesty1ApsotpYll7vgc6rHcRlMZWE89CAFA2dx7w
-          w8smCQYuUZezd2th3TWEyDsebkvfWwAI8vOvHms6vFMIk1/8JI/j1XNXUJ4bCGLlMrPh/KLj+2Gl
-          ojEGPpF6bsvpY6wY5FyX81yc9GKY0Lk6gMPKukS1z8doQPFiITbQnHkRRhpNr/fPhz39uMR2tFfU
-          m2ZcILWlD//kHwV3xa9bLwYrWTDGXOqSP75V0LHAjwu1B7aHJAEVwnd/PYqE0ncn+vBpwfYfv1r8
-          XF+AUPgbOedGSuc0O40navcBkQu/rUmv/Bh4uRpfrPfZFtGlABr8ToVKzrAb8+lUygvqPWzO2711
-          1D9+DfA89D7/k7ZhwWkX/MuPsfO/VZDDDjy741+8c7DJPyGDS84GRPk81IjN2XCB1eWdEjf6HEG/
-          iIMG/+aBZYxePe36CsqplJJ88Myc535CD3e+vuPrOyddLXbwctYsrIDbIZo2o5BguNgu1vLLJWfy
-          8NrAzi5WYg5ArXlLDBv4FGKXXErfjyhvby06LdAj92q51gy6hxJsXodtZmakAuZ1qA/QDC450fHv
-          l29seyrFedLtedPZsV6JdRxFP4pSYmY4pVuWsRU8GJZHAvDVci7v0gMoi1XGBunMfaPOidDCSUTs
-          4PgCG7WfFqwEoSDP0NDBUvhvB+38dj5szsFds2sgoMC5jviq0xvgamYyoPLVfCKfq2O9Fo0/QsmD
-          K4mf3qwucaJkUFOeBZbvqUC3sXj20GCCiUSmNdbzWbkbqEilB7bcrx4tS/MdxcCQVhKqrAboITFm
-          uOPzrlc7sHi5fIAtp0k4S63LMIeYcSAYZnVm+YDkRB+6Ft70m4vdIunzramvHrT4KyHKrqf+6gfs
-          +u6v/6MlSr+i+C30HCtxa+TbzRE2aB/uBvHKnx/Rqzyk8CbMR18sVZaS7Tx4gHutN+wfgTYwSyiH
-          UJsYivf6yRnvEY/g9rz9sCXaLvjjY2C+GA+inJeAMuLTL6Bul6x/4h6duj7ZogV54c/EbPQTpaGm
-          awDPyQFbO78ZGr7x0DkSOf8IfdllB2vRoNPqnL86z5+7ouLQQ309f4h8jyYw/fUb/SQhOd9lIydF
-          Zvn/6vPOvSN14td7D+izf+0vZf9y+lWFBUa38oKTkn+6aw8eLXI7Lvzjz9HyyVAIVc7S8KW063zQ
-          wbsEHxzfsH5tWncyaqlH9RJUxF7e7vBPv12u2heHzqL/6VsRHQqnJ1aZGruf0PdwM06MDy/0N9Cn
-          mykw9bKKOKjXAPewQhFlCTdg729+TBluoVF+HOy2yb1e25fpwJPcfv3TY6jVUTjbI9AFMSOmf3/V
-          63NbDTjf7+2OZzhaLr/jBn66vcyMnQcDx3+qBl1kLcV/+o3b+wN9voqz47lbT+H+zovACh+sKZ9c
-          XR5YMKAoJHjv3/W/STmTtWVhJApfkAuZJGHJPEMUEHEHiAzKhwIJkKvvB/9e9q73DlGTqnPeUybe
-          4uqbAI66iMiaqxSreBAEsNdXpAWH1sNpaX/h7aOMeOcZdD2PQwKu1V9Lgsh805nptvz3eVGoD8W4
-          9vvNaXzFnZEhC0dtyrSuh25u2KHUw80jc5NsMJtMinxdaLpPBHlOqlt6QXu/oEv3qA7/6qkx5qm2
-          +7ED9PU6Qb5wrenWMJ8IHN/dSIJWucdsDN66FPNdiJlZqrVf/Ya5yPNh90f+PBoa5xDs/RAFytmm
-          TBE9XrDzhRkpm0609bkxopQxRkSQDm6AvxhPDjiYbFgoWwjw4+aYPx6FjPBjdJxjGi2AR5PBp8+Y
-          ajNDQQr4m1OEK3ZEry6fqyyh8DsjO24/I61y24efOsyRy5J6JDuvgILZyeFcQ7fjq7EQICPMGpYS
-          pR0p74g9lIfUCk/3pR2XWrqEv3odLjf5A1a7OU0QcbcH8i6FWqw7n5IGgUvD01/fxXNh9jV80mL6
-          1euCPCb8hedUw0iX6DBuLLdV0mXa7sh5jVuMyciI4ovfExyrenrLft4l170zyLkvasc8lGCCovAF
-          xKifLF3utM8kGJMn0asLpLTwhwP46UevSt2YxVZjQ1YmbbjO+lhQ63sZoFLLfBjs+p1Z4kSV7r61
-          hO/bR+j+8QUtr9QQJpTVJgfzCRw+T4xX5ki6+Si9Fzh9E0p8Uy2K4XnkMO/LrYGQDniwxmDW4f1Z
-          /ZEwrAJvuVjnDJbBK0aWfLcplcndh2bSlyicXO6//TM5vN4EpWs2bt/qzsELswnI/R5Sj935F7RA
-          oqO816SC5pKI4TdDLDJTvvcW7tj7oBXEChk6/ngf/TqHcI60LuRMk/FmfntlEmU/AHmb9Qf+8ayf
-          X5U9dxk3SQsH0ToeRyJLbePxGz5y8EsBHwZ0I978e72OidbwkN4vxYJZ/gW5bv1DwZvXCiYAHgPI
-          y27Iv/Ow8xO49xfkhsF7pMIHVqJKigxZldFSbPPPM7xr2w5kjWokQ3zC8PWEGymCt1AQP93vGDhB
-          AZ1bEHoLfdYt9B0HEHWL6njtA1r+40v3qyiMw0e4VlBohBT9eM0/P7up5wBdee9VTLFV+qLj4v53
-          XgG5HZ0BovTjk0sTRt63o1EKtMMjJMol+3QLdDtO2v0+0nn1oC1DVrzgGgcJUs9LGXN26n5/6yO5
-          E00Ah1ssiBFgEHKfdgkWlNXRP/1rafcXmNs6j2B5iqN9wu8NNrdHGSzeYoPQeV09+hRd/Z9/QjY3
-          gVWSmgMsHoJK3ItteuvdbPGP/+LTa4yKjb82A1RvmUiUA9t4eC4WVxKdOEU7v+yGAHgc1NVnRVwU
-          OAVdHkMKDhcuJFbi9jEeDwWGUmG1v/PX0dv5mokxYQsiv7lwZNx56uH3KX7xIzadbt33D9z9BJLl
-          BMQLulMVHBkhJsZ6/qOEA3oqHUzXxxwZ/uKxcUMRfuTaQRU2NIrXU5bCbNIpycKpoKtNgxxmcrYR
-          c388PcZNCPnC+yIH9yGgaTOfYXbWTVJ9XM/DzOCfT85Bz8m1zTvv179h8/WSff91MRuLUQp+/s6L
-          30+w8uqci6N4DJGa9zPYjiuH4bHpSuKlhalxjcGqP77/45WU2MVf9M9vKLWjFVt0qDMocfQVbmw0
-          gvH+EBi48yh8mI1ao2PbTfB0jVxi7vxgC6Jc5j8ywyJz19s0UQ+7n5+ikII3GqfPzebgojFbuPRA
-          iik7VS7czLeNrASt8do0swx2Pkvk4x8dsWWTUIyWB4+0n37jOe4MuJ3das/TTDFn1BjYxU1Bwbz9
-          dft+gTDROrj3C55+fv37+cyfxJDMRsOuNqjAZalL7L9s09Y64jjgXdwaA94aR/rqsxYuHhOT9FgR
-          upa6IkuAUUJi7Odn94v7xH2NkEwmWWNzbm4hgfYxFJ6ZCnY/ZkuSZujIM8hT25zXwkm3ixbs/cGh
-          /Mkev//0hfV0LyO3uX8i+Pn5wDz74/Lrx7s/QW5z6LSd36nw6WVnvHLHP7A1hiTDpBISdIsTIR5L
-          iat/eQoJrPMSf45Bv0D2pd9R1sddvJ7WawkZ44pw7SEwYj/hW4Cn7YGPQuxri1E7LRzbw/LPby3m
-          YflCdtJ5lK7dy5vqIPLh08oVzKYfBqwX0Jwl4q858qJ61QhbuBXkK+ZMHnv/Xvd8BkjfyiVBYLbe
-          No8nE7JJYez69dVNAEgvuOuPf/pk49/fHj7KwUEpgka380sZKvZnwWzxCrplYsECDCh+sLDzrNWm
-          RiZN3K0J8St5xGveTz2kH+lErt/rXzzsehfI9qFFho+F+CPKHgb6Wt8wT2RRG7b4yUC3t7hdX1/p
-          9tMPqxX2IZTyA1jFwyJA6kC45yVpPBmLW8LWYNNwXY51/K9/vSyY/PRmPJ6ssyj5MNBIqA+gm/sn
-          cv+tR43HoOBjBglgkT2EXY7oBTfIUgXvne2h9A7LcRk+JAKaZ+TIPFS+Nt7ksoJ87PhIk+2moMkm
-          uT89GMLms2nrYX4xYO8/eOfT3mbYXQb/muVIkofVedNkoRpM2y0K1ygLYgYB9BXvRZmhX32iyz7h
-          +cvPKPN+FQtdGF/ceTHJjvpr3JTPkoHyut8xtecZ80lJKyhPo7rX54u3PJQAi5hZIArEGXXboZEW
-          0QKpjozl/NJmIcKpmPQiDfkrYr3tkRUcEIUEIatx7wXNA4DBwbULfDS5t7YJTLbAnUdjDs8POhVt
-          2sLnbf7iwyLLI0O0WIf+0NTIF3Ma7/oT/tPvyj1sweKWZwGQ2o2QiyKm2HThm8FV7HOiNYzbbXcX
-          9lCYBRNdf3rGPAiDqMXWCVmn9TWuqjGUUJUVBqm3Xi34sckOcOdH4SL82cVG3PosGU04hKfdzyyY
-          Pb6gX7Yb0RYixkPwEWwIrWnFgFfrcV6sCUI1zkciewh0S3g6mTAgoCOKnxmAdc8fBuznh/i+sNDV
-          btZJ+vETxMZ1tw6wq6F6NCfyyxPwFCgHmCWFgmRkvDsqy68DnL6qhbz+grtpk085kJPhhdKU77U5
-          gkcGfsqaQ+7rEHbcPK66NIfXIzFdxfGoUvcMrDVLJlacZMU684MNX2BgkHH5XLxtDEsMeZdTkHLJ
-          nO7Hv2AtUoI09ku98XZ+ZPDClD6xfudp11vSxh5mYrJDVWzHfWJFwuWd2OEEwOofLwJwRN0LT0Ij
-          UqxldSs1EscjvbO+44qtjw2K00v/5Ukdd36ukfTLT/c8Wls09W1D/9VdSJgnZrH5CV/DHNkCuv14
-          wy+PrnF1JVpzunYTdDtGuqpnB6X1tnnTzjPEhW8iFMReMuL6DUyYE4/F886DJ5mCnR/zyo/neDPD
-          mjqMfJri+exl3S//hanz9YllVUcNP4Zk+eVb4ZGN63HR1NkGO+/DXc/gYvVRLsJocKfw6IJnvEWN
-          xcA1FxzkhGLgMcvrD8M93yH7fi7Wq1jZcCn4M5IzOaDsQ/cZSAR7IbpS+zvPLyNoa0lHYvimHTmY
-          wfZvvW0xliM9Ce0XmMJ7INYErI4ctTqSKjKF5NxynbadkmgC/8dEAf+/JwrU+58THldD1NbV70t4
-          srY1ZNj8SzfIhCb8y7SRyP138SYSN6Zkj88Ysxyr0gWEWyu5UDoTtfsQQL/5y4SvekhJGQy9thz8
-          ToVbkq4hlHqn44dN3uDgEYSBbITxxim6LLFpOmM62WfA6Po5gpnVY2Rd5whQcY10qOoxh+z2OAEa
-          D9kXaqfmTNzzpy/o9Z4zUOmPJQkcP6D05FkuzFqaI/v27YuhT2UbergxyaMp53i9XxhXYv5gGfJH
-          5+PR66ONpGl6F0QpJDvmVXkS4WyYCbLxe/XWd2D0QO/5Lwl9A2qruZBI3J9P9NNojwuLVBOuL/FD
-          FLemgOJ19aWQPDzkZbxTMB/vvkG1XDBJUlPtmFiKXNjjw0ysIn7HdE4CG3Dk9EHJCfbFegI2hkcV
-          icSc5APAiyOW0rbNKnlYfAaW+5KcJUVKezytxOrYI1Y5Ccy1S6L1phas0QxfeJkchtgw7kbuqjep
-          dJPJO1wRO3SjrU8bTBQQIoXCr7bclUiXPhMJkBH5BuBtaDBwUwMNyWqHx0VczVxyk4OL/O2ieTyu
-          aSkVD8Yk1mLII5v8ZSIAc+tifi44uoLCcGE+JxaJWi4duT//UkvxR7SIPB5fgC+/jirZ2QUj+wFY
-          sPX4nUlCGgLiFh6Nt3P9ZmDmMW54ZIU4Zi/v80HS/YlF2X4H8PqG1gLWwQyJ0ki5h/VSasX759th
-          xpjfHgv+QAWn6+Bj8PiqIxf++Qs8p9yeYIaiNomrmcENKQ9iHQQvnq4NlmFV9D2Rc4q1tb9O1ckd
-          CIMUkbDe+vg4Z9hT3yUXf1C1rVLdCMjnl0EuxzYs2Mc6uFKohR/knkVd42uAXtCbioy4j7jpaPle
-          XxLRsgnZ8t+x+NRXqMLncXPQrWBlIPDfowr1tbyQqoAPwDcPzZXKIArJNcPUW54QqDBPoYTKg21o
-          TB3ZETzx/onc7anR6IFvKglX45nImvHV6E052GKdvEek8cdopH4cq/BVtX8k/NgJXcN8s6XGuZlh
-          HZP3njBcE0n4ZDXKUn7WtsN66KG1nAdyB/yocWI5vKBdb3/IO4p0pCM5nSFcKYM8+c+hnBKlWGJ6
-          4U2ypVQ65lUdMygeTJUgWTuOW+V+XMh7LwvZ7OcO6PW8J4r9LISr4T09DsqwBJ2ymcTGCY2XUdBL
-          yAtCg6IpHostJ3777/1U8SV2szz95RK3KRlyrMGhXA42Tno9qgWZt+Xo0SKhjJTcE4TOjS6NNIuA
-          COc2w5hv1HvBzcC3AaxkJ1yYTO0YXvb3/7yxGXLnook3plxc+ATtjE9fJ6OrPJEMtsjNSNDMnreg
-          R+/CQDpdkSacl27RZZmDZHruN7WwZbE156WV5O4iEGuemHEBqu7DzP6+kXZ6G8U3c4UJcoxjhtQK
-          WYC/ripLXC5fyH4eKOFJ7kL+eVvDg6HQ4lfvgLtPAGWVTzusGWdVimzRISH2PcDNWmdC85w0Ie9a
-          YcdUXZqD0MxCzJ8YJqavMEhhlbcEJY/JpMvfO9zg+1Bi5Ey07T6/erNg5UlMu8gAy0YSB5w2fRDv
-          Jgt0Gji9lVhFyZFWyqs31+03AYfr8YYcpS4orVK6QcpXDyRLycPjs+c3hLpmykQudA9Qb32W4FXm
-          p2CrSyfekL4IUBmUmISdrXnURLEr/dZnVJzi0fEvyiV1Zt5ITZJPx2htOYkJfeh7PdO6IXblVrLJ
-          RPb96WmsosAvPOUwJbff9z1weg2vQvhEevESxqnKjzXc+wsKL2PYLeOYuuIAgxDpWnL31kF9hvBX
-          n7x2nwNgknsE47iV9t/rXTCfePlKMOdeBAU1o2H9XWbwVMoBun5nTWNL9BokfTRMdBdsR1uud8uF
-          5PB3I9ozuI305bACuP/VLLoc5Y2uSSWq8KIqMnm0qO8GF3w2YBCmIdFtC8Cm5NILrucxCD9Vx4zf
-          ryH24v2L8xCzTRTTA1gE+Az9GJn52I+rg58LdE6XiVzcdR1XyDSbdECyheKDJVMa2tMAE7bfiOVm
-          U7z9nWsoBQmX4bdymcDGBWIKn4Fdonschh6D8ChAK4LhfApuEliF5qVL79fpSLSGbwrOEJte6tL3
-          HzHkinajWTqV2PwHAAD//6Rdy7aqPLN9IBpykyRNBETuQUDEnqAiIIJAAuTpz2Dtr/n3TnMN11BI
-          6jLnrKRKXgIcD+3FXdxPWaJ3UQ74lkSfnDdHViFDhVGQbPuxPJo9D8NzfcAePT7AXJ3OMvLVBmJ9
-          2R10wdldStSYB0rt3z2tv/ixiAhSU6U40uiw3CCRoRmmb5xIfKZLn9nikbe3FbzFQ3dNC5OHjv+2
-          aFQWfc70+6FE1eew4KsRcvkc8yyS5fky03RUn2CO99qIhjYwsGOsRs6f0D6EF9HYKlBjy8S9eGng
-          dDQSen03CRAPd6GBJ3TU6PF649z16bwdVJfWld5kq3dXbQAp/L6bnKZ+sLrsqI0i3NY7kI9DGi/J
-          N5SRsZYWPquL6koHSk34pcWIQzu91awilow04yzSQ2rK+pyKzEMr9/jQk7Hu2XDxmgqOAXYIL9wd
-          sKL72UT97/vaunA5+nxDTgPvU3qiZweLbvvbGwHa4id12nLPyJ3FEL2Up0hPX6TW/MkMPejVUUWD
-          LT/+nuaZoOJ70wmnfPqYVZ+fimaheOP8+0v+8E+ouL/zj7z1YnC/gXEJ0G+7I6c/h8KdbeYokNgX
-          RJ0X5w1roPZP2OLyQ0/s0MfL+hg9+Ba+PdmP55MrOPqkwkgK7zioDM9l6S9sUFmGNr5ZzmeY10uv
-          QuueF6Q97N1BQj9mQVfgrGC/fT4SxcvgFv//4m29xhepgH/5zLkrl3rh3kUJc8XoCfdgQTwL8bmC
-          9wswqcvWGCzcwzf/fo8m5+brCtepC2E7KSI9PsW3y7I8mhFNipo+uJfpCs2SNaCpZoOGTPcHtsJQ
-          hYkqXPGJHex8/sChAFo+Sjh8td+cluDUwpI4d2wUJp+ztLUtmFy9N74IiwvGqBYqtPkbPn9rb+CV
-          jlZQNMUR23tPdYVD8IPQMiqDmq3quwJerz9Yp98veRbPff17LN4Tlim9kX0SfeI1MmceEWXPqLZ0
-          qjvatzYCXpPcsMqdxJzOqiSDmmgdtdTvK17PDYoAsa/ov/fV1FEGdKYmVflezUWnmDhwG+bztj4n
-          naGnJ8LRaC4Ug+cHLMLVD8GI4wM9HN2dvibtpp+g/YUWn9RxRVe5q9A8CR158uTmsruER7DFN7r5
-          uyvIN9+D5NmH1D0zdZiPcarAP3vH9Kuz2QD7BF6Lu0b25aCx+aMqHITH+kxWy+vBpxwOATofg08g
-          XaYVrGkRiPB7BAgfSCoNo/0bNdh3jwVfDpUz8Dd75uA9U3nyC/s2ZvcdaaG63fGLlmsVb/gmAn98
-          YnsflyXWw4Sh8uGCfWmfcqr5uQefe07C9skk8fqXz52dwdFXbJr1fLlhC6pQ/NJD2sX10v6WDnyw
-          2tKAPOyB92aQQY57YrKXby2YB0PhlVMntQTtc+zO7XkmcMeWOahfdDuxeCMhTAdFp24fmmD9w0u9
-          4JlEvM4vffYPSwDLHxRo2Fr8sLxZZ0Kp/gbUSK9lPp80dUZ70daw+pKO+rjhnf1hCGR8cMoYzH/4
-          91irJ+wUpRWL/nnXgqcZx1t8KIeFNrqD0uiaUPvJRzFfKMCA5j5/Bugi2q4AkdWBytd+JNTqYFix
-          rSaoQlUfbO/H1pMvERAmpKX4iax6KasqBXa49WyjoQb4VAQBfPlOgbWUiGx9G852Qnp/pfHkfcEs
-          X84FMvx5pHnzmAb2ft9baD4csOGHTTF/LTNaohngcMMfgvWdNbTX3xcCN/65kmwOkWQXN5p0r4Ux
-          VhQFFK5ZE3Bk99HnYudU0JTOMz1XM4tJE8lPRLGNKW5GOIwpcH7/4ldRz9e4cwyVQKj6XDDeo3c+
-          14nqgZYPDzjkFDpQ3IYEXk6kDOYNf06V6//gLl/7ja8o7hoMhbe/DeuZuo2HwdhpqwwtQdOx6ssX
-          fYHy4kHqH95bRTDPZ/wWWpA/RJMa4eKzOT4uBuzVYsAHdic13eGTCpv36GPNepwG8b1TRnm9VinG
-          h5M4/D0fzI4PFRuGMjCqlmWhuJkL6WnPJzHbXScFmnVl4kMWmDE9I87b58Apqd8ums6bubCCPzxx
-          9q5NPjbfNEQflEmbP3zzmbPiBhoBX2P8zCd3XtRYhO+iGogwTzwYHSuzwFTGKtXl1dOJEeIAVoHx
-          ppcfp4K1cW4N2GdiR73l2QOmXxQDIL8441z5PMF4uWEH9Vpz257fZWI/TykYXvw2dTE02PIo7pxS
-          gN/Wpdw0avFUGwpS9mJLVS0KgcRJfQGfedPiu4SCeOrHpUX//JGGFVh2Yk+geZK6YH8LthsL8ycF
-          kqy8fZYMziBy/ZXA7zMTKL52lbtel0KDfgpNnLNQjpe6+UV/8ZSg3eEwrPdtBtcSrYDq0kPSt3hH
-          lBcOHZwLJ4vNjhVayrf6RQG/GxR3VpeoRNneabBnNVjnQxAF6ITnDmvR4+aOJzPzwH2MQrLf8Pp4
-          42sCzuKa480+63/7eyTz7x/+/7d+Oc/SYAVXxpbTYTJhwK+vST4z0aUyVJ/QXeeaRrIz5dRdroUS
-          aM6KT3g6uNJffrh8NMVH+3Fk9KwdCURedsdO11k6A/sZwuKs74NFXVR9vVjziu4afWMs+ZY+h7v+
-          Bw++8iS1VTtMPLyOGuw5EWz8uNFXtvNUZVeIJT18a68WDvk5RD3O2IbPSf3HT8CG/wjnpJYr/dn3
-          qp4u2JZ/Ri20NyBDBakujvd6z6hanLV/fMC5n2A+R3Mtwi99jtQMot5dof9bgf1yJGxqqKpXWtMC
-          bPoVNduzA6j9tO9Qld4JNeNU1deehzOMk5bHqsaPObvuXx7cES4gaL22Mc3fPwu+OBJh/bSd0D5w
-          Fg9BvMuDQU09dz0vQwtd2qBNLypcxmKhhcdpZAG3/b1qct9Br32egiIZCet3u62HziUS//H1H0rt
-          7fPihJ9bvtjiV/Snp2Fsci2Yw5WlaPt+wudNNrBFthr4Zg+A3awdhn59sw6Fl/VBqDn1+nq77RJY
-          fd0Xtbf8Kez0zED+RcbUt0CoM7MMFWSNb4cM6rdnzD7nBpR/xZUM+J4MbJAzS/nRVMHmofOBINSs
-          Rc5sifiwq4JY8q6/J6gHWcHP8CvV2xiWDAYCn258Po7HilshdAVoYetoK4Bu9oO+OQ7xFl9c4U5L
-          D+ny60Z9YloDVcuuAB9ncf75zzKRfQk3+wo4RffZT57EER5U1aced6sYJSV4wudL1LD+bvVB9E5q
-          Bjd+gp3OHIfZAe8VXW/eA8enFxvo5JJwf5grD9vl/eOSr+WYEET75o9P1ZMtih3yofTEhpHNOQm+
-          xgo2vB0IG19iRAlN2OqrttlbkYt9Zz5h9p5qrK7v1e1fKhHhq7tYhJVnsv1/ZsC2+9T4oC6lu9yO
-          Rx5wvQOpeqOSPptN8QNS/QmoVb+tfG1QHcB6DY/4j18xpRnmf/t3jB71QBKamH/5kij8fAKrCGEJ
-          07aZ8es5OmBuntL9T1+h6un7qedCYQa6nd2B4izeub8tfoEKWxnOzipkbL4XJUTKOmDtDFBN4y7r
-          wCHgLthsmZnzxeldoM0e6Bkajv7rX98W7IT9k/rBqsb8+x0oUDRGDj/a0ohnep1L9JboNZg/UsqE
-          pLafsADd648f6f213EMAcyUPFO2W5eSinhWkPp9psAeS664a3JtwRENEvVvg5HTTR2EuQbDpP/eB
-          EiU00AAOn0AJezMXHM9V4M3PVHzAse8SCBsO3D75KeC7mQ6bXhdCu3m65M4FCPzCchLhVN0Jdhvi
-          5qt3su5QNPmR5tllAuu9/fxALnGAWrOrs+W63UF2VEKwbeV1vOrHTIX9oazoVf8cctGiIPj7Puro
-          DxyvfNaNUJwYCET5BmOGqiICv8K+UG+8R4zh2+8O8722blNd55hs8UahK8/jK/TVXKDjTwMbfqW2
-          NR508ZvMd/THN/TkPrJxMYYIiuyEsT81kz7N81JA3vlpG1/MXDYMbxHop92d/vEHnnt0M5wNWP17
-          fvFepR5Mk4dG9qPa5+PreOOg8vk4OOQzbRA3/QoyIaVUk50pZmqsOBDJHcIJdoE+Qb5f4f7OpcEr
-          NtthpnFv/Om92MCTOHwG9TTCLnbOhOe/Sz0PsldABXAJ1X8V505ChHjZ6SYeOzg5xwyenyXc8Bl2
-          7r0EFj/KI/jQPpQ85O7C+C3/gy3eYdf55vEqzKYHtTRvqBH0r2EpadaBbG819PhdQT1t/BREH2fE
-          mrfT3SGLmAKuh6XAnrXd8c5VLQFxL5/o7fUd6hU9u1Ixj1ShDssjxozp1UpsYLstvmiMn1vJAhse
-          o0auIXctpViEfDXY1NCTvT5L7CHDTR/f7pw/c/LbGx4EpDkQtbiVjEWv/QpJOO+wUyyty7b4Aytb
-          OgY7H6cx+bPHwxjRf/lTgEj9IWv6HKi53pxBmlPfVHb3caXm53keFvkDKojeGAYAw9Cd1/O5Rd+y
-          aggrmZMzzYDJHz/Ep0tx0ZnddiGYbnthw5uUUU1tZIg/4XObCi3XxP14IohqjSP0dzjH9FPfK5hG
-          l2Sz32LrSRhu+X16U4uIBIzN9xn+rT/2w3cSz/HeGeEgAyngehC4//LlVr8J5twY2Np3wRMyfd7T
-          /HsT48lJfjOcHsd78MPJko/r+dbCDEQRPjaJlE+/5aABrjsdN7w46ostir8/PQMf4OXLCNe/CNj0
-          +mDv7/f57x3KJbRtqaHmKVBzadO/oZwlPT1dXCmeZSMnEN75hhYH3NWzbA0W+Ksn4Of4YzRPmAgp
-          97lideNP63ysVfTHR463kNRb/nOg16S3QNQcO18D9V385Qd6sj/MpSnpVBhzU0cNFsF6vQr18w+P
-          4SglKWNLxe4IxCgP+D4adPKLSgWdzbCgR+8iDBsfSuDvfvgSGcb6sPHhAAqxa2HLu6juhk/HPzxM
-          Wnf/0Oeo8zg4tJ6Bsy2f/vE/BNKTTQShJPoSZlWnnFSxpMH5HgKmyUf4D2//6V9THLkjECnoA4UP
-          vFg81MkTgvRoY30EP31JhjJBqpd0f+vvskAiPIh9d6VawzWAUP7jQcl+3ujpvvSA5V7BgebVPLEr
-          nCyw8U9DieT0HAgzF9WL/Eh+0NWTPFBOZHJnaq4zNHZp9S9fLvdqr4DOF1HAb3osbYJjCpXvKNOA
-          vst6vtkyhOu1TMm/9T8LTIXLeTJIbaeNyz6zKkLcJhdSZeJH3+olKtTPpUOfx2hx58tHeIJt/3Cw
-          2efaSl4Io9arqZsZoc7Omk8A9/PYn14Zz6PtRDC32x/+i6djdbopgJvuWaC8pgAs2/g7qKW3hlSg
-          O9f1X/wCIRODpZPf9fKSsQyTo8HoOb1eQF31S/Ev/xmK37jsr3621dfouUlbfa5W7wefL16jydHt
-          h6UyOQ9ob9Wlh+1z2sMd/KufYhybbU2jvEqBt3cV8uOCB1gK9+IpmSs6+LSv05r91Q9myzphq15N
-          /U+PBL5C53/xdbyQIoGnJX1hi/aWLvTj0qAYfhVqBnUL1rJ9GtDOUfQPH63XJVHhx5UO+LTlp1Et
-          jimQ9lyGsWgvOnvNWQbBWMQ4Hm9BvaSnRwJ3pWZSz2qo24WSNiManDJ6dKvCZVtDXrTlW+xs+GVe
-          aKrCCj5XrL+eQjwX+l2D5sPa+MLpBxaaNyH4CePx3/pKf3yV+vqb2q90Yix/3Efo6mlO1eUBdMIX
-          swWtVblgn3uZ+rp1PkTsfYxpcD2/9d59nw1kII/HhViVYInT3IH5+7Nggz+m7qyWxQo1ysVk6N/N
-          MHEeTuCmx2Jcz2/AnrkeQEE7nrHbhy0Y09NehRt/2vQeNrSR1mewTR8CEYODm69MgS1oqtWgx0aM
-          wHzSrBXe706FtbLoY5pd3QQWcExJu1ojmOFrjhCujD1N5RvM//GFyld/+Bg99FoI6psKj6V8o65e
-          r0NnLhqHEuLO9Dhza73ulY8GEm13DARPOruL9Z1VaL8sif7h77GM1AjZ+3gkiJtrtvwWW/3/9CiQ
-          //eJArWWKmpKrlmzMP60sDaFNpj5/V4nqusEcE65A7WWNs5/snwWEdDvIz7a2Q6MJ00U0WvGJjVr
-          U3X58uqF8OO7Iw2EibDZFI48dF5HNRAQutZC+C3vilwHZzKm4Y41EVlCdGWGQdUVGYA362WGTlWH
-          RPlcgnhZeqGA3f5aYju0X+6iv1ELT5Z6okkL7zmJTjIPz29wCcQdaNzJT8QUXMOdijU4KcPMFjMF
-          xbxVpJbXpDMjvWrITo9PbC3TfuiJlgQg9ouEqr9K0ZevTCLIwaanPuhywBJLEiE7tQLVxzyo2eGx
-          ze3eWTU2gVHG0xFMClRGrg2WQ1KD5SzcEtDuNoRW08oVo9MsQg/GH6pqTlkvpqytyOCIE3ydWx8v
-          lVAbMOE9HlvV2YmXzpgIhJJ7oW79PsXs+/h4CCtjQWPbOubCcFdkaJn4Ebz78RFLjSQ84bPB29QI
-          yg1TxfYVPHGFFyz9+Mil00/5obs4EvKZH3Bg05PjlehadPiRBF4t6vxDhTCMJLJDN92dy5bM4LVb
-          R+yF/meQRs4K0O8g/3BIbw+wtjszQa9yeOFAXF/u4gzDE+7pTyMsSzt95UQrAduBFHpTQZGLdWJD
-          uOz0gZqv6uYK1hk46HDk2gCA7JDzvjolsNGnheLzQxsm+zUaMFW7NthFVcQkUXuYkFfUGGeKZeRi
-          /PFCKI6KQ4DJ7Hq9VesMn+HypLkXG0CgQdwi6McqPjZMBcIteK+gp8uHXi6jkbOfUENQ3mtG/YPa
-          s2VHyxA9rteKqqgXayLbE4T7Ghk4Z7ySL+Vk8lBdqyu1VT51hcQPHZg80gvFn7wCsyten+hwhC2O
-          x4pnLN91KmSLtlCDXT/6En7LDB1ly6SxqFGw+LG3Qp/PUnwF3jQIX/MawYTNOX3c9XvOf8vVQW02
-          jrSQnzJYdirMFNf4JdiRnmXNWk1P4fBOI5rKmlOvtAQqYO7rQYA08fpyuHAefDn4hp1BvzAmrmuF
-          iP8zqM3Zbt6sLyVSLk+kkOVy++XjfDs1SPoFBHvO5MT85i9Ik/kHPWrPdy0YRSkjSU/CQL6Ju3rO
-          k4OBvpdWopbKqL70B61Fo/BO6MvfTjH0X6ai06IURLp9PVe6fd8hOHbtK+jZ0LqzsP81sIBNjk9r
-          LeZMqDMNHvvXmTqbv0v0mxsQfDgbh9rZ3A41FBACNN3wXVepS0utjpB1TAd82H11XXy7H4Je9qJS
-          rJv6sLy/q4UesfrG+JNrgG+8tEJcb92xaqFskKbxu8I//wPN+aOv4Hlo0edFbbLfnYR49JHjwBui
-          bjA+h1/NvGecoEerXak/QyeXyHVuEcVxT3bLkauZcXQgEPyiwl6YuIDfDzcCw2Y7gVCN52HmnkoG
-          bif6CtjDHxjb9ekddq7j0GOp6/mK9kMBryI3/rPnkcohB6UF7bAdG8HAixejQusjOuEku2KXSmGg
-          QZH3NeqZ/DxI2W1OUUAmHQejHzOBnEUCcRJd8VE+1MO0o10Edy/HxOfLzcnX+ntq4SlXKD5+dlK+
-          KCHHK1v8xOr5ZQ/b+zVIC6BF87s8u0T5YBG+yv4VwKPYDsv303Xw0RYz1unDiNk9uUbwFsEjPn7k
-          sysIj9SBaXYTg92LP+TMu8sBpPjcUxfiT76oShL+8+9GiPf5nJFbipJy/WBDrxxAmLEPIEy+MjXX
-          CsfSi1kE7kMUbP501Kn6ygOIh09AE1szdfFWKTNo2YvR4LhTGN31z7sSh08XHxF1dXbKMwui7fS+
-          8jm9GbWUi4UaebtzBm5d/kPSoqBOvXDUOc1eLel7t9p/4Z5gIy8Hl9cfJ/NffHU1BbM1WfQO7oyr
-          gm3rVcezhlMLke7mYnuUCrZ6VrjCr9xM9FhkorvWF1DBrr388Cvo9/rsq58UVcXuTDXJ3+XMDX9b
-          8+7yTC2zK9g34dVWqVaOI70uumwZ7qsMLtOPUtNK0mH5i//Dx1Owdrsv+cyRQobtaAF6m/swX6/b
-          meh39Wuw+Uy1gdcf2IB1B39k4XUYr/73MiOzvRZkl/qqK+a+5IFcAkogbM9LyrH8wam4rVQ/NM9c
-          vKtBCGd0NwmUnxkQC2VfwXcVuDRwRG5Yr6dMhrLxLvCdjb7Oc881Q5f8FGFtOQ/5olrHCFH+3lCn
-          9SJ3PoV0BnLtnfH1/kn1bqL8iIJfpmHDuAVM0DS5QUbxNcnexyobcapUUFaDnPCbfYgpOBfo0HEF
-          zoSsZfzp4IdQZTcb37b4Qv294KFI0Y7UXPPencOLCtHTvwQ4/rXqwPMgsaAZX9t/+GGa1n33F59x
-          2q8LYKyYErDb733qvMoyFkViaECZ7juKW/Ia+EvSa3/5jqpbfmf+aERoz8qBZlt8JF9PSuEnzyz6
-          fL8rfT6/1RQl0XHGG75wWXv+8OD98hJceHnkSm7pKXsqnFOcIo4b5t1vPyPFNCyajZTW49/+gVQE
-          ROp1EDfVq6zA75LusT33YcwsWyLwOF33BFyU1qXuWyyQUCsdjbd48z3ksolWNvr0xUuRPr/564he
-          mRJQb4tv83XWZDS63Y66qxjm62XqShAGX+OfP85FaRVoy88Elt8nmIu6boAmi49A0fYWGN/8laDW
-          /jyDrQ48sL61nzDbwQCrl0zPl3zfOvAjfjrq2/1BF8T+UcLLadCxDzoAluL0suDvUJypKd2Yvpbk
-          7sE9MyJ8IR2pJ5sGDRyCld8UbQPwwGQddLndIeBlSc2Xz8PIoNz+SmrbjyleCv5WwNAfI3qcH0W9
-          VnKpwfl3D6kuGEcmZJGnAHK5XagnVk3dAelmIjO+tP/yr4gcqYH6awSERV0Zrwnfzuj1GDqq5v7Z
-          FT63PgR/8cz5XEg+3Tw5UE4hXrBKo+vws+czhOQiEWwluxz0RVuWiNMtgjX1qoLp73kv6XzG12P1
-          qSfG6xFSa2Fj2ObBFfeXKYWUF1GwhFU7sMthaUHK5RU9svMHsBWVEfqHP/h5YKsxL7ISJ4uHj7zV
-          M3Y3SQL/8K65JhVYlWi7Pfh57whIrWhYDmY+KkTLX9TD6btet/gHndf7hm3lKoFFi2ITVn7+Izte
-          ityFHV8ZcFfpR11+eMbr/vJJ4et9lP/lb8LMaoX2eX0QbsB9Pev+2ALleV7oKzd0MH+OYQiMTQnX
-          lMWMpTCeGvjJxTcpmXvOebV1Q6jXtktA/T7lBNNVBQ/cnsnsT6k7ByBskGvfyN//x8vYKhGsCnQO
-          lMGB7qg2NwIHR/bIN+krd4WDUsGHvo/x4akdhvF+HQOg3MorvkpxDPixFhIwLkFMdmp0j6fHcWrh
-          hv9xUTVGvL6ha0H2fMVU34N++MhCogDBV216cp+3Ya5eXQm/AcmwHtG2JvgQWFB00onaj17R57/3
-          6fDLpYGqm2ze8ITidRmiR20MXHGeLyXQXTGmeobfYLXnG4RbPqIbvhtmMp4bKJrPI7XecRkLTwwy
-          yA/XJ0F3XYlpVcQNvHF2Q6jcfwbWYhjB7FbZ1HzlEViymHuCm549afaO1VwUKq/7w5f4ijW1Zvus
-          ioCQsPuW/6xcSJ6jB2oAS+xu+EP8ertU/mXMosZnrtn2PC0S3773n//YNGiVDQ9Sg36WYbmFcaqc
-          n+CJA79+A8KdigJu8QEH2eHrrg7pMsDmNqOW2DzieTYPHNjyEU3eqpnzoj4/UcZPPbazNY1nRosV
-          Zm480eOpcIdpduQnyOaso/knFfKPFvYV3PAlNfKbwdbf+lUhCg4M34qHPkjz0inQ6tI3GTf/ZqU2
-          hHBJwg+9tUbDNnud4bh4MY63/VsEzzEhenzvVP/z93J2V2QwwJFEz5t6ReCiQYd4Ac24XafPu14N
-          AH1nhIC7HLodqh4lgPVkBrwQ3/JFlasRhrKuktEEImCpP3PQjIBP9kGyuFMWi0+wfMOctP0+c5kt
-          WE+o+pKAT8uz1Vnh31RAnRJTfc/ruTi/DgVo9tODTOElYosxQQ7Cg9sTKekrfaxeXQWvfjlj/Zie
-          wYqpooETkT9Yv21zGb08zf74AvlFwcL+7AX8Lsl+8493PeV74sB3baDg+z193aX9RoGS3WBFoDvw
-          +fixwB0E1XfEmOfkmJLe6BAGjhCg9/ed//FBqPp5EezbaMmXVC5KYL60hqruLshJiXiiSOn62v4m
-          +UgBbyIpTUx6M1q7ph793sF4n32aZQuO5xoyCN9V12xddTu3P23ax3L8ztiVWp/xfpCGEPxiB2uL
-          c9RX1/884fKN8gANgVZLVn8wUTX8TGoEauiKOawNyITLiouL0cdrUNkm3PBFsC8e9TA/pn0G4kOC
-          MI7ycaAqlsf9Fr+pn0k6EN7KbgUf8dsFS/+5uizulRkaD/0VSGKg1+IwRSnc27+UMLVN8mls1xD2
-          VkHoPReYzpQ29mA0XEN8ao5LPgyH4Q5w5uzwpo8A9uLXEEExrLG2WlSfuTl4Quu4GPS0fb/w820O
-          gGucUH1onEGSuEWGpqCecc43mb5Wt2cHhFBZscM+Jpv/6Sl2Z1Ez0p463d9zEW58mKppOsVrzhUd
-          FOoC07QBxB2GVRVRb+1P2Nv2m51fIwRJdJoDpeeO9SR9lDvsjcbEr3J1hy6x3h1UC+OMN35Tr5yo
-          pmgI0hy7/aEaxs2f4c6zQqKcejMWXrmyAn3hVazSG2IbPijgKZcp1SNq1sK438nQjJ58oJz6Nv7j
-          Y9DrhJRaXyvQV7V4hnCnXFscBG+XiYA+Ohj9ZAknv++3ZqI6VDCyrndqev6cszqbCwQQvVHTd+8x
-          C8+8CifjV+FTJxA2T8+ugy/Hv1FtTUFNDf18h+b+GgXSDg7ukPozRPDFP7D2hlr+D6/8bviNj635
-          GFZYZJrCXYgbTNVniGc9f5Rws19cqDsN8O/a/QG1MM80iF8X0JVtu0IojhDrvlwBmgiKBWk07bGh
-          50Yt1l/cgjO/d7H6QkG9SN8BwrGqn9h++wd3QTyEcPM3fDDdU73cvDlAj/Y5B9+LkA2M78YRHsb7
-          QPg56PX+FrxnRBrDoTpQb2wZbnbyH58qI5FtfC1E1ybH1O/PD32MuXhFIHchkWP9Wy84f6towwP0
-          T99hnHPLFHQLb/h+nzWw1tfJhHFYuP/w2NrU4xPoRLTx4yH4bMl/bgcC77sjLGmSeoU/lYN2M8eE
-          YeOWb3hMhX/5pvjmlisunFmgTW+kmLhlTqPiwMOwsTIcLWc3nif24qH35Rqs3i2VscpbDLRqD0BP
-          zs3OJ3df3CEJjAd13Kp0RaEyfko7fhK82b+77q9PBSo3bsa23cR193I9WdnwDAEH+GEsuocRuif2
-          xkfGIl63O3lokAcb+xn3c9dxfyNge1/qbPoDu+uuB/znvSeKtlaDEKWYU8rTqdnyl5izf3x1m+rg
-          3OeKLfmeWH94iezeszAsWeTJSjGDB1b5/U0fVemVAvyNfOw/mcTYcBgyoK7cjE3SC/XsXO8F3PaH
-          6h+xdsfZ/pjwET8ViuOyAPPr9J7R33qmF34ZNr3pDu9ieKE5yN7xAtzKQnt/dyWwarYpj1eBg89C
-          1XFUH6Ka73Y7C3bqlcPqQ9vp0ys4iJCb90PAMnwAa18bJfzjp2kZpeCfPryr2pQIf/xafRsrioF/
-          xvGhecbscQszlLnnCRuig/R1BG4E3+rpRPGLP8RULo7Z/vdVUozjEoIh35Uaygtyx3hPLmz608NQ
-          +T1Ta37rrpAUSge3/aSuOdbxchkfFsRKqAfrpUNgsR9rhYrp4QdC6fTDb8P74DfiK3Xviw2ET37n
-          IenPKGi/qAbzdYHPv/fDap5ccyZ7lQIZcix6ciLdZe3Whf4Hkg89iJzLuo3fw8na1di351M8gwwT
-          sPkbxS/wHaa9flTgsddkelhPnTs8xqcG/t4n2L+qeMF5r/3Ta0UrEWv2+I1PGIgHQObzy65ZHk0j
-          iKWkx9dez/NV6y6dIiGvoaGF5HpJ/MwC82+7g77xd4qP/A8OfGrhI5+KMRHCuVFmLpIpLvYLo6kv
-          c1AP3Qv9izez05ESan8njvm8ZmLfcAZgLhCpYblVPF3GhwPmnOb433qfhVsKSMLBjd8/B8LXYgCb
-          T1b8w1fLUw0VOOvQpae1CF2W3eYEamms4lOvvNwli8VC6TP1grX4o8fLJ494RYvuWTDihrjUsncE
-          Lp3rY+9Iy7zN4iyUzlMb46BYjIF1OeiA/4ktao2iwP7hCayKFjbS6qAzR9sXYFH5lKavfAXrZPEh
-          DEQdEE5cd+6oOOkKL249YFUexno9ZbkHh/Q4Yg9dirrIL0oH/AcMaSgP40CvhJZw4yekOIrmMIkD
-          NpTrQVqpapgJE4YxgKDbX0qyTYcH6ymLA+jv3Zn63/oSz9GBg9Cyh5EeROk4zDV1VVjfzwLd4rcu
-          zA/OhOeTE1D7CtZ8bNMx+YuHOMKMMlqqC0Hj6xRgr+75enHgroJ2kPqBhDV1ECbDX6H/OVvB5XQ5
-          5muyuB3IHCnZ+KE5kH0Mtqma55IaLVRi+uwUGU6D8cXPoIt0dvv24T9+cEyXpGabvgQDQvVgr4vb
-          gbTG4hUNfaVAQUdNX29vO4CgzSqipOoxnofUuP/jm97FpUOHU6VUFvAp8KG7lUN3/AwJrGhFsLfp
-          Ayt42g0EH2iTj/GGOtn4BAxy6lNnf08Yu+a9omz5gVqN/xzY9PBDecOn/+LnZD+UEkTGc93sV2M8
-          RwoFip8jj003HbcTv1m2/8rthHXb+uSj0nfWX/2FaqczB0jy7i1Y7r0OOwTSetPDA4i/oU9fn6rX
-          58SJPajHnwYffX3Sp59/gPB4afugEncRmNfXGqKtHkZoXBZs/tNzxXJl5HvoB8Bukjv//T8N3jQY
-          Nv45AqmfL9jEO72e79ZhhJ3wlLHTc596w993NARJTo/6T9bp+8WRv3xMdr/IHITjHiaQNKaDnRzu
-          8w0vBnDmQhk/hPEL3pu+AAQNo4COwX3ryu+t8P3NOey03uqOF3M2/8Ur9fzqa5L7UgDPz8cx2J3Q
-          KZeq59GCePcSaaBsU+lEXS6g71QHbBsTZcv2/X96GIEjqOP19j54SIKrS5SCoXy27b6Ez1WuaDTU
-          9bDw+BMA71C9qMGfKjBbWycvEMGYHrRb4g5HnK3weCaHQNry82wu1QjTtov+6ls5e37t5C9/YG/S
-          +mGxH0oF/ZI70QP8+DFpUTGDWedcwnbSo55T4dr84zuzzxf5nx4Fr6URYO+usfjfei3JyAUCeVv5
-          LzrNPERjhKmaJ1I+d3PRga0ehp8WM3KeLUEClfvNpof0ea8XcuYIvKuThp/2QuJNT7rD4TPE2DI7
-          CEjOJT94Q+GKT/KZc5n84DgAD3aPD3jtwKI8vzPc8g+RNnxJ/T3yYHDiP/QeBWfGkM+Xf/ExKGuq
-          ueMljRrUW09C3atZ1It3Ot6RHP9+9NA/Q8Bk7yeDzV8I/yMT++NDaDyt1z99vuYtMFewSK0Ub/qy
-          Pr2/igWDa+Ztei0XE+U9ElitkNv0SU2nljy0YKun4uDbwJxgohbQKLeWcNlu0WdVuiZImSVM//S6
-          LT8VUHQWkwBNoWDulV8B5jzWAnl5+a6w4UmQGuiLHetmuKtQxgYc5N7GR3k3DbTLwhIpjlBRD4Zj
-          PLr1mUf+3p7pI6xfORtxkYBN78H4aD9dQs9ZCW8/esHa4eKwZb+rRuXPv6wg/rnrp7QTsPJOS1By
-          ard6gEoQB9ue6tzqxyzFvxLUc1yRwaII0F2fZtD87HW85XN9Df1jBZ9FUdBsmW7D9MfPl+/I4+ut
-          /IDvVt8Ep0Uu6Ok+wnyKDuH6T5+x5+qnd+/4bSGJ1DTYXaMwXtjxmkFuGw8/b/XQxTvG4l9+pRY0
-          snpKv1bxtz+ExQC564WDLZhX9x0o+zsPtnrQHe6USxtIm/69lrM+//ExHMLGjJf2J6RIr8Z148f9
-          MHdkMGBPFIZd7z7oM/dQebjVv4hCBDsf5+tTg5K+9lTPUssdvcVZoYN3a1D0YepK38fkwVhK+60+
-          X7pz44wc6JkvUG3FQj6+XEP+/5wo2P/vEwWxJGv03qpnfQIIJ1A4QzEQkz11GZ++KmjhT49d8/GN
-          5zG+ptDrfxo+fpdvvF5MXkTyU3KIIE8HII4iEIH0ri16OldrvUjj4w4r1l2oEUanWLzWPw36DJ2x
-          H0qGvii6r8Jzpu4pvvy+2x0HsQRhbUw4/IYGm3Q9zSAf+B7O6ngYll95dlBliDI1fbWoV21YIKrP
-          4xE//fI7LNz6KkECjTd+IOUBmP8K7jBroz0NyiOI6d4UKhR3qYy9dkfjcdxbnFKLlzs9nEKRzYba
-          PKGujX7AP6bPsFSon5XeiPNN4Tnoc83aFWGi+GS914D1tyv4wbAIQhqcpmgQhMvEw0/26XG2G3h3
-          WR9PEcSSohEJJX29vNXyh4wGFoEQ9QtbNPMRAC/0O3wcDko8v4wign7KHamrUjXmhzS1kHnyBmq/
-          IHXXOJoNlPDPLJBQYtfiPVsDBZ3uNmHmZQWk4soUiTy30sNUVq70g02L5ji2qF7KLZv9MoGQffU9
-          dhte1XlxSQoUW4yQpB12w+T2RolqdXhjzJo8F07PLEAwiwWsLb93Pa5I6lDzqCZSaAoeeO+0ajAa
-          K0z943GnL9aMSphawYHGCwgG8WWeM3jst4pXat9zSeX7ba7a6mKrfplgMWonA3slQNQ+Pm23G+Nr
-          Auc3I9Qf3seY16zHE/b9Ltns75TzP1VRIRO/OjVqXdTXfHlYcF6ND43il6GPhXeMEOyTG/bvPj+w
-          9XNuwLYf9MGZe5ccfm0FR4o8GkaJU8/fn5NBnU4ZffTf2qW75cFDe1Ue2MVMAqty0jJkQrmhYcd5
-          Ln8CigenqTGpX3crmJXjMUJEQzYOjju3Fmnvb3OSpQd1+u5er8ZLHCFo7j51H3GjrxN+ipB/ljp+
-          Vfufy/f6LMLWHG5UHdvLIGX56EAsqIhGuVq6fNK+DJjgysWH8qizeRI1B3H7X0avxinJ11qutjOY
-          hhnsiIbr2VmcDFz1COGDnWs5U1RJQV64u1DnYjcD+56SrWuSup2gUMyYusOhQx3nDPjwnXxd1Mi8
-          Iu92dan3yC5M3J3FEaKXFmHds05gvnyrEsnpNrdyOV/rOcx/K+AnraeGtQzxupPKBJVGtce2w98G
-          vpH22t48cRl27pI7sPOlUpEpqha2GuE68KnupZCqjU4N5RjoArdeS2BBe4czp/YGSbhWEAGfawMx
-          3Y3D1Fqpg5j40f/iUd6l5tbTIZNOwfrs+mGtr3oCX1xmU9zgO1i3/UZh0p6w+j4nQCr6awSiYa2o
-          5737eP5+3RZ6Dv/BDsOJPotL8oQefSj0dMuPLv9+Hwr0uOQPqpnGEfB6dyjgGAVJIG7xR+Q5O4NO
-          16RUc9tlmPOLTWB9JkfS99epXmdXNuGd8w+BWCVTvGpEnqF3mDQCIHgNzFSGDnL7LqO2w+9rWqXc
-          E9o+DKmusq/LFHWnwJ9dPnF6GHJGE/PLQc2bCuqueu2Kw26OkFQ4x2DWr2rM55GjwMeHJFuPjaM7
-          k2ByoF3/HFzUZgEG/2Xe4VlqdHyZeZovp/3CwV0Lc2ysXj5ICJ0N1O/9L72J68Do1bQ0eD44D3w4
-          hSmgYZI28CwaPbbMJ8f6tOtN2NT9g+y0IgbrkN9N2Mf8h7CfVuhzs7A7mgfPoqrX6sPC5aqKkksH
-          sMPNoSvlIUzRb+ddAyIcLgOj03bHmpKcnnaBy0QpYHd0CScp+E1TF7NiTE1YNZlGw2A86AKSHw4Y
-          YmGiIX86Dwuf1wVSzvcz9o9VHa/N+C2gSIwLjdu1d6fbPlZRPxQE+8rPAQsnPFNYl8EQxOOtYmvu
-          bWPJzZ+57fcHCD9uGeHAMomq4u2Xr9IcraC5OBLWZCWpJbg3eSTNLMfu7nmIBXd0U7j6z2uwP0Y+
-          mL8/7Q6f86HGYZT8hnGt3ynqhINFze6k5eK1HAKAjl+ZaoM01uzN5QUsFxEG8JEJbLEu8Qol+dbQ
-          o/I1XMHXVojGgzsThWsSwM/2CtFdlXc04esqZ0MR/tCC9v52ZlQDItQyD/LHU49PN2wzltNERjf/
-          6VG94oxcCq+PO6htZuDjybqw9cPkrcs7LqiBuHKQtFMgw1uOj0SI+jMQEIg5tOVXqhWV5/LRek/g
-          LYcrTtlV0Jl1DWUkuKKJn1CxXOFYui2E5zGkkUz2+colqIQfKTex77W5Pl+mzkLeTjfwKebdWFgf
-          KY9GjTsE6Gl+XZK+rivgSp/D1kh+7uo4LwP+dpyBU6Eqawk95xLNvpXhF9FoTac48dDs/jgyO3nm
-          rsaLI2C8GTPWd+kIpr2tdZJ5CgaaWf0vH1PTtVA1qm4AvUvgSpJLCTi+q5Ue7fgdS+8bCuFdkFyM
-          N/sWH+OPg9bLYQGIx18+T9dt6tFlKug17mZ90dVXCxU00O3OA443e3yiWXnmREbLGcxZoYxQtgsr
-          2GEpcplRCpHif4oB37LyrU/xeV+i+PwsiBKtD8acoChATZ+IHvRi1bv30/kpI/da8WEVo2H9w1u+
-          X/+Cfe9uXT7Pq4deohbTl5EbOlNjzUGH9CzS21d41/N47zJkr/KDvhqssDHMaogc++Bhxy/8eK72
-          YwZH7rHiQ49v7hJy1ogc4VsReVQ+8XfauprZ+WTTDT/lo94dnujhL3Gwl61iWINXHoJ+j7+Eae9A
-          X/LpuoLL7h3TtNH4es3uUwuehx3D+Hzj3KV/HUe0/R51pn1drwV9z4h9D3v8yr2IrV7jF1AoX0Ug
-          FEaYLxdNeipQc03C4kwZ5rdWyugkU58Gd/BhbEeuKmyGc4vt3l3jZbr0Ccw9+qXbetW8wLAIaErW
-          bXZdOvzZB7rebiNND9LVFefTxMMND/29v84iXmuQJ2WAXjb7l5Kt4qv9mIdNqdj86/Yl4JtbFHt8
-          sNajq39nKOJIxr5rYHe98zmE9s+nAf/6ntiMVVsF1kspcVDvd4CgO3MQSCR9w1sgnyXy+ME25ULs
-          7WiQj9F6T+H1V4TBZcPXPFRz8+/3qXa+tfVqpfdE+XHFDwf43rO1ln8jmCB/C8T4Wus9iaoKbvmU
-          Ho9aFrOj2ifw0mcNNQZLZ4IpfEf4cyIXq1cxzZfgnVdwHfyQsKu3yydTbkvgjrpF7b62GJsb09lO
-          tB3pUS/HfAw4VEl2fI2JUJdc/M9eqZkD6pp9Nax/+TFPLY1G+rTT54tnzdAvFoR9b03zlR3U8F9+
-          c+juoNNvx0QgdFJEuOKc1kuwgwacKz6izhVX+pxnagXF1L5ge7NnepdvKeR07kX1LhHcP/8GnnQH
-          QflxvjlpRvoEG17Ejt11w1KWfAOxVXypfrs6gCXjrYV/fCE52m+2fuaFQDfJdv/5L0CnFFjhL6C6
-          cy+3/KS1MPJ2RxxUxRu0xdq1yjTWFVXHuxIvKv00sGluDr0Dq62p670NePgmd4pvUTgs/nG/Qn6M
-          4J8/5aODBxVG+Wv4w7v5ErNQgfhEFOoJ8FYLTD2Qf/FXKX/lsAb6UkKdYI6wLX7Pu8QVlQd7+YTt
-          Ug/MF09d4eHQvem9Mbc7ksPOgvRkd/i04UUaGLIJTdy+sergAgiZjBtgPq53avOyXQtd5VXgDz+9
-          JKGJ1/FSPGF5H7pgNLqULW+167Y7EzyRX4PmCjuxn2HQ2TP+e/75OfwfaeeytCyvROELYiAHJcmQ
-          MyiQoCjiDFAREJFTgFz9X/h+wz3bQ8uyTEHSvdbTSXoOoO2mB7KfjQqwq6LF8KkaNlV9vWPLJnI4
-          CCV+Q21tXs/kHEQNSlInjbfT0Os084UC7ZNNRr32swXzF40BfCrnhN5f41ROL62R4cc4hevvy25E
-          m+MZ1YOxIVfrNrhMPPd/eocqhzTWp8fhtoWCngU0S8U9WJ8XDz0toyS80Fc4tcUxQT9/EqKAhtPd
-          slvkFrs3UffEdpdsyWvoXIuWGiWZw8k5PhLoPypMbeN5YNMajyHq9QcWxHEuB5Xcmp/fxACCTTcH
-          YC/DR5PeMF9LMmsUYWvAMufikVv96uhb2wi26qEhuKvXPpcBccAar/A2CkpGk9t2gUdnqUeRlkE4
-          7/ZaC9nLU+lRLQgbOCGKgPSVOYKHWeuE2y7V4GcHvTFf5hwsunHmANaLlJgeODDpeCk0cDkVDsWC
-          einnuQ5PcM1/9HZ6BYwpgRuBnT5fqfLJcneGRvCA2/fwIfhqv/TJFvgK+f6r/eljfRkFJ1s30gTk
-          sNO++heqxxbhJJEoyZnXLfPnJsJr892sn/uy9SrzgU5l3NOwOQv6vPP2GeSmQhiBUZo6VX0jh+0G
-          GuTIx48uN5++hrZOnxItlotS7EXGw59f2dtV3a3PuwUCpSOWmpiGDD2mAkFJ3FAiCVU6CXIsA05f
-          bwlWwRxO9eRokDsOMTXfkumKZuVAyM2nAsP30UlZh7YavKjPimqxrHWCrj8SuHg3RnwzatwJVHPw
-          i5fjSx07fVr9zU/fUeUyxYw5M/Vg54+IYu2q6RMlUw2vTbchVnjV3f7jdBrcipNLkiL+dOynH3/5
-          ToPFwx0aveTgwpwjUQCV2QI4fysbuBConXexu9zkWy0/UZsRn4clG8IhPYGbvBDMyRR2jD1DiCT9
-          HuPNruzK6T3yNZRCraB2NZmMEU5OwJNUPT3R57pji0M5WN8nsbbcCKYFbRrxPhoHsvIGnTrXWIav
-          ayFgrshUNg3nYgHHfRGSX/wfcXGy0FM41UQtXDWUemlJ0N3aH3AUv7blEvCbBA4CDogCyk0555v7
-          AYh15o3587qECw+0Cr1sn6O2dRv0ARazhh7hYlBiGffwu2EtBEIjnEjkx0xnr4w3/vzcT99PXKI+
-          UK+CO7U9r3AZf88rdDONiAa/93GU9onsLu4WTzg3mXTokfLTE2uXJVy2O1/R4OKlbLwE5Tmkgqjn
-          aIBGP9a+J7FvFqoZUgr7SpyvlXfzq1AUYABxi/uTfnHH427fQCmcCHGOGxkw6bR2KVn1kdGInT5M
-          hzkHe+8hYjRJottXM4hBXlRfXOvo3BXGUxyBC6oTObr7b9iL0skC1wRYBCvHqlv0AuWgcbKFkGIY
-          2bI87gnsk+pAot2hdadTOE9we+VG6qXt2A3ya6gggV9rnJa3WQo/fqZ4/fmPp4hnb8nQyfisXTLc
-          tz4KUZUgc/+eyTq/9KmcPhDa0+VKlTV/TkFaTPDaPgJcASa4jIsXGWV6ZhJcB3XHcEYqiPrZ/eN1
-          y3YIIYx6+4v1q7cJme4kZ/nHZ7C75Cm7VmsXmIK3iUoTIxVmg23RY9JL6hzZpuynJi3kA9rExLtB
-          u5ti7mQgB+56YvQPvvtsu7D/G/+BRyDt/TJdIOHDgLpNUwHe3Q/JT1+Ti96vPFCIIgjazbLyh658
-          fF6bFuS7p/nz6+WczYWGbqYVEas4++mqL3K05htyT/xzR8F6gmPlGyP6fnRXQDe+AisfpOHVe6ZC
-          Y7UL+OSvZfVnlbscr8UJrvkDv+/cqRwUOlQ/Pz6C65mBSe1ZCwzzdqK6cM3YPFxeEfj5c2vZcOWQ
-          hfsMKLJ2pe5xcfRZ1x/xHz9d9Vk4hX02glkor5TcLCGdgriDEHJDuvq9Yu2n2CeQKNa88rHI7fV9
-          s0WTMRXj5joZ5epnR/i8YpH46v3yj1/8/FPUqkbH9svSQFR2DcHZ/cjYffjm8OT3C43XfD8h+f5v
-          Pf/m92ilXgGTg1KQI3fFHa0uuoFOHjKpcbZ6ffm8pPaP70TxEIIFK/IITOn7wunK14ZKmBa06g/M
-          l2lfDu7Xy+FFvVd0H3xwN2eB1qA13uLXPtVC8VEPZzh0s/DHW6nm3g4wb72F7OHN7MbdJ4ogcR4f
-          qkmenwpTH2/hhWM1FlvVKGdRuImQCjkhXvuJ2dJ8nRGa+89M99FFKAdueRbw1t5jGk9R2E3iZenB
-          crFeI193z5L9/q8LpYGq6ti56/haSIcPpvrVe4a9g9euYB+eJ/d3IYST18cYvD8CpKR+heV8We9Y
-          fI3pke71EJXD9sBpgIxOTrxx+wLD+5jUcqaNMVVSYOjSZrxqUHctndieUoMmCTYZcK55S6w8cV2K
-          M7v6+Z1fvgsZQjcLkDq/0rtpPt1FOekVFOZcIWZSVCGfeWYAss91oG7YtyHTv7sesMa4k/h1+6Zz
-          VHQPuN8tu3HmqxtY+ToHWQAOxDXKt9td6pQDhVzMGNSjls7jqS3g7CD15zfcld/UoMy9jiiXactG
-          cru34PJaUmr53pWNLzmywGLymMRpzadjB+cMfowgXE+EdOXPz4ENmW0arPptkc67dR/T28bcRjG7
-          xbndHLh531V66+6uu0At8ODQvwrqL3Oc0rkJErTl4jPePrtCZ1uy40B2q3gMeoO5fcnGCQaTVFNt
-          35NwxMrSy1/Xlcbqi9VyXnkSFHm40FV/lHzFyx7UJy0g6/jBsnX5GpRqite+2Qd9+gx775ffqXZR
-          A320L9YWXufzQBzVGcNZV641XAopoea50XR+qRtHhk3dE/woSbi0VyLCsxU348RZain0dNMCAE5f
-          vPPdPBXk17uG/PkCqVnoSjr+ePp4+JyICs0La8pb8EA/Xr3qDzDmelJBlfYBee6uQTj5BvNkcbQu
-          1M6fGZs+7lcEe7NO8KZKI32a9jKEn7Co8Xblsb/6xB//wor5TUe0C0f0em47qunDU58Tdo+hfvQY
-          8d/FJV35ogMqvleJdvS7dJarKoaZa2Fi8GURsunjcdBTqUYP+cNMZ21/VmBaNQ0NVr85bYQOw9r6
-          3ohh65U7Ky5JYHhb/V7+eIcjJ2YyDFxwI+rpBsOx4JoI/PTRbTYM9nDxK4IRuEKqiD7Vhx//u5+/
-          069eEPKNWWHoYAn++d8lo98JfNtBJYf4fWJDkL3OEMZHgfgP66PX/jtYuz7jYNz0jh/y2VC1UAoX
-          ghkoGpctxWmBaz0ED+oxSZe5jRvYRCimKsVftzv0ggKz4Hkk9nR8seXLVSISXOuG56+OWDO/sxGe
-          koGjXr8Y64lLLwfxR6vx9vWuOtbCvobZ5zKQNb+EC5EFDf7Fh+cjDFm76RKYe65G1nqUO5eBfUDS
-          1z2ufqRlIyeeZWgfujsW4Mvq+Ld/8cCsfFyKSbJn81xmGQgp+fHht77ykQgwtZX+1dO6ndxAarvN
-          OLsXo3y/cTqBH5++vR7AnYpU5+H8mR5EF64QsPvUnOHOk+UR7J7HsHt7Vvt7n1hiO1fvh2MHgX6I
-          Imp88auckymq5J/elxrv7I5X7dDCnef6eMfnTO8LcajATvbQyK31unHlt0AVq4H+9OOYP+AZ3FJu
-          IToKSLjArC7ggrGBpdX/zU8z4eGS9S21n3vP5Z/GOYDaBCCxwZ5z58YqJrS5jg7VDT1xe/uaabA7
-          vxNyvg/vcoqqrv/NF7z91e+ioszgSQ4dcnBlNZUGnlUwO0QvzB86v5OG4J6AwqzfuC6cxV02UnOG
-          zwvNqGZzjjv+5ltXyUd8mv17ynzlzaNGWARqs12nT0Xsa/BkrjWQQ7p1f/n/l49X3qCUc1MYOdot
-          FU/IiFs2PfygB2s9ctwwOAOahsCBSbA/4a13DNks3iYLOXFwJfrKx5bZ2Z2AP4ivtb7B60OI5R7O
-          gXMhxtnydLE8Gzla188ou/IrZSf+UMG9j0v64+OLvu6i69XdnagTDFJRId6yW+vFuFzn35SetC2y
-          oowj+3MappNWiwoK9LLEj326A71zfMRwdM8bGryO57V+QSGofCLiIlVy/U8v/vTtGl/cHljBCKXi
-          JFDNOJQdXf0EPBlvQvZzkOtCvrkcYOKUDlUHpytZEUoNoI8I0ENRaWzuttACynfIiSlGQB9/+mOU
-          wIXuccOlzcf98vJ3+5SJUaZeN+p2usBGMSAGQtd333R4LkDSnzHe2XEXMob4HFVEs4n/CbJwbGp1
-          C0+eH4+USMufX5NznXmUaA3VV57Yy4L+CIh5507dtDnrPPykB4rlR7gDk4OfNax8XySH+TW45bm/
-          VWCt55BzC51yfLOpQOExy+jNjt30x9PhzeR1chfFWmfzyz6D+lGtd45aajmJfHhCrBEPlChRFE6b
-          q9YiYS4U6n/TDxDePYxA3o3eyp+mbta8hJdTb1jv0BSVbk7YJf5/dhTI/3tHwcU3vlRXH9+S2d8p
-          R7yiJ+TQ+mXH7ByI4HE+CtRz2o0+9dJ0QsApD3j4tnM53sEHI3JUdjTjTjXohq2QAwl6Ns32+Mh4
-          73maILdkA3Eoj0shJscecblfEC1BpFs6Dk6Ql1a2utFfoKd46KEy+wpezp0fTu5mOIBFvdjkvLi8
-          u3ye2gifO1Ecd/m2BE37lmNoXiJCcJ/o7vy5PhY4FV9IlL74dKzv9hGckaxTv8VVyLTse0BmaBok
-          vO9xKlj27gH1czxjtH4/Fs9Ogx83m9bx0W4O3FKEKd26oxy/PulE8TBCj6oScUnogsmr/AW2z9Gg
-          TuTkKT/t5R5CNiVEnzkXiG79PMFXnKv0JlcnndpSjlGI44geyXcXTiIoWmTNEyGRFQF9fl9PJzRp
-          DU+fdXMG7PA+KOhxg4A+2T1Pl2hMo91FHo8UK/OxFD55+ED70yhQv+BhOsbY4YAaPkO8bPXaFY0a
-          WKjs5YIajXcHs6MfOCiMfUUMDw/6Mtk3D8oAmfh13zXdtN80DewTmJFsPCcuf7fNB7p5Y7LeIrL2
-          LZaPPHq4+4549eesL5NwXWB7nRm1u+hZjo+c8GAjopAez/PRXWKswN98omrudq44BPKIbl6fEH05
-          SmDqYBmgohhF6uAxDqckqCE8c+FtzI8jKydkuDXss71HEusYu+J1L53gPd4G1Oxbj00C9hMwvEaM
-          t0pmuJJjnh+omRV5hH3ddsxgiQbF52TTxO6/3SxdBB4Ss6uoHxeOzm/9vELVxyL0SCI/Has536Kg
-          iijxPPAGyzz109/vU8vlU3be5BPEFm9gufX1bkpth0Pf/haRe9p9dT64zAEqjyWmBn8XO+rlYY4C
-          M79S51Y9S3p4axriFTUhjzF9dBJKpxHaTudTh72KTvw9z3TvFpQgygMmtcqIFFMsqXH8Tl1P3ycL
-          ctn0oFH9PnUDn/QRNO2BUvxSvvrgB6czVNBsjJ/F49mYRy8RmWMW0yDdS/rCtUILUjcZxpcg2Wkv
-          7aMtGr4zJImafEL25owIDXFl0DTJnjpfI71C4TXPqZfIo97oiqqgmyD01FHopI8yEWtUapSj+o5D
-          jD1LZwJpdxNGlBPsCtox4WCkvgOSCQcaztJjyVHwaHJ60DeXTpKkKUE05DN6swzCpOsz2cKdLCfk
-          LGwbl2V9G0C5DiSSUIMP6QFfWhTBTY9336wpWXvTH2g4yRyWdjsZsGoKHNSrYoh/8YQpkBUIdT0g
-          9ifyO+l75zIgg81aEZGmcLYvQwuvx7tLHMTWvV6zU8vs3CKyrudwul7uGTwzLqcH/hm7fFUnI7iN
-          3IDHrf3V2byNRyBYs0uJZ9pA6rezhhzTMolO+286P7N3Dd0cx+Ok3hs2mXsNI5onL6r1Ue6yIJIV
-          mEmSMG4v1eqUN4YM02jRiM0d1G7+fs8tui9qQR5v6QaYHXoHWCepP8qC/O14XdlrMG4zmxCzeJTz
-          W3nLqMNFRpR7uit7O+IScNufniT7uDVYjmZuIb4qtkRjT9Vl6tYpQHgtcqr0hd1J500zwaYgLbUq
-          pUqbX3wqr/hM7GaeU0bSQZHj2INEO90OYN4eNRFqi/MgyZ7rdbbBZiF/kpNCzbOThrPiHCcw6txr
-          hFacdfTi35XfeiLOKe3TufhuHyjkzYEQuI3cVuu+NXxFgUhPmpPrwvWuJuhDOYw5a7LZ8qoemsyi
-          rUOPqc7C+Th5CZyOmUGDLcvTqXuLW8h7nwcxzz12+xq5NdwcYovqhXRns+DxDWo+24l6ljl1y3NR
-          MWqfvUFVa3FK4ZtWUL4d7oQEuhunwjd/ciD4LC1VvDDQ5881msB4ayjJxC3H2FsgDexfQkpiNVxv
-          5eYtGV46oI6Bte3T+TDlHkTf50QUp4Z6c6jUE9rLDaZH9Wq6gjIPZ7nxypTuLYMA2ln5grTrZkfM
-          O2+WwvO6VPB6dY4jt79E6XKISwOehJNPbUFr3b98Xa1NXh2BLOlwiRlGZiDsiKU0bicYydlBplQX
-          4yxIdih15zhDzsPD5MJyFfB7seTAeLdacvA7xWXrHfng9zwsvkpdXtm3OXTb15EekJ4BIYuvMegc
-          IVodjgKkJYQNlOz+S8K6mtnCdgOEUBslstcXPxVaXzihJNCOq57Qy1nZ3JXd9ykcyW3vz/rc7bUA
-          sTLz6T65yCVLLnOMNu3q2AeSh/PmkyYQmDbBTPd27hi25gEOl8OePB9NwURCLAOZgbQjprdp9L4/
-          XHM4RNGOql8NgD6UH2e4v6sjCdPDFQghkTWYaxuFXuyqSvsz1yjIEvCB+N6RpOLTHw6Q5vGLZuns
-          Mf5yWfvCB3FO02ivulLfqRHUsvWM3+kTAr4haoFM0ig0u/qy3hf3YwE+5txSp7FlNvBJFSH7c3+T
-          w94e0uUoTRG0ekMhAYBqKoYSFOH5k6QkXPPhglnroNfoERov/wEAAP//pF1Jl4K6Ev5BLAREEpbI
-          PEkQEHEHigiojAmQX38P9l2+t7rLPt3amlTVN1RISYgugppb0DYvpb8rtlvadJpUcEU6QPYqT/ls
-          Tv0Kq8Ob9ectf9cc9CNc7PRNchQNoHcNK5Pk6jaTK0oxwJbx9iR6PinkmSZlvUjmm5UA76yYB9ms
-          jhf1+gGfzn8g9XK3ovWE7EbiW/FMlAuS3XXmnjPI3vjprwFziObDajsw3hiyYRblQFfDY+AleE3E
-          tpea9oB9OlD7jl+izB2jLrPVQXh4RCIuvX0FWq/RZyiZpwcW7Cim0/gEFsTVekCKypMB6zc3hXym
-          H5G+xSvHqXsID2lvYgHfJJeo/nKHox76/lxyx3oc4BCAkAtOKH6LIJqDncdCK5dU4tpHu+aXeVwh
-          v8o5yUp/yil3TAqJCbyMbPnprp6QtpBahoA0g3YR7utUkKLJwyTqzzpgdVkv4SmMP/jgrrt8Tswu
-          EDvtISE3b5phuaTUg43SLEgWqBxNfXKA8PFcfKTJL9Od0fnbwA8jDOi25ccMmwsDjEc4I8dmRnfR
-          L50jPq+f2pfYUthG+rAOPITSkXjRyXJ5ETEfGMXGm1j4Jqlr90Q+HCWPEHc1zxR/lik7lEUSk6PZ
-          fXNSXS4fUVmdgqAtfuaOVsZvfcjj8woithg+Cgw1vydudn+6S9+8evHUnRq8E47vnCoOWUWBeVTE
-          Qu+9StXjcZUudouRJ8haPSVCuUK9pSHxE+dFKbx+FXDpOp4U5T3J58CDDRCA+v3Dg9nmayjZr0rH
-          tNpLYA48toGMqQWY+wxJTrnUYcXoMFOC4m6s1+a8226VFp64n5KG4vidOZBnPk9/BwXeHe+Hmw/Y
-          RZOIIfJSPiVGWkGVzDoK5I/j8vv4bcDzC15R9LW0YbaXmZFeaXUkWt0KAL93DwMO512JnLoXI+rZ
-          xIEcCLutm5YM9MarM9yVxsVv0NV3aWvqjaibhPi87c0RjfhRA6+QB8THuyVaOHXPQOH7aYj1Ga8q
-          fcfCDOQFyeh4BgCss6f5UMmQhi8E5e6qKlkJ3HNuEX3tVJdnhHcMcC11xF73nsoyT3KH/ZVSvDwu
-          F4qt4+rB3/oBwbCiEWVjAJVOicnpC190w98S6Nz1So7xaZvyIiYxUJixQk4kKdHGb0XYLoqIrEsz
-          1bSZUwdO++yO9EvtUWrfTzH81V+NO3bRylVpCLtLmZLsvekaQRtE6OwcFqlLdnDpfi9kkKWJReTl
-          JuRL7d802JdA8stHfhtWwF4tYGjOF0tZcB3IKax5gAhr+cHx0YJ5X7ohpGek+MwlUlW6yyUfRtYU
-          kPwRsjmG6X4VNL2Iicc5KJoXnIfw6ismniWDid53os6w7rsrenJu4O5VQVdgeB0mf+P/NbH2zSg9
-          v99h49fKMO0LsRIj1pzQCZ4KOk/3VABbfJDIO2r0tx6S3Go82vgCmCd9rGBY0j1xi+lN593S8iDp
-          ys4XivGtkvV+iAGdqIH332Qa1ou9zlD5LDYxH9d9PhWCxsDlLa3EKy8Lxc/TZAHVDb/+ZU93YDKL
-          PoZOTUsstMNUU3q8MSDN2z06ke4LVq4KQsh0cU1Cw9qeOdvJd9HdmwmRF8lVyb6XR+i6+OAf7KUG
-          i8h9RVDI/Iqp8zm6bKqIM/RX//XTg+osTPsCytsJABczcz3DBIc/vueva6vkZNP7sOkrh3jEFQBO
-          c6GFPZdFCHGkou/f51WnSkRqIZ/c5Z4+UyjTu4tMW3Lz+WThTDQ+mUgM73Kux2uROTCQCmHb+XO9
-          1P7ZkN7dtSdyyR0H2gRfDW76l+Rb/RzxHs6//P7xwwg/APEl4SgkRM8rVWV39WIBseIhQpFv0eUa
-          exjEc/lG0eZfzKguGclWD29/th0f8N9RTOHCLgrKIkOt2Q2fpa9bzP5AKzXfb/kJyci8kWzEt2Ge
-          qR+Ci/AwMGvDflhyp/Z/9cpn2UcyUNi1KTAdMUXG2NjqftNL8E59BRlfW8/JYsh32LVSTq5padTU
-          wEMJr4bOI7nlono0n0DbpiZSZI7FO1+cG2WlQRjtP76wDOuDh6W9Zza86mo8BeIIXDrkW33eubOg
-          1QJ8CS8bpTIJ1NlvIwaGN1HAh1WeonHQS1k6G28FOZxo12Mez/1PD5HTlbm4a8fMgSRqSkOU4/kY
-          LQ+43+JztyIDMArY751nCz/3PsUzgK+INgExIIu5BakPG+f4XqoJHL/Vy+ce7LtewygOpeQaFX96
-          bI64IRGbV25jadvv1f1cA7AeotaHW/wv11gboWsFDPFRPNYLL96EPz/hOEVKvXiXzV11xpAUeHfO
-          11G0Z+C/2zt6SiSmeykXRvhKQh7Ze3ynS/ZRPnBVm5DcDICHP73+IukOiwnSaBsWYgHA91IhJWf9
-          iNspuxj++LK5XHiV8HprgfZmlnipv6naGaI9QjuxZoKOSB440tgxeAuWuNULhn4P90clPi0a4aV3
-          jhHXnoT5L58CdD+7f/zxvT5Zcqy/gkoWclfEIPfeSBsf2kAPve7B9ivOSGGTVV0ri+/hl9+f0ObP
-          0JnqngfPT+IiKwpddz3vhRimt3xEJlGjgb5Gm4EDe6JIL4YmX98njwXb+mOxXVe3lc40kLjVbYlT
-          +lO0cB7swVbPyJ0XCjo/AsWSbtx+9MXc4Oq1ydjNweQy/5V/DfX3faHAPCuy4QFoCyW8QzbPMELo
-          FYM5u9W9JHKaiQyBOwJej9NMrJbKQMkaCvXn8OmhuHseD0iVvz1dzDGUpR9+/PTk3B1vPNwX3+8P
-          b/N1Ns+e9HSVghhivUTT7laPUOUHDdnVYZeTpDkJcPPLUMidAjq/HRCDVvEron0KrNLjBzlQqS8+
-          edr0nW/rA0Wnu39QwdyGfPEHxv+rTxs/rRfS2AnkH9XvtAxfL7P1YqSnf78hfU+flFrwOkq/eCv8
-          +QaWn/5k1P6C9OUlDcu9T1f4hMYRl6+xzLndnBogULQ7epiTE/34kMQ/VxOdSirT5XmfGhjrlkX8
-          9N7ko7wyGujh+YzUMiT1uq/iGXofZjuh+2XV6ecXyHm1/Oo/pQThOwS6jrD0Nhjw9/6RRQL/lXls
-          Tc5TJUjX/h2j61Z/OeWp9FKTj1v9Vd5DG2cqlL48d0JOzNJoHetHBlBBJ6SjJol4Euw8SA9Ojk5f
-          vVbn72v04Ht9sH9+MFGP9ubPSoDcxndHacQ3BixG38DQAzrlePEsSjU2Q6R7lh2RgiktSe21HR5M
-          hao/PQAy+H0Q1YsDum5+EQC+FyCjlZMIb/xMuu7dG9Js6NRkW3/w4xPKA63u2quEheHCh3hpCaHL
-          w1+2+55Uz2d6Sx6oBZ8Y4qWxkUPbrzqqGlSAn51NohnUzjm+nkrozUaE58jWXbp3nj1AMD0hOzuc
-          8tbJvQ/0LgAT2TlOw1xKg/inn3Qm6f74GTzg5IL5ds+5o2J/FeDp4ofY5vOVj2bCZ4At2I5cxqtM
-          8WAuJZyVnkXGG1O65nZWwGuEE2TujmVEZacOYW7CFSm3inGbeyxZMDuiI7I98e0u9ncoweWiqEQl
-          u0vEoiVtgTuud6JnTBAt99fd+PkZyMx83l1yCO7w54c4X02JaPfg75C9uTVeziAHNKtfmnQwUoil
-          jR9y67m24PmShuS8/Z5Y/tk56JGpIatkvsNyOkwinDoKNz8hzEl/c+/AenUhunD+eSCzA+5A6eSY
-          ePPnPiyTls/gWwCKzOnSuMuNigG0Yyj7h/g7qK2StxngKqskkcC+3PE8Vg5kdsqCNFWK3b/8Pgky
-          u+l3yx3LOEig3Bo8BmA7kaFis4HnlD3heCUfMABqpPDEfg5/fvIcWHYjbnzFv+0sDawWeonwhMUe
-          IcNYInpJbqPI7OQF+flbz+cjdjWw6SH/kM+eumbbFEaJ5XRMbVJFGJDRAW1z53xu85d55XEQQH24
-          f4m/nCOwN657B66mfCL3ZyH/8KeBV182SfDeFdG65ZMoXh9f4sYY1Xv2LVaQA0FHTqQzKZ0lFcOd
-          2/DIBciOZmaXluB0mV/IT+9atD8SlQfULSn58SleDRwfRn6W/Na7Xs+pGYLyRVOfjUAJCNwHhvQa
-          dpq/Xqpxe+ICpDBZe4rkNdirq/y9CvA6stvUlGyoN784BvlBvqBA6hd3EjkiAuVeIH/Y+PbSgl0C
-          gqIvybGaDxE1O6GS5mqA5CS125QORgjgHU8DOm54SeLloMGjAXwkfzMEsEHYAFIlXzHr0zyi0XOP
-          4ceADrnEVp7/8cXDaz0Tv8DGMKz3QwI3Pxd/OcVx1xQ/WbDpHXJiS4ES7d0nMCuPN4SOqKzntStD
-          QM7vM/rz83QGjvCzJluHzKsGek6Y8pcP6BnrN0oFfezhHvomURdmoLM3OXew+XUYoM+VTsaiK3C9
-          qhfkb3782j1NT/Jo5RO7OjyjdTl09x9/R86SJGCquvkOv3XqbH7+blh2twGLL3GOifcsSjB/JPUD
-          Y/Q4ot/rZ3tX9pJf387Iynn0w3f/D69++mmBh0MDF9POMBM+JrCmuhfCdU17VFxjpsbjZcJQuUoH
-          4n9wWi/VXV7h96ClKPecHqww3a3wfvlWRBZZQNfvuGawb3kfHaGdg8WZWx9sfgvR8tYAGBjbnX9X
-          zUdns/tGa+S/gh9fIt7X/dD5wjSltOdfR3Tq+vOwvf8MgiYmJGWhkc85o3mA9d4FQauCwczJFww9
-          GZWYS2+KS5eUtEK9R3e8+iMP2gVHATiGQY35ZecNv88HbBIc/vxWstVvsPl/G74+6s1P7YFRMTvi
-          tP4uovGuXcV9ydOfvs7J/uVtT5SC4OcnDus38RWxeKS8v9v8LHrSoAX68iBhzpSXYQmHcwpPXn9F
-          hi8+wYpVVZbEhzVhZh3f0XrGoAcjkz/R6XL70Pn49lOQdFXnN3p6BVv/IYTfOnOQ03UBHT8pqWBu
-          25W/K0U9Z3/6KVkKFXd8I6u/ev2rZyQv7Ej9+ePA61MNg9LqBrKQWJa2fiXRVIP7wz94CVeBKKyo
-          0b/+lKbfY2Qul8Ql5eWmwZyI7o8/1uzGZ3/6wBdfuueuHskhLGpT99tNL82YHldJfw4WbgZFBkMu
-          nhl4b9IVeZdnX//V51/93fhMvVBvtv78PY072tFcH1dL6mvbOgGJ3YNVNo8yEF6HAimo5gaKmm7c
-          pvZcfv0OOrUa1KDsTS8U+xREa8GznpQns0Iu7HX95W8MGw/0mNv4FYZCyUpz/zlt/PFF17CXe3gE
-          2YgUabnUdO07GQSplBE7/rru1h8Mf/oTL5/HFLEGCAtJ/ogGUbspUnkg5SysDl8Wg3JnRcveMqwf
-          30Lnl9LRLojLCronRiOoFN9Rm7kUSterc0aKM2pgzmOhh1BjFH9XtApYWxF/4H7K9uRXP7GARA36
-          7mP/x/+LpeI0UCziBakH5gHwz+++BPVEzLHQI77ZBQ0coKP9/LFh758EFhy9T+zffn7tfK2yX/3w
-          uZxD6qrEkgc2P8Nn5OJdj7ul5KWal0fyaOlKx48hx/AYhjXSV0hc/E0MGX5gV/7ifaAQR5lUfouC
-          /PzGzV9XIPMWWqTuDDXnl0rSoKykLXET++VW8Tu0QPWQv8SrFwZ0mx//X04UgP99ouBw+aqYkeQv
-          wAAYs9SfdxC5c/IYZjWkH+Dr7w5p976li8zfHNjIJsD1Xb/QWfGPH0lpyj0JC9BEAzrtP9CydiJB
-          As/WiyHbLWx3Nevv6KFX6bmILOmwTot/gJENVqw8WvFaxxo6+R47tL7XxYD9uDYyG2iq+2JhZXiY
-          ze3M7ADzWb++G0gwS4mvYQjWhtgWjOhw9CnMjjl/C1RHagOp8Fn1pg1LJjgtoDGoiMxhiY7HB8Kw
-          iNa7f8gZxqXy+1JCiaw9soZnSWmsYBEy6jf2I989ustyKQupmbc5pPtCHnjmExqSCDWInlp3GmjY
-          UgyLGTi+cCnbnKugVwL9s0JkH5fXMB8P25wubmJ8flWbmlbXYJXY62cl+oMVVSo+CA/L7OghzZbP
-          EY0FP4QkCB6kiDRSb69PJZ6mOompvxsWL7M9mOm1gqFy4dUlP6m8FIOmRqitMvDhq9SHzjzlyGb7
-          xuWDe5pJF5J7BGXnAbRzeMdQzIS7XzM9E81lVhqSIzMNkefsli/Nx2Ch0RWez2j4Drik4gzJKB0L
-          nUqODIuvuSW8X1ITBYuWD5j9SCvw2pwlxyu23PXE01a623ZHnktjq8uzOqfSKzMigvD7Pewr2ClS
-          axQHXwTWASzNTFL4pE1Grihu63VMoxjejVeJ/HnHq6v2mHt47mcTXQTu5XILMR3YKYWPKTgK9bTq
-          twTe7HrCs9kcB7YpLpVkaxIi1vQwAWe+zyL02htL1D63anpPZQHCXBDJ8/a1a/b4QCPUWrsgCdFe
-          7vxOMQT3nTCgZx5a+X6PaSmxaExIsGhgoMbuoUFkhQfMqJmg0skxGKmJlBi5d9OkHPJHDe7Hh0yO
-          p0HKR1NIA6g01R7pinB1sRYFBhzizkMBMeKBf2rrHUqf+4c8y8PFZbs6YuCbFzQS7quPuxB9mwvK
-          TDqxi1TNWfNRNvD7JDxxOWKoY3q5eFChpu8HTq2r7Ok+3+HLd3Ry+tixuoxBpUjgXVjEB2Otru8u
-          bGAJTiNydDHJlyFJG6nzsxzlcyLVVDGsUAoM4Uri5Puge1AIrHSQk8XvVxfVxGkWVjrZQkPc4m7R
-          ZZeuvhQmT49o7NeK2K9d8lC1riVRuJPusjCABaxUG6Hr6If1CoA/w6MtuUibmWrYL06UwVrct8Qb
-          mI9Kj2LRA4pPH/SwqzCajm4YS4yYJCgrvas6w6FxpNcd3/FhkhI6gyXEkmUHDrGPy3HgH85Lls46
-          eiPb1A2XU9+nAr6hlZDwXZ8G9uFoFeA+250i3v1Tj0paObAqzxJCcfRW5/HKs1Ay3zsUnhGl5Ovs
-          NFhRPyH62XfcfWRdClhKHEds/lKBZRGFEYYRpxL57qoDd4dsL7l5hfCC5RYsfM3fD76FVnzoGRpR
-          g4UVXLJmR9To+R7W5XGeJVFvHaKs/muYoyBZf/8Pc0nlD6sRRpW0Zr3hS7slzJdz+/Th4fJWsfhB
-          L8DXmaUA83ZSkR0dKZ3N+/UjRTfrjoow2zpWL/cDzafAoGd54Ny///+J5RUZpoEA9l5rBX/ri5wv
-          m690bGKpKpsTSc0Xqemp5kJJ4aYr0ZZWA3x5A42QYKVH6pn9AOprpQIF7XpHx/6qqvNYPVM4Alcn
-          dnXsVPou2RkEhnjFS3+t1XVdMwFyd9FFnpqsw1pbRIT3nK7IKy/lsJwfTP+rP+TB2V207/Osh/C6
-          t/zRSo8qHwU0hO2O1YgHoRKxu2H+SMe9eSBqYvqAfJ29AVvlZqKkv2RgCSypgq4uTUTJzoq6PNe2
-          lFyICd4H7RS1uxIpcPnUJ7xja69mNWW8g/OUBij8nBDYv42TAX/1z8zQAYxTdAthOEwFCt2zR9fv
-          Q7rDhXBndKnqMsLbfkj5d2djweLYfGaVUIGHlSzIP8+DOu3oS5R0+8oQA8jbfWY9q0COMQbiyfLF
-          XYpTEILt8yJDLgy618XaAF/n2RD1up7qda3cGWbMVyKKHn3c+VDJlmQpDSJ5bQwu6YbXKtnu4CCL
-          1c8uJ/NnS6omqCJDM8d6WdFBkCgaEFJyKAO+k64OnAoJo1PMPsCYvYUVhp158eftFu49nlsPMneQ
-          YuDSE+CEZyDC/UIiZAEtzvdC0aTAUj4IPeR3q9K9PQQQidBGPtsCsO6/jg/tlgr4sAePmrIXOQWF
-          yfsoL+4t7YwgneGnO2t+GT5XSptjrkGbkxPiak4bjb/9eegOg8zbez/grp4Lyc84gK778kH3kkUt
-          yQltDTm1m0XdnpSyROIYkcvltboTKO4xcPMSobsVPVy+ZsxQKi1+JGe2Dd1VXw+z9DILh8ju3h7Y
-          4iAlkIqFRp4WvNRTee9ZgJZeJ87yOuW8cuVa+CyZM9GKizmwdhV+pAd4zyT8CrieaTWH0r1vb5hG
-          z3dN45BXYEW9hNzj2Af0aXA+/Ft/gXfAessGDMzvkqEwamLQT3tJlp6FuicWM4nqnHDgDjFDR5/l
-          YRcRx1Ecadi5M9KrWs45LeYTKJyGEslnbh4o0I4O+OFt3k5XOte3oZKGnT0TJHhrvpr7YYbA6SIS
-          jbNNuZSmrXTPlxVnz4rdPr8QiOzIaUiOWVZdg3uQSq8XQ9CxHZt81U+vEkpF4yAHCQ914WumgBHt
-          jkjJzpVLP7nvARc1RxJt/GK0rqICk29CiHsloKYVfMmSHxcBcnYrVOmcVSWsys8JC8yUqVO47ldQ
-          HYQzOq/94M5ySCxAX5aMQnooc/JK0CyC0VJJbCXE7eGr1Q6B7N19ODAfd/1i7XPwpzVG5rNmohWO
-          VvtbH595ZUq9P0fiB3JRY5J721pDvyuHSnr5lk7QSzwPLBpVWXodOAd5W73Bs5W20tG8KOQeJl8w
-          7wahgR/uFBIdP7E7cU/vDm+rcEBuO13B/CD3GWTJVOHuNJGofd6eGgy00Md0URc66usl+PFhdATf
-          S74aUpjA5jYilHEOVTG3JAnkn2cXU39+5vRuHlYoC/PgQ+9oAPa0Lx3YvPsHUafMU3nnaPLwfC5U
-          zPbnS76qIz/D9DWEyLKfSb0mnM+IKRvlxMslNp+zhsyQRTghhUZlyr+77AOTAyOQk/xu3bXHNxly
-          0XmP5/RxAAM4tBUQfejjAzfINUd3FZZKa5uDJ9CULnHhGsDuDIugMWgGfCl1Hmi3t0kSn3DDtMdX
-          BTSzqxD3+Xjn48ZfJEbeuTg54YNLIwmy0DQ+ItbqbnTXh/NSpFKbRWLtczyssauNkqM2k8/Vp6qe
-          8BN+4NRUZ2T5tgBWfzlpoKwvZ6J1gQDo+eDwsARoRMq+MlwWt2EKTJVZkMEe3WF5w5SB8UggUb2y
-          iqbQFyoodQZBxnRTIh4NSfXj1wTpUqhOc4M/oOLYp4/vx0xd4as1wNR+VGRUSki7k71q4KGUHDGd
-          3ojonPUVeLVtT/Tm9Y3G+FMK8Be/UZ5HOeuEbgG/J3Py90qY5etRYSsxOz+DreMsDyxz2Vj3XXCR
-          e15ius+TnfDTQ/4sl2KO8erfYSBdbOTkB0xHGMA7RNrJJBY3I4BDsGpwh50dcculd+d+gqVY4EhG
-          rmu7w7L23QdW39BCR3fiozkKihnIpPigIx/NdG2KVZbw9Pb8WlvTemHjQYSKGl9J9LqJEXYeVrWN
-          L/r60r5w8tk6ne+idWVfSN30F3sBvQzbHa8RA/SLS9vLtYU/viU79dvd6sUdyCBlkVbsbNCZF2eF
-          JiWtv+v4XbTsUtGHgna5+zrQC3etm/ID3p6JCOr4Z4RfqiTC9MS2xDy2gkuKBcogX84+un0GN9/f
-          q66AXPQx/c913IGp5i9QKpz7lThOk9ZkETUPRnCn//CkXs/eMYR+Winkcno3Ecn2iSM5z1JEJni/
-          wCy9pQ90v/cH8aRPM9CEOYZALUtvO8B5j6ZN38DiiTukNdIM6IbXcJaKM/KY78uloLul8Fd/vZOi
-          RuRUSyHg5pHDO+8NcnJ91Ao0xVtN/G/aRQttqAzXgf8g62b7LnsYWwzVq7/dGv0OomUtjzL8fsWK
-          HBtGHN5cN8kwBp8aefzhmlPR+owwHieIfvvfNgfdg/VxnfC86YMeBogBQU9khDZ9tnfjCv/4FNKj
-          2lWpjWYsHfr9hZwEawRYnmj8q5fo2fHbCaohFODGb/1XQBqwCks1SxHOsL8T+HjY6mEhbvGHaZHx
-          ADfxrfj5BeREuaamPz617T95CCtxJ8eeWqjP+pHkZMx//L+FPEcYdLzcTHXDc4VhFjkkx3cLKSX9
-          0sNc/G5TlZ55RCfHZ2AX9RgZiaGr3HXP+yBboUz8SCPDPI07CORACNB9gpO6HoVIhL3ofNFp5I/5
-          3tJMB+RL5ONpynfDUp91DIrn2PmHpL8N9OUSH54bzUPOBx3BXNtxCsIlOKKEHU505c2yktDj5SL/
-          EBbqQtD7DhMs9+QhPiY67fxbCxO1NBD6+ok7bXoXSGAUUBZ81WEfTe9REn3GJ8cwq+s5ynYJ3BlB
-          Q65DM6vjztjuHCyfIzE2PjuTamCB+hAUdO3FE5iXOxfCeNdeNj+jcPtU7xugUN1Hqs/V0XxUX4qU
-          3lYe85Ei1YQTg1aqq93JlwIH1PhX/9neA8gnR+IuhJVG8HnEPAkefBXReryJP37q7xjn4C4ha83w
-          80h4csyvJ7q4umVIbp0OeLfj9pQ+4b0Rdbv3kTXSd73uBs2Amx5HXrOrKPEFJoXS5Lz8sW3bejQf
-          bSPVyKmRnrlxTg70XcHXCxJS9NcRrKLMpZBi9PGZ/jrS9ZbVI4RG3yJ7u7WYbn4CqLxHRLLkvU11
-          a8pGWpMmJsrmR236DMOEhhLSV4tSyl6sFJjhyyL3y+40zPYILBjxZ8dn7IeqUsdxLEgSmCKvFBkw
-          t6oYCLqcOOioZn60nHa34qdPsXhrTpT3XciL4ZMGyKleqcuOdZv9/CSi7+9ChPf3Vy8lQ/ZCR0e5
-          1uvG72CFzDtSNv2wkP7hQI5UMUJWGEUUx98SsiDfpkjgOJ/H6pqJw1h/kVkwssvfepEF+VeyEUJA
-          AItHiQhE/SLirc7TZW5wAymdbXI1PwzFlx5n8C6FFfGEE1TX60dmpVNaAMxseoRGEsvD+yu+ELXb
-          phJdFRr/9NjPr6mXxz1J4ab3/DUWDgOVsJpJaWQAX1gXL9/0bQEfQJRR8KHj0J8/M4RPnJhIfr5e
-          6jgJXQnB8YKIdhv0nO2sJoTgeXij28/POLvjDMe2gkTp96u66PrsSX58D5BxNY/1yoDcg888x0je
-          +Egb6owHbfObEvdkhjVho10mnq97TPyfH2P5BIu//UsIrvPlx/9rkWuRYramuoqfUw/dcq02PqJH
-          a+x6GKodGxD1dcvyhWc+lnS96h9kHHZhPv3wj5+4yqcxdfIf/kN1vN+QK2B72EsWcED4cE1kzXc3
-          +sxW2kPt1B+QPUk8XTxUttAD8YnoaXkZ1mRHSoC9bMTiFq+k5tdCpCU3E5fT6+2shN2AWDmlyILp
-          QaVXBcTw3vc3IjOzpy41f2H+/J5FOO9Bl0jjDOv3O0HxvbcAd6GeD4/tqyKOV+0G0rtqKXna1yUb
-          2Ea8qXgp2OotUhX2XPPHx+hB1xIMlAZeTek7uXtw4zdk47/uJNqRB5HYHci9DGA9K2nvALS0OlEO
-          +280mU7l//lHVuCpdPObZCm/pSvRztEhooVXz/CWpjG5lmFDF5hOM/jpTfMgqu5i73RF2ql9jjTq
-          P+vFeg4JONliQ374P+JTNUpOQ0+Yr6dzzf7wTDPgNqXBK9054gUFng9J4XP4iVXylYsSXCo1Qf5R
-          v7q8ez1DKOmBQfw9c4vGyZox3Pxt3PeHRp2+nzyEC5O3xGUU392HafOBHlAlX9z0HufRrwA3vYnL
-          Kb/Q9cHsUvDzt4+J2uVU7yMGvjItIk/nG0elu0wWhKekRsdI2+6Eub0cqblhhFyOfNT5WKQYvhL5
-          ta1XTSnXH0v44xNRt3zymVpyI4nnJkXy8PAp9zlnFnTrbCDH126vDpAJKyiP7IOYW/3naJYV8AyY
-          L5J1lgH9FDYzaFebwRyl4YCR4QcQe+lIrPOpjPbVa4mh3rOYxI0UgJ9/CZJ0TXHrstdhuu4ZT1B3
-          WkUsVl/c4RO8W5C2lxO64w+JhncS+9DmlMSvMq9Qp2R69JA82ghdxByA+TC2I5SvEcDC0cP1aEaW
-          AjSDYYjxGS/usqJFhJHz0X3w2p4oGYNeBhjJ3C9/8uXUdKvYXtSAuE1GKM3CkwBh2PCY3/gk2VU4
-          gNc60Qiih96dD8MTi35aKshi1CvAT5TL4DOAEzJPXUdnz/jE4HSvJp/PM3tYC/HGw83PQub33Kqb
-          3xJI91dywQem7cH3Lt0cmEso99vM6Wua1LcGRiff//Nn+MaeRWhd/JTox5MQLVOuND99RJLrUtP1
-          /BGYHx/F876Qa46/XUT4avses/FVBt2zumXQ3SWCL2poX1N066xfPBN984/mS3L7Vz+uVeO4bJYH
-          vnR6OjzxK62h81iXKWwPju3f2m6NNrz9QLW2LkR+8Eq+PxJGE93CSsnf93N12YCjm/lE3/QTfblf
-          H+iXzx3JVbl1vLRSFlVuLnzeJ5dhbumgAcmFR/I8fa6U5m2qSHJd3Yj7ITOtVe6gwf1D7H7P2EaY
-          W4oYaJfXBZkXeAazs89H4L3p4B924aSOWfNd4dr5D7zyQg2o3ZeplA079peP+frIBAdq/MQRndJ1
-          WE7NawVNJMfk/o23/HiWEG761d+Hkh/NuXgQYPbIH3jm6IsOG58Ud9+nRlRG3yyc7BbAI6NpaItv
-          Olut7UDzSzO8Di82Wg7M2kAlbTrkt6OvLvX5hOG9lQN0tq4pIJ1xLn/8n6hh0amYQCKD4NSE6LQv
-          +nzmTU8GpgoXLIrFfqBBY49A/8yQ2O8eg3lntAnc6jdRP6tC97/8l69nsPkjirvkX7Dpp9vXJ+96
-          n8+H8yJLFmJ0X1rKetj8URFSdVJQAMxPtDKJw8DYAhFxD+nwO2eCQR6O+42PP/NlL+w+EMfRdiJ9
-          u0LwUFnOn/+y+QvR6J19AVw7QojlU+CuW38P/vIlzTk7Ym2zEuAuL/lNPxK6AF/3JGiHC/Es6RmR
-          hdQ9VOAtJM8J1BHNk70Ii07GxMwDPl+Oc9XA+irvkMEOE51p+LJgFus3YrV2Ha25qGdQOQuYIOcb
-          54sRBKu4rI/cX62wq//098ZH/HXzZ9mxLjPI65JJjkWyREsWiikUP3z/698MG/5Y0OSvC/rD630s
-          JDChgeRL+sGr9/dUFoFh7u6Yc92nuwzilYfcjLlf/xLQx+T8208096VEF6ZoV0hTKUZ5ZpeU7lJ9
-          hLyaMFhqzdylMtf0Eg9HG53sao2mnz99BvCLLKOMXNaFOobr8bgipfa9iMuVIQRrBF3yjCUzmoXF
-          D8G4cMdfvuRvoVvuklpWHjllp0++Fs/eAX0jXogj2Hw9d4zbQjOsLSwxRlRj+Co1KRxIgVRGvqv7
-          I+EN+Iunzc8dph3tBLgSMSVqTPtodb8vQyKsaWB2jMscn+4uAxO1MjZ+zUbUjJMAKjoskTbGZTTv
-          j4/7n9+a1CqJpvJesVIjb08sVcfO/ePDmx7z9+ZRHvhIHrepsYeeyJqvqdOwyIHEi5qMHl92pfPF
-          cFooRfyDyGbzqmmfh610L27AZ06GBbb6L0DdvjCbnt9Fc4UcBrZ9VCEtG190scRchqQSPsQLhs79
-          9X+gEYEeCz7NVfanhxx6bfAaC7ea14Q2hpF3YPBKD3JEFUMO/vi1v/Yx2PxjWVjI/kycryjXY3DU
-          POl+8Gek7tMVUF13MZzD7Y7lAvjurFqRAnf1LSSuzpnuxt9baCQ19RdHuQ4L6vIZDtfo4i/hZTeM
-          +HFgoapFmJx4tlUpBtYM6vc38Xenzx6s6v6agKfh3n3p7Arual31Cl6GS4bc4RapFN6cAAqXF9zW
-          t6z7zd+G33dwI+d5N0YvD7Ut7Nw5RFcD3uvZrrLm11/c9k+O1oYcHWmLb6LvWC2f0asv4OFcn/35
-          ZkJ34RlsgR/eu8bTjigoZh76XD8Ty+Li/I9f/ocTBfD/nCggI4NO++8UzZybVHChd5GY1qGL6G5y
-          DaheSplkl57Us7b7fOBSx09klCVPF4a7FdLnpVFykU7mwCb2d4XhS5hJNAEhmi/9MYBoPNbIG5WK
-          zif9/IEzFo7I074HsCaKOksS9hFemtXI+bQue/EYeANy4tFQV707CrCvix6dUDS4yy0oV+mVhQU5
-          Wch3V8AXLGxYa/TjKDgNrCC9eLjH78AfmAYNszZ7PgiC/kvMEVUqUbJGlIzccJF8uCQDlbt3Azxm
-          dMnxeJ/q9UpdBc4Zznxutkx1fu7mEPr1WBLkXL1o7NjRgOryRf5rB6/DxGhUhJo93DHob+42f4OU
-          cO+cfVTszla+H5WDA7XHXvUPZ22hq/bleCgT6hNZukrR6vSYAUnseMRwL0fK9V31keJaKcjjW6Bh
-          6ezbLL2tOCJPOznk9CjeFXhyGBnZSzJFsxPNH8k9zjpJAU9zcgrGWYgz2vvsUFY5e0osRZKB0JI7
-          sIx6KesLAwe6Bz73AjBa55ltpLesl8RnAi4ac8jcoWPtMoTW6ZKzdh0ykmXJIUqGJgKTwS2WBN+N
-          i7xjfR8Whj2loNCCPTm5Zy0nK3cpYOYEPnmcY39ge+40w4o/NOT+4FKX84Q+lLSO7pFr2x93CfoR
-          wgpzCUHvD1aX+BHK0P9+DQzNblaX3VvDEnEUC5mRZOR7mIsQfLluQA4HB3WWEJfCECs20T+yN3DL
-          I1Ik/1kJRFtfvEsnOqcQnrOGXBZjHqZT92ok9zWJJLgq17qvbm0Jj+jQkSzt4DC/yooFzsM+ImO3
-          HgGvCUojbfFIkvS2A2tpBRnchaAmar/P3WkURSxVIpyRSWymXvd6GUhpFwKipmsf4S9IUonGxYk4
-          2dTlNG6wAokYMOj2cFvAG6nkwBe+6+TOmNHAG4c5kVgIBnJs1k8+G8d7BjOmCTC1Oq/u0p0YiKi1
-          JBLO91c0E+EVQj2rYjw8Yqf+7RfUjmJATPbcgnXfvBWpzayIRHu1isY7o5RQFmSFnITvg86IX7f+
-          +GtGWUEClefpG0uecYTEkV+vmu8fDxmGeguQbpDGnVKistLJZz9EMz1ZZYfPyYDxLgyJF8/Xeq+A
-          UpPoPCp/68NRXDtSFiZ3/8UdTgPbJ0dZqnbOFclX/eVyli6FUDsKATrGjRAtYrh48DP5HsojL3Bn
-          fP1o0GrPFiqAJ7j06LS8JAOxxYmdvKO1Et6hNAi+5R/E/AQ41/JGeMhfV6QvRjBwZIywpJ5d6r9n
-          6+uy5TNZYcfTG5G5VFbXSxhUUrc3MuSt2zNcnZ59/uJLeQZdTfi0tKQXf5mIkeQ62OveSRRHOUiQ
-          L+OYsvx4G8XpoR2RdnjTfAFRJsK//dvtji7bxSyEK6AS5pXTnq7n4GhI4og6coxeTE6HqgwkqzFf
-          PjCtql41wfnA7/nS4Vm6PvL1O7CKFCmxh87XDx5mxH0K2A+I93eP5puzPh/w0tFyHv48nW2V14Um
-          lD7tfuez6cXPV/OW8HDx9zx6CrRz6VC1IRRoQnwhab7DIpzVQjIMwvifuYsHTngtguSexZL4xOqH
-          SdcWUYr43scruX1zvOtcAXpiu0Nac9fzhS2STExG74bulXAGa/Ghd5j5Dkt0xGoDbj1NBM+32eB1
-          V71dMutWA19HECB/NOIIi3vbl5RmqZG10C6np6OTwZ2zvxDlHUWUe9IKQ3nWAnIlXqvO2Nj10Mg1
-          l2i74UNXgYsKSZStnJxPdTMQ+olnsNVD5JqWMizY4VmIEBMT7YScaP/oUwjv0blBpzgCw3zSiA/N
-          HaMRteH3dAQFr4l2A28oK+do2HdHN4Nt6L+InghHum71QowcPUWXQf1Ey3C7sBII2AklwgCiBQ+i
-          Au/5KUGIDgwga5JiqAg7B6kH7U1J3HwUaV8998T51ZtHkWfwl0/R4XjPaTU+x9/PKETmAIYp5Hzo
-          PAqF2JVfDStiZkHiefOL/OJY57NcZCmstmcibw9rl89bvYfqI/KQHh/0fC9YkgYZ5lvi3W690OX+
-          EgTpnRYY+Z35zanJnH0omocvctzqok7uF/DAH54jcZvUBH94Vr6HnGjd/TKwzPlYSAkbfn3R/iju
-          fssP+GisO7pY87de87egSYcv9dBxZCeXcuaCpUw55r96oI7LLecB97iY6HLmWneuq6CRwm6AeOR1
-          j/LSlc7Sq3C1v/1fJ59X4PA9mMi8Huth9t7yKD3PRr3dCfGkfPl9WLAt5AO6tecRzHNqztLjrvkk
-          lz6GS1VRGuHv+yqH1wtws+Q4EAlRSoKIyfKVVpdGglKbosJhrZwdeVrCB2JvpFDWV8134wWLxt7O
-          iHWtzi7b1lwLs0+pkqdED/ms4KSQ9ORhkctbqdWZ2+ZEb/mJ17OForXp517iLzsPd0OkuqytOha0
-          JfmDrmug51xr6aOY+RaLcqH2IjIqxiyV0tRiYQfUmtu9vfHHF/DhZgj1Ih+LDH5u8hshdlFyzvr4
-          Gsj6tEMOCwR3TaNIkEwyqFu+WOoWH5mkpCwhz7yuwcwEV026XQaXHJ/bLW8f2o7wh29nu4gjcnCO
-          DvQUO0FONtn5HAvPBnLTo/KF+/dMJ6DOFVCl/IvnygI1EYGVSXktY//NBK+cKtkoAGt0MXGrqxat
-          13DreL/Rk5yed4ty8gIglLCHUPg5P6JxW+8/vD6tDa2XZyPH0o4mO2Rv+DgHZif/8ocY+vnirlnT
-          yPAaXHk8h4qZz8HnPcNr+umQo78WunjtpRTH1JZ8aI2OykE6jOJCziE5itCM9mW+QFhl/IVoVSrk
-          49y1H6mAc0mej8bM2SfKRvi0lQvODrIV/eGjsctOxI+hAvYfcithdNoxuIdSAejv/Rak3JF8OcZ/
-          9Rxq3bLH4Bl93UmN1gbiWzQQ87571FP8yGSoVfKA4vza1MPlvRrSxh+RewtVwA78mZeuZb9HqGtZ
-          gJ/JB8OZh4j4Pk/cpeZWB6ZHIyJOw9V0nXbR5w8vZT6xh3nkaQVmo6LEFWovX3WhCcSdL7xwvzsO
-          w5pzWSEyftyTYMOLvdiRD0gn9+XPNBzzSX87WByggZEsZQZYrsUjA124psTrIFK5ri0qOEne3V+R
-          6dLVL1oDHGZZxvMSETpn2zMSYFxKkqJcdGdesizw0w/Kxlco3zolRI2++PDBCe7kCFYDnwGQ/L3D
-          3fMlOWEGlLfx67OmJ7t/ePPjr26Xvwb8hfIKpF5ukQX4KJ9Qr4bQvDY8UkMiu+xn5RO4mv1rw6d3
-          jn/4h+bSQw++U8H+TE8y4M4nTAwrcuuVrV+99IjlI7o6gk7n/D61oAvnlIRs0br9xudg59xsnLpF
-          Xdf9zH+Aaocusoy8o798gmbeK8R9xH1NHLL6sN6NPlFr3oxoYe8LCNL1iTx/mvJ5MntFnMP+RHTO
-          LcHcNdcYemFxJRs+gFW5Bgbsv9cAr6nn1azDHlMJsemKlObiRIt/u7MgqoovUtGg0B+/A6zMP4gh
-          DHn0Fy8JGmpyUuwsp87TneGAxgJ5lXCmc2KTGbzLCiKkRsRdm03BF95Fw/AhvOrlxXop1KZdQZT1
-          0dSrj50CZn3WEQU8WLpGpe5A1Q5c4p8EfeB0djIgPqLChw/hOLAnjXgABPzkr6YcAxrFLwYag/ch
-          WpW/KF2ar/LHx83ROYPldHCZjWtaGPBVC+gPb+v6npHHhs8rSzUZnmqrI3nuJ3TN6icD/bwN//jI
-          8k7GEcruw0M5uYF8Lp+/O+rm1hcmfFC7QTzL0ndZdJKuSwhoNV5HeD6rgg+bwyVavP1OhNoNPki6
-          /T1uJ8uQuKm30JEXy5pyppnAVme2ju7HiZazI/HgplczMmMpAct89xmxq96rf0l5ANZruHric4qv
-          WPqHtCvZUpbn1hfEQDpJGNJLm6Ag6gwoRFBEaRLI1f8L3294ZmfoqhposvN0SXaMsAX0YNSlcuE/
-          2zvm24nD3e3RwMDjTyS4HLVifVzLCgCrk5EwmhkjO/gqYdPILbaSeBrWpydUkH5bfl7na5qsuVUg
-          eL7n07znz0rC4iZ7KqcXy5Agen6xXO+fCnLf2x7b2yPa7DVHJbDsVUS7Z/9s6fnG17DixXnmBlNJ
-          qBNzVKnieiRIQfPApJsuwkRpeDSsdRP89Bv8g+jz43NzTo9WqU4vH893ld2K5VTSbB90zmHz612w
-          BG44Q/2JR6KXA0nG1U98KG1plWbJYjsEWj2rSXeOSFEeKRh//u0svTNy+AQWY7cLv4JFqh38Fxyf
-          xXpEqqH6N1sgGo0zU1yXLIabH8ZR91cO09Ope7BzIMGFcMfmcp61Si2T0idBmenBP76JefjBxVZf
-          d1GWOVgOzYW4f6+i+OkTsOk/1Or1Ltj8BwLq1+iJsWvsQGB3l8LrdaeQyI4fAXNnq4SH7VWGbLfq
-          jP/p0eDMHPQsLtZAd8spBvWkIIL5pUnWdclO4LTqAUYvu26ZtlcyuFI/wJbn58HyzExNLWbJQTwO
-          1GTJoo5T9VA3kXTc39r14SoOfF5kad61r++WBxwpnB9fBbsbvk5b/gLC+eZjJy29YuGJxateva74
-          AM4vRjLH5yDQ2pGY8ZAUi4D/RnhOkm3H/VSbq+3uLfDDp0NlSAW1dnMHjp4fI6mzuKD7jrscYIqD
-          f/6M7T+2DP4uyhsjNSpauu6fX5VTZmVuAts2f78XcjN1sE00yNhvvGO3u+HodU/M9bTjY0j92434
-          DLxbCneOrLbBIUBU339aFu7EE1TS1CD3dTondM8FNfxkPYdzsq/ZGleXK7xO6QFf5AS36zAHFZxe
-          Lt70mj7w81vh4TdJMUmv57mY0r+TAdN8+eIsvuzN5edvbtPmYy+EJsvdtF3of67i/OaPLqBwh2Ro
-          NwVEzGllQLXqdIUvKzlg37mDYlxvag228UfcQ6PJamSv8efnZ0nNvEIq9rcvbIfmTSw/mILR1UUH
-          Fr5oE79dPcAscM8BiZQZe+TssUU57RHU7j3Dh7f/SFjcuhCOrtlsn/Xilw8p23rD/qFGxY8f1Doq
-          V+xL6lysxYs60LiKBOsvShit7xlVjZKciZ0r7rD6wTcELp83Gz+AZOWNrlQ3vkJ0iF/t0jl3DmqS
-          VGLzPEemwAnHCuDYoTMwzVMxPk8mr6QdsPDPz4pL/RZhu5sRjo7pPJDxlDTAeHIeLoXtTvqWp0EJ
-          BTzqR4pbqk1w3W/zO8/6yBWrq5wMyMP9gL3+GAKB596h2rmyT9CDu7FVZNP44wesD0sS9B87f0LD
-          HVPEn05Pk97G67h/nPQBW7fskdCs0DqIEzUjKBQCwH9e+gyVZ/WHvZP9BPSdHEIlliWMfvhZc/Ug
-          gqwAW8+E1zsYN30MnXz/+ulHMI8oTdXhDQ7Yf8xqwJQ82pRklqKPcOWD+V1Ezz1/NizirWkG+F8e
-          uvlPXIQuC1h8gkipX6KOVGPbYafh+IT85BwJOitjQUO+deBVrRok/PwC6dMaLmZo4Z+/mKqTzcPn
-          4SVihKXGXEpDp+o9VT1i75AP1sLiFWg3XUPctFYYsaBO1QzqDJcRMbfvM3ewQPNAHPJ5mtM6XSw1
-          fvLon5+QfvncF9QSzq+fsuUl+EdhzyUE46QM2fg4aiWk1VtDyl7ri3U5RiXsM/tArn+9BdZ3efvu
-          s8O0zHx4rNm2npRfPoaP5/fJ/PEJbE7vZt7vqkMr5bLdwVKi4T8/NfYWasDP/2sa0IKlPD59FU7n
-          C8avDgV0uSUifPiuSW5/gQukXZRAVY1WRhxxu7pMrSqE7furz2tonwfWRjkHK56fcWA5YTI/3qII
-          fn53//PL9seT4YDnClte1A5094ga8Cw4E7vbeEybvoXFzj8QFxd5sJCxmEE3hSEpG3kBtIxGCPo1
-          /b2UtTD6WssvrMreJnd0YwHz8SOHLr1kJGQ6aRkvBQgs4DpiY/M3xJs7H255wsw3VzlhxTWrIHq/
-          HHJoDu92GYdrup/PhY0j8uoGBoe/FBhP6JEY3MxE3Ds6B7muTYj2DN2WvbbOpSt1A4y+9nafpqQp
-          rF+fgrjoj7bjuh+/8AUpRQ29SWDZ/CH03KJFiv8cWj4flXz/+/4neruAn57+6SFsGp1RiIKp1GAs
-          G5+gwcwLlvcDUr5t+cUxO43JeHrjSlHF098sp61V8JCMDQAl52AD1gjQN29cYdrt/9OTtGsLCI5T
-          dSLhc38u+t96GRLZwac6zgNaEY8Hl5HG2IwHlkyh3MRwe8mOuKdGYwLylBLm+aGf4cQJbNn8HNzy
-          beJaeTdQb1VTKGUP85cfJSMP8lxZ/F2Mw+vKJZtehPJ56Bvsi0o6rJt+gc4Ov4n+mD/Bqtl/T9iS
-          IkIf/pwX62Fvc9AQrCfx5NxvhS3vhJ/mvc7sT7ACnqtbHvLLIfnhh7mcxnyE8eKY6J7X32C0aBjC
-          X/7k0avJhFCSFLjlq0T3H7M5ISXWYLdECAdzsTCaw3KFtcCVOMoubsLD2vf/6YuI9togqWHcKRIn
-          H4m3qw4D3fwpqPrnHnux4/702woTsUdYS7QY0I8gu9Dft0esA4UypnkphT98jaLTG6zn07UGv/wd
-          3dUsWZQm6WAtwBJrNxMES3hlPgTq9soHfbXBT6+rnCimRNvqcTkB1YI2x2f4/ulTsFTM0yCrbzqS
-          PGkFZPdXj9BNTwckxnnTrpcTFuHmZ4nf8HVB6Ot2BQotjkTjTxCQwuUddRs/vOm7gAZpfgKVlh+J
-          v/H90ghxpfJX+Yst+mrNuS/eMfygjmK7kE/m7A3HDByK3iBHjkPBVq8hNNw5xc7tbppiu3c48MOT
-          bEfkZJWFooRc90hw1D+8do3v+gqFW/XAbj0a5sgUJ1UFL3Nm7q6KxVQZQwV01/0jRoEawPrYcoBp
-          ox0J8OEO6CydQ8A1/TQLsh+B6VpJ1v6Xb2D/Mm6vJSSZuvlpJMu8ag5QkWZ4/AuOSDpzHhB+/uV5
-          USQcGZ5SLLrxrSB1aoZ/+f5cvuwZtkUeop2VfUyWvuJc/eVFZTBr2/83leqNckA0+RKZJBNEH2qM
-          2wR+LBTLLe6pcgguOrFY8wkWeUljWOzcA74Wbz3h5XL+At6aTOyQLm+p2yyx+sm+3G++ktFZFg7O
-          x/pOLjSU21Euux4oml/M0j5thpUtEoRn0rqIWe8X+/789ltvMMoluTJJH4eWUp9bDY17aw+oc3y4
-          EAd7F4eq9w2os+w5+Jcav/w8KSYjuSrQS4frrCqD34qGVMtwWq4WNidw/fkvHoRNt2LtoIxszAr3
-          CfaDeMX2Dn0ZKz+WLP/2kywcqMVaXW4WPLOZ4kMEzWBVMnpVf/znqOyVfB58mMPxmrbbeNbJwiNo
-          wF+++y+/dx5BrXyx+8Ya0uqB3ivBATl3+2IUS+vAuFA5gdAIMjQm01/LimtVKr/81Nr0yqp1MVWM
-          mDYEO6cy+MzDqqmB/R1meFz8dp7vz0odLus4s5S2w7L8JRqMF8ucOT4KikXC0lXdMZ3DzquNWz5T
-          ogwoY/TB/9bHBTED/PRCVN8NJirdaYW//YpAqk/FJNCrAp+HtcT+l6nt0rvRCMlNarDjpzNbLtX5
-          qj7JziI/vTSbIlfDJwAy2ut7dZjImIzQnJ4RDprLs6ABuxuK2Fgd2vLNgoZWVEO9DCqMZeaZC494
-          TT16bky0i60HS+9SSwXhxyap+s5Ntu3fQa0VCyRbedeuAXxUsLajHcZ2ehpmVNUW+Pnj8L63wHIK
-          PQWeLyWYIW99WumtaYZadCCdP/cwNOlt6FJ4UoUW1Zu/nJeaiIA/axbOyMcKmBO+O0gq3SN/234x
-          /eUp9fmh4ezTa4z5d3MFX+l7RB2+uC1J7U+83z4Tt+kFk5wtToYvfv7tXxvsl48A6KUZvqiZl7C3
-          pmkQXaqB+OELm1s+PYOvqSMcfK5ysOyPk6/I4ApxdAjTgF8/dwX65wjjgHuSlohucYWbHsWes6sB
-          vQHm/MtLXW/rSR8UXQzLS0qxcdBStvaTa8GkinfYS02OkZAWCvy+zzFxDvHSzr98wDVVB0fK4A+i
-          cQhS6LwzjVgoYuzH3yDQV5tY2375etZ24//nRIH6f58oSAd4I3jaN+2qddwIG2jBWYDBh032YnGw
-          /xtGYsUFbnv4F1XwHTQFgta9S2hu7r9qXmdHojV5GNADvMkgeWqEOF9jZgvv1y7MX0WKYLvoLR/H
-          twrqnlxg+9JNxbTrb1f4FOt2PoxNEEjyH+wBeHm3+c+Xd8NQT34Nj1u7bffokIL6pdapWeX5xOQe
-          h0HqRfOpysGpJw40HsN0RUkHCzIpSEU7I6FVE3WwcFGBkS8YplTr2ar22xk1jTuJw+IfvS9su3FG
-          SmmuyUKEx3YGC4izKOufgfq0XIFqqYgcydNqWTl7CFhQ0+clPB4LmuY3BC+v5TvzrWMCKbN6HrLQ
-          4nFhxjngsZ1sXdClE9GdZB56/bRsXRAtgRjPr5pMRPjIkAfQQ6LvPxOG8jWH82Ro5Mjx1PyU47tW
-          q3RJSF5c5GKxyobC/u8zzo9OOjEpLtxcHSNzIJFzoi0xTrKh2A9kY5S6bcA/5GulPm5KQHTz+zVH
-          56SJKjBmFYf0pjHp0BU1FDXJwIGvNQGrRlhDoYtifMuBNwjwLeTwTxJtbK0ZLhbzOKZQ26UePrxC
-          HrC7tHKqN/0txC6jtGWE477Qdp4JiY3mMQjT91zC1ZEXUorZUIhfMXdUV6MYX/LeDPjotabQnss/
-          UtxE2LK40K5q4EOfoKtxTYQTtk6quW9sfLwmj0Kqst0X6p93PI/YuLNlPV+fMBmRNa+LWgXCSZEN
-          +DL5klyb3A3mT+hpUINdQy5v4RYsXTZAWJzrlFxvOhxmdr5aqv+Vv6Sy/9BAPqFbQ91rKdb22zt8
-          UQ6u2yGLK0nuygRoI6oUfgyBkkg7d+b0fJS5au5rGxfEPyf8hehX9X4KQ3IKiou5zNyuh6Ko2zMc
-          RgRWJ777QF94B/+l6isQtY6boRsNkBjpLAcseJUibLCFSPB3tAd6u/UN2O0HMPPW7d6uz4vsgrjR
-          PJIUiA9Y+tJ81ZvuC+L2J7GYiFbn6pkKJ+xeS3ugr6AzVHhqVqLLwQDGzJJT8CoGB+saAOba1DsZ
-          6jurwn9qoycSUmxDHZQvIRYW6kQKwW2ErBs7fHp+1WL8M0YIu7Y/zvTmtsmwJrMC9QtKycFO/YSq
-          fSDDdfFk4m7jL8TmToRdrHzxgXu8hvUiVy5cwuiOvRO6APEmP3311L1FYl4iaBI/XnuI/owY33Rd
-          B0IbriFInCHbuvJeg8U+rpn6KZzXLFrOxaRN8+mgcr48cVSG8rA2YTuq+7gOiS4eqoIK7GKoH1ln
-          CEYLCaS7sZSAT0iKZF002Xqexxw6NZejfLj1Cf2tDz5WJ3z6KzxGEZUpbG5HgQT6zg+kKb0/ldYZ
-          ENa+51siSVtXuyy4tRgrtTcIpjMpUF9Eh6SfjzqsXTBpIM8Fg0Qj1lrhZNniD3+QdDseWjb6jN86
-          qVbExZ7VSsvnb4Vx2SjYaa9rQl6Gu6qeVj/xpXv2yfy8bGd60bcnWHCzgHe6PwP67qElgVIHrUS+
-          TaX2rQ/wth3TLueWr2BG329891KnoDK3VOrYay6utnpZv3rcqVo6X4khnY8DXxkKhOWfnZDz0H6H
-          9bxvRLVFfIsdnxQFvaHGgngsFBzd33pCqSDncP/6kk0QHAq24Tfs2u+ROMIpL+axXkJ4meCOHBb0
-          BOtrPKXqJf4ciN/bTUL2PXDg8qzO2Hnf/EIyy2RVLyj2SR4eIlPU3+EMv+pXJ3iubsPogQjBs3jU
-          ZvCSw1Z4Zwun+vZ6IJexGYKJaP0Vyq3PMHZeF5PFt2kEG1+gY4+7dknnUYajkOrEZNIQUPS8f6H6
-          Xs/EjydoMqv8rOAkaTpOh/EEpNTKRXi43vczY8GSMIGzOWAN8IHvR5wMW31pKndPeOwXFzmh7WXv
-          7h/7t0aQu46A4dtlhPHhesHh8j22U5SzXKWSA8nhcOQBhSufAdl96CSrRnkgSmX0UA97A8cH/pvQ
-          e5iN0IKGTpzaCgt6vKkhfPXcDq3TBZv0LUYWBLEmbnhWDivd2b5a+ZE+q/uCG5gPzBEW36+Encf5
-          z1zxM+tVAm9fjIMjBeyMbycoZlmMzQl1rHeWpILm8RgTJBy5gv7wuS39cob7wRuEH74cQ0yw1g4Q
-          rG+Nf/7DjyJ9Jox9z89YLRoxQnzkHAvmLkdOPc0swGayq4feXT4lvLAS4bKgXjuf5+dVrZ+Xv41P
-          l4S9HslV/c1P7lXuMO5OgwF5tvIobfJvsHjUiSG9owWXoE9N4WKeXXXDG2w95rhYKY9GaN2QQg4g
-          PBTiOY/Df7/3tJJmEN6D48KsCnxydpJxYEyVoOq5F4R/87Uo73cOc/vCk8hP+nY1XdeBdpe4M+ue
-          fiFxfyaEXy3+EOtjOYNUjXyj2sUsktAWH4BIbGigPGQuoto9TNiwUzMoJt0HO443Jn15TXiV3sMF
-          Z4+zGlADvBUodDjG0aZ3mNBfLBUvwJu5td+zBSulBg0LG7OYO15LHpfiCwcXJdh240tBetHs4A+P
-          /NI8FcxZTAgd0AzkXEZ8O+/+8kr91bdeQbsV0jwe1WWhiOhzeDTXXgwt+DLFkgRvT00IeskiTBIV
-          Y4T8OmGhE4cwE83tDtvebYfz3MrwM90f2PnObssKk/oqfbxnbATFJfjxDaiP8DGrf88wEFiiKlB4
-          uNEMNnzi9WPFQ+A/3tgK+L5YucP5CdGfFpOqvc7J2oiQwhS8540v7HZ9bGdyq8lwsZ6Bafj39+Ag
-          yCQSV8WcTjiMIVEshH96YeODp+KwUccH9+O24p6jFVTv7X1mYk1a+jT8E5i+jUWifVEN7IKsTv39
-          fr+9xoV4vMEQbvyHHZIbiSixp6aO5/hBzr8eFVJ6vKpHvrbJ4Y7WYHk0iQXXeQUIbPW/dGPPg0vR
-          mVhHqGATvTMNDqdiu0/aeOZ3q3eQm9oVV/NstXwtkhVgUof4dumGYnlmfxyEhP9iX/+jwyr3oAP6
-          mEozWxRtEM/WIwObPiVBwlntmlpFBnapYhLLuvsJrSS+BNevOpHj0I7DbB2HrVvhX0C8kd8Pa2ol
-          KXhYVxUJxdtORIn7VjDh3hrBAqrB8h4cHyIl9bAXH6OC3S/2F16uYkmwFAQJjzstU4GLAhyhXZNs
-          fFJCzrgeiBO7B5MXthPElyG6kvTP1Quh0dkKJwg0bG74SUrjr4Tia/ciXhUfAH1JR6pu9YHtfam0
-          Y5oftxNU3TLTD20K5sdKD0/dS8TmLtKT5aURB0Yvq8PoULdg/YS6Bu1iFJFSW36xfv0uA6ZX2DiQ
-          pHewDGLfwYC6CF8WZAFRP6wOQLxSEATOQbv+SVMJNrzCZ94/MnZDpyeEu/uOePd3W1DxHotQ/7zi
-          TR8bAbnsaQrPzjNAfH5bzW39yvtb7C3/1s8gpIYM510O54VVNzYfkbBC+pw1HL2beaC+E4xw41vs
-          toPMRo7TKFTX9E3OKLUGXtgc+wPt0Mw3uW9KVKhlkDlZjdEtDwBLc28GP34xzPxuMkScJzzqM0dQ
-          Tfxgod99BYy34RGroNPAwDryMNV2Kg7O8J1M98msgVR4Pc4d4Rms6y5x4aankKqPj5byqXUFjSPv
-          SZUvh0QsozIHx4M/ksOmVwQlSZBiyYDM6v2tF8JHzDv45C4R1hWkB+LXn1NYCJGHPTF6BYt+Wgxo
-          ZfcE8ZyaFSyy1BoEsTMjFj8lQHf9UVE7LmyI3YgyGJV1zqAXOALJht2tYHu21DB96H9k03ft+hDn
-          Fd5zTcSuZ0vFimO/hOV1u2N0TfRCktJYBuMB9iTc9F7/yp6lmsXySu7OoLO1DfcjjATXQUAVPoz2
-          dYVgX6Tx5g/2JhuGh/+vvn/6oT/Ao6I+rFwloV/GjDVhi8APb20XLgFr9XWFSp3p2OZcvljP+TWE
-          6xN9kVzQ7dWhslmVH98YJN0PEybnDpqhPxJjJU27ztpRVrfx/fk1k32HxIdNB29YozsBsOCofeEy
-          nnbEg8GBsbzhV5j3PiBhYl4Y26v49MM7pISsA2uYPxVYj82KtY/VD+sUGCm8BIE8D5foajKf2jGY
-          G9si2ri9UqYVuxVcBc4n3hxOAeU4bYWydhkRl9lRyyTvVAI2sA/GkvQy17XWahUfvgYOLpEc0NsE
-          ZyV68xH2HG4sNj8IIXgFN7Lp14JKf5wPozW/E0855MmPr9Rt/ZDTpjfpAR5l8KquE0F0IMnrHYAa
-          JNxLw+Y7tgK6O4yduqeRjsgbOybfiOqqjMrNwv5w6wv2FY8diE7pDleybrfSrB1zNQ7/cnw4qPOw
-          6av0pxfxRVq4ZKwvZx+mi6Vi/YnQIIW0iEFvHgqMB+EcLL4ThMrjQTMkRWApVtnrOXjfBbcN70/B
-          d/LePcxiZcWOd90xZpyooUoCy7F2Znaw2h3SwKGXLXKG9s6ksyd+wds8O9g36xdb24etwEPm/uEk
-          6+2ASml/VZ3k+5hrt7YSwenK5h9+RmGcFGsVHkb42nqw2caJ/28+vzmvoF2g+sM3dOIROOyS4Z+e
-          nC7mnwsP8YCxPYfE7C/yxYcNIi2S3sAEjIPVE8pCSwhy5nvC3KJ11BPHD/i2+THBgcEVEp3tZ/7z
-          UdulMBdf3fzL5q+uLVuTToZPL/DxYb2f2+Xnt7f5RsvxYzGa7M887FRJmiVpqZL5KfkrVF6FhbWH
-          GbfTQ58ohAeunqWmpcnsvXAKnSx3sQ/fWUsPxbODpqleyMYXxbT5P+V+P+yxFq2Tubpdo8HLgK/Y
-          HcaYbfP/hCA2xFn4+UMiPGQ1nUeNxN/zvli7rK3g0ZMTnMlTm/z4Tx1qpyLBGzzYGNLvqhwftMLa
-          eOvAOoi9BW1AM3yI8GVYq4f4hDspu+PD4ioB5T2bgwJX2WTTX8mCb/MImnkxsfHLHywHKFAEFcPm
-          QaUtpQLN4fw93ZBw5R7BMg6yD+OPyDByPLcY9WMlwuFch9i6Tddkra49D06PmuEAhK9ifRl6DBMd
-          94jUnc7IlvfA/asn2D5+DMak/uXse9MuEE9vFvvlA2rhQ4NkzUEfWG4cSpiINCN/x75jS/dormoq
-          1gMOXuEJzNeGnCCXhwu6fovaXNdq4sG/evn5kzzCV2XyqycOp89toCFWSygmzw82/06PdrGXECpr
-          xx1xTHdntoavIQNIiiRsV7HHeL90n+DHD96+MYvVdkAONMnnsZ50ZOP3tVf7aQqJL34SRu1yiWHc
-          GB62iL6Y5E96QuhcyRNJeSMni3S/pgDspu8M7P0uGf8eUgctcxo3Pce103f4ukDfORX2x+FYLHf9
-          vsJzhmusb/5iJZ7wX360OoJrTsa7QPAz8SE+DHETkGe4D9XH7VETLZjvA9Ug7KATKwrGF1kCS3P5
-          fuGmL2YqXrxCyq6qAYuGj3COfKOQquij/PgSb3jfkk9d/stTyaHF32D8i4YchJMLcGjTi7k8msIB
-          fLyb0O4RXMEQ0Uf8zw9bWNCS6bCcoColGYcjWXy366c+I/jM7ua8R9GO0feDyqqyU8/YqA7mIF6e
-          Oqee5/JMirvi/Q8AAP//pF3LurIwEnwgFnJPWHITEZAgIOIOFBUQkUsSyNPPx/lnObtZnpsHSKe6
-          qjp0A1q/HVmT6X1C+gOvCXvkRfZP38Ypzi3x8y5l2FSyhR5/+XX0YApNBqRZdITOXw9Lcf+Xfwv4
-          zcYpRsEKTio6YspM3xIM56r+079Iff0auuzuLswuOwG/WaUw6i6/Cv7x5XhfBQ25ztcQ/oC3D2ld
-          TD4+Ib7SHHbJkL3FL8nbYwV3zyolbntdAZNjLvjHL3PurDa0f2UBzLZJNadbw/s0LA1V0z3XQY7n
-          2eV6An0A9mf3Q441fI+snqOXdr6GQShA/8hYmQcpCON9GgIUe4BeybEAvzzlSHRaT9bidkMNTneB
-          EvvkLCWxvnEATx+nI6derZOF/63rHz8lx0eL/fV1vXtAzNKInJMOJctztmooPRSEAj4R/V+VvzM4
-          Hbgei7jX2OYHDjCpZB+FwcG3/vxzeLRiH/lD2yXjdFEj8DzuMoKqOk14I15qzQzyKznd5bVcYpWa
-          sC+zaNOTQSk+mlsM1Fdq4C1fj7MHAvwvXyBj1/vLskv6Pz1BbPX6TXjZRZFKDuhL7DUjyfiWXQ6a
-          QXFFx9rTfBY4eQA2fkSKW3ECy8NsK9jK0g7p+f3XkOd134NfX1QkcG9gXFH4peDSyDbxxpubkLhV
-          Xfi1rk7Iwu/Jkg7lxVbQIbmS7fOt5TPFItj8YFSIvwRQ1H568DutE9r8NYtdFZrB83JjGJ6dOVl0
-          +DcXWwQhPTksIUX4amGDm5wYPerGRew7G7a/YEAu+1YWTZQHD1SPxcg2KLCWFMUx1NPLHALu3Yws
-          yzUdPqWMEPP6+Ppr0lwhjDlxJKfe1QAJ1b0ORQpN4ujBm/H22rvwt0KKbGE8jeKm3+C238gxTutm
-          2eoX4BtGJqlcJ/EZFgoZ4NhqiWc7gsWkfnagvsuOeOn5R0Od2OUhZ5MGmbUhs/XP3zJEcETeyx4S
-          tvwuFDzW/ErQJ6BgnY9kgLIf9cgRfwLDSXPl4PO5V5AbvPRS3D3iSlMGsmJRS+JyFZ4Gr73sREOm
-          lqSJoFT+dk63OuEdkp6AXRU5g69VN1DOnYtm8UGBIWfPTQgRb//x1xDeTitP0DV5seXPf39Nr5Xc
-          VMZbY5cN8E8vosPvNDEKV5jC+l4ckNnuuxKD6ub++bfE2O6XbH4aPLwmQh6/p5KQcyhEcIs/vDzf
-          93GyO2TC9ze0Qy65//zp42MT7nH1IJY5Htjykc6x9ud3NXPYAQKqvP/jFyhcrj/r++dPoPv6IQZp
-          23FVmKqruU4PyLqdj434rrceTF0F8BlK+0b8jUkP08XRtvg6lZPnWPjf9ev3hzEuh/030Jx8brd4
-          fVh/9QhYPS8t2W/5u6kyqYdv5aP/80sXma0vZUZyHE6jfvGn4oQKoNf6GVXvsgEEak2sHXrVxn96
-          nWX2S9TAnOHNj/mMSzHDAE7f8RHK1Y36m38gwm2/IJ1PeZ88r6dBSZxfhnlOE5NF4moO+pq15Q/R
-          AALHDFGj7aQjqwx5a6t3tWD7f+jYq3Yp2c6kAkn/Xolr2LtyNGNZhx/uUCPvlzGLneg7ls3bFBGn
-          WJuEsh+I/vwXvLRqX66MrzE8k81f3PjBEn4UHtqMn5D7I04ymd9kgnEG38jim4+/fn1gQvOrH9Eh
-          /vrWcojdDO7ueA4X3mMMj+OvV7lvPYTqNC4l2zurCtqE3ENpjEyLbfUd4MU/gej7PUhYgi/yX3wQ
-          JzVh8g3AedLSqn9t/Kdma1J4AdwpP4D05/JtFiNWdJAYp54c6iZKaKkYrhYdimso6oEB/uqT8E1M
-          /KfnLTEsVlXzLP2FHq0yJGsrmSv87Tkz1HrVLP/8N3CL/GXL57eRhuVRhbu0ycP5daIjDcAvVOe7
-          9MK7SzQ2DCRxoMp59yZH8bTf/OQ3hVu9i9haV4DZK5459LSkIyd6vLClNsAKp+tXI6FHpHIK22cP
-          5ELRsNrNeMSv4DJBp3z6CGnH0JK8SB0gvtYr8hQu9v/h86YH8cWM+Wbzgx0oVybCbGba9nzM8I/P
-          IMc9EIu9gkcAXxzAyBYhl1DhmEBt6k2XWOdoaYgWn17/8tf2PMf1JR/vihW4Ezpt97vKD8GDf/Wm
-          1RF6f3K7UYeStp3oHiPdEpZfRjXfXAVy5D3EVvs+RBCdpR+W/+pR+a14QVkjdihwLp/MQfHJoVi5
-          KoqU6dAwIfVUkLOgQX7RG77U6dj+V7872uBqCWG0nZA2sfavHr3VN+9w4zd4SeNf8lcf+lfvOJ3r
-          u/+GX62ACPJ7Um34siDyaKFvUgGlxS32WZp/dXDXXEa2enSypui2QsGwZ/KnL9bh1QSaVPo9rqme
-          jOz0qXnQh/oVZe4BWVNretE//zRQr02yDryUw40v4++mH6kNNFW9du4FnUjjjILy82xA7eGFjpm6
-          jET4LVSD88FC9t63/Y2fZLDL0pw8s+/Xp+f9W9fc19NGjh4YTPKAP8HpZIzEjHgTSPfGXLUZKvo2
-          xcBohsOy9Sza2RnSzXsGxNvt9oKKFSTh2xo8azpEnQf529EkYbF8y9X6Fv/XiQKB/99HCm5v6xxO
-          5lg00/G1YHDUb2EIqzlMhP3r6UE9SM7EczBtGJ57ERxz+YtO/NFnvIRzT7vVsU+sXbcv8cmoJnCr
-          Ix+Fas2zRThdckjdwUKGc9csyrUCr9pS5SCjvVN/bWfjtbXZaImVg/e4jFfZhed8RihQy2QkL2JT
-          CKzXGV1NIPsLjHIPjtrpQK634QHWUrvLcDNcibWLRbCsJVHVXXu5IMOh4zj7ZzvUrMIAeFFOWTPH
-          HsJQmNcSoeC+lhQrX1WLJpBiTUdgO7RdqxrjZIfsW/4H6OcdDCBouBMp38dt0OVxKODf59uzHYH+
-          0+mxhkJ9h9WGWIC9YcbD5GVzqHwnNpBEq8u00te+xGA33NDTMNfq06o+JMC/qaEhZIEmRR1B/v68
-          WKyKPR64/ftM8vpn+ezSQl0r24dBHufqAITivmawG8sI2evp0KwFNUXN3wan3X98W87jJU6B+BJF
-          nONDOYr4E3Sw/tCBXEVd99nInT0IF3ZE9vllMv5rmREcT88QHd77Ollh0/dQOqYIZf5FLNljbUxt
-          vFx9ZHZeb62qrctafLQ/6AjVHjC+pI7mTW2FUNa6YyeuoNNo9euJDqvK5+NjIMPanQWiIz0eBd78
-          eVriIiOEB64fl1SXRXD35xfxBB+My5Xqdy26WDmWRstI+MDfrxoy2xI9vnY28igYPWiVkUAiccgs
-          qQxRD01q58jcXq4V+mbiQbaz3yTHpfffn2fjqpOblX3BpF23ts+7zCf7EwiApNlWpOnfByGJkF7B
-          4gjeAPAhUfE6Hg+jeLVuOiTbIU/j8VwSfHJlT13APUNu7WslXbTR0QqBM5FZnn4Jz5O8hU8WVOQ+
-          XCprqdcdhd/SP4ZcaP0svJVHIGCRh+70kY6LKb9ULfVlmdyg6gLRO8odhMXrR9LgbSX0u+9t+EDl
-          NRw1/50s31zuoGn9dHKORcOiyUMutKHdCVhbRa3p1dIzIfdWc2IukFjTnQmiVle1h1UH0xHbzY6q
-          UlvPyOlfb2vJd3IK7ToJkeeKB0tKL4ELXmIkk1KPz42gGzsRekTdIZcVx5HRu/yCMnjvUUC/I1jF
-          lbWQaWOAOfhK/fX9C1O4ey4KOSRTNUrL9Cq0IYQIxVn4A8v6al7aPO3v6Dhxw8jzw52DtXLwiWO8
-          cbMMesfD3MFXdGneYKSPaMfDi7qc0D6kF5/PP2EB+WEyQnC+Oo0ofH41vIPHDYVHoo6Llnet9mWG
-          g8KKR6MYJm2htTox0L59leV2Pzz8Hk4+CW30TrB3pC08tFKxrecjoUkOVxiGcoOK3ePoU2/8QBjl
-          WoaViU9KXhrVDOBXvqDTemmYeI/1SvP4e4nip6wmUxK6NWzC2CLp+274wtoYmbZ/ZiaWa19LtufV
-          aX0l6WFDfx1Yrj3R1abzdBRakdEIN9THoK/pDplv322wXf4yTfnCB7ouht5QdNjaYg7YCNdOp0Ba
-          YRRrRgIOIQByYgnBi3e0Z27qWxs4hZEveWN4CXcxyrL4Yq1DvnKQ/Q4LCqrbDlB7B2t47+mFFL/Z
-          8sV2aCYtkdsrSU5cDVZTabBmC9+tLbSx9RsPnEDbfVqEUu74ARPcVzZI2zxG6VsrAO2jNIXJ4dOS
-          U72/+stPzlMYiOOZnPyg8NdiPAQa2KU+eQZUZ5TOIITRr2P/8JNHQeNpvbh/kRLgkknSBZqQuNGM
-          zyfJGAVpEjxt0toH8XsnSCTpwuvKs5htch0Uz187z6/VnlQPZL6NJdniR4QZgntyW9w3W4KqiqCj
-          Rjd0SEQBMOnCm/ASn3nc6ZOTiNl0GCD1yIC83fdVSt7DcODJ+yBynUzDGj6StgLpEZ+Qda3uPonc
-          9QXflcuj9Bu5TJj41VbgIXBJXFrbIMhkl4J7v16Iq+Q9wxmeK+iTQSb7m+T7zLntZbDhHdGLQ98s
-          1p7p2pZvka0fnslaawuvBZ6UkZAo7cik/MxrnNwU4RKLb39x1laEZ0f3SX7NL0BSVo2Dz2vR4N8g
-          pj5va9UEnepe4eq2+ybUBw8eHql/wEC/u4Ap0tJpqb+dT1qUm0VjLmrVv3yzp13E+mNVquoYSAvR
-          n47rC0BxVi3sjhbyx3MDmD3ddfDg3izsvu938jP7Rdd2skrDRbmeAN9H9wzGN2wir+pOFv3bH0vS
-          FuhW2rM/R6R5adTnWrwznknzD68XlVfIw9/anr2HXwaRwXHEyIefvzIGdbieHxCdDWAlAnps9zOJ
-          H3R4OTVYVfgtwHTZ2+QyrD37vfKk1w6dnaJYsz6+kOoy/4fH//anSHj+ru1OtoCuqXAZt/jHcItn
-          kqSPkonGD4ggu+zGUEoFYaR1fXKUwTfPITl/jr64F/fVHz8hz2vTW/Nn2WVwXxsH4uajbdEv53SQ
-          sKePDh+alHTHfybNY+iN3PYIR5qURQejSUlRWqbeuJx3SQ6XTgXIVcKlWeU4ycFn2l5SQC5m9Nkb
-          Hqz0wkfG9nyFbwI5UH/WAZmdoPhtm6gt7KbBI+Xf/lCEZNU8ktTkxJ1ri5IbmqBzBBb5w5f1eexa
-          sK0nscvqDSr0UR0lMtMU5TtTSZaNT8Am0tzwC380mYHiUPh1qz2yhL5Lftf+a/7l9209mpKevygC
-          azyJoYJix+cHAh3IPTJI9A+RE6btUKdiS7thTXHHkql7f4IdbY8kESri04sWuVo2vJ0Q74QarB9J
-          oDColpQEP94uNz7TwlqZeuTuzFtCrCJM//YTsoMTz6ZfXetwPD1C5Og1sshNNGKNe8Ab8q5PyVor
-          bo1A2mySJ1qf43pJj5mmXqQdCQd17wvttQzAknQFQngbtHtc2hxmo9uQ+O27Iysvb1urjxzB85y1
-          5WpM514bw8nFa2avzYzP0Ia1gvtQzKO0nHcHLgKX3j6SNDilYL29Cgj273JAZnnVG7EYDyGM7zcH
-          Ob0Jysk2iwL+qlYJC7fOG1ZGcwUH1Z2Jed4zf8Eoi8FJhR8SDTKx1h7IEDpHxSJBNYclxkWeAcvK
-          FeLGaGexYq8EoPC8A5b/+KQ+iAOcl+iCDOW9twRpoavWnLsPQjieAd2nvQM969mH8zf/+IvbFhkM
-          7nZELicQMD5UHwXsWHJCbn69NMtPWF/woesCsfv9kjAui6DWO5pNrvB583kPWyts2uSHfHQ7JExu
-          HBteL3NL/HHVxnXt3lCDRf3Dat5jf1msXwfHELtkX9XQp9E9MeEEdzourpcdq87xCFWu/d5D/pyN
-          I5H7I4amHT0wrx+e5eKfgwCsv6scXsJ961N+SDno/qiPqr94Kk+BCE4PQIl3mKk/K2+9godAe6Cz
-          GogWVhawzXYcTWL4Sl0u512Z/8Ovgx12SS9eHR6ePmzGq717NGvzuPXwmes6yfSpS1Z4PwRwtkWN
-          hMh0G2H6eCHYtdcL0kX+VK7Kwlyoa5FNkP+Ix+VsLDEUXaiQu3V8lVS8m5VW2dmRWO2tsNYwM0N1
-          +33ky6Y30r98ew9gQXSRn0uirAIHm8+rIZfdLbd+Rnp6AXIYI/LHh9dinnXYH9o3upDFTQhoxzvk
-          M2qhq61cS3FXOhQGUowxc8eU0adQBqB4LR4547ZLGN+1GXgAFIf8xkcF5R6If/gUKmIP/OVRKip8
-          7UGNXJ86Pl9d2F0zd5WKXGfpGsxx9gpPUm8jt00YwOe44cCGH3hFBwaoDOoAVIjOxNjy+9J35qDp
-          3ych9/grgxU19whu+QcDz7klDI+WqcH9ZUHOmvYA95Nha5hPCYpCEZTr/S1nUHM4g4RUbsstfnRt
-          jPWBFL70LVmFZh0KN/4RLr+SjEQ1ZxVeh2hHbgtb2RKtRqaNwTY4vlJeDXtNmIM+H6bE94ovo9xP
-          qaAoFGesnasvoM3CQ1C/xSkUhtUFi8eJd7h/PEYsNdAvpeNeWuFxrkxMo+OxYen1zcG+fD1JRMUw
-          2T4fQiUpTfwhD4XNvO5n8E9PhTeB+muTVyY4wL1C/LOVWxs+bvqXb8k1maqGTu93B96P9zXk7P3g
-          /9MTUtQSlH+rsVyZ1UB41T4vHIsAN2z4zT28qOz0j/+ysIMqAOswY4nOsTXuqm+hNHAxQypuR5bq
-          03uFdfXyUHwPjpZEtHyC2l3rkL4PBrD8Xe/PMyq8Y2Xb9OSGMAhDtSH7Aeq+lHfnQuMLu93sLASW
-          8PThwO3dY3L1TOwzPs16oF8KN8xZtJbrkt5jcJ7lQ6h9bXGcNn6qmZbVEJ/Tj424LlUAZmqV6CBM
-          y7iIj1ME48ego1J31WSyd3ytgctHRmELHV/KG9bCwAE58tjbA0xz3o52XHIDheCq+YsbBz388ytu
-          RGDN9Pd89dPLR5Yhn8pFLL4VtEx2J871nDbLyx0p1FM6ok3fNGvbn1TIfnOIuZsQ+Ws2HXqomk1F
-          Tnoz+JiaTge1/XxAp+xxKJeXf+zgxreQfz6hcu356+uPD+Mef94J1XXB1Cy2IuTV1C4pe77wH36E
-          tICUTbroTpDWn2/ImWsJVmq3BRTLn4asE++PKwc6CI301aKMdymb0r4dILuwiOwvZs1m4/1s1S8b
-          GqLb17qhCLkc9F6vB3r4L3fknakrgM86iehbfNKbsqiA8/QSHWhTMHYTj7FKym+IjLp3rXViVwcO
-          IYeI7l2MRsjlSgXySXdIZNl8QgU1EOHZZudwp+wvZf/aWx5cwtsbY4pn61++d86LjtzffWholTwG
-          uBe7OqQ3bUkYAq9C9axHj67pbmLsdBt4eJ7jAR31y7dZtMcQQk5Z01DOr5dRgovw0ja9gSVrKznd
-          TmaldtdHjIqmFBKGFK+FV7O0CNJUrRkCsc7h+/3MkHlGXtKHEARQ7/EvBJ12KbG1NSH447sooDoQ
-          ZaS1ShDpiBwW49WsnKXfoXbfdeHy+LkjL+HI0zZ9jnwvihp27uMYIi6/kL/8TjJDNOGqbi/Zgd+l
-          mRS/8P7pEzuwHuMEhkUFp4dCiWOMH/Dj1NZTJUkIQwpYACgQOg+u3/gcLnyOk5k8QxccfoWAuU3/
-          Tov8vWv2AZ+ITqerte3PFPa1X5MQvlJryy93LXlNJQoa8WMRLOxi+K4dhvafp8rW1L5V8AGrACte
-          4ozSwywgOJw6FdnZ0yz5Y8nZsI6+KzKUk9iwKTzagKfzF0V/8URuBwz2TWygw2S+rXUoI6rxE+tD
-          L+9Da+pYnILyY1fI+zUjoG2ythq4VgiT3pPGTzLtI7jE1xeyM+M+zu/PZ4IrCBH6y3ejXvfhn38X
-          sit8s8VI9zWU/Jv4zz+gxNd6ta7vR+L50qFkq6JW0LpyKULGz2PSKt0H2O3gSMKLfmA1/tidtn+m
-          5tZ0gTbzn/7KvuEHbfhlreVuxbBwlRPyDMUthTHKMNz8OmJnBhxXXOMVqki8IcucL828JucKfl+p
-          SQ4fykriVMZL2zeRQeItH4+j8nipsixuL/XiXTMfxZf+d/1YMkFusaaJbPB+NFd0OKhS0/4eOIeb
-          3kOO88LW4l71GHLu8sY05Ra25Veq5Sa9odRrbJ/nSdRpj7SLSBgZbslEx3a1pTtXeN0/rISH6nvV
-          ChecUHgMdSZwspPCDGX95t/1DcP9J/j7GvOy3wEqSESH5dl9kdynnT+dhvmlFpW9J5cfOybM0OxM
-          Kxup3fDtm8yD60bw1Dg/curUXULNB8f9889uunGzep7kHZSjrA9txXNKUXnOFbQOmYQ2PBrJli/g
-          7tMhclw0z+fN153CQv45xDs0MniPlyKFd4gTLFHpWLIXsVdtUT475OStCdZJLjCYfsMT8/pn77Nd
-          M+ow9X4c8lSv9Jk/JC7onesX6R+SJ7RKLgOUo7Qnya2Lmz++o4rHLiMWvcjJ3N4WHabRPSVHzX5Z
-          9O/5mpO3vYLH7Up2mJVCSYP6QvYbv9r8OgeQFz2gcxvnjH4qLKrK96rjIl8/CXUDJ4ScIGZY5q59
-          stJjb0JmSDEKTOEMqK+fX2B33l7JMMY943Xu6f3zt5y8GJvfwVY7ENGTuPFvYWTe5/n640fkOEyQ
-          YXhH4V98EHtRFJ9l/CgLG58IlU0/SlzCPFChdSaWKwQWw4IUAymwEuQ+hmZk1+xXg+ZTNyTAzALr
-          /Q4xxLOD/vkdy+N2t8FOeZ02fq/4rFcvnGqtzQXv/OXtU+26NZ452QLZ8CLhBYmY6ubPkhgdEsaO
-          +x0Ff/n4j//iy7vM4edKbuj0cK4Nrb/6BLC1u4WtKMvNsr7G+g+fcP+7e82vVm8BbH9hQRxn+JTk
-          3McR1B4fi+zDfWutJjReKm8rEwoPC0hm9AQD1H+p8+eH+otxM1wY5Vdl01OJz7gsh8B2nglGzmqM
-          0uk2iOrQiQEGdz1j/KlHE9z4R6jukgXQdBsLvOEfFv1fyza/6QU/NGCk+NMvfdOK4Ne4F3SMDuE4
-          acWxg/OcZXiH3ayZh/1wB9617JCx6Y0l5QoORsa+QE7Qt836O/qOuuVTvOPyplzokojqJ1lKTFH6
-          9pcw/rVgUKmHKpcCf432Z1Wb9PePBFeKxrVLdxTYIgpQIHR+IuF+DkELQIXV5jON61DmK4iyKEbR
-          S6XW5ndS8Cz2XyycwATW+5um8LzbYeLWzxHMjWd42tvdQeJcD49yPN3dHjby6RvuzuSVLGZXBrB+
-          8xNeNj357+stf4RUHDJ/mb69Cd+1zUhaVm/2tz6Qe0eA7CPLKmmV32JI3d5C1efxZPTAQQe+K48P
-          hQG+rH/616mqKpTz0fbnYqrDPzxCSdssySrHZQ4v32o7UhabI/+k107e+ElIRX5OtnhbgcdEDovf
-          U8yWjZ/ATT8S3fOHkuj1K/zTH2Gzu2OLpDFOoQIiLhT+9G7sDxhGv5bhnfEbGKs3irCvh5Ychm3s
-          hjSu2V+8oz89yeCivQDKXjeU1ofYmgdXj7U/vPvz637o8RCV5OVwWFU94OPf0bfhr/EuyO/ck48r
-          hRugTbojsv26GZmtYFvisVmjP/62SKTg4FbPQoH68YAU2LsMnvcFRiY6j1YvqDYP+cJp8QqJONLk
-          QXNQ3CWfnB6FY/39Pdj8RHTitjGwY90W8ABJFiriuibUEvRIO8ReForK+TAucriuWtLcnuRsLJPP
-          mI0pfAljiRfn/rAWAKZOjbAX/tX3/PUDlQIsyneHwnbFDPf88wU+PZdgPruMPn4MzgrYZYnIUd6/
-          xjU1WA/+9OhBqNUSZxMa/vQp8dgJJit8b/Wwn4e39a0b+tXbCsBZ8DZ9pjf4ecQtuPTOER3f5pQs
-          qU5FePjcJaIDNoEtviLluX6NcLleVUAXcIrg/fmr0L7ZGQm/8VHtAOcMnfl630ixbHra+/3I0Mm5
-          LoB4+b5Sdi23Igu/abl09smB6/kJQzrIxF/r02+F0GlHPMVLZDFX+8XgkXCMbPUKa938dBgT8sXd
-          /HyNdEUyD9WrrRBDzeZk/atfzLtSI4FAzZFf6FnXmjg0/8XDEq3HFJrXlkPlvnQT/vhSMMy+wWd7
-          vltTmH3vwPLsvUK4w+/mn3/iatORWHg8jJ9aPYeaLd0dpE/1ji0bX4RwPs3oQKVjIoUjff3V8zBs
-          ppn9YBS5Wi9OGvJOb9knG3+Az+z4Q04dJuUiqYcQCjfx8een+fP0MQNNSW4myuan3rR3Men/+Dzx
-          +OHN1g3fwSNcVKRf4lfCvK+WgZ8f5QgpN4vRm40DtY4+Kzp6R1QunrV6cLxwPXKv25hVu3u0UDqH
-          AXEee9isjvXs4P05VuToJd0//vmn38jzTPREVI3GgSfQmqhqGPJpzosx3PgnCYV5GtlfvY/w0x6d
-          t3r00APKQe41ViH1Lu+GJSZX/fnNmJJZSqhzF1/azf4oIe81rcXeRvyCedFfULamPVtz/3v/qyfi
-          7ntaGUsOqi3/4a/dQbWccim14b0Vr2hv8hkQdPsy/PMX/SA/+LSPVReq1ylDSb4M5VJMf3rGlUhS
-          fVZ/2fQavN3o8q9+vOmhXNNBOeOleUQlcyqj1uCad+SUJZHPbqy5gz98xpA3R2njO5q6Nd31Xzsx
-          2a63gvfk+yAmKs1kBeeCA//PkQLhfx8pGCywJ65bvho6sbgA+6/jkWB47S1RnZQQBnsUk2Pk5Mmi
-          7EYV6qy5hJDDViJMH7HTxKnFJJQKBfSX+AOBTLbILQ8Gk66yLwONVNtbV4erxZ+XPII0fR2Rm+7w
-          OLsnQKFwvJRY/p0mMAWX8wtu/x/D5wqs+SDcXzBrvw7y9vM8UjGWV0gI/yEXKaIjm1FB4cksLBII
-          dVau44JMcMvOX6RPhIzsB5UO7r+2R/bs3LL1CAwT8jqkKAmrOcHfPu80lRykMFouncW+Bb9qF2ut
-          kJP3bKSpmqxwHs5HcjW7gK3qPq9gVByEEET3p7+ou0qE0jmKkdGuu5FpjWcCa7z8UKQpqS9O9b7S
-          ENYNcj0ETbI8yqsK2aIbyJRxnFCVT+6QinyMwueJMPq8zY5SnG2OlN7lCWi8+IEWW3dIyh0U/MVc
-          TR28wHNCB8Hlk9VZaKVR0yNEZ2sKWh1cbZDkhhbuNHQaJffimNryO3jEyfx2pL2SV7A8Pn7IPesm
-          E6pWq1VuiIyQ6yKhnPSES0HtnFbkVW+z5GVr8OD7ITnI2Dd2sq6fJ4TOQneY+UUyrnmKQmjfM5sc
-          j/bXYm4lY22lnkv20eUOJFWCLRzueUwcU85LUTRGUfPv7BKKy90t6bvWehjuqjPmzAKV0uFS2pDb
-          PyBW+eiYCNc8rjT+a8UoDC2/5Fd0C+GXyxVy/uWZxaKxpPA+TTDUVv8O+N+rDeFuevbEXLuKrd6o
-          5vAePTLydA5kJM+OytAW94T4/cUA/PM22/CGxT2Ja2tv0SP7TfAQ6h90CJELeKhNOmweqJp/R2sF
-          2GhxAO5ZXxBDbC1fmr8Kp90wv0fVl3UJv8wogiUOHqQwB9cSNCtqtTGlXMg770dJu2uP4TuBKoq9
-          1SglUVMHqI45IdE3dICUalAF/S+MSDBpqUVOR8WBjKcFQpYlMJYEYg2Xu9CQpxi4PjXDuNCqp8ph
-          9j6qIzYv00ahpxtWiX231ttjUbXB02Vy37tmOQa6k8O+eHMo8H4DmJS7kWpYyn8ouzpZI/JyEsJM
-          xQoxukvUSPR4xzCKIhUZtp6UY329BHD0oInirPhZbZN8Tag+gxt5js81YX3xnWD12b2x4nUWkBxf
-          q+AhND9If7ZZs7S38gXv2VDgHzNfpTB2fq8Kl69JwmCzBKqeTUCE7wc6X/sC0LNu81B1RoAOXGqU
-          ErgtlVZZvxztPcFN+KnYStKv0w/tU8BKJmZnWXPj7xe/+4ftC3p0CqH1fj5CcW+8fGHHH1X1FtzH
-          8Okiy2/ftTZA8SCpyNsluKFmt3Tam/oRQmc6NlNmTTmMsmLGkDvkIx8vVgiXXaghyy+ShvehGoNh
-          uhp4N38/gDaPG4arf0tIBppDKTw+XgcFf3XwGiSDtcYRj7V7ca1xP7OmpM3kBKD8jDssfi8/ax08
-          twAwv3Ohsj2fhf7GUNvWE1XF91bSw3B2ofTN+BDglY7CI2gKLUkzAekoyxPplM8drPGIiaHoD/aH
-          53Db/+iuugEYtLDkoDCOLTpInVSyUXI7SHYsJd7lIpTLt37qUPSiD6mcG/DZKf902myH35Dp17DE
-          XBjKkLpxjWzW0ITR45uC92Ddkb5cOp9lEQ0gWkYF0/x9LnE3f+8wHbM3ht7PA0JjGLlWp6pNzsrX
-          bTCYWk+zj3AgDqdnJcvkiNN+z9ObxP3zyST8bWW1fdUW5vV9DVgJaK0tgcoT+zqJ1qJ8Ayx9+62D
-          jZOl1vLU7hjeONKgA1r2jZgYzFPRVQ6Jf58+7JG/ZQdMt1sUdqD5luyemlgDoiuQ5BW6Df9zT+a/
-          +/FDvS2F2dVrzVC7mBSXi5DgOjiLcDW6HPlWrJaznSy8dg8PAO1/x6WhCeVcSG7UIg9XqMeVflYH
-          VvCQhqqB34ye88XRHImzyCm/Zz6VrdoD7XfXkr211uwfvob2dwjTgcPbt9QW4ovzIpYL6bjcVC+A
-          h4PwJhZ/d8CSikMPMjPSSApO+ig4D7PQ9EdtYdlCSSl9zUsL5dt3wrDkrUZsgoDCt6375CjuP2B5
-          P4Ma9kXDYfDoJ3+u+8GFu3iY8XdqVp9erOUOqfouwpd/DUZxWFJHI8bp828/LjwMUuhYCBE9kC/l
-          KF92FO5fihPKGF9KoX/onlZ8jy8sFA9qMQ78QnhRFoIulp9bxJ0/gSa9TwbyL6bQLIUiuNod/kZy
-          Fdw0+YdfWV7Q7X0R05f4k5Fqj0rj0aUx40RYi/t2hMuNUaoNaKQ+O7Tw83Ev5JzNPaBqsE6aMjw+
-          KFmTp8/bO2uCfcc+ZMPrcuUPB6r5sjai9C++oLEPwI6HInl6TplI8SkfoIAdhLUMnEpxyHkZVHIe
-          kK2Vzshj/1nDpE1agrLMtKa9+ytgGswvrG58h3A15ODVCg/oQtkbTBc1nGBxDRZ0cPGrpBdLqeAt
-          OvIos3huXMbO6rXvkWRYg4rDhLJUdegLcROqNzCyVW3CGk5WinH7kpDPc7Lr/ONHG76OaxKRHuap
-          o5MbM/VE/FYfVZPOcYzFV2yBn9m493/rdWTamjCMTpOifdqK+CPOADu0qINPTa/QfTJO7G/9tQ0v
-          wq87Bz7128SGZqMjVLU4H4ff7chtXWt49GTHelwu4qOD6zzfEYr4fSk27cjBv/y0qhGxaJhKOUyM
-          XUmsevJH1vsggNps7Ujo1iYQ4obk8DzscKhs8bnxm/tf/scwCWd/Ka2YhzcnlZG7v3YMJwZwoZmt
-          PXItxEo8LkgHoUcxsgJwT5jXG7GmN12HHKKOCV2qmAK+WEZiIEseJ6/tPe3ZySO5Txrvr279E5XX
-          HcCQlSbYJvndc5intk4ypkj+erjZgdbtLYb5jBKw7G0aaUbsZWhPmcGW7fpBYhcOsr9qWk6H8wLh
-          5SafCQqnBSzhb76DyN4svHcDm+V4PReaeslOKLwfxXLB2ZeD1uV8+OPjI4ksZsK//LDsG7tc0v01
-          hKP7vpLntP9Za+ckFAq8t5JDMEkAQ5tL5bOELsS8ZiSZlvkQwR/jMnxN6Lnc8jOGljZb6LLh1erf
-          qajBQcyR12oPn7qmXMD8O0vkCPQrWJAhDdC+KTdyk+uUrd1Mqn/4H9Qiz7A6KYGyvp8iMQ/40NAZ
-          Ul1LJtEkkbl6pUR2Ox18ApyjU9dcLDLr50CzdaciwclRANUOxqSN4c9D7u2z9/lULSnID+KNoAq0
-          JaumlIc77/jEYwiyZNnxhgo4g/IovFPDou9aGOCiGMq2XmW5XpuHCLu9wQhK4r5cTj+1hlnaX1Fc
-          wj3g//j8NZM4Yh3ljNHr3e/gz9QAyi7tjdHu+prgs5guJP08PiM9nq88hPj9CiX+VJer5TS21jSX
-          OASLIFqLeXtQdcvn5Oh1DWPleuJB/iUShin1LHE6dlu3/ZuOkP5Zwbp+rhAGDTOJDQ/cSL6BHoHd
-          9OhJnm9zWPhIrSAx0IeYjp81dPd0VvC3/+14aa0lNG0Muav6/cdvhyWjNRT2fUKq0BqTpXhfOBDV
-          Q0qOqr2MhLeyHqI38ZC/L2vrfbzecjiaZYts0tqjMAjbZOI6/2L1xS0JFdXbC3xA5JKbo89/8V+A
-          gcMlMcHqM34RFQ4GD/zEUrNrAFMcLoT6fF9ReJsOJS/GlGrHUxWikx1KDAuGrEKb60/IrgrdX/Ro
-          H8KMijNxc6svVz1Ibbjl/3ANaQ7ois6hJtqHN7mfp6jc9MQEFfq9kWP4IeO08RltNhaIrKjTwWpL
-          77umlpWO8i3f8KFjqnD55Caxlnz26d595xp3lb/keiWa9cW5nEMXCydk4i+zGM5poZWx6ZLnRxMa
-          eok/nLYqmoXMeTIs4R7QCgLjGRPPR145S2KZgUcZT7hO5alk9JhO8JyCAB3Kwhil8+hMkN/VTwxu
-          P72UZlSs4NmpY6jML6uhOng6cOf5T7LX54uPP07P/+XXP33sE857rLAdCoSCjW+u5J7eYXOrM7J/
-          vC7Nog1KBls140hg2n3DPkrAwfBqXpA578WSPo2TCutqjZBhPr2RhnPVQgsdExLu7MGiqHro6p+e
-          PLl8mCyI+bbqxp8vcXqpsybrSnW4+5gH4qa7sFnVNOCA0AYWivB5ShZx0iYIfiBEpzVYfHaNto5Y
-          2DQw/7ys/neYBwjdnBTEV55dSa1dtGqG2sak2CVhMwFte0NMyn+kuEDd55+e3ELa3mW82/TMevEt
-          DFPk+dvct+fIvqzkNIl1JbEcXrMYV/MQdqIeokuqG+W//TanxwxVj5FZY/KpV+1yU894HfpDKUgP
-          YkP/p0N0DVEP8Gyb29wBWwqXZzz6NE1cW1U/MURBFVUJK7ZJtqGu1njtzAjgaaTeP7xy0KNo8O46
-          1eAh3RZkacbPWqar6fzFayic4uuGPxceemF6IFFm1+WiSzMHNr6H5encjeuJe3VwejYWChyLL5nI
-          Iwgj7+iE3ZvXGwn0zNReJMmQTtjRX4o4LWBd4V3Yqp3A5qrVXmBt5DcKF+Vc9tt6gfQTYtxx37c1
-          WQJ2ZHEdz1s+TMpZDuf6337Ib5+PtYrGyIOPLD2IL8tls546o9A2/olcTjsA8Y+v/+lhJzkMbJnI
-          L4bLpzCRp/ig/J1eBxHaENUoSOlgTc/1lkGlqMVNv5OEbvsfFE2Vk72DFkah1ppgxwkmQq/7bI3B
-          9+tCXeD08FsWRiMNF4tqG96icsNTGu41cZsUvQtrYcgbhnO5gH9/n09np5Hwd5LhTf5SpMu534hf
-          pOTy5lfgPz9BOvNLrtXB5RPSI1+OuNUsD6Ky/6L73q23I8b3F0yftkrQ9Zz6U5N8dVg80jYUxW3y
-          MgslG/7FD9gZjc9sJqcwXW42sgtSjn98El6MCyNG/7At6SPvZGjatxALw/4KBOM2iJAzVp7YwS9i
-          88jVIvRM541lYSnAN/h+PZBVdYOOzHwla6NF4d9+D/GinJPesHkKlVc/EJ2wnz+oqQ1BPqsA/fFf
-          +qdX88Gt0f7VkmQCWZ7CaP1aWN3wicpx1msXGBaYJbsLoyLteCgj3iBVGbYWffx0qnn89RZmdiEk
-          S7p/htDUJ4cc4pBvqKTaFVQqxIXzxnel3yUSwa3/GejihdpIW0GWIUm6HTp09xisUmHp6k3+UJR/
-          QsNaApm3//Y/Mv/0bnAfKYyPHQnB6aY0+HXWKNB/+Lat3ziuvJwEmu9MLGxXI2mWOFXlf3gWFOe4
-          WaND68JneMfhFTuNxRTZGOAcVSE6wLVt6LLHtYrK4Yu58+FjjarEd8of/90/3kfwj//NQ3JEbnTL
-          Lel9HWJ4/Z1yYtbc1V+qzDC133XVQ3ieaPLHT7X3K3ijo6PP5Xz9Ohys80+ALFezAf/ymhou0ec/
-          AAAA//+kXUm3ojwT/kEsRBASlkwiMiQIiLgDBwREZEiA/Prv4O3lu/uWfU737UuoqmeoUKVQ7IyW
-          u+1uRAcbnhPWESpcT4FIS7D6fdiab220fB7vnYK29Z6eu0Hvl49lDTBolwMZfvVmMXaL/PMHKDjI
-          xurvFJCE/RY7t82+Gu5aoSrGqyD4MB/tfkynoYHlbGVE7JV1b4vHB3/+TUyUK+B/+OhJ8pbu9aBx
-          6TQpN9hEZI+EI1DYfPN2Dygqtw9+XIdDtFVC34TV97nH9nBqqtXf3QFp2AUYX6+ffrr4ugqrKK1o
-          uDQc6GlaxeC8ASXVttpiTNwzEpS2Mu/4/lK4auhOXQtXfU6WlQ+KfhmESl/Tke7zjcfYqF49IGQ4
-          Rj//adk5fAbDZ3/C2DrQajkRUYBKsrlTa81nntupppK/vxu6X/XiIi6lB83A8IloHrY942+6rDzX
-          Ky1Ytw32/URlCYRyKKnubxswMdWp4RSXR7Rf6zOjm40OFEBOqIl1xSVpfPB+/Jruz+K1//MveCkb
-          6OEu3KKpRqkN63xUqEOOjrHqUwdyXahRbdyoQIg1YoIrN1bUDB0pH5bPLoXvpIqxeTvtKnp51gSO
-          sr9D7+24qab8uUPwyu9aBMkzYFMs5wsg3ifER1179d/CkjPQwWii/nHiqwmJxxCmF2ZTZ8XPnz7e
-          4f6TYO0jjvnkj30Lz9DL6O3JcL6EuooUUXl8KNpsoDvMj2wCb13wkOi61BAUlEPQlYVFz4eZVOOG
-          13bSJ0T1P/7200fNRs6oz+VVteyWuwOHWfSxmmz4n5/zkFf+SI9jYfTCqpdk3uEh9Tz3A6Y4BCrc
-          owYS+Xlvq2X1m8CHL0aaOGVb0RXvwPEQOdi/cyFY6nOXwmJuZ/x4n6to+PlhPt8BVMnStpots0cw
-          cxSHGrZSs7G6tjXMX5xIj4eyc5eh39lw+4E3ipfywiZkrVPCbNnBevV+sXkLNiFwhMXEZp5Y+ap3
-          eDiDR001uSJgbqP7qse0iNpiJPaTB0cPlOHzSsCjehh8ObQxPNnfN6K7r+hOTl3YcPvMlT/9PDnZ
-          9gaPUTHTS5VKLvvCuYbN5fZAHL3p0Xy1hQVqXNxjxHLksp8fPiKLYSRVWiVq0jeB7hOpv/5BNStG
-          0MAHFUfS7rZ8z5qvkAAoswNa9kHdzyt/BzMNED4d9prLxyHQoS3GOY1rIQPk53frTqVhNX3N0TJG
-          hvzzaxHZqVM//s5T8e3XH59bkH+34QV0xo9P9kvspgu4cgtP/fN1qliTa4IybD8NPZyismr6Miyg
-          6jpf7Fjtpif4IHlQb8YLPtyAE71RNkHF5qWJfN3XFiyTvxdg28xvevMtiS2YQzdp9uCeOvvR75ev
-          2i3wvvccfDEP54oCLQz/1dfAz6rVb7B++LzGux8x27qWMJNKGRtmIhszPZ1MZRnpbe0/HFzh1igp
-          8JODizb26phv950O7udDS5FSjtE87S4ESvF6pdDGlUG5mSuBRfMavZcuc6fuVLaKP7gj4sLv+lUw
-          Li2oasEdyfb8yT+XoHeAsU9SqrnmAOYwXmRgvfgA2+v7Gkpw4wG3f0JsrfW2NdA+UbA4D2j51H41
-          lRE0d309jj89BYaUU1PluBVk7B4zH0x9boeQXhcD6z66uWz17+R9AC9EvqUyWzLUh1By6gO+H6J9
-          L/z0+PcyqTh/Oi2bA9gOohBcbLIr2q7/+Yng1//KR6GL2pWvgSk8PLBbHfl16osU/uINH4zXg83W
-          3UnhHOktEe/CLZ9cuKz/X3Ogt4f4dd/hXoqhJFCP+i8tjdZ+xgBupawhfra0f/746j/jg9Dvow6i
-          rwo+eHvE+5G27vRykgfQPL2kbt3dc2bmngVIjyIkW+2zn1JOzRS9oReyXf0UQbxuOSBftTNFndCz
-          edKdm9wNZ+2vfgwEdhxQ5nWKdi1kjBm6xMHlPo1ECeGrH8J0UKFxNAoEjwfkisr1GcMzhQzvgzNk
-          y91eQuWn/6P8axk8f2wRTGuVYmPLVdUwCzMHn4FwxHsprd1pFz7WK4lyhj3RZO6KDw58c5sA9Vkd
-          RgviovLXD6S3lmuj3i/TAAbDE5M2uO5cao+jB328uRC5bbWKR7GYwajLLxSN8r7nhYnwsNTLDkkX
-          58SI6t1MWJP9BYdBeGb89pg78BLZAkbhVwPzsfFlMAlCSD5x0hoj+r4fsO+2Bdr2H74iX/wMYV8o
-          Mfa64m2wWIE7yH+0EGPBsw1+rV/KEXAj9V/Ko1r7lypc6zXpFnsLlm0wpkAYGoJtelbdOXDvIcwu
-          aMZIzK4/fssr/Ri4+HSfUDQaJUuV+HU70+dmDCrmqm3x45NEwHe5J5uNH/7qPwL7vHSH7iBy4ADr
-          Mz6l6FSxZJdCADbKhA/m4dwvSfyNYc0fZXqwiRptf3xV7jO6+td7xv/8xB+eJfF6F7v5CjEI2umA
-          z5drEhGWPFJZ7A8+tQ7Pyp2PfjMott7u/urf2EbnB1z9ZLSrP0nEXs1jgaeXw6gp9RVbzMfWU55S
-          uEfCY7LYnBsZD0RpQDhehwbOijQK0q/fc3zwuiEeRZ2He/t5xCv/B4sxCgu4pkOI3Q91XfGq3Qog
-          IllEnJhdwdTtuBjCSQbYh1Kz+hnfB9TK/Z4+u7rIWfPlYgC0e0iU94ew2RxOqrLyY+q86w9Y9TiB
-          hw6B1Z/o8um9rQTIncYK+7toyueXOE3KxTcjInBqEi0QvXRYgPtADUU7Gv0BkkX+f64UCP99pSBQ
-          O5m639BifKN/YvniFiZ2RVmvtuSWEujhdXXXQ9XY+H4qDtzZd0BV/bGJlleghRDPeUNEWsVsLqwu
-          hbaELNK/c2ow3gExrK5zhrWTKxiTqN9aAE4Dora53rqWh6hWwiCOKTascl29vKthAOwv1XjFzIcd
-          97KhX8MUn0i470VR9BL55ZcX6n0zM+cP8taEul9jej6Ed7bUMAuhcNh2+NB/s7zVX2UCbVS6VHNO
-          32o5eqdSkS/LBu1oi/uZe84PxT1bDJvXuGBdd1I55bgBNlafPZ+zYLAzkAuySjNb0aPFlSIP8MnW
-          QJUB3Xz4jgkHqFwGRB6cPJ/3qQMhts5nnNvxs99Gk5tAaCszjSwpq2Yz51IpHKqG8PAhGrPyRgEE
-          ekixxQwtX8i312Ebpx3Nv56WbyV7NymvdmI0wfo+36bmokIv/3SEl/IymqYNK5UU2mcabW+jOxZW
-          l8EMTs16qzgH2/cjX6BCzg223AzkdPf+POC1EQKMmv3OYL5+vMHl4H4RX7hO1d8Obg0eZLNBSika
-          Lo/BOtpIe1mklQO3n8aoLJSciRu8/1w3+cxEU4Avv7jgA0NatZAuIspNLxYaPrnQ5S/cnYeBPEV4
-          f303/bY45Jai+s7x7/yWNFRUuN/kIQH7S5Cz7t7V0DD7Fh+pdI62VvJqlD4/f/ApFXQgoqkuoQhb
-          jt43prOenxXAw/74wHZ0rlxBc08CPB0ukMhsebujdZ52ikKOOT3NOAbDh2M8THUVEuVY2xX7LMtO
-          aYvrQk8nlDLWCE0MEW5DnJqT44r6+ZUqSM025PvWXTYd5K0F7mJiUeP0gS4ryO6hWKhtcM6RbyXU
-          4FtD3Z8ymmaexwRT0G7KqNoQm7csjZb5bGTw0+33OLmOtSvmx3sH01MQ0Jwp+3zb7Z4I/PLrCb67
-          fBANQYc35aPjgyKpkXiZHAEGz0Gn2deswGCiwYJNBGtsGBtiEFSzBl6aqSDFs+ejPhjUTGnMak82
-          VdBEU/gGE/TPAJMWv4hBdHmsoRYFPc4e170hmpq2KFd2legt1xJjm6QtB/ebz50ArcrzJamvspzA
-          YMIHHLkuOxuzoODk6NDYAWE1F1aZwuldv2j8cW/58vj0HVSUef/Lh4psOk1VMrJn2B3Oeb89324Z
-          sCQ7oVfubEWCVQEZMO1k4KQ+q+72Kjx5UIY7H2fJnLvL61Suq0sDDu/PvWeIg7cnijT2b8QN0tNl
-          y32uFfmycbB52m7cRfUrWXmEukYYeryAwHPX9ZbtzafmiV6rRY5mXbksnzs+1JVhsOraOlDfBQ6+
-          bw4DWzrjWUCg93t0iQ6Fy9vbNgDzsE8x7g6TsUjp3YZUzURCHZX0ywCmRnm1dUtzO970C+kVC9po
-          SLCGXNPYNvfKgt30HbE3nYjb343U3oHQqlHTP2Rj6t87C2Z6KOK/emhN5zW/4wLnwzmvWCNedHD3
-          +IQGzfUMBF5pOajokkpG3i8q8W4EjvIIkUcdLM8R89knhIH6oDhVNyKbYE4zuNUShpFHZbBMl/MN
-          IlX/Um32FncCbWvDyl9Cam6ncz9RtQ6VVOsajDdZ1zPe91Oots2Ej6wujemWHEI4dryAbfxCLu+d
-          9jZgmrWjtirgimjuiZf55DZS/Lyd8uGj7SGUYJJS3Slp/jL6WVYexhlhFGeHnk/8giiXiy3QUNq/
-          2FaWk1qUG96meOCuYPHEowNJPe3p7RiVTCyzVIDfbkxoolYzWKAeO+A4RU+sb08eE9Z8AEKyY1jb
-          iEE/PVUrgMoJHvH9cub65YqPDew2kkqt62QxUXGdBDpostb4K9wtzQtdWfGNeiK6uwuqMgQNeimx
-          o+UKo6amTcprpx5wmh0do+eAtYAyMlOae/sTEM31y5CDOds4cKxvxILHg4Ne1zIar/E5sU97g/qV
-          ldRwBBEsxi4voLvfPqmfe0L0G/sMh+lyoBYNqcFyZ6PCpkoNuq/9suqjwxhAwzwcqPvqIFiSpuKU
-          4fzUsFU/2nzpN7UKzZ2cU8MZacWeS7+Darux0Ha7nXImVfcMqqJgIpBNQc7P9BBD9Vp9sK8JV1c8
-          b5/tXz47e8vphd11LyifL6GEm7o5b6RzloJNCCesVR8LTEIY2ArXBluMrWMC+N3mJCint1Bjq05L
-          QF7BzIGHcTTwyahsY/5a9qBIsKzIRZ1jJr6KJFbqKI5oInigIj+8fnB+RES19XqR3VIPmkg/40tP
-          N2y2dQ0qJkIF3te+3gtQ628wuX8UqjWfHZiiUo2VpsoMfM3iLuflIWp+9YzGy1gA5h6SQLkZ+wgn
-          eG+7oqlXjpzBOqfBrJ0Yr5eRCoez2WP39ZHdSUmHDjb7aKauzGNXmKNSh/sqc6n/JgNYdGhxoJc6
-          k/7wfB14HoLTOzBwvvKrZV/XtfLiZ4AtM67dmSvdYhfN4wnHB87I2bd4x8ovvk3VPBr8d3xAyF39
-          gZr9XmDTOageMFDNKzbg9ZYLm2+QKOfGwdh/qm4+2dsiUPiLhmh8GgTGKlcgihvLLTWkyGTTqfV5
-          yIB3oh7H6mitJyosjPeRJEXH8kl/dQnQW77H94diV1vvxq2rE4UZ69JeA73inFqYwrLHzl7dr6uR
-          KYFyWvXYkuYSzBo3OUqeMoLgi4r9DMRcgEpox1hD9isnmgc8CEgkkCmQ11u2s+VBp9tcCbgwzZ1M
-          fJugpOs8NlD9ydn2eIIKtiQBKaWq5L/4gu/sIxBuc6ZgXPEBKooSoIlfTJcAQymB/Kl57L+Jx4iK
-          bs3v5yMwHYNqaZPzAp+tU1KbW4cernxL2RivE8VgyN3t15pTsPIVemDxx2Cu9f57X/Q23K/uUimg
-          hdFy+FCzg6ifNqYywGi5K0TSnNT9vvZHFa58GpvX+9GY7r8WseoNpGW3um+PzfkGq29uYkfDC2Px
-          Pkfw0qQOvpXfwp0Caf0qQoYPesg0r+9yaIQAQPjFaMW3xW9iDgrN0NJgPogRe04nQeY+Q0xz0dpU
-          o2rEsdLG6Ek150Ld5fGpWqi3bYH9qWRgukBjgpemPuNbcDgyYf++h3B+7vY0FW5JP34tlcCN0nX0
-          cOMAWKZUECBQXg69B2FnLJEcysow0Tv68YkBcZtCIudiRlMX+P1QfpgN0/F0odnggHz5hmUMn2Kt
-          kW7988Daqw0fYXqhzy93ALz6oTwcn8hEO7Or+3nlP1DlryPdm9bLZZVHUyg0zoFsQ16PfnwdPgzX
-          wPstOANi4ziEql8+0ebIV/00XM4cDPpyg33vJYKV3z6gZmw9bN7iEGz5WgzB+vepHaV7Ntmi1sLF
-          aVSci3JZLfMZmsDJL5T6+w65c2pFGRSWS0Y2ZXIHs+RkNuBas8Nq8wTRLIMNhP7srquL8RCRVxjF
-          itDE4RpPwJhL/raAVb9Qo5LNni13qf7hH3kfQgVMqSQN0iPMLtR0uG3OBHmzQEd6aPRAL0fGG88g
-          gBYy93TfO3M1rHwciNok0egjvgAzDof1+eOY5hw5VqxHzwVMb10lskzjfLZ1uYUrnuIrhTuXgofa
-          /PFr9IYYsNhyd398w1KbuGLy8wvBr/5eZJsYMz6KAvxKTUz1rXuItp3YFOCyji8eOc7sBc298nBJ
-          PjaCgTP007V/1vD9ffCIi2PdKJ4qCtbBsDa1mFUbxFE2iazuVEyvcvp12evhheCckAktHU2iOUkL
-          TrmyXKKxqllRy75yAUVtkahxfjCwZCiQFUcSLmi2iktO9g3X/um9bDDPQMihEYANeb+od1wH44eX
-          YlJq09z8nsdtV/0CBVvb/vRLxJ705SgaZ0/0suqXCWYCUfjGupDFmHuDcPQswOAW9dTZH/18KeJ3
-          BiT9c/7VP1f0zLwAuwZWOGU3s5/5pSVyOvYj3pte08+rXgauZXSEze3OJUxXBLDiF/U3ZucO8pDX
-          f/rJ2JZL1W06TQfny96ih/tW6yc3PsgAz35ObeGWVL/fF+7Ni04P0ubN5qv5G6A46thmz08+3fHV
-          gmHgHBF87TZgcmO8A2Z2ZVR99nG05ksANQJu6IvpbDBRVHQ44hSTuRm5/C25Di+v+p3upaE1puyd
-          meDHH7HlXfuvn98GuOI92brG5K7PZ8Pj/l1Tc/sZ8xls5Rs8u8GRGttrl69414B9hVSqd5Wbz/w8
-          6T/9SI9RYeS8yhc1SE7qmfrBeQAsWW6ynJ1IgviPC6Pho/mccoizBW0sKza28Wg3cG9KFT6+YpiP
-          zU0jcDlUG6yjxIy2PWOWopARIklOj+5WnxIVJppX4HMR2e74PZTWn361V308dsKWh9Uu1PGKt+50
-          +txtWWZviN1M2AJq0UGAK/9Ac9Ude5IrdQtWPoj17WW9Qk0nR9kO8/UvHlY9eAPXTzrT5FcPTv7J
-          g/AjUGyi7egOMP9koOYWCaNN6LtivI88ZdUbaFr9FHY3VUG5Nt2eyML2CpipLY91dfGMj3IdG5M0
-          iQR+smeO2uzYGaNtVDHMPbSnHr+2qPFrx0M2jF9CdKCx+f3cOlBRUoqmQGHGgt518Oev/PyBnoiN
-          ruDzwuND5L/7xeQLU5lvT0Jdbh6i2bw/F7DyJzRViWaM5H5MoZ2BL7bosQHL4+vr0MvZFvtvfonI
-          uRV0wIZLRMDKd8S39ZLhnUl0fV+3iE7VxoTZqXCpQ7nAEHLoBtCf3g4+qD2O5vN9b0H+sHbE4meZ
-          Tw9UpNB+xCo+b8GZbQvbmKDa1hM+GfeyYuWtkwEY6Znut9sgF86toCrkbedUN2bXZQnjQsgCW8PP
-          uO2qkXuFAuAv/pXw1/vXJSZfWCAcLjmBl0Pez0/nzoNYsycCIqnpF4E8bLjqB8L3euz+8ee+SyfC
-          H0yZLefuksIPekSrv5LnYnhpFxCo2z1F0wdVg1Nl69LcZ0KmmO/AVMgqgfwSddiWIxQtUNUtOB32
-          ErbUgrrdfcAT6GrLRPCx2HlXg1cD1/j/1Yd+iXeq8PMz8Oq/5WOwzwf5/iE3BIEt9uPTFT3l00UB
-          4Y5LX03uS4rBaTFEalxxYyz7xofw1aZ3eqw3X2NOMj1Rgh6cqW4QO1r89qODPP1U1FKI6oqiaMYw
-          HrCK98XV77cHq7HgKLc9jpxpvZK6Swiwv3DBd33SKn63ufJwjR+y4nk1KRcjgPpVGem+eDJ33Gdn
-          9YdXOL17HGDitoSKjQqXZq9yk49j4dnyca/I5AHybTQhTizgFuojkrIMMyoKeQtZ4Gj0UA9uvtSK
-          w0EyXwiiz6RnbKo2Fvjp/1WfAUaZ3EBPNkOc7a2uYnfTFqCdSV+sbY80H7c1amCg3ihGv3xZ8VK5
-          hcmHlGenz8kSSglUwrIjwuqfzT1zMrgJPZOa14vRi90jD6ElOQnWrHtcjfgwLgC/dZu6opq5szPf
-          B7gk3kL186Htf/jypw/0JxcaizKhAhQhf6fG9knZbF3CAAZP2cdGJycVc867FG6htyPLWm+WcXw8
-          5Ne1elL8vJJ+PBlaqwBjndrCkNYL2nH3gDHgO3p70Hc/NlkRK/c2VKjdzyKb1noL9yaosMdBr9ri
-          44b/+Y34EDk6275fSQGPMb/D3vei9fyKH1CE5p3GhDXREL4NE261kvvpTbak54OlnC/R6efnrIsc
-          hxg+PdmiKjnsoy31uXXxFH3Tww1IRpcrdQeaTe78/Ip+2lhPC67Pg3H8PrDlUr0QfIR1SWNEVz9A
-          pfBXf0nheAhQus0h2DFXIGOQr4t1jH2irP4Zklqp6qeotGNocIpHtad0yacWviZljQ+Srvn28yvB
-          7jJibBSqE82HvTfI2Tje6YPPq4queAimWbMReJ+P/erXpUAaI0xkz5DYUMj2AK3sWRHweOm9mF1P
-          Djwlnkw+150ZMd6LB6iERYf946MyplenyLDKjhPqBo0Zy5ZuStjuXwN2ZXFmpNc4B7CguODw8C7c
-          WTp+H2Cy1sH+M44Zf/p+ELQlz6Iqv7VA39275udfYbz6TVP1/pp/fOqikmX9VlLuoIc9j+bcW8vZ
-          aO1kILtQp/tzPxh8Ve9SQPsAYucdIDYsfSTDW2jf8CmHsjvrxIXwcrloFAV3J5omLnN+furqB0ju
-          ooNHCGMtBtRWpSWfGv2TwNp8npGy74g7lW2/+r1LRA9meTB4HzETnBf3Qv/0UMu5DoiDGuNEEu2o
-          NQP/IcPP44pYJyfreUMLpJowk6+cfn9+h/6nrwQ3y6MifLsmOOwrnqI8jsC8v9Yx/PELnPdPg6br
-          qmn/HBMcn4aEtf17Z4Knt7PwacXvwfdPBJq7Yv/nVw7refz8VSJKmzeYVEGDkPN4RI/vPaq2Kx78
-          9AoZp8QG/LW/NODPTzlMRcQeLK7h2i+ht2Oy72cgVDaIZnpCocVvwEsGIgf5WYqxy5Fj/6e/7Wx8
-          YMxXmrH89GlDlR7rKadGfLDTEvDjX7jyjV5QuKKDsuf41BK8vF9uIGwAqWGAb49BBzPFT+6Pn5u3
-          4g7a13iAMA4yE9vcwYiEE1oauK88lTpRb1RrPS134kn18W06IWMJwkaFu6b2sdW/IKjTXZDALWjX
-          95U10WgMXQPtR6JSh20/4I+vnhsb41h1SLXcpHVRMEdPqL7GBRjeYckrfNLuqKn6p3wOJ6hDYp3u
-          NJHENpqcKhsgTN859vNj1E9O2emy4N48fH2drtXiCekDrvFCJCos1Sw85gVKp9DGJvqWxsjPkwrT
-          8VliW4VnY/7Vt+N0elJLvXu54MqwhjdSSahdn5+f4LRA/Xr1qE4uarTl2TUFK77QFX9z+vN7V/2H
-          pua6BSsemwr3kZ+ogaIcrXjQwWPx3VJrjddR1fvV/3DOVLMKMV9eYZRAtZ3Ayl89RtNxt/z8eZzK
-          Qd8v+cbSlWZ/mqm7+t9s6+h/7xPBd0PzIY4aDy6N+iHyq3zmS5GIDyhqWYWNUy3lzDNK4U9Po/hS
-          s2EMZwHmbLvBBultICrrYq7jPr9i44ZTY+0vZUD+oAfWoSvl7dJVHtSim0BPK78f1noLdunhjXFA
-          AkP4xgaBautlOBVuQjULz/KmlDTqiJhpQz+pn48Avfzd0Wj1J4eZ4hisft4f/pHkc2tlC1l7JJmg
-          y6eN6a8lwvPw471ZquEgmxmYA/OE/W/vuHxYZ+kPf4lIM6+ilWvp0D/bFRLU7FgtnjgLCno+Dj8/
-          dP321glAAm/gL76X6ylqQXS4DtQo1C4nYJ/sQDedLmTnjpIxSfQRS0MsF2u9bAC7G4Et/+rlkY1O
-          LpjC8QbdOj4g5an20Rwu50wR9SPG92+pMfF69xpZUg4mWvGxX949MaXOchV8CKO7u8RSiRR/+jjY
-          Nj8vMB9sfIPb4Y3I577VKpGuLchTct3Ty/p8yy77eGDVY2R6vhGYEr8YFKd7G9SbXLsS01G1lGdL
-          Dbo3jxVgyrGFf/1C5+18I6benjWUTrcI39f4Ei7lroR2V0XYyzXB/fF1mAxfa/XbLtXyOnUdXPsZ
-          ZKmWjzFt9iSDvvWcsDupdjQf7MMNDtP5sPYvAkMIORTD3QVk1E7OW5eIXTfAmmsIAe8X3y+fzxFC
-          BtAJq1XyWvHp9YASjFPsRhvPmLpvX8DrR81xdPWMaGnP6ye6NuUIf/7GxuIomxh68jpFUsr1fD4c
-          kgkIF4v748uj0WsIrn4rfsQXkzEiHVKQaLOGXdHa9Mu5e6bQRiShVjEmhkj5PoRRYz7p3XuJbPrV
-          uxRKGj1t5ChitXOO4fTOEnzMLqPBFOfUSc0Gbf7iZS6Fhld+8XhthQLMw7RL/vKxVg8um3/9iGPc
-          jkSyl9KoBClJlbXfjA/K/v7PL3W6j4G9x1CylpzSG2g3/Z4es4wyNhwdCMk5btAuytuqX/1DuB0+
-          CG146rl84reDvDEuFtYt/gmW1a+ArOcUNG+toSe/fhRMrwFNilFwp1//NDnpZ7Lrebla/cTbn18X
-          3g7Z+ol94Ci7hRXUfllSPvzwafVLsfvqbmBKNqiAj3AZyc+vYe6elyGX9i71S/W+LjrRBOn/uVIg
-          /veVAs7NJYrmho+mPrM7GVbfhkTh4xix2UcIvrXRpgjpo7FcvG2oRAbbIg5Y22qpX08BPr+mTk0h
-          fbiL42ct9EE6U5/urGhr7RsbOHPg4jiRL5VgkmsK3E8XUVTWp3xqz4Wp7A72sO5CcvOJ2w03WLwT
-          jJTMrtypM9oCgo+W4CsVdy4b8zGBBTYPNBxr3eXh1lSVJ0k0etRynNOLbwvQvbUBftaIA8MwtQv0
-          SPYg0Y4/Vax8VLqCbqqAUdPMVX99vxCsXUshzHg01bTIGqd4UlmSqrdIzh5ED+B6kZimLXOiYZGP
-          HERWwLA+xCWg5636gM3m/CSctGwBk80PD9+7c4gfYxH1/LRutZZNaU/v2C4iEl55BEVVfeCjxXC1
-          FPunCqHOntiHceeSoys/YBnCD03dTdUvF2FrK/L769KY3gBgg+mbINheLxh/u001iXplKU/iqfRe
-          yLVLUJsnoMiQS+rOzHvxo5iOAtOqJLz14F12PWkJdObQxUfvemLCnkprCdRuGB1QD0Z7QTF4uXWC
-          0Zacc1GIqQOfX2RgW5yHahqmYlKExxThlHv7gO+vaqlE6MRIFl2fxgCWiii7pWX0Orqpyw9WsVN2
-          ZbkgniueLi+Orq68BisjilR/c9ZdvAkKR/tEDeNm5WLlVB7kZ1nH6G4HlWBoVwc+Sazh+4isXijM
-          xw6+8uFC470TR+zi2zxEVsbT4+d9BmLTuBxEQyFjJAlJPjDQ2dAJHJ3Giwd7ooa7Tn7y1p4aRTAa
-          8yFvPEXYTDmNP4EcsVtyhBBGbonX+M15oesI+NqH9RbQUXHJ5XW3oXvrAmpbb8PlecVJFOWZE+zE
-          87MSyuUZQPc2nWjq7c1KTAS1Voq3hxD/tK7GzNWBA5vH+YLTNP/k4rmwS4htQ6OHHVB7Xsr6DKJk
-          t6PPMJPB4q7XPl8zfGJ/Tj+Mdck6KVWfn/TeC5YxLfjIQWBdT9jcqnc2u6LFwScn7PERa3POro0W
-          Kp6/XEeWtN0abxcVnl7XkkjC9+pOQ65OyvYyWTh1FSMSu6FelO1YHOl1qdJISJ43GbbJdsbYMnfu
-          6AgWD9fzQMppp7nL2Y4JlPF3Tx/5gKtJcQcPluPwoX486LngeFkhw7av8PH4nqrZk7AH3+M9IGXr
-          vcDWHPYNbF2RUB9bqjHJVlVDuZuPOOpt1PM9LVro2UWG7w/vLx+c3/vHvix/o2VPr+XfefmfwWQT
-          WKpBEeosIR9F03LheU0EpRW3Emrp23WFw11WZRHL663W9fcJNpiDDRkzbBokipbsdBfgWp8wtmWp
-          mu0HKiCU3A+1i8rOxfvQWbC1PhTbt/uhEm/9F8KyQl/shPtNPjsvS4bf6MBRZBZuL5KD/4BPHlkY
-          5xys2PyyVOicF5PaFbaj0asONpTsdkfIJhAYjeZCV8JdNWGDhRQwHjSd8g73Lk4Dd5+zd1W1ANra
-          g9qn6t0Lmax0UD4fD9QAxanih327QDfUT2iZUiPie2rXsDyQEt+rpjaa5Ak5WNGYx+lzvVXeXYoY
-          4qqyqZ8ulSsWN31QisxzacSTCUywMC0lQJKP7TXf57Wew/d4xjiiiEbL3ryWcnEUbHxkIO3F2bcQ
-          8PxdhpH5cPNhvLUQntwop/7XadzFJKdUEVAXrvne5Ev5uDyUYFuliNUpcRf1s2ugCNWWhuVQRIJ5
-          kUL4lBpr/f0MdxpvBSeXHnzRXz0Xavcy7eRAO9I7la2evQ4dD5/RQ8Xm0YsrfgkcC8q95iL2EFaL
-          NBlr+NYUB3GVPOcsuqm64hchQVxZnyK+VVULVley4OP2egXi2WkLKBdHh2bJcQO+HE4RFOegxZl5
-          vxvzpglSJQyrHqM970bTvNPW/Owf1J/j9avNpJOBJAEd64TTq0lQS0ERhbalV/NA+4k8Xgg6oLVp
-          9CqHit71zoMCygLEUrOvFtO4prDVDoTa19ECszn6A6ye1oJq+fao+vvgqDDUjQ/FuJPdxX6GSMEh
-          MJDU5ZdcJDc/hU6uG4TdPO9XT23YLm+ZognxLmPvpgYf51xSFFpjv1jTtYVbbzGwcfK+YFrenQn9
-          aVqo3noaEM/bolTEPquwPk9Vz16sg/D3vp/qcGEsZkWg1K+Dge3j2QACCkECXxZa8WQ4G7OVNSYs
-          26HD0ZcnbAnUfFC2d8ci3KVBTNyGZAedorXoVZ4zl30+g/OLZ4oGh/VTT9VaeSXNFT+im1+JMbM9
-          WH0hjyMnaaMZfw87+BK9B32+m407Yw7zv/PEKZrrnv/hQSRHHI1M4ePOG7BLlWoby/iRpZS1/bWt
-          AXq1Ek33uDLEJ9VT+PUvMuLKIgDC4b7o8NlCjaI1fiZFqBF0iuVAjbD8gr4f10G/TVVQ8xicDSae
-          aQhPTb/umiMTWOxjZsPqiRhSmsQEwq8eQya1+Be/wkNLCuWv3gj5UC3hNbaVN8MxtdP8EHX+CSCo
-          INBh25g1IJzSmw5l3O+xcVTbfDk/4knZshLT+GmmjAmbj6MIxnKi/rOoonkDglh5cpxFNoXDRYu5
-          TiGQvhmHyO4hGMPmbRL5fRsdfDbgho16NHpQuGYRIWra9YttyybwHnqO7eLTMvownQxK0QSxj+1i
-          rX9I/uEZYu/BB8JDf3RQsjOB7Pr8y+gVfAOI22qPmGBO4JOI9vB7fmpW3hWwnU5rCBT39lePP21b
-          B8q2DR1s3Dyv35rau4U8bHXsPx+Vwc5PO4Vy4TpE8Xm1n7GCCRBk54rKXW2CATqDBbxqeuID3GjG
-          lsMBguBmRNSu+2M0cb6XKWVqvdH1xs7udoOnAFaoIYiLd6SalKK2lfbzFuk1/UaApfXLVIrMcsiO
-          9S5gF7GYFDG3Xz8+yEj2Vogs46uBkUacqr+CVyhVx9uEdsHasjkvfAtfexjjRxIQxjJFeyiO06r4
-          2SZPMHNSYCmi0LWIezkTGDfyFELxPL3ILj43/SQkjgAl/qXTqLdIxH7n9zHXqSUlHA260z43WNWC
-          QNd6XLH76DjgrXxORNa+OWCp9FWhqwRnfAFflYnUfnfwWQnGuqu1Yi0ttEx5e3eb4pOVu6yw9A7+
-          +FG6RH319/sJT/2MDRYHLstC7QaKrRAhRSqPxtRd2hh+71uOhu918c/xiTpQB1RFuziycip8qQlL
-          DvXU/3pWzs87bfjxI3q0xMmdFLdG8utj3cmmkR9gcbywgDttGdBD2SU9SXf3BMomM7HJFRuXlMsl
-          gOFYjSSU4NtlO+3zAMUbIew/UAWmb184sOE/T2zvfR2IefuqoYKkbh28KBukej4bID+lA35Ej5Et
-          iaOokFN7iTqX+6an56edQbcszmjdIp8zYvv1j09i4yF40SCmrgk/wqGnxjlwInEWiQ3FXC+oDoc9
-          E8x0zmDITvOqd+achATaslKDiXzJa+cuPsxtiPV0Q+PAFNafv2/gS0QPaseRFfHL21FhQTlEMvOu
-          GCTKYALcpU3o/Qxf/eKIWQmV7jrTA6hUg90SDUJBtq9EkeVjLrzypwmrB5zxXaqPERvIW4dY0o44
-          ogHuWfxoYyj2Rf2nJ9iLlRDU+xrSB59tKzaXlgPBLToTOTpzFet6s/mrp9iydjkjB/8G0M0RMdZV
-          sV88nO2gL+gzNu/Drh8U17NhFGsi2m0UZCxZsF2gPPWIopc3u4uVSwuEXN4g6fgO+k7vJk5BN12g
-          2cq36SPVPcjDQqMxmB7RrFSBrhTm4BBR0bRo2lRDAZvrPsVprlk9z2AXgtB7vemxy8WcBEPOw5Mj
-          pTSL500/7onPw5I1A/bLuK5mD2EEg/h7opl5hdWsnWpOeVZExzYQx4i17RDCT79+5ZSE2JiA7U6A
-          U78S4VkNwRLP8Q3WQb3Bj2AZGV3rl+xestNaD3Qg/vALJcUWydH50XfnBXYQvObLT0+7s6aMJnCy
-          nY6+u6Nc0YdRJjDAGqKGEx7AyucmudSTkhr8coqWa6U84HrXnhpiGlTLTx/hnWTjozBKPXvunRJ6
-          X/uJcSlwObtfdEve7XRGDXVqoumrtg5whM6lKCqIwS5lIfz4DNU7/hXNwebAgaJrED3wouqys6s2
-          sL3sGTZRu81b8HB56Mz6EXELx/eMPzUTfA0oR5K8lfvFA5EKBZqmFN+ze7+ku3MCPT6sqC8VLWv2
-          2dwp/tnuaLTqvaXiLyUMD6+WIlgm/eL4YSuv/gG2lRMxhgU4llJMdYzYzfEi/jsUAfymnw1aCFf2
-          S5AwAXyCQ4PtDfCZcBojTgmeLiaSuse58DSSCfrbieI0lN7VMJVfAt3PEmPsIClf3t5lB/G90rE/
-          IquinWYOMJDnC7Unpe9Z7VUPiFNXp0flnefsLfWDvPssM8Ydt4mWw0HiILrZIjUO5cGdVr4N5Ky3
-          qX3fftjQX4sawAO4I3rT7/my8m8poH2EUfQYwRLlcapsw+WIdmu8zVaGdPh5b7/oa0ezMa7xKa35
-          S9jNHlzK6uYGgfjK8PFG+X48WJYHPv25w8gjfj57NQ5WD7DFR1VD7KtzI4Lb3YKo/9bfxmLlVwJe
-          bhzT5xXd3ElISlkR6umKH+9wBstDexQ/vUVRTjBbjGFGsLE/MTbqlBjtXGkTXOMJSdoFASYRSmBb
-          7gXsmPe7O/Vm0YJXMmQ0MzRoTIR/hdAv7JEsDSkqdn6qqbLyb7rWb2P2Ab5B97QueojOj2pSPC+E
-          5aF5EUm6Zzm5d48AVIifcPo8HioxsgpVWfkdzoJxYyyGOlswvMwdNe+T5C45PPPKe8SI7GytNpYn
-          SRb4wWKF/YdjRIvWTgN4DWZGbckfe7qFzQKjx2tLdaaWYPjqaqxs9WBPzXN96f/0QvRlMmJ4+bps
-          Evbwlw/0GB2XiGhcMCiRfOKwHeA94OXF7SC+9wZFHQmrad4dCVDkL8OGeJuN1Y9wgGSnAl7jq5rV
-          kycDoS4veB2/aPRPWt6gcG3P2M9TPZ+GvOAVz3AybNf0yPr3rm8A1F93bKC07WdN2atQod+J6ive
-          sto3CuVZDTq16avrCY5BDVvtMlE/JWU1q0HNK6veRiOVrWpa6henxHjeUEOwTgajxz2Bv3pzPLEw
-          3+L2MEGFggXjXt8A9sy6Arw1amO/599sEhJdgNuws3G8vs9pCXQTvpL4gm3uS6J/56UcdoQnNnSp
-          bFIeBPI3wcZH9wEvjpWpRNfXjhqmcDAmrvJuEFkpj32lOVbL2b4N8Kcnjv2YsMU0pBuswcakCJlj
-          NLoi4uBarxG5lCRavHqRIfo46+Kdqvz5ZRZ87/wTDqfMqBYV7kpY7riBRG2GosXdZR1UTLf742tM
-          nj5QWes7tj+XY/7+8a8fHvjYVnPBrcIByoKBqHGzPVfwYeTA8t402AzMJJ/Gw1cH0bXaIf7o8VVz
-          EZRVb5w4igtLAX98beXrZHNFN6PfDZ8bDOTXGe0kPDIW27YHnX2qYbMdQERnr7Hhbqcyauu5xsRv
-          XN9gtHkp1Hh3Y78cDqedgqCs4KMjBTn/1dVE+eDPi6KynN3OEZdakR62RI9BvhgsZjaCuH3tcSQ7
-          p0rsDQ9C0WnveK3HEb/IGoT8vGhIeZvvfFHDoFa+d5GjtlXpgF38Ygd//lB84GODbduGU97K+4Qv
-          O0mNlnSnpLA8CBXFanuJFqOfLaWKb4xeFO2Vr/4UBJK/E1e+UrgTWHoCP8Wnwz5MnIg+c72BXN4D
-          am8U5ArGMHtKU+Mb+fkZEzuVAXjl3JXs6u/RFXAXNQoPO52QxCPRz+8Frz0XU4NDy798CEzDQ0zQ
-          g34YLFuAp5fxRFw3hWyx+dCG1dEjf3yahFeIoHDUA8Tc5lDR6fnmgfQtN6Ttn120mOMplFtvP9HV
-          v+1/fjlc+Sb1t1lr9IlQdHC3LDz2ebM1ptX/hcJ3SsjyStRIvCedDt0yiwkzMst9c/UOwVPeP7Bd
-          nex+GvbFAottfaIp90GG+JaM7i+//BtfRRM4ujIM79oXHx1lAjRTvgV09plGzXbIo0Ga3Aw+pdik
-          YUOKng3GW/35U+ShLb0h3JPkBkv26KnjvR/GzEmTA8Hpm+A4dkSXLahJof8uR7J5kHs0/vBIfs8e
-          1kH4dunD6JKfv0d2V/px2c8Pc4rOwsgNvVy4uIoNXaW40JV/RCs/eoBinUBhW6+yYjn3asGK/zhd
-          //3o1TgEa74i5SQU7A+flOeVUOM9jIydHfsBvY1aYUT4KKLbD+HBql/+9MWid8EO1hBbFKk315gW
-          +QjBzz87frWETTzn3MBLHc74+Mz5nqb464CVP1A/M8t+0bt0B3GVHyh6PIJ+YqcugP52oRSj8Mq2
-          to1uEIZSsfJFsxdXvAN+VnZEVAytmt0S6/LPv06Xcw9mxZ9U+C7HgOwOYllNQtKtU5qlhR7FzdVd
-          nuljgJ9MbLDvL+uVhFRHyur/YF3amZXwuCT1rx9AuGp3cqcNrglElrMg+bi9R5PQ6wt4weGJH3dr
-          ypeQ3CzoVctzPe9HPvW0beHKN7B9eu4N/pvYDmjSzWXlV0I1moPfwFcupNiw0Cf6+b+wFjYHtPMO
-          NVsuwt2UX+KjoH5mlWzk0OQoXtQ9qP9cB/E3b5eHZWW16/seIhbbRaCAj3tBfJrsqj+/QMzDkmze
-          Jfcv39ryINCVb0WL9p14KNmFSBb+VoGlcp4d+OXLkdeuFXvZJQe3iuxR37sV0azBgYNF9vCJhDYn
-          lz3GDgExaCucrvg+LfUXQmGr5tRIVdcVtCgYQIUGin1/Ug3xfukcuGUZonYo7iMxlV82lDaGSnbK
-          GfXjyvel1V/EKEtwP42LChXoXx/YDwqVCWu/C8rFySXsZB+j6TuoHoTtXFMjI31Of/wZhqBAXCwT
-          NtBCS8H7gI/4mGmxsfK7TIlstpBd/3SiWf/6CGImrVNCx2s/b95T/BdfK57li9buyB8eHf3v7NKv
-          ajZwd1B7eiR3sSr4tnz8i5+fv27tZ6Jsx8Wl+o9f3ZNSh+JL/R9p17HkLIyEH8gHkyUdScYkSyTb
-          +AYYY8CJJEBPv8X8e9zbHqfKNQWiu7/QkvpN8C+N+nlXDymoCLap3d9db+4mtYAb/yGx9/40TDnQ
-          HVCy3CKna6Q1/OlFUtgobxm3pwsPVisPJxh9tYHeJEfpWVv31R9eUBzZIxuN/Sv567dQUmx7hft0
-          2IG//oNZVTIYplJz/9VbqXh+vVm8uhBu/jKWppMLZsXq37BOzdckTaT33lt9hRh2gB4lVLE/fxm0
-          592e4vc77KdLdVZlIDYZXgFum3kv+5XcKKWE087p4lWN0vdf/2bi8MCDST8HELbqBRGsrb23Hu8h
-          RMo5dyb24/A/fY6+2nkmm19orK13Xf/0PFYMDXrv2n5kf/4V2fw8j33egw5Fdy3JUQnVbbAyK+Ap
-          i77EYJcgr/7W1z/NN5oG5MVmfu+2QFCCjJ4uqu6tThUlKKDPaJI3fi6eH1WJyHHeT/LmB0zH4036
-          40s4PRDDYImjJsg9VzpJUYgNgSTsjbxBiTZ+PwHWCF7w54/QUy+8jTU9oQx+TFRjtPUbado+N/++
-          98ip+ZrxsmtTG0Q++xBDt6xY/EBTRW5fHentLMfG5q9D9PyYd4r5jgPsuXbcv/puP2U754DjSUoF
-          cExMD56NAT4MCKWj3hPjdvG8CSOWoG/Ic5NcLAlgb6Xh4DP0b9Swgk//50+Brb4QsvmP/TgffFg3
-          5uan34pm1D6jCjY/mzj2ac5nLnITCLQwpk7tnA3up9n4X/8v9Y8mE4rvzgR//gAOuMlYreImAF9e
-          W7LxB+OfP7v1k//Vr69kND4i6c0kN6eZ42EWfjX0T0o2/fRxYuxcqTWqHm+POsqSGethWgq0+RsT
-          46aZzfAxqPCvP3p7hxdj3stDCzb/idqXvRWLN6glEBROQsw9V4LleLQSRTZckeg79dCL2/sqYQ8q
-          Yrwm7M1c7BYwSPqQmJu/MSur0aGvx0/0lp8ZGI3zWMC/fsGpq97GDPlh85uu4yTv5Nz4slDHaOPn
-          //JvJOg4wTBY7tT4cRhwe9msYFWZAXVJWHrLX78NJf1ETxVs40WPX/j/Gnwg/e8tBfsvmanlSB2b
-          70SFCOQwJLrrxDG72PEOtjT4TnUdq0zQK6NGD85WyUUMPLYceiFD/m46UtWpnt4UaglEHI0G4sSr
-          bvCCfqug+Bkbcrw7hSF4a5zCYppzak+y7YkZ6QdAeDsivvb8emw2qAq+n5pO3Ik2/ThI9wuYj6aw
-          7cqj/TpI9zdMpgBT9yR+wBocaALFMztSCzzGeBZ1NUJepU3keCye+Sy4Lx1J7/JEzMZPG5Hfpdup
-          d3iiYTXjnP6+zwsS+e5D9L4p4vbv98tPkPDCwQbwwwYhsncHND8WevyVD1IHedVtyIFHOzYnNo+h
-          erU6nDBn6JfmJiWw+ug1Iandejzd7yxI3yeVJoen5bE7lib4y28fzHgeNL/n+1MAZXS2i4PSYzN7
-          a6/AsJMv9FpJa88E3VFRluE7jfBXyZe/338/FaXEOGmMy2L+gqwI/mj4naxcdKcmgpNefqfRfj0A
-          x+ksQnfuM1P38k7y8WpqFnoLQCanitTxejWdBPYP40GOyznsBeX86uCtSxA5/B4AsONHr2GNfgU9
-          BboN6OEzVLDOho64ZfIGW7zsUDfYiJrsJzYrCPMI5jOS/q0v9yzjAlUH0yI3DmiAq7vHG8rd+UBc
-          nAqAoWSVYFacHHr+lhqYCxJCFChaRQhvvA3hbLcThOXckGL/PvX8Wf0NIC6qmqaHqjGW+n0pINvn
-          Pl7MC8gXvdAUeAjdcVLquGavBH4DiI2nRc8IXLb9v2EK6fEsUJs6p4aV0c9ERl7bNLLcazygs+cC
-          NBcquSFwBfxFTQcou/cXlvliBm9HPM4wIbk5yQGD+TyLnYVa7cYTv/TbRlyNkwpdL9jTYn/5Gkui
-          KgPcH6IdMRBjzSy6hQpWb0nIY3ndPGEY1AJqHYgoCVzfWAuiFchDcKF54HrGUsT5G+CwdOkBuXU+
-          SgdJgYrjExr1Dcz/xR+bHhaxAPmAQbS0DpwXNSTltLzBbPWjC9PbT5xYkEfN4na/FAja7jG1ff9o
-          RjWtK7SiV0C2/xezYbAnpHVyRKPs0MQcM8wIahmqiVkWX28cJf8CnvY2a+adG564C6oa6dlU08O3
-          fII5QIMOHmZQUbNEorekvscpU5NOpDijxljlE5PQ7Hjr1LJb5PHf0cdwjPqU/tWL5cNfd2DX3myS
-          laaW89Z2Sg88jIaYVqQ047FRO0gELE0o0Uwm7NNGR9waUmJlggToJDIOnWtTxKJxf+VjqCUTSspL
-          SXUwn4GIeaaD5cdJ1Ch+gTfwrqAjpTMPJH9dlYbBWxnBKS525L74fsx2BfrCzI9vVI0esceF2qsE
-          J/a5EKtMDkwQLd2G6bs+k8MMTSCc1WAHS7Op6JFvmbeGWrID41pJ9HTXFLbyeFLgX7xoxuXlMfu9
-          1PD6mij1Fh41Qzc8d2hUu4HoYhXkLOBDHe2itCCX7NDGozOxGardcaDesahzzu4NF4n6eqB/9Y3H
-          /Prve+C9iqGxPi5bi1rbPchDjo69sA3WhJo0WyQvvthbm45Y0Blynqo4CHvOmUAFX/wRUeIPe4/d
-          r2aJavgWiLbz8nzdp72qNKeXT8KWdvmitvsOCqFCyCGpF7AYrd/BYjd5VD0sbr+oxd+p+KM6SYfK
-          8IQqy1p4enQnar2bT768Xn0JvcqYiOcXUr8q4RBB/yEgqsLjjYkT/Vow+Ogl8d/dp58lbGaQmkFL
-          Xfa8xCLnCjNwO9uhaURIP76foQW/58rAH/N0MbjU1jGynopBMFPteOXwXoL7M56mVQw6wMy8ypCU
-          vm5YAlJmiPvqrcMNH8j598j/6lUr7eJ0pckbEG/mrf4LZO8BiL1tll/kE/hC+/s5/Xv+Qi+0Dh4q
-          6zgpp4+ec2dYqejxiWtiirbdLIT/1RDe1xM9GMJi/NVjGOUDpGcelWz9jmYJxbYLMbdKhiFUl5sP
-          X4FFqJfehp69nmELdeWtUz23UD6KrheA8dmu1IJk5y2Py3OHtnzH+5kze64gNwjhyiCWk+gUCxHS
-          K3TLHwwrlwsG7JlpAtzxV4wHYNX9cs9PF+WujRi3TqV7f/iHdubzS33wWfp1UkkGD4F7xfyPCHkX
-          S7ONrsn6I+oNVQZ9/cYEzorxncZjUnlc/S4xdOxiJLGxi2MWyFkKpvZmkuINFmOd1GP6h9d45bR9
-          w662C+G5tkR6OuslWIv88oZTmOkT7N2oXyaRXcA7VDwSXaxjzF38sYXpsfeI1cciGA6Ppw5DZx9Q
-          PTvbQFiBBlF4VRKSTcsLcEyZS2hf3k96Kc0qF4km+uimBCPJNnxcvC7NYE6OPdVSTWNcTmqIbt8k
-          JF4rkn7O4/YCrWj3w6uoHnKhLl0Ia+XtTvtCSprO+sQBEN/jiTrf7dRXP4Ql4ledEneLB95V9hI0
-          H5eZaIl+AKvxUGr5vOghydNf5IlGA23EqXlDjx4cYqZXxoz0fJC2+qqBXgmHGj6dfUK0GX3AuuUT
-          OMYzIRe67gCFZ0NCQ8ivND1Rs2dq+3HR8TwD6si5z1Z3HVL4g+cRM9Rgb21/OxPK+a3HLc0uxpoa
-          0gVMXMKRICuJ0T2iikNavpemCR9xPo1U4qCeTxKxaSYac+Y9WyBo8EEup88znyOtKGGxMo7o2aHJ
-          Wbq1XJI9TqguqoeYivjgKjf3fqfeVdbyNZYkF/T4cKP2u+/6lbcGG/oBaMmpnZx4DPigQouzTycO
-          B1HPq+3+C4kRWVS16oexyufdF65ZV5LDcprypYzkN0y/jUY8427FS3uvJgBnqlKteA5sYYo0wfRd
-          nenl8PvlVGsvLtJvg48Bn+D+H95fVu5ErD1d2cxbzRe99FNOb8udA4M31RVsyEealP1EAXOnJoC8
-          fnpQ8/B8G0svTQmYpb1OtKqywdz+UAuYVyfkQNjLY87EBYp5yhKiW64YL8s+scBOLwZy5v0uZ3rh
-          ZvAdwYX4tVf1zHlXNcrHHY+56af1XJF/LPh6mi/62Or7QgEcFEGHPL3dtRtbdomSwuqj1vT2DEC+
-          JupzB8VkelLD90NPfN1TH56f1pPoNFrZAkNXh3eNYmok7s9bsmNVouqrS//q/9A+FQFK15dB1Xdv
-          N3MRxxe0uPuJaEXQ5L9H9G0hSYKChqZ1ZsPxE83octH2tDykGuAvwQfDt6FpWNljky2ipUnITn93
-          ev5Ox5zzusCGlwiO1BlPNGfts3uDc2V51Lxgt1/usdv+8Tni8Zzc/Iq8vMC8NA4bvza9VS1aG+Jz
-          s04s6SWPoWCpYKUYLTHFL+uX1yhi6OaYEtWZrV6MVYCBrAQ8tZoEgI/ZNDZMrS9PCjENjH96wtvt
-          jxSfeRWIVZYJYPsbv7ODmQ+PEhXQvF/MaZedPcakE3TBhtf0EK+ascaHUwJfgUmmJTAcxkvY/MoH
-          I5bxm0YM0Oau+FADYjopfVPmcyftTfC3fnF7rXtW1dP7Tx9QHRCLiQmsVtDucUrMSf4Ywx/+HTQf
-          0aBEF+P1Gc8YCqFEaOwXQsOaF7cDZZN11Egy3PMJrALEhtQlYVV92Ug0YYLtjClxxWBgFAVyBYUY
-          ecQFjyne8KkCxlBgeok+YrxcgwuW//iMCa6uIQpusUIHnAl5DLzdCyCMI3Cq3gnxBVHzuBguKrJ0
-          vdnw8dUP4+wk8ESGO7klq5TP8KRm8GjJCzU3vbQ292X3L74Otjmx5fBpdTTOtU+vP2v2+kkEHPzT
-          p6e7lrHV+OII/rI8nD47QzHo7bov4B9fjADZBofBaoYbvlCrbL2c38YqQMtQZRqaFxAPqa37yLjX
-          HFHdkG/munRLWGAJ0Fg2DENs70+MzD0OMGpLg00j52Gododh4iVQsfW0uimMfFZj4c18wCLnCoFV
-          KzeCxTkFS6CNA4AVC/BbrJacpcb8RqI5L/Qw4iinkTRnaDkqBT3+Pg9vxfw6oz88lT7Fw2BGLmdg
-          zUdz4spx6Ve2Uy7gfTBGau7Fs7HqlVfD9fa1t5bW86/+dcr98daITmsp/4tHyMZiJnq23Xr0Hu8F
-          eMfoSDHAL7beY52DYQcu0z79xcbaDV0JhoXVVFOv+5jOIMaKFxx31PM9YqwS9iW4fxOTnIJoihfl
-          Js5w4xvES7K5n1GozegSI2NafoTPWV0aPro0EqMkyJaeta9+B+VjFlN72xI8PLPIQpdaudPjzOk9
-          FzvYVxYoKtNoJ5d82QdfHQY3pyPaMhVxZ/VFh3Z14k+gklgz1JljAaOoSixlu8JbywvwFYM+KxKb
-          5a5Zm9v8RvF87/GvqFwgmB8++Pe8918ps6HIqQXx616TRyFxf/XBh2fu0BC9brucJWbfwhs67zG/
-          6d+pyqI3ejeKSA0+GZr1rKYQKqdQ3PAhYvyfnkasibCIdNWbXr/RhF1najT0ExZv/DIF4Y+byUG1
-          BNDVpbsDp6pNiMMPQj4E/DNA9dBb+DWgL1uwfpygnF5udHv/fCUIVVC7SgGxtG6Mh1VULWjxco7l
-          F7nG39Puqiqynf0m0R/uxuqIZIbJ8/wkWHezeNHaqQOP1g/wu46f4HcJqK8AfLxTj28Tbzk8nhES
-          VOiQ/JnuerZLfhKsRzpQT467Zr1ffQzDj2AS/DG4eAWnmwTJMd+RQxraMdOvkg69HTpipeZC1pel
-          gOGkFRGNj4jmFMtcBuVv5BAn0I+sn1lawPmjv7HQJIDNentOkak9ENFa6sa0+waJEkJxndZ3887X
-          dW9cwDlyjiQ5LF4j2rVkQ3MKXHIL4iVeds9jrXz4b0MxWwowa+07Rffne6JaEJN4+X1lAYpiH08A
-          1YOxMkWCUFda/a+egC9T5kLe+OYENv332tYHSFY7Up1pv3xGoTP/0zd+WbgeYzpU//EVX9dFY0b3
-          UoLeXe+JJg9JLsTSNihjDkRqbPE4LdtguXX3E4h+ufiALsDF8qVRGFE/xd1gHTeWStR3AznKsdPT
-          21UslD+/wnilRb/YvZfCjR9Ouz7iwSJFlg4WF03kmP6WDZ9XDvKh9Zt2+9LLl5xkPvLpPNJ4vCT9
-          Am7TLJNPXNITYR9jKYjjw00fE2N4Vcaspl0Eycuvqd4bS/zjdk8LNvtHSo3j6ABOcj0VXi3xRXxM
-          5n4e5ugi1903I+fA6MD3JBgRvDyVI3mgejTYapx02OSF+ucXGNv6c7CsswR3l8Mh/5c/ZngbiSrK
-          0FsG6Z5A7birqK2tQTNe/JEDWz0h/qYP2Gt8D3Dd9QI5Il01BOddRfC9CiF1PwaLqTEcFbhdb4v3
-          i+ka/553/qqQ5Ch69ytlT4zq4WeRi/N9bfoj2k4lzgdi7i214TsO7eD+7E/E37/HZimjpUV/ePrH
-          /1hi8zv4elovzIrf7K13PJdgtr2CHneG4rEI6TPcaeWFOD8c59/Py8GwFPKBWN1qg7ku9QL9jhmZ
-          5o8v9kvsWCUESBqpJucdY/Zbrv74K3Xk7RYgzK8VdKG93frxQNtF73OCHpfQJ1jYmzEHT2oKf1Jw
-          +JcvTMQHG/nzYhLs2lEuelMUoctrT7f1nHtmfvgKtSS3cbu/WAYHw2gF4ZfLSELdoBFyLzSRkts9
-          Oc2DFou349aRbIczVWnGe8xf9S/KgBnT67b+C++OkXIKO2Xi233A5hFeShS8VXvTL6u3ZKTxYe7r
-          N2pZ2Z0xtXqtMH9oOfXF77lhv6Gb4Hw71hhsepmJlp5B55ayf/7Y+Hp2FuzHnYzR5jcK8KRK6DRX
-          Jk3DSW1WHr8lZRdnKxa/75e35XuG1MyDU9rnubGS5ZvCfoANddLwl6+HppZQTg495nT9bNDVOOgA
-          R0U1fYwdF3/9KaxAEfApvY7v3Ljz7iuCyikWcVCOczO+x8mHz0zO6CHR32yRHSUA6NkKVN/06jfU
-          xhLg83OlFiBHwPZF+wXR76ti0L4tY3ZXqMBxrSWMkvUWL9QQK3BtJoM4qMPxmsc5pyzSGpE/fTQv
-          +1iAaC5Vqpb33pgHLp7+8BrvDt9j/C23WzMUdDxiGXU4Z67wVuHbNK5/+g5MAj4r8FVbgGKxYvHy
-          x38OD1ulaZ2/DO7w+Kl/z0M9FFnNUG0jKZdebIkf7o4ef/ULDoY3Wd302iuft3j+51fqTHPipft+
-          BbD5YZO6ypa3chayleBqTsTM7oPH9MqbIWl9f2Ibn+LbJ7PgFn/UMc6GMatF/1VudrGQeOa0fpYd
-          RQebvie6e5MYNQaiQP2+cFs+fP78ggy2VbTHkq5fvJWKdQtvdrlQrd1TMD0vtg9jmOkTB48KG0Ot
-          GKD1lAxy0ZRtpMTuWsHwuGsJKVKnnzY9Do8SHalzDx8xMx+/Ga7bFkq7RKm3wHOSgq2eUTLwdvPH
-          J9Gz410MI5r3wuGR1gDXWUA1xX/Eg5aGATDLJKMF7e79/A+fdb3BIulrNh36sYPZ3gPTfEOVtwZ8
-          MMPgYqbk/h1f8Xr23ybc9CJetv83G9vFGps/NomJ23mMd8cVwhN3nxb7ELJ1NfxVgXSViSHHdi/K
-          h28G5evFoMfj2LHf1XRMOJnPFgsn0WL8+xWVIKFRi1etG/J+EzpA+lSMWNFxyKfPSxuUCPb6NFv1
-          w5t5vYtgsvcTmrcibZZDU3co8hZ/WpYJ5sPZ7Lh/9T0o/Wez8ZUBFfvpiXf4OMWz855n+EyVidpM
-          3hmzv0YKqN1ZoYkzHxuO00EAHbsciXtIK8Au/jmRB18wsHCQJ8Y85ZPBDphXeg/0L1vRHWcwWcMP
-          1fFXiafX85bAPz1TRBQ0jOmcClVZs6ZfS/2ct6m5ws1vw/OkbIN3do9K+ct3CXyCZrmYkQUr20sp
-          BpQD7PPSoMInHiCHdBVztkueEtJS5T2xBKDmr78B+8r46xeM/95XeToowZ8Nr7ktP2X6wt3EvXPT
-          6+5XE8NPO9zJoziq3rLFP9rWn/rh9Gxm+TB/wew4K/Xrvu/7TY/L9vntYSDHbs+OjePCK9c55Cgb
-          rTdt7w/+/P4D73fx5gerEDWfeuODp1x8XtQS/fFFPbfu8boCZwefV84mgZievOX1rE34k6OcHoeX
-          ZszyQfoqGTR/5PoV3GY+0H0FdPRLpnrLbyGGS4UalBZEPdG6Z2J0raEt1Wjrtx3BVGawhTlUCFWz
-          iWsG6xGo8HE4W5SMB+zNo3pV/vgv0dI6YOsiqm+o3pz31GU8D2blLNhwPlgulp/qmY1u98xg7qu3
-          adZW3Cy+clbgxvfpgS8oWze/CWzfk/rUmXp66EcJ/OQgJ6QV/GasLulWrkcBhwnL4vkMK/1v/ekh
-          iYb4L9/QvX6ZmC+qH5gXMTQRz4wDbo6jw4RlX1zgs5dnkroqNsRdtdeVdehdLPQ6jpdjY3ewVfMH
-          sTb+PI10GxS59aPO6XMEq4RN6U//E/9QtV53bGwFvjWY01QQK2PxpnCGfGDGlPDt2Rsuduujjd9u
-          R8BUJn5Gf4IzMHb4+acXtNRRYXWwLKxQrc97bvd8Q57YPkm6TDX4XbJmKII/nUTiIgM20sVEIvfW
-          qJ+NcTPBW6n/+fd/+MmYaLkZeHxjiRwH/t2wde+1cCTdmRy+O58Jr9/LRKs96DSKPpd81dq79NfP
-          wGsfC+AvXuD38I6JgaK2n0+rJkHxNYhES9dLPsD7R4I/eB2JcbybMeevkYTG5lPSgwwzNgMnVeHm
-          r2Lpgr1+fr0aH/758ZoMb2z6LgcOiOIvJmnpGz1Tk+CLamdKJtROTi6g20FF3XSSqcP7Xixck4cA
-          Y79daam7SjwXOTWB9hUu9A9P+X21q+Uuc0OCD+mTzUQGX7BcBZ8eK2lthqupcijzwxtJ9iLvTZ+f
-          YcELLz/ICTlVzFxh0mEqN1fquwHXv46N+kX7l1tOAgsJmLndwkHNhgL56z9TKcIVAIN0peflNMUL
-          1skAjdPLxGAxHW+pLsEOGvP9M+35uO/XEZ4g/HVXjaiCpHtL+bCGP/+JEgLceP7r5zyPUkoT2uX9
-          99vOmz49NriuOcqWffJLYZncA4pPog3GP/225fckrVLrrc8y3v35+dPsPoG3JjY/QFtu4mn/syJv
-          1VJ7VZ6WpNObbVIwncZZ/b+2FMj/e0vBs9wuuv7tXjFb83OKfst+Jv7Y92y+VuoK6ZpCgo3tBDCU
-          ggrS8vSY3sJFNfgjZm+Umu6Bulo1xlO+C7F8e7jiBMgzMoQ93rcw+khnEkr7kPGpf4CokJBJTNd7
-          9iK4NCo8JleIV/jwveUTjRbcc/GA5/3x2bBc/ZZw9wtk8nh5rsebt+YL7N1engZR/HjsI4QrGODs
-          0bxifb7QX1pDoE8aFrlm38/DsmTQXyuD6F/gxpyzM1YkRfBB7bre9ezzKyro/l4i0Y6Xpp9Lt+rQ
-          Sxt1aisKn6/5YLtwkGuR4qt78BYlpjsYlv6DlIrxbmYPVgWqR/2G+c/T8oSYa1RIrZtCLBbUuThf
-          BwWWXyeeupNb5Kty/7agrQ/iVFaiwhaS/mbYPbFFHDUCrNFsuwD2Dsk0p8kZrNLDkFAumhxN3d/Y
-          M30RUmC89uVUq3LbLKb/zpDnlxdqOe+o6V+dKMDrY/clRuXXPV9eHyVkc/aiXsW8nH3PlYsWcjkQ
-          W3hZxvIVCYbfN0T0BKsD4I5jqcNb8OCJR/warB9SlYhetMNEUaMyfjmtA4wKPifaOcrzlax3AWLx
-          VeHrJDhsVlF3Ab1uY5q5EuvZNWgTKMz2kVj+VfPEu9Ou6PW48ZOwe+ke96D6DP1IkmjSFylbj6Wq
-          ou6k/4ix90JDMDB6wwOJSnKIHZJz3NQlsDupP3ojHxCz4b5BcFsphNzDay70V62EBcQp9SxUsTGH
-          ugqjMB/xcrw0zYRB3SIsRA09eJMeCya7DkiYTpQeTdEFXBzsLpARMSQnjYw9a6Z5RqswUHJc4Kdf
-          91hs/74XPUA8xIvb8i6qzz5H8p9Qx+Kpryb4tr2GHmKHxsOtWSJ4e9gisfNj2M/2MdhOSY86OZ8t
-          G/Dni6DDQNEMGhIlAcvui315e196R7/OWKuX+obCKf7gne0mYPXSfpYal54oTr8pWLQlhdCBsUQ2
-          vR0LYybvlP2pOBNdzUo2SIbaoT1eomn3e+nN6jznGtzeuzs5mmLHhp4VA9QUtSHl3KsxP3yMC/qr
-          B1c7ujKxmaT5XzyTBzN6Fv/gTplQeCUaf9h7UxNRC172akVDt9Di9XS1S5ivdUPdngM9Lb/AgkS+
-          z+QOrlosTjFMoRha2+i8383jIvmWweP1E0z/8v3Z2xEc6+BFCr2xc96qPztQFYpOHmszxwzfbRt0
-          SvKlx1MOjQlFc43gHknU+dlCv6rvWkKBFl+IcVoOHv/bzR3kQ0mlIS8nHh/ojxlyt+Y4CcnxHq+d
-          CCyQukZM8OP2MfrvSZig6L7e5Br4XbNybWUj9+FeqZUc7zmPhPcKhXgNtvWsmza4RC26q3NHDsZD
-          zblCXgUYfn/7/37PF+xL2Bb+lxgX9WOIb1PJoJWkBjkqntMsfYZNSPIVUm9575qx9j+p8uQ/PzzP
-          wMr5VOYE9LvWAeZneegn7rpJ5Ov9R49ilnsi+gQror/oRg8nXzCYwB9LqGHtSexGa+OVc18d/LzM
-          kWz5BXjj62MIcq4hmVadYs48GBlKy0+LZTikObvuch/OeF6pmwSZ9+UrZ0VStpp410yXfN2+N/gM
-          UkQKfLrmKy3rEv08+0auJ0UDYiidKzQ+tIBGATbzWT7fJrCOrT8Nr8JhS30bvrAqJJ34gTuzzkv7
-          FX5cjZEbiXxP+HwOF/h0sUbjZiwaEeQ7Th71X0atD9Abof9lFaxKf0/UuEpj0Qj9Fdq3R0W1sYu8
-          1WSPCT572yF+P0UxNzWpgNzZrKi2D5V8wR89QfdDAogTWFouTpX8BXJqGdQ4+5GxjHehhTQZbHIP
-          5KLnxJ9locW3FnJcQ4GNJ73CKJkkh9ytdGFDiZoEcmhoJiUGbr82aqnDZpJ+mBf0GczWebFRIwyE
-          lB7/yrvB/ipw7LbZZjg9ARF+oQUw5leqhvcQ/IvvAAXXDW+jhr+eswE6T+FGzdpc+8VXhAEitlb0
-          OLMmHq2aQqhefY+GVeQB/s3KCfrdJmncp9uL6nSS4PV38Ii7N7DBl9drgZK+44luJ0s/k0VKgTTm
-          u6n97Q65cIW+Cvc1pBihn+vxB0neKefzWFFtu0VImPxbBrLQsIl5oREbjkocwIuIAdG0SW7m/PNb
-          QXdLK5pDsMtZYO1cyJffnuRBd2kE7nAT5CMuRnK5riewfo57FbJWnUhA48JYadkV8J6+CbWvQ9tM
-          wT63oKNAQI4bnq/7nBOg8xk/JOEeszFbNlNgvAKbmNwj8MbnMW+hx+KRWk1RAC48HTgE7wMjxTnm
-          PeHVFAKcjuuOXvfJCfBWF6Zord1y4h+RadCr8bHQxh/oXWn8nHmjr8KvsN30B8Cl5wULuLB6qy/q
-          LNe0Z9P7VMBD5X9oKSl6LEZRYCJb6wHeOyPfzGL4w/D0tPcEw21esKyVEvDn1zDJ2NlmYVnMRuk9
-          +BDD7T/5yrVfG+HzJacWd7OYEApWgo711yUnrd0s8/MxQj30JWLEJO3F1jpPkMiPmbhN0OVze15t
-          dNnrFX2EHztfrudsQuNBmqgjJgtYUkuxoeDBjjhy2Mdvm28ltBuqEzH8r8p455QGwCg4QMpFJQ2f
-          iPQLf2TqNjxevOXKPXVo/PyVaC+rYkzpHhnQ63kb3HBoAPt9ZAy7W1YR/bO+wPJdQAbM+3ZKYONb
-          k57qCfpG6UDx+9GydQDNG45H0uP5sYP5qu2OM+RMKPzjhzMxtQr9mHSkarWG/brVv7/6QYvA3SSr
-          LHLQ4i8e8aJI7cVuf7FgjxWZ/uM30dQksNU8OnVnyTT6PFpqtPFZmrSXPGZUtXXk8dgkJy5sAP85
-          hQEyy9GjVvJuDOZlCoRfrl/+1Z+ZHoG6XbqDKQmh2M+VFkvwhQaDGGWb5otzCiJUrVnx933Z+p4j
-          S5nFqqK6nYTNutTZDmz1mhw5HMSryzlv4OayQe5+e4r5QS4yWJL8NTXmhD3GfChA58ndMBr0J2DS
-          M7aUrf7S2FMvDf/OZQF8RflA7PhUbvFBAnhYni3VH2bpra8m4eDa3HOcP+2s5xXHvgD/8WuIERgj
-          m+3br4NYCBpqCRfV42mRB0Dbyws1IsXJ563+w1rtmmnFocuESgw7lA2mRq6iXDH2LA0VnmUFUqN6
-          hmCsk5QDW/2gZSVmbE0d5wIv2lDjy8aHF8/0KvhLxRO17BX1jJfVFDWS4eNqf3z2jKqqjhTj0U/c
-          5ySA9XmpVESwZRPd0m2wjmspwM+v25Hj49l5czQU9h//wOAn6PnS3u8usPjEIxd7vfczbpmL2LUT
-          8ZIph5yJurkC/2L79Kw6e2NKLGiDEICY+CczNpb9MWmhms8Fjb9zDNYU311lm5RH8DRZYLl+mlX5
-          WjqPpR33NdhO/glA0z9sqgkt8z/+ClmKOIKVlGcL2IMKNCCoCH5e1F507Z8ELHC6UYe7C/lyhHUJ
-          RSf6Ud8Y/F4Q8jiVN/5ATubZi3nFURPoBo8zXhrnEzN7/QbwJMGW3p6j08/d+1rAXnfxtDgob+bK
-          ogPc3wR/EhKwxhSdLgFkmeSRi3mf4mnXnTMl3WFKPDtPGjG9P1tUUO5JCl0SG3ZI+AiAdm8RswBT
-          vEqR9QXg62c0DlJkLOeGRejQf89E+z5PYPld7i3Y+UAnWtQbuXCW4goF9anEClIcMLyPtxYugldM
-          3LqdQg8FnAB6MQ4TCmTY062J9cfvqKX0NB+GRc7goowjvcGrYwxb/AE/ITz1BXI32LETBHQemwMh
-          0Tw0jB5yDux8WScX0vSA/RQ3gHyoqITwP+xxdskXSIxyf9ty3uYDlNIKvsvtlP5b3RsLz3U6jOMQ
-          0YNlSfE8NYHwh7d4tjMFDJ/lWANpOGjTco5AvtJGCP7pIx2HLhA+CRfAVW9icvzhBxvjyl2hf304
-          xMqkvp/vN82HtWDDSTHfnjE/CnmCLzQZ1CuFBbAoSi0Y2bFLsUzCZtwFzwuMMe2p5YRGLq4X5w3B
-          F2fUROIl/uOLYONj1O5ffLPC582Cm54kjpKRhsXQUuUllq4kLlspp00iBCgoEkIfwr0x2AktMzx5
-          3xcJyspmoveZMcT64Yjnj0QNdj5WqkKtXKHEJLq3YDlPwSlZbOKFTAHfr6eboEXujHcYNN6y540C
-          OffoSGx9G6T1S4cZjp3lYuF1E/Kl+p4wHHffnDpKRhvmXN0SHkU/p6W+4nxGdDt5cagcWl7dl7GG
-          s1LDl1JSvHteqj9+bKEtfqijLlqzQF8S4Glox7/1baa/3//pk0RZ0oZ/aWKi3E/NmVhbPKx33i9g
-          edzUbPVNAKt28As6rQUkjPomZ9xguJBQPdjw6gCYvz9b8NvhkBi7a2EMkKASPlAkUmfd600fMm2A
-          Hb6+yaZvmtGJWgspipL/4x+Tent9gV7EX2pybsVYV7YFuMv1lTgGGuI5XUNO4R/7L/GNYegZaKoA
-          1a6nTFx//DVznrU1QmVkTdW+mMHUbdd9h937grnM3bPldDoKUFVpg1ELTW8BY1cCf60NPOQv1My2
-          32ZAHY+3SdLCTzx6ob0CS7Mp1dO7Fa/mZEjQGqsX8RtX85htyhn6y58oHK5gPrThgKa5GuhV8X4N
-          azVNQTCKD396pV+KPEhRfbphchD3rrfe9rAG3/6JqbPh3cCf5gma73qZpPwYNrPO3d/K2Hzqqc+e
-          9X/xkguTEcOY1+LlhbkWZbzfknSYMoP+6Re3rzANOlcEc7zdssAM800KhpJ4vSnfAW7vh6nmmLE4
-          av1F+avnpXXK43/PnwnPJ92+n9f5rodhFFgz0dXDNqu9QQFYzqeEurlmx/O9POugOlh3jO56GLN9
-          OvvoD/8IalTACWvGwTN/6bEsh16+0Cc1oWZWLvUsp2dMX3apsuEPwdYpz/lN7wHTAm/qPPKKLb6y
-          myBn+y4W9IpjS52kgpS+zBMJvkMbj1s9hct0FYir/8JmbbKgQmGGPHry2zGnAk9KuSFxSbTEuXqc
-          7Q8ptI02Itp41gEnGXYHNz8N8zI0GlGEkoTOfovohn/N2M2Bhax0l1B1/g8AAAD//1ydSa+zurKG
-          5/tXHH3TaCuQBpszowuw6AyBECJdXUECJDQhNDZg6fz3I2dt3cEdB5SI2FXv+1SZAm04TCR1ACqv
-          HnLcg6muQf7MwQ4jQpyvnrpNQw6+evExVBlYL0Puw1o0Z3SBnhOuRmCtv3ovWYcwI03pe2KfTg90
-          U0Un5E5g1wKy27retr5e7GVxBQyLxzDi4wfaA9OPEtgthYZM7drYM+8eRmgmVkW+8YVA/hjBJ3ce
-          PcrTnq5fP7Li4w+Sm2xW57HmLKD2T4LXcjQzxj82EHeTjlP10Vb7osAR9NySR76n1JS8axN/eRTy
-          4qEIl0F5euJ2/5RwW90hnR/D+QylqlWQAmJj6IER16yFM8CLVnT2VD8ulniU/Bxp/nobOLYeQZTc
-          nkS78fOwyLtxhQEdj0hOhSactrydw/2lY0diPljt64VP4JcPaemnGX73by+3AEmBqVTUrbY+KFTH
-          JVHNyyoHZqJBXyt79AgSJVx2N8cBXF3LHn/JDupSbfQVwiF7Y8azBrJ7vGrxEqUVMvLkBPa4XHqR
-          xTdyTbJxoEXRRqLfYZ0YwQXT9fpZIeisRkfqaX8Of3maZbU6kul7CFe2X4Ci/ujexg2PNuMxPhCy
-          0CeekFwAzU7wDN6NPnm+H03VvPbLThzhaiNZvvVZn0ds1jCmCtGcRx8uqnGuAeOjRLW5YegsPxKg
-          sM1HJBvDma7509fFulY8pPqJaHcza/moX8YeH8v6lO2J9e5g8xosvCaPNqSjddCByyUhcjDn27vy
-          UViQ3l2TOKaf25Oojhp0ftYLcs/PdqDeQzJFN8xt5K6fW7UorzkRvXhIcLfhTJW++PdB2NmbnmjT
-          bc0WcYdnUGuZihDabtSpOh18qPqaT+Qqjys6qnoHjxFHyC1Ap2F9wiaH8AQ8ojzl67CSmb6AOECB
-          8R7DXgGSFFHDUYwk6Ovq/pt/jOgCsURWGdAmTDeC+tO8kZyyU+dkBi+4c4O3t8peSH+v//pnp208
-          e9ztcQ8FxcTIUQ77gd4UdYXf9Wl4aFFHJ3JTyPIzUbuZ0iV4trpgyFghhorbcL+ohxLeZkMip2V6
-          DmOi9RAoqq17JN3igY6XYmRjNo1fXjxr1tUT2H7wdne+Udf15nlffevxup5keJU/8a8ft5TPMqxf
-          PcS/qYc54f4Jl1j96QHTl8Rk+aizua0ivPbFESn6NVHXhasVKNWHrXcI3Zyu0Y2dWnX0wFvmrKzm
-          a2nOEFXtC/OH6FbRe6PUUDX9Hsm2vgXLGgzOMU/zAulDjIbuuVJfPB4lG2XxG9FvvoXvU/FC6uHY
-          DcsZlO2vP2Lxk86cBO6wecYuMXynH+ZVVASQFthC0kem1bLxn5H4OqY+3sxOOuwS78L8v9Lgg+FJ
-          FQ+m/g7CMBSJZnpmtvbaIQE3/8ET5ck7A1GWXQrWHSbIuJ5Nm493WgxlQam8nUViyn39JOPBmM59
-          W82mkdS/vIH582y9k7mD70+3IT/kROgwzOILMJ7rHYFXD/P5GKTwdUx8cg9jUBGdU1sQ5GzwHdhN
-          4eqIrgD182MkjkNUiuPrR4Czt65ICa7msPiz3sHTVdcw3UGiPh/T5w6/vCcfxCSjc/2yREEft0i5
-          slPznup24Mv/z7BwVEqoMorxrTQw03vh4iqlA3owNMQUnY39cfheANle54hCjWaYEn01f/mby5N9
-          NkHnwMFz9JFxzfQtVdighbGGLYlY/JhvMZQA85tEVl5eteR1NELV1/3v/eGv32V+3eu/60+6TT38
-          6i20GRuVmnuowSjc+zh7BRv6+7wq5UWJl/sK06v1CoPawGzQsgaWvjpz4odHGl59rw5nTkdnWFU5
-          8OgOIpVCfongRTR9Ygj2Z6C8fSvhY6N33jFAzUA39roRslsqYX6bXlUSeYUFgflGxI1ySf3qIcB4
-          sgf3e0Ndf4yNBBEkGlEN4fAbD6Aa3Fx0miim68FZd9C8Y46gAxnAtGoTBGVRQWSdUsneMZ4Ak5XK
-          eJN0CaW6LB0EYsYBUlj8+K0vZYdsIW46mxXewvYOmq0socLJNXX3jDtFYPrdo1Pd0XV0uVH82NaN
-          KFxzDLsccblIgpwjkrSd6PLl6Z27S4mcZkVFmn67O7L9ipnfowur78Hrx7CRUlQqXfD+qMP7+HY9
-          Tt3q9n6Y+RKweIlpXh9C+rSPG/j2yztCYlXS5cp9FCgllcnqbY26olKb2RRyifz0QQrWtLAECB+Y
-          oi+fWfoq5YTP4Z6Q9L6R6V4ZXy386rl9/KPS9XRuNVj8nDKklGOXrZFcRt//C2mfYlvRqAt6cb+r
-          OxRbqVHt5NNyhnl6L7zDuSrDsev9VEw088T49pPOr+slF7/8rMiiB2C/V4L8e/HQqWUlbEl8xTDo
-          hi1RiXIPcTZKFmR8wBNh2dAxftaa+DjFgOjO9WkvI2e20Grklrhr1IeUt4NSTDtuIq7htIAeWnyA
-          S5J3iOU3dS80dgu5j3lGoXJvMsLqhbCz3jpRigNV128+ZOuHaEZKVQwUa/3laTZvv9TFII0DKxTk
-          rD60obRpRgkgkBgk67ftQBOyduJmJ07IWMzzMOtqp8EvD1SdrgRE41sB+psmQhIn1zZ1Z1hCdj/y
-          T5GoLjQv76J43iVEkuyEfuOBuDkcNCSHnDGwepgPMyE28ZHFMyoB24JHuclQ8nrl1XD1x+jLN3CT
-          0zybLyiURFTVL+Jz7K0krB4tsvxCHM/dh6t4PrwEuOkASfnBoPv0meoQv9Mf9AOvH5XCDmpw2Pdv
-          YmHhmOG8rQ7i61kPKA1BPyylX8aw6MaW+QWozos9vOBWmCNSTCTLpkvwckTpEnLeLITOQLIT9OHx
-          qNjIvb3rDBuBNX95Dg53Zqou5B76MOjrmJiPQ2njSO5iUPwYmdd0vTXwdiDNX72C7O16o2u2uTlQ
-          LBqVuNa5ZvXSMYZbnD6IjiqbLqnYRPAcxhpi/L/65V3A8M/EFAeVUj8EKazSOMFzSyzAB9aPBz7O
-          I0bKDdwAvfdxCq3Pe+91LL/SIk1L+NbMhsRjwdmjKwUS/Kel4K9//et/WIPAn7Z75A1rDJjyZfr7
-          /1oF/t7/PbZp03wbC/7gMS3zP//+pwXhz2fo2s/0v1NX5++R9RqIwoH/bTf4M3VT2vy/j/5iX/if
-          v/4LAAD//wMA0gD2croFAgA=
+          H4sIAAAAAAAAA5x6S++CTtPl/v0U/zxb30TuXf3sEJA7NAIKzgoQFVBRLs3lzXz3if7mkklmNRsT
+          oY1NV9Wpc07xX//xzz//avO6LIZ//fuffz2qfvjXf36vXbIh+9e///lv//HPP//881+/z/9rZfnM
+          y8ulet1+y383q9elnP/173+Y/33l/yz69z//Ytpa9tdYf6DVqw86Gk/cnSgby4w49uRV0G77A832
+          ey9axa2TQeIPIvFsM1BnAVsVxlFc0uQQL/kS5VUIwOMd1Y77o7roi9RAy80b4iUvR2UidU6hfO9y
+          omnMU/14vduD0eTJuLlGu45fr6CBtTGO4yqc6pUG5i5Ej0PJEDXUYme5TxsJOrm3f/vJZ450EyDu
+          LfjT0347vcG6JWjvT0GIu0m6KXBGH7K7jKheP0R1RjPRoXoDIUenXbs54icJ92g5jo2MOXUaXouE
+          nYFoRG8ulbMoGFwUDUVI03fZOcPzGlUg+5uHjwRm7Jbdbqdj48wb41NXlJxt9sGCXGOOSXI0ts7U
+          TMyCa/IeqR4ctuvE+h8OZFWxqAzIVZn90EkQiKtGDJ3ZRDPL3jb4uN3ZNKScUQ/l2ZSxmFUTla0k
+          7qiwKAqwWL8S584J0dywSoV3Gy2m5zjxnJGxXFu6pY5NzKDSO85veQk7fc5TXWNO+XQOpQp9vxO9
+          atWaJxKRAZTdkXhaNK/ztL5d8LaKRdwkLR2mfaw2PtRTTI7ylqBpJww3SGUmII5PQ2ftzrcGpKhf
+          xy6qk3rt2e0Jzmm7p9fklec8m+UBNHs5p2nq1jmPxM8TpyxQ4j66DVoVRZHxdW5Leiirp9MdmKcL
+          61XE4+KcuWi68NUHl0wqkFzibg57u74WGEczI9rGEuq5iUcTXZ/RjmodeXSMuQ9HzDruiWYc5qJB
+          0ioThMvBptnxBPlS9rWJz52UU5fztHzeHrcB/M77ikXHGWBrBWCjGBEFRQdnEfNQw7zX5DTpDqQb
+          d8GplG7x/kUtmfhqL92EBTN9U5PrbrxH/OPjyLBJ3hPVVLioc3fYa7jn+4VqzXnslvRSKfhtyA4J
+          rve9w3dWmGFnuubUeF5shxHVu4+xqzQj7MMAzTM+LuJMzJDKKrWiodmnE9JfsKGXoTG7v3x9Xyqe
+          qukgqPNZl3Q0XQKLZuqlUpfWW0K8WQqPpmSUUZ+Ynom+5010mzPQZAiHAIdKapCDqLPr5BpZCdVj
+          f6WqMmU1r4Spi9YrwsQ4O0q+FOEtxcv14xN32xbqmh7uT3xIM57qixjWnNTdAsx99gI1HT5AjJ5b
+          Iz5SLvEntlQ6BreHGC/scSR2odYdV4p9BarfBFR+y3LEvKMUYP98s+RaxUo3TTQvQCuDLTl9Qjnn
+          j6K0YFF1+ZFfYa/yQ32xAZNQIr5y85wl6qYCf+t/5JqL4jBMwHHY9qoj2YF1yzkjsH00D9OeOB7q
+          66VjuAymvgAie75WM1cl0PGTHM4kvj3v65oJjx5cImXjfP5ccpZUJxPR1+lCNBmfVGZI+hGc/swT
+          ORZ2+cp6Bwkul5TQaL8+nMlUPAnMqrSpa5qtOrmnF4fziX6o56ECNXeTZ6AW2834MlhendP3p0Aw
+          pDOxy92Mll4RPjh29T0JI3V15sLZc/CwbWkUhMec8y9myjAco87nrosbMWW8hDgZkp6Y2SZR+1ju
+          nyghtU4CUQNnyshjhEyIGVJMqe5wWFZCjAI4U5vfEIfrmE2KOrd90gTnt2iIDP8EgbDl/Q1xqm65
+          6gPA9iIXJCJ54KxbX3vC1DQqySZHqeeGtW/oi5/UlUsn75dzX0i38c1RPYptZ7FfE2DX2X2o0R1I
+          PQ/5Uv76B4l/9aEJUYbh1pnUeJXKypKlP6GMORyotg4vtChl8MFf/KW7TjajadreP1gDYaEnSL11
+          Yc9hjOLIWYj67R/9eExHMPhu8J9Hq61XrCYuVPq193ER3tC0coyCFakaqP3oxno1S60CPej2ZKcO
+          V8Q6Q7LAZz7YNGbuXj1Jk24C8q7sN3/VbjlLYgNLye7IUTjFKqfa7AeBv2zG6X6qu3kwryUEsqwQ
+          W70oKl881QoHhoSozk1KR9daFqRDvcRU3TiV2j+GF0DgQEmSBbn5bKSYQSzWrmOjwsWZasZqoPVk
+          RJSnLeaUfOaP5EXtQMNA651p3CsVbl6XZpSmb7zeKJDxpRkt6t9NWeXqnAjQPfO3D1cyOqtHNjr0
+          aDpSI8ecurwCP0D2NiupxW5ezjqoYgMzsUPqgVA5fMGdR6ifWUsMnMs5PwYSgCkuqo9N+75OnHOS
+          cTzPD+K5GyNnP/FBwpmfNP7MXWZEv/0GKWnbkPP7pqFJYJ8A6q3MKCnCWV0tFGuYUeSYJk0u1PMc
+          TBxmzrHnM+TyRgt9wwT6leVJAtmrZu5nXca/fqho9TVfgng3wmPSXvQgxfe1nyFmcFHsdXLpqhNi
+          mObMwb1DBU2pzjnT4dym2KwKm5Ta0el4JQx81NNHR+NDaEasFlf6X72ar7fb8UmEG+QFWUJ/9c5+
+          6lf4h++XLx9a1nK18S0IDvSM1kc0fW7hCdv7nBC3Hd9Ra4aFBs87TOQKwpKvz6TlfnyJFPtwQpPX
+          VDLemyEhZnT36vl+ySfJG2fwPxf5rK5xUftYHpqUOBfydNa9flkQJ8V76qNd5iz7h+jCwGxSGm/t
+          OZo6/JbwtGzf1LooVb2+r7sUC7bMULdQ13UahJ2GgyhMqctvrtEcJUUFibJexlkdrut0XSwGvng0
+          Tmkprd3haH8gRAn4zFHM6s/jwYa4vAgOySfr7EyPsV0gEasTMYWlqbv31UpFlZ8i6rrg1ezYIgYc
+          vU+JiRohn7ZhFeMvflN7cqp6veIlwGe09/07WHI0HKudAKJo1tQWn1m+7jlyg5EtYqI4Zy5f+Ei8
+          wWl/54kyMJza1+pVRpPHW+Njv7LRbBTKB24Sdxunl7Nb2Vn+tCCrskXMOzPmP36IHsdbQhOcYGdl
+          JjsA1X8GxHeMbdRrt+qGkTcq1FtS32HS/Wxi4o8CVd7y1E2ienfB2C4rsXJPcwbr9GaAMJ8HMZvP
+          vpvD7B6jHY4Mampak6+f3p6gkUEgJebnaK4yfwPjR+NH8LCe8zdjlEC730/Ef8C+4/IDfaJNZ0cj
+          s9uMaJ3y9gSvpeyp/5k6dWXktEJffB63l8haadYmN5G9U5XqUk3zgQunBriPIVDfNNN8vgdZBnLo
+          AnUkTlb5XSNlcOCVz8iE0d2ZHr6WwWUfWb54Y2k+BGTdiG8wP/SwNV/RHNq5DfOsFET7xn+6lk9O
+          uhX+QL0yIOq6XhkNUgdfaTmlp5wJ2a0t/PgcjLVRr8Y09jA1R45++W7+4aP5hgfBHUfmUvM5HeVm
+          gvDcy/RXDx+pzQtIyF0fN0a81OvhEXBgipNKLbnY5WvxdCpoJDB81vrs85mfTBt+9aZRjkf9uZZH
+          4K73lpyi+OPM/ebD/PQNNbd5Gy0lzCesKWJNUg7rav/lj+iCXJYeNih2VqvbCyDY3G5kMxiczzk7
+          upCf4EVsfxCieWxXBsqHv6euHr27DkePDOLEO1KtcfNoJiHZoN4SXHJ8FDuHPS6pBri8Dj4vPSx1
+          eFS3GPhNl9D9dQq7+dPcdGAZfKa2djBW3sq2PRxoP/nMOrzWReraEFH50RAnrDbrqEg0Rfo+aqnT
+          XiH/8W/0i/eu3ZzQtGuTGN7H80zsyVG6Od+dA9DXOCdJMSP05QM9/PgDc9yz6tyPcQuDqhF6jdlX
+          vTqiqcO0kwaqDYJVs2rNl9B6ChpZ7s52s/WcFFwO+sZf9EVzuF+8vv2ZkJh9dZNuPHtU5AfRx4sj
+          oVbYs67Y3zLxez4on0Ldy+CBPM4Xi90btcyc+4BZ/0XjZy47XDi6BToJnjcKUSxFVFSCEDOKEvtS
+          kg5oIv4qS4PvleQqaC5iv3oUyuO8H9+jwTttpdcM/mDvSd2XVHczfxUBqjm0iOsYVJ1v6lmBH38o
+          tWNXzyLfjnjLNz3VX8VbXdPZsyHOQ4Eq/K2JFoeRTqCu4+QzRbXp6Ld/Sjd6UYmFnuGvP/ZgqkxC
+          CcFlN3XeZEuByqfU9HqSr/qp3iB2Qhzdj+oarQV36IHhQSbR/C7qYXuaNFg2wY5k/iDk0/2sK1Dn
+          XEr1NvDXZVL5J0hUIzRBz/s6qMatB1QILDHdjvvx5RLx7uoTuTY23WhrxgRj1Dj/E1/zdxpDkhkn
+          qvjFbf3mxwmqA2/9zjef2Ue2wFcvEHc8MVFD/FVBfGt3PtyFUF2+/BDOu/2WKD1/6+b4KFdojbb3
+          kZr2ovYZGUaUVP6Fmosld+svXmdKZ+oacZUPh0GtfnyV6vfXB63G9OxB3XeYqtbzgPq7WKVwQT5L
+          XMu6ofUSFBV+DsilO+K7NSvTWYD5VLpkp0yffET7Tpe++EmS4Oaus2dGGf7Gw5+0mjqrqKQByO8b
+          S0pluEXzaXsuQXxIHjU0xe4ma5kKSdocFKId96yzYi1Q8C9+JhlvaKmd5QZqE01/+283MRfDV58R
+          S/WP6sxW3ga62on8lQr3fG7FMkB7ze1ppmlMTS1UaMgapYoo4lFBKzrrvhRN/ZsEsXDPFyndBzgx
+          Wm/kPwG3rhHdZziQFYW6Hn5G67DWJ1iyuaS+bb/RJJycFL56lBR3IXTmUVQUuCt2Rgzp8XaWzXCT
+          8XT2n8QYjec63z3vCXWuOtRAzzuaKYtdEHZopHqOIJrfRlDhu61GxJxQ3PFbvDmh9WY6VNZqos6W
+          6nPQRZsbcdJcrvlvv4E5zN5EG9mtOi2n6weubL8j8sA81qUU+xt00tWl1vMiRdPU+x+0SmNGNO6a
+          r1+9PEFVNgvdYQlFC3nWAl6qfUrMzJ7rRfSDEfwX3lI1EZdobVxUoJ/fEP/wqYmfJiJD3JGICvdo
+          ObIPBVh1H9DdFlzE5NgpoLckl/70Kv98cj6qwvZEDi+M1z99KCAJU6vYvdfJnPkWnkZ6I2613+fr
+          T7//+OTuhTGa9+WphdlQXaIuH/bnn9xQbTbK2Ft7XE/g7CbASbYbu1qw0NdPUUAcZp2q6ubZ9eUd
+          tZK67Vx/4xjb/NPttc+fH7anWq+u+7OwoH3FXP10Nl/r+Jo+Lfz8EX8+U2cmitSg++qdiMHTwzpd
+          y5GDbnOjRP/qnTmY0hh91KL5+jeHdaKhKmAcncrxtaA+mtoHMsFHRUVLx7hGq3o9NDjzjw2x4NWp
+          9APBBiudMRAf84foyzca9M0PXxpPmrM4u3UBSS4uxPmMd+er7yp0TTSf7IyGolFK9yE8ei4gYc/L
+          3Qwj0oBLWjrSyqDRIlxGDeqrPhLCHGSn54V1ArBpTvaK3HYz754Aj+WQjbzYBNEy3GQXhC3cSF5r
+          Wj7d37ELeRVrdB+FQbQc5kOAvuftS/tQrjkxD3UJ8dc9tWZ71zG5d+jRp2TksSco7hb2favANPSB
+          qCNv1lwyuT5IyXLyH4e4UpcmXaUfPg0bzmPqeau/BXxJ6ytR5k536CuOFRiN/E5ciTl1SyJnJXzK
+          1zryr7JC3LO+h+BtZYsc6Evq1vJWjdDfUtFH7O6+Tto09RiYsfvir1IveQwp6nL27E+e39TLg1OL
+          n96jWgp7dQnqREfVQ+qJGcVZ3h43wQ1/9Rhxvvk1JWl+gqIwdLp/K1m9CAMTYHPip7/7TGVNH6y9
+          24IqR/GTD4W0bSDGG5/oh1Co51G0FcAzHEjx5WPsp2k1OMYlGpvL9KlnMXHjXz2M/JcfUCIZMkwL
+          ftMYu6ozffcD6Sy61L4qn3r6+YPdVQCSGGyiDp7gmXDeO9U43a0g4szzZoGkBJ1e66//MRdiiMJt
+          fvUlW+q6x6O6nbDSyDuSSwusVFFsBfbJVaG6f2u6+X6JJvxdTxV90dT1lCgpfrnVdkS18F4przUh
+          joRX5Uuvd4H6W5FxSGheGvUt6+DMnxUXUNdmSNT3dovmxA2Fn/7708/LlUtGwG/eptagl/lMqtJE
+          Zkjwz79zFnQUA/j6Xz68pLqe1Xpbgoh9nZp9XOXLIQDhV99fPwfWkQ/6Hr79kmidEv75l3BMq9Lf
+          LNlr/eWXGOn07DP6Ekc9SKhCT+FaELVoWmd56IcebgvNqPLZWRF/3vgVPG7v159/2ydpFIM6+Crx
+          lHLN51ByBcCDNND91PNRrxvjCBrRXborK91hd/2goZfW6sQxwrbmr75cwd2L3l8929fjg3MKycrz
+          7Q/fO145JBn8/KBzcmLUXrt9KkhGUyTh+mai8cpde8CnJ0P1Q5h2bHePQtS4YkzNTxh0i9buXciN
+          +4sqgdar0zi/SrQ3A0KvW/4cjbtJ4tBkrU/iX6N799W/DGwPmkwOee+jBZQyEPne9/3Pa5BVptmp
+          MsiMEFJt6am6nNaDDB/2NNCddfbzlWdOMpy39jyiYK7X7quPpF++W8Z+r7Iv6HT0rWef8QlG808P
+          2du0pF99lw/dPQ/gq1/GL7/Lp7dqL0ihcBl/fOen92F0gwOx6FF1GC9p2z8/u3zAo/72vxZ46yZQ
+          k8O6s36cZwqFiHrqL5mxTg/M2b/fU6PJ03rCBgB8+bs/B8oSTV6cc/AuY42cj+ljXequ1KXhdHqM
+          o3HfI17q7z5884PKgvHsBtw9Q0hlLvDBw898TSL8BCu660RfJefrb3cm/PpBadvvddkAzoSva0+d
+          YK4RrcI4BJMRI5J99cUfPkYn++wzytbolq8/CtiVG6K+vG4dN576wZwcc/SqdBzq2oPhwj6vHmSv
+          M6PTG62jSfcXK377w07t8x4H8FDCHfFFtUOrnXYjLEZtfp9f6N5TnSu/+FPD+hy76uvXYEs7RcT4
+          4k2vcIcbJJlQU/kzo6jftdcT8K3ZEa8Ib+v0SsUFlXCYyY+fDarR9vCeXJ+evvUxh5ImwLFodoQU
+          1yaaxFpboJE3AvWGxvzzk/7mQXL+YdXh5++dLh9MfvFfAnW94X4361Sp00c0WvY7Rd/zGMf+8e7m
+          g0tt9ONnekdY9JsnwJfP+6Nc5t1iJ9WEWo8xScLOar3I6lHH33kV2alz7fToUm/g54eoja8h9nhY
+          eyz0F0wUyr26SZuEESJ9OP/xuYndvQsUhYJKdIeXa9ZIMQccF1TkFBxr52++9bjkwXcektYruUwh
+          OqefPfUP8s5hwLEWJBpBRveJJefr20cu6lJZJWWtj9GPv0gJW67jRlAidSaRPOEv3/jjv7MsyyZ8
+          n4deoPPXCRxrgqt4T6gcqRWa3Ccno7ZNU3/+6WWvPmuSyi8RsW3Jqccvn0NgD/koff2bj7t5CmBc
+          9pfx/fX/2CwYP+gCskL38pai0Rd9DfW7VfclNRjX+bQ9FOhSvJRvfuTrLMVTDL/zcgVFWWfnYMlo
+          PjwKot4SZu1PeWui37xkn+yxM3HImOCnT7bXg+DM6bsqsHi6eiP6+v9rVPI+yMLQkhM+P+vJPVEO
+          +vpNiL6IS0c3h0MGX/1KrF2go16N3jbc+/E9MhtL6GZFHEuQOrmn+qAx0dTbdQy/fuA8hjbqfvPM
+          8sy1PrSjlS8/fcxxYfVdH0f91z+Bw2JJ1H1TxWEevptCW3gjUZlrhCj3eRfw7UfE5XfWyi/7+QTl
+          cd37T9S0Kg2VtoSv3+sL4Y1Z52WUM7zb6DG1RiNx2s8tO0GpVJRYP3ypxkrAy3NjEuMqyzXz1eP4
+          JTf1iHZBoi7nT51iH5UV0a0Wre0eLxpc2XFHfnxibfqrgtZNPfm5qY/RYOl5C5M1P0kE1lSviUpT
+          OK76SM2qjdbFuAUplq4d+caXoPbrn0GIjuB/vnp0+OY/5OGeJ4T12XqVikiW0i69EhlDg77+iwTW
+          9VpRzVgHlRvP4xM214Ubt/v7qM757MhAhNGktlJY0drrXoO+81ByEKO3Mwy5VKBHuwzUDG8xmq3T
+          ncHGAXySnJ0qosa6VrC1Jp0YR2Orvvez+AFlF3dU8cnQDbXRTFiA0Saq06r5ZB5bAa6Z2JD9ZdGi
+          WSe7EGAVS6pZbVf3ctfY8L/2ryE26TYF8rfVnih1us+51y0uoeXWjX8N1+Pa5lgt8Fefk0iKF2e5
+          H6ITRtvGodfGujszicwJDpav/c3bZjHRYtjdxc5H0gJo9l6bDXz9KqJzU9WNu52lS2l0OBM3fto5
+          r8zrhPWbtqfm+8Cvw6HAJ3RJ71dqrXDMl4zvJ/jX762A//6f/x9vFLD/7zcKvITZkZ0XqCt3LPoQ
+          5ImWxJwbTeVuR/pBo0cuVA4D6FbZqBjgMVHG/tBX62y55xEXeRrTs+DLOeOvtyfus2NMjf7o1Yzo
+          LgIQJprHR6buVGa8iRs0iCePGhyvOUt5aQA7u6SndgcJ4kg0BuD4huYfze26DhV+FVCccECOtjR2
+          a3aYQ3ESvJ6aZ7l3Vrb8BOB6Aku1Fs3OXGtIQq++BKq++iwf4Nm0EESFSGPX89HkmrcMn3f8ixDR
+          VpzZ35kS4L4/0GLrdCptWykDR7lZJL2/XmjgqqBE72ixx9kLVDSCmS9g3cI39YxFR1OsbUNIBeFN
+          L2WuOWwe4Bbx9b4lkSCFDi+zXAWMrY/ULkJV5ZQLl4H1gpgWOAgipt+fJjiJkUGUbtyq62lOJdBv
+          7kAznr7RzFoJg48CN1DLTbpuyKq3BGFrmiSxxDKaXurDxudZodRKN698YHNXh1KU79RI5mPOZtVd
+          wpcdjemu5Vl1rRnVBUu3fCI7nhoxrT3fgG+CvQ/UH1HnSpOGEwtNRJurHs03oTrhZY5lcjhkvjon
+          dcjgSi+vPiuhoVvDZcfARwwOvmCpdH2MpK7wVCgG9ZD1zLk+u7sYvCGlcTM/EXPrbB0HJ3rzW35d
+          80kilwKcq2D727Hcqex9tAuQwhsi/r1I69XkeB+PeOeTc3Bn0bzsBBtYH1yfCT63aG0T3YR3r5dU
+          G9Kq44v85mMmKSe6o+eum9WmAXjOekATAd75HCRLANEk8FSmB12lnxvWQFY0m2b5zqiXiUt6kJPl
+          RnS1eueLnM4Bvt8Wi9rEP3ZDpRxi0e66I9Ub/tVN/sgt+OMlEzGma7KyeJhl3JXwotf6KqzUw2GL
+          37Pp/tXPWF76DYzKwSE5051z1uzk7He+tBiNC+JSo6+gLEWdBvevo5Acnk9Y9vmVmA5s6hVvDi1s
+          lLSl+dUanLGLCg50xVVovjnrKvewSw4gnHUqX0izdqdr88GnTtzT4IUe9dwLkgZwrUqqm9Yjf+vk
+          4eJ9q7+IAeF7XUEzFojbNKLZmw9XduUDF/xMOhOzOPXdNOdXF4QpyomVboyIt4rlhDnvcaO7VCcq
+          s1ZdA9d6eVDdy24d10omh7PXcaTak3MQcz8fGfCVZk+KXkkReyKujOIaE+JAvFP5/F4G6JNLhBwu
+          ySefamaWIb01R1I8VitnD1nB4U3DJOTkZXLNPYj5Ab9mE2Io56jj1IeyAA3pg8h7MB32Q1UbF6+H
+          OM7k2CImPnU38MojoX5RNvVU3nMbSB6vRM7UsV5SqVLwopgC8S/hGE07e3X/7isZPjv82e4F6Tg7
+          ii/UVVWzZ0nzsc+2l3HzuuNu7RQ/g6ceuzTjWgHNr6X8SHnJ21R+cLd6YTtfgtHacGMWg1Iv03kK
+          YMjm0UOnPq/no8F/RLUbeqqn+q5bn17i40SXb+SS7+t6itjMhw2d3sT5xp/b+ecMLszcExPf5YgJ
+          7FDBqH5sxg0JXDSHBZ9Bpag5CY6iFnHXyNOh7+2MeOrxmK/2vLPxKzxW/kOUQ8Qp1c3FNLHf1Do9
+          pXXIH24JV+ttUzIngdqf/NGFGJfyKDCS1g2er55Q4xGByKN9VNvgbfYwV3iiWn3rouchkz5wD8ln
+          lDgkOyzT+jdIffNBTOyL6nTO+BF/hyn+29tenTXcaiFuokKlh93lsPJF3rrwhM6httrI3TJc1AVv
+          2XD+7mdS1/oYVzgSPYeW3CdQB9bftVDQqCE6w6hoZnshBiajiS+exEpd0ngEaPtwGNelq7rlYp1s
+          ONarTK1bGToLC8hGharJJJ0ukcOLn3WB+22yaGyqWs19AikAZ7YuxLQ3dr1eSmXCAns0SficHJXZ
+          HgUBQnrYjowzXZz50dIRiAcz2ZeeE81LmTyxyJoivZw3Ycft8ogB2rsn6qpJ5oxVI6d43yU+cddO
+          zhnsWgCBaBFqWWyMxoqIjcQ0rkWMp/RCi/BaU/zxjhM9ONs2X+f+8kFml9vEwm6cz2GxTeHdayU9
+          H5fZGXZ9FWBVMXVyoe7HWczzDOjj3zwS4MjolunoF6ia3Zrub1+GuxN94a/fGBnuOn7ujx+8ZuhJ
+          Pa5nVmqit4K8tvwQ/QN9PWWrEONeOG6Jya9RzoM5PzGEq+4/7rWDlud0kOAepQYpto7jsCCkArxE
+          /zhKLENUprXFG3bLuqTuEDHqHCW7E+zRlif6m0XdEme7HnPCsSGnYLqrTLHfAtzyYEcuvV3ms5xG
+          JeKEpKGeq+fdar5CF2dRUZL8vdNyxmDtEsr8c6XXdE3VPq80DXf5KJMkejkOV87rgtoX/6bGeldr
+          Xl06BZxTfqb+cnugOWIFGRZfCGiZ8Ga3sl0zYlW47GmxdnLU8/Hk42NJQ+rr74OzuFIWQnjC3Tj7
+          Pr8+3hnEsM5TRq7VvUFzlFgnWMNHSCK19PK//kvs0aY/PJwS5h3AtRx5Yh3a5zrxqrpBI+3ckb+/
+          ++6HB+jCrP24HZ0zWoIX1iBmlyO95o4WsZVyPmH9JaT0SJKbOp38p4tv7VeLQJhE0y+/vnhO98aY
+          5suqTRW+7rZXv/34db147VPGJ2r1RBYUN5rpej0huTyMxNm97upgT/u/9cT11dxZGmleUHPqP9Qu
+          k9aZZ6kewWLjE913S64yhzMvw6EWIyJrnYum53UOcNFOJ3qg566eJCWS8a6eOWJyleAsWXmUYBOp
+          hy9fbJz1fjkzgLK37NPv+a0sTgEaPz3+7WeN0TSBxirEP5rOR11ZOyphuPoLUTnWXSetFQC++Tjy
+          Fxc5i40fCw7Ks0PzewrdNPHdCSn5ISS73jt2/dkJeyz3nEcVmuiIHVDRwtTEBbXeew8t4dzISMsk
+          xucOjY0mayOmQPLT6icQWutEWXmCX37oVOfWoSpKFyrftgn5xmuWyBri62y+qNzuA4ePj3iEthQ6
+          ooVF6YxKsAJ0b02hdgxVN/PJ9MQGzq7+9tlRlR54XoJ9pEj+Ulo0H/cfQ4C62zzHZTrm6iJtmgk2
+          /U2h16Xt8uExHCo4WZeTf58+hsrkx90HWj9U/Mo9pPnEOFYIV/HrMDOC9+VLUiCV9E7pyfc/0eTL
+          tf6r/5FbiiSaTresQBfdTv0hWyCawWA3kmFZHxqdZbGbCzEPYHt9ZqPw6nvU6cGNga3in8dV+aj1
+          ++6KLigTfhNyGhf0fR4Bkm7kRnTvVLQoW9KiLz8fmbAo1Sk5Sk/UW2P3zSekduh1C3EtnlriLMUb
+          DaPy1kFq4h05Jl3c9ZdSeML22mTk1Humw4mXvIKilzH9/t865yryYbUDTKLnqEfr/lKHEFnnjCZb
+          J+2WcDMWsPhSQOWeSHk33jQTtDbgiE1NX+XbdknxTpJlmmS4q+dk2TJiO8c5Me5vt+7H2wyATFYi
+          RNx0aDY7MwV7RDAKZ4ZdG60hJtLrS0xVnNTdLx/gncgz3X9so+bnQmiBg6klfm+K3fj6HwAAAP//
+          pJ3Jtqo8u4UvyIZUkqRJJbUJgiL2AJFKVEAC5OrPYO2v+fdOc401FAhvMeeTwpesgEx9OsTyzz37
+          nUfHhH9+ystQDOjnht3DL3ymlJyiUzZz5jOHVDMhIc8rHtb+AUfYPrUdMVzbGTjpls7yX/84hYXF
+          lllSbXRS3yLFdfNt1v3lF8JQWheqfnemzq9Y7kBWfs7Uuf08wPviWQZvdTpQ/3leh5kzbzm0ouhI
+          PKBvKf+qQ3iGpzc1jkPijTHuMLzB25P+q+dW+OvgC/Ip3j+v08CgeAxh+X50NBCWGTDw/gTwi6ec
+          GjiMAKv6PITCIsRUTdUvYGh3/8AUJzNJ81D3ONjPPULPmGAheNwZXZW2RMLNCOnjKI76qlysHByb
+          oSE2J7fNv+s7nHeg298DI5bT//XTia1JFi0f81fD50cIKYmSNfrlb6rB9n2NyHG6LGAZUz0Fo5kb
+          JBR32iAc01qCl0w5U0WUKjAn9mzDkEb76fcFE/tEFxvCmR5uWFKbl74CzpCgGVNATs098RYxMldZ
+          dfQVLzfYNsMzqmVYpdVu4s8daVZ2edQwc9iDeko563/+AkrOLiP6d3kOy+YnYP2mC+Yfd3XgmRGO
+          cK/5dyxu+ngspmsL7mVi0ufB1XTemncFlMbUJa72IMPyFu0CRhkuiWLOZ8biXRRDfBRGot33L318
+          DExBcmEJeCcrwcAitZNhktVv6hmoGubzMISAfAyfWralDMw7Jhq8nvcyPUaVG6377lyjf365LJ1s
+          aa0ih5t+pNpDK7L1Sz87cFJfItHP5QTW43SYkPls9WkfPPWG++0/PdRHySHPbD95y4rXFvzp/+j2
+          isC0v7xCdDWTHRacRQJL4s0l6lKSE1X52h4lrzlGW/5QFZ1rsORvqoD7Z0rxGn7bYdkbfQ7u8N1T
+          +wvybMleRiErRTTh9kVT8KfH5WKJBYK1Z8cW4yiEUOf9YtoBSx3Y1XtwgBvLB03BgfPYWdzLoKNg
+          JSc1TZulFZcUNmYY/NNT63w1c2S4oUf+/AEHO1NAIVZ+WHLdki1ZwPd/+pAY58L1Ft8rZ4jj55Wo
+          ZVM2k7a3evg53mVin8oq23hEALnHfodl+VwODO8R/tMf1Ceur/8U765A8enf6GWeGv1f/RNKPSDK
+          lm+bX/mATe8SYoxL8z2NsgZ9NHZYBoeLzux36oNgFgDRsqCM2MVmIVIy3SJGcLvri3RwBECXwKBH
+          9jrp4p5aOzg4kky8sl3ZKFfOCP3m5VMLmVEzf5/KBXXpKcdjFYts/qZFDbr3q6D2k2QR252EEZqo
+          m4geL73O7qhP4L1MTYITlekrYz6GH9pc/q1gFvKPwYH0U3r0dL7FYD4MVQfsj9uQ53stdTbbXw6E
+          ZfHA8HlemzlwU0XOQmvFIo4JW/76+y+DHnnAVAX8AYwJJIvqEmV9dwOzTAVC7pbPxNXPh2iZu4sE
+          V/4UUD+z7EzIjLsGt/wgtho8wHfTX3A6Nus/v7GgL8FguSUcKRKVeePLqwsE29anx1N+jBZdj0aY
+          FAyT0zIpYMl0hv94G9HorQPzjvIr2vwrnhP4aEY5f8rglJkf4illoK8vN+b+9CPR16FulvHItXDT
+          9/RC7i1bF9QaKHOWB24S7diwV/6RodM0AXGnixGJJ/fqghl/jtMSRMj7vQ13d3gt4ZP88Y51fDc9
+          fGFs4vwXXbyZCXIC+LSySeQT1qysekLoRjd1Eq/jHK3TkrYAjIWJD0pKsvlGxQJseobchx0clrD6
+          +og41Y+q+/gNhv1VN+CJxzohU1Hp41QeIGhi+UCd4+A2024+JAC+BUbcp7ew39qoKWTvfj9do8rN
+          xC9SMQRjbpKjObcZa+3QhM+sCP7pVz5Vgx4GqI0w2/TJcO5RAXlHflItx1wznSDHIcuNLeKfRw3w
+          +SELYZN6JZaFTwKWYflNcDy9wbR0KwH8xgfRVb2diRWWsrf2+yOE5giP//hVG0eJAg9QaPH3oLne
+          bDvnCyxhFVHrInXN/CjmFm16lqisuQJ2hGcIdM01iW7QD5s3foqkW/zAP4Bmb4a91EMIpZKmS9+y
+          XwUMCaH04JHCM+th6R+FDJexqKb1+BKjSXwFHCJlWlEN+5HOXmh1oe6cETmttduIy06VoF3mEnHv
+          stD8In5WYEy9EX89+QDoxmPgHFYxfUiL7NE3SwPoF1WB1+gxRTOd4x0SKa+Q0OphtHqG7MPYecYT
+          //4EjOPHsEZbv6Uk20/673TgcjAdq3USdzfMVvNxKiC7afE/njRdUmeCu6et0Gss9mDtit8HgiH6
+          4irlo2Fd8pBDM+6PlCQs8VjwSWWoFdWBHMP0kNHITEZ5D5Y9XhPp4s2ylimgGWBH/OC1gJk/XHbw
+          81VM8kwl5I21CmpoNs/LpkeygZ1yJ4Hy3HT/+aFH5iZybsZvogYR0tfBTFOIeK2j6vENvfUWGin4
+          0+8nmKqM8/RdC+RnGBFXpF82Kze+h/liDBu/KJu5o0UI6RIaE+aFLqNX7yrAv/z68/Os+NEZdIsR
+          0HjNxWgeLp8CWg2Ytnr6876/KcTQ02oHL398U2yB8m+8Yq6CzSgoVQq5OZum0eK8QVB/igFrbP9X
+          D3grPrSAtuVACbZu+txPvxhs/JY8vtbRE/fssMoan/xouL1fKvg7DcI3x8hljYXmd4uTBBY4eND8
+          7/s3vgvT923C158eDyOehBnCG3aniaPXaGUKHqFwepcUi4vqdcL+MYEB7sO//pEtNDSNP/1KvOir
+          RQtWlf/4mLzF4+bnVrDFL6aphPQ/HgstCWX0tI/mYZSkUYEGhdW071Xa/Ezyw+CMfUj+/MFKoi6E
+          AeoisvF3b3XyvAQS7Rfq2nszGvb1qQQbD5848X0G85xw/b98Z7Kq60Ii1wpMn+qdkCplgAVx9wHX
+          ZlFoau+7iGnn+wzlI6unnczJWX9j1w/46x+uGu+a5am7IaSRrBFN1i3G/vyNnlk6cf1rx6i8Ki66
+          q/ybaI/rsRF867KifShqODn+SMY5stX9x6fq+wj+9NufviS3Ihq9iZODDommrBPNFdZhjLR8BIbW
+          liRk1aKzhwMgEJ5JR70qyTd9Uxd/+UsuTelFzPREDTqZDzDy4Fn//k72DuKCffFB714ZxXOC4aRF
+          HlWL/YOxD89xQDf1mKiu8RsG8aKlIKfn9q++ZouLCITBp+7IxocA++M/OVbCrX6kHhuitQefBt7p
+          ufA471fPLw3ds0tCIq3Xh1lcbh/QjplCT68CRQz7+wniyJ+IgbRbM1PjJ4DUAHss2M4xY7PAl3AP
+          Yb7xuXaYT54L/+YzqCuflUZshdcKj2UCaXG6hjpdn5EPH7OUU8XmRX280X0hC7yJiDcaaOg/ssIh
+          tWHCxne1Zk5PKACcdtcndkVKxsn5TYKHZ17TU3Zsht+K1w6WWahOu+9RiRjTiQGiMsAkBcc+ml7f
+          Y430Z2n9PV/DvtJt61ef/TRbbTb843uRJvsEJzwEK59VMry8iUCMoFcyfm2cFO7o+sXcUvtsHtpT
+          AaPiYZGkqeuBZZ1twipKLWIXZytaH/SMoTL/ionpxS9b2u5kg3/8ue77bJ2BU0JLfeyJEzVqw9+M
+          0YXoE52Id4Bcs1bTJ/ir/5PwzCswqwcsi6/RX+g5VR2w6fsA6POSUoUbPcCkuN3J1Jnff/rPo3tK
+          doBEypMckfvwFvueu2DTh/hgiCWYjMO5lTMJvuixvvHDPGleCE9RivBbHH5svHFV+Dd/gpG37WCo
+          K1uBTRsCvH8ETkM/nzWFI39M8Fp3CphfXl+AbcU7FqP3oG/zJS3c8ocquc3AWtovH7JDfiTEcRpv
+          eS4PCBdK7Ql81TZj+/pYyn+8pcAOYHOqyALwUOCTE3DMTJwiZ4RyyhNycvQ62/pDCjOuqibhqy5g
+          yx8FAi0eqbLN15QHYLbwT89FqdU3P+ngcMDhnMO0e3MG+MFu/NfPJkjcUZ+QK8TQGfiWOk1ve8LD
+          iW1oY92epIOB2XiJmxJdjruWuHUVRtv8aAjsspAwT26lN/7Fg+cKD6Lw6K0v08ytcOOZkxumA/j1
+          tT3BkC9dEsl7UWe2sPfBOOsDFq4Tr7NvEZaofDrnKTjCjz6vanhB3APtJnHTMwxeOQVu/QwfHJ4D
+          M/fCLti1wm1KlfrCfrzz5LYnccl147ssHqkBsygk9PQ7k2h6uQUHl6fxpKlneNnPipdOXscwpJYp
+          vPTPWqqa9Mc3FXru9EXWJAG+PtAl9kVQwZy73xW+My6i3t346KIbhgU0opODD0L/aOZz8eFgSZlM
+          vcs7bhaZgOCffo/s4zf77dedC+8nGNMA4Dpb/KTx0VGiAfHLetK/G18+cPN9mlhTz3rfB3IITT5M
+          qRNUV/De6jXatx+bPC/fE/sZx114YKOpE2WbT+F76rlg3zwHqkZPpvfw7Rawij8T+ZvvW4Q6KUB8
+          klOibHxdCDixhpTmiJz8e73VU8jBZzO/SMK+jida3OUjVdmUUecgRIAJlc/99SPi6zdZn9NnVUNh
+          4eKJmyfdW69TUMP7olB6Y1UzsAnfZKDwdk1843MdBAjOAfqbrxrifQ3oo9BW2GbrbZo/x0Df/PX4
+          /1lRIPzvFQXN5Lypal7f0SLVnA89PaVYmhapmeXcL+Dl4/2od3y8sx+9XnbwzlpMLde86ot8fvZQ
+          WRJr4rjDCmaJ0AICl7SU2CZueIFLVlDBa0adFAQ6L3K+CV+Bsqcaev/0VSrqFjXm0hIj3NuAU4Kv
+          AKUBH/D6qpdsqu2mB+eDfyI3XtiD5b6vezC6VkAtbZWHcRZLF+Y7Qifh2grZ7/LzBCjpgoB3tdR7
+          vzTmUvgrTxlNCKvAOBSTC/m7yoh7jShg3GxpKP9kF6o+RYeJ+XCI4aD6M16eogPosxxC4D31F/UU
+          RWmEyzkSZIE9Ie66qzisODh3SCKtRp2OuAOXB8YHqpZfkwfnj9nyaPwJmuNgUiU61/paRj8NCrkl
+          UafiEVt0VbOBrS1nEgPJZ2uYX32ovMYjzbqozVZBdRPELcNMNevzzgbQVQFqvSInRFk5fQ3zB4bp
+          /VpQ7FwvYPZDpqH0pZfU6PobEN/744TQ451MvFW9G2bligR9W8mJ4l+pzrT0AOEOHSxqbuO3iI0i
+          Ico+KVVyTR9EfdV7dGzmjhTvs6GPC14kFN3zhSpVIQ4LLJwYoP2RJ/rlBJt1J3oY/o2PFK3EE9e7
+          q0Ax/txoZN4rj6/4oUbtrqzIk2jVsPrrckFfYX3jsbN71ioHL4FjNurk+LkZTNTSZYfehxMmTgo1
+          IAy/bwBDvH/Qo2XJbBQJtMGvDDp6XtRXJgqvMEaddROx2U/nYSTPQEHcID3p2ffsaHxyigKec6FS
+          YpIULPpXSv69Dys761t8GgasukYhzvlIs1WHc44ufqvS4NCwbAoZ66CwD6MJqdYl44M1MdGqjQFJ
+          9tBs+DR+j9DW2Jlu128WOyohugmGQt0+l7L5FE4xfGRlSwLhswLu/Dq6SA+SleZ52HtcrQQpmjPt
+          QXO684aFe1MTZiBUyPUcX5n4jiUB2O3lQCMklNlYeicTar/LSJ9/CyyZWuzA4zg5xAqfRJ9PvTHB
+          ae6u1FYbLmLW9TWis2nb1P3dJLZezhGH8rqZiH3/zDqzjLpGhtBaNLwksc6LjS3Bo/lY/sUD673Y
+          gNeRpZj96F5nwbd00Q/AN72yYN8sbxxI6OJ3KnV/iTeIql6GyNa4HMuewYHlb3wMvRDJKQzrTACR
+          xYGqa73pI5hHxp3C6QI73Sbk+rm17Ct6fQ0ufi0QR750GaX+YUIWUhQSovdJF+Y9SOGJ007EYYoH
+          eCbtIMz4JCW5/+IH9jmWO3QMHUatfHI98dKWLexOwYdoksOaf/c/3HQPM+txYqu/jj18XN9skk97
+          PvolnwAifK17/BL8EKzn38uET6u8YpFbI53XZNDJo0F6qs/9c5jJjGu48FlGrXOmZcsJnXNwSUKd
+          kOeCsh8dwQUwiUupVjliRhumGdCyz2g6hM0HrCQuTXhXamdaTl8VcF6ryfC0szJi4A6Dad6zFPFW
+          E0/c/RAPYvVTAoCFPcS87igRl1V5Dfi+76ZFFjtvOrSLiZzPriNx+OUZuz8rDWlH+J32+vT2aGRS
+          BVSS+cHSGBw9PmanD8jr9kDT3y1h6/qIeuStWJ3E4w/o7PTwbLkG7m+64vyjs/vzq0EzADKxSAPA
+          8lTaHYxOtKUWLiswyvndhRpzbKpfTvmmIC8SxOt8oRdvVL0FUKNGsnoz8BrcP+zbAFmDYGA+XvfH
+          rhGGXxVAY/7G9LjVr1+wBiYCihvTYm09JvRvbMJ2V1dU5bY9CCGRYnhmUU60U2FnrHB8CR746UQ9
+          rvKbr7QRXwXHKnFJUQ3LWDYa0g7aRJXs+40+4ykOoSnd7GkS0AvwKvcyUa9lBXXKtQHzNzkoMByD
+          kJzfnhEJkimWqPXynDxv6YetIl1z+CtJNonI+3k/on80uZ73DnGPHwXQJylN5An3L72BqfOWNKYj
+          uAmmQu0S9g17pFWOxJ2WUkOPLI99jnkA55I7EB8bKqDX1vLlxJMyvF+T/fD3ecjhV0+fU/ryWD4s
+          MTpoHSLYPtv6mltVgoSmDWnxDs7RYkeXAA1Tccffh9hka4tlF57y3iEhbJSMK72jgZRU0akpwQNY
+          zaVP4OGLFEw/dQFE72PUyBItgjs567b+3bfwcHFfRNXTjq3XW31BLmojEp2iCQj7TtEQkkVtkph3
+          zITXBUAoMlklsZIrA88O+gctw8UnW37qHPCcALFzfqMxkEZGbeMhwAJ3X6rb8gUsySG3YU+nkBR3
+          NW54ybtyCO0tnpxK8wHW4qW26Hh4qfR+f8TDHGmyAT0nuJKn+bUGzgPfEMq7qqbP9R5GW7xc0Mnb
+          HLIssGw9DpGE3Ng7UudafCNBf6UGXO5rS63C1weRmy0FNmG9UINld0+03YOGMloymvLzChggrgz/
+          +mfx+UnZKlm+D4B7ajHcnbpmPVXfHB60j02uTYvALFTBhLr5MpFHFFN9y4cS9Xw50aPY+5GgPoMA
+          pTCfCHFa21vJewihKLkcdZZqP8yZtMrIOH0ycvxxwzB1ujRD83iS6PXZNdGsfmyMUr3Rt/7W6D2y
+          IxvlqpHTbLb2bL51JAWK4T4wcndvQC98syKC7y71Qmw3dMt3ZD3nmDx71dOZOOQ21C/fF16yB/J+
+          wtGKUXydz+R+NfKBdfs0heV7PVAfTnI29e9whcXyodRV8UVfE14rUZUKO6oq4gkIylzaqM3uZ4r5
+          4tCsohVI6G98nme43ePX/Tde9HKXvxk7pEYI0TdKJ+DP1Ft8XKUgJSHZ6o3K+LU6QJhwnEX9rX/+
+          XQ+ue3nCMpXPbK6VIAEcyi162UZ73u2Bi/zHMtIsfteD+M1fO/gaZo/qVr4dtnIpSoi+l9e/frtc
+          fkEJzWOkUzVo64zdn/kFhOn3iWHbL9kCzH4HEnXWyFmRlGF1uksMDq2VTocU1oDt3USB5vgGxM9D
+          11tMwZ5Qw6mYuFx4yBan/+1gcSkCjK6nGoyXVJ/hLIXWJIUXfmC7delgWq4/LF+3M3yCQughuB46
+          qiBU6vMi4hl2kXqg5lHiGqZcZhfOUmBRUkx6tNBr1cPkkUrU9198w3jhUcpFaVnkZOfVsFbLXQYn
+          c20o8Y8ArO/MzSF+qxa9xLtYX+++1UHxaFCqiiAEy+Uq26C21cfW//bNknwSCOdMedB41XyPmR1q
+          Yf2Kw4lnwbOhrfsp5foEMTWW3ZrNgNUcKs+wmO7DUAzU29cyAMHRo97P30eDcplt9Etzd9ohkTVj
+          Vn0xHG6qR63vsWg+vXUpkGBGD7z2e8ljt1CM5U0vE63fJ974p7f8aCxIfr85g3CbaAycH7ljmB0o
+          Wy8fGUPnAzsSnNcrm60PPwNxLFZqMjNhU7XcpUN97wOqorVrplL7chtRfhOy3tdoRmfPgMmj0Kkr
+          4jEb51Kdkfyh1oQeL1dfNeELgWHRHoPZXodVfSYB3MWPA1U2/cckbW2Bbu8bevRR9qdHU2icXgF1
+          HqIeCTLUZuicq4QeWbAfvg95MKC58yeCpdRnYweSFdI3TGhofaxM4L9SDze9TlO7qbLxELsfqJ4M
+          m3jveBf9+gtWYON/XYLfwTnjX/djC6N7sVC7eeoNL8lhC0+fi0czU8b6dn8d+pbQp1E4XxteC0MF
+          slu4TnONdCCau2gHj+yXUTddvtkydYdSfpuOMy38fWzYfX6WcnXZJ9QYy4Qt/P5YQLXNfZoPnZFx
+          Nk12wNktKvHN2crWiHUylLniQr3ZegK2gIcBYdoW9MH3R8BSfHbR2XTtSazdnfevXnDpeSTFa71t
+          /fpnwt/C4b/62KxBjEc45NJCC0v6sTkmWgp77V6QP/27mMWjh2v7t4Ibj9HP0e4mkpwxJ6eVHqNV
+          bm+j/L6NC31QzgBrEJsTbPZrSpxNry1rtezgwF0UEn3nM5huj9aFn7jWp8PmH6hBpA6+z6JHzDru
+          M7b2Rx+O4pnS02tn6MvpBQyw+de//I/mBxF38ml3zMhJepnZbJ6lFboVb1HL/uJoVdQhgffR3eP1
+          /OAbWizABkf/eiPuAIyBA54awlt/YQQ/Mj3jT/gygovCemJu+Ujdw9cEM4kz6lc0YCLOdwUcuIDH
+          sqt8dEZ53AFzhyeqSMcDmzc9AE55tyP6efyw3hZbH9n+jsfQwKrHUXs24E9QBSxctz1VQbHr//oD
+          9dXwB6a/fNvynzryxcx+58rG/+LzBn5RRElz7wE3yE96zE7fjAlQGuF098/kep8e3uyDvEPyZw3/
+          9HPEtPQXQi42DHJNhsb77kBVApLaN6JJbpeNx0bp//wldcajN/zzb2d2eZNwvRyz+dMrMuoLt6f2
+          HprDct/3Pfz1Vw6vKfkOa+ilGjz6aU3Nb28AERBNRmtSD8S6M2fg7+99By1DdQnmi3vD+EJ04ZnF
+          b8wnQ6PTffUdAZOEdMvXZ8MOjmrC6H7vSBomJFqzL12BzzvzdHgIwcB/XnWL9mclIcHDDCLxVb1q
+          dL1p7tbPvWaybM6FVTC+iG7lmb7sVcmEO7MjW/4uEXsF/QXIakqo6k5qNt+uiQS2+krc8F5747Gx
+          e5jH/DiBYu+wVcN2AuRdU1MsoBcbu2ZN/56XnJxnNszBdnJaGHRnchrFvT6und7BekYOFie4enSi
+          xgr10WTE95wq+kD3LSPOeJ7J1i/YEpFi96+ekts91rlowDtYL8qFYP1QNqu/HmLQHpuBmoJ6j8QF
+          XA1EeIaneetfayqLPcQ6EPFL6d7ZMmezC7b6S+xzYDOBD08txFD6ET/Uqc72ayuj8bqE1G7BDXB5
+          RQSAbnSdxOsdgukbOOYfjyLK+j02wjHia7Bfa34CVzuOpv0LxOAr/HIs72zZa+7PSkEYyj/i5Pcv
+          WKvU9reNAHdiqbQc6KhUCTz/ngr+lNBtOH7NJFjfX8eJuzzqYUkCuYMfJJuYv84b8d9muDd9h0Gx
+          /4KFpIcR1rb+IHY4aRmTbS5FN1uIMcgfTfMjl7iEeSyO5Ng9d9lvHqcecq/9RMyw/wzLW1gLFA3a
+          nejjsEbj7nPZDimMX8SZuqe3/nJjRFs8k+N11hruhtEIITQBPZ5jnq0TixNUR/RJT3EA2SztJBn8
+          8SCzjt2I663Khivn1Vj8Podspsv5A+ejLWJ70+9/9R4hX5/xRD9Ss/xEboKGXmtYdE1en27T+wLh
+          fKOTeE1pQ23jysHh0zJ6dLUs63fPUdhODzSI3QLxv/r+WOXPxgNPHgstJYEyl1/IKeBsNg/C3Ye+
+          reXETlxhWBb3E0B1al/UnpZkWKriLsD2HEp/8eZxG/+CUhYYJCp0lLG7IE/wJBGemHLWeczfxz4I
+          ErnY4ukD+j9+xojY4AUtDIzp2iaom34y0dzmDZbim+F/ftVtnHlgS7Mb5fEJBno6VpX3F29/fAvn
+          oU49Wl8fGILrbsGs48/ZeuDfLcya5f2vHovc0+NAztkVdZRryVZlLl0IrtuK8/X0Gua7DmP5+7wl
+          G++q2RiLzwAur6KjVu0W3vSMDhMkFK708n1J2UTijwGVO5RIbMXYWy/nsobT9wiojdIymoMTy1EW
+          cSst6EcaBkM6FvCtSjH54zX//Gy1BicSfop2oLINE3njB3/5Okxk73ygGL98Gm887RuzMAaRjDBV
+          E/htFiOJBLT5faK/DzudhYnXws94jogX0DwSE9vt/+6PprIzAmruIij39oUQJ9vngJ2TIPynf+0x
+          eHnjtUxDaNyjkGo8eYFVTEgCc86tiFGdFm/RVdeGNpitTf+MYF5RtYOtetEp4QLT++PDMLBiYUKU
+          hd66x9UHGjInU8cKKo9uehhVThQT912J7HO1gQy1Iyg2/+9k6/r4xECdXUyVQ9vpv2iXTbAGVv2X
+          f818LY+tfNgvGfUcHg/c1p9g/JT7qciQ06xy2yZw8xPEMSIQLbnHNGBKSUQNE7+b6Q6MGCGKvemg
+          iW/9u7pYhneldEjccDobayVJ4Z+fvJEmY4uhoAL+8UHL8t/RKkcVhrnn9f940LJWvwDapWHS63vn
+          Dj9ZyfNDs59Tek7TbQ+zcFUgB8QLcR5iE4lxX6fgaT1cqouv59/nQ3nYZZiczNcPLEsWr/C1NDk1
+          5czUeeXIa3/+gVwFf2W/OHuH//zGKbf1jLG5TP54D+aE25j1t4fEwU7O4CSjY6mvUd2M8HsNXUo2
+          fsBKEULxvTc4QoLiymaPxwbUGj/Cu1tLBrrxTUguwYIPhYyi+Y/XTxxvE7UlSzQr1U8Bj+qmUGcX
+          sWHLj1B+QCQQHezHhp4m8wLkm99T3Xr8GD3uyh4sIFKJc1zfzRYvEG68jGr9XtK/11F0od3jJzWt
+          W8lGdLjtQKZFLj1Rtm5+ylyBT9PnJFBrGBj3TGpYAC6i56ygjNnuoiCvfGFqdrTNJixNBYhdgxAL
+          Wkok8sKjhtcY7jGXJRqYA6zYqNOPxl+91hernQW0L/cnzBe2wzgtHnpIlcol6uCeBx5OogD+/LzP
+          c/4wn2FTw2bCKyEZq9mPSbsd3Pz3JHnpG7DZ/ylQ20kXsuV786VIKGEwTQfqJniOPudTN8OmNe7k
+          vsXDpsdzGKiPYPo5TzCMEnkXAGI5n+Q52sr83anhoOKZHsfrM5pffjlBO/mIJHw1rTdymeb/8elp
+          7344MHugCtA230LI67PoVLy4BewKLqDRebQZO5JBBr+0cCkO+tqb7+vBhMUlO5LT59A2o+D8PnDT
+          H9QWzBdbPq++g4fi4pBYLI/Nxi8V+KefZr09NevlnHFg47MTV0RNthqvY4LCVW9wlT0eOoPdlp93
+          dKCpdXw3w6Z3wd3ta+L/NKnp9ROYwM3m4knaPeTomx2fHPwJuoDlQL6y9U8/bPwU8990B1ZbnCX4
+          cCCkxuaHJ5HKOYTC6YoR3JXRv/7VCf6FhDw5sl60AhkJvqNv/Q9ElC/27r/7Oen9KROguofg2gFu
+          MjtqZGK7ohQ6ke2Ra8zyYd7iG6hTlRJ1LXz9k65jCtn57hNiW1W2LCtyoeYJBC/TcdXXQjBmsPm5
+          CXJx5M0lYxd4Ps97+qeXf8mB5EC6D2c8I3iK+DSmk7z5R6I6wZvNWjxfUKmkAt637XZmR9Je5NOB
+          +9DLYLTDvF6VBPz1E2PrP7+JmAWM8U0juLyevfmG0SQvtg+JHf9IM2/vU/bvnUH8cmh1On52oXzr
+          Y4YXh/Aea91MADuzJZv/uWfr3/uYySWb9lP60ucOJDNUXh9583uFvumj+o93TkBRlEF86LoB05da
+          ElNOWbSyB4b/9Ptp4yd/8y9AaLqQEOvHZf945lPu0o23us0MXLjpNcn8ly/La5Y+chfph794HFbm
+          Jzk8jRUj+PbSMj75JLs/HoUX37OzZXHLAP3A7o0l4Rjp64GnLfzjBxv/ij5lLdlwG69thUc5/Og1
+          h3CK4oHq7g00bPNjf/qD+tztCER1dDhwHZeUqtVlZmxfVSOSSKdNm/5slnPJSjhRbaRWckHZGF7U
+          HfwJlko80Xg1f3oYcqiwiIp/UzON+T0Fv0vbkuyw73SaQ8rBkUCBHO0dboRDuxhITPk99ePc8dZ1
+          ijm40G6/zVcl2arMHxv+1WunXM4eQ1d/gkRMVbL1y80fjwk0ZEaJKT5X8GXBI4H3t+T/43nDprdQ
+          8dr9qNPORTYvXRLAW+mnVMUjAPM2nwj4uTzhvUlkRrlnUCP/IYjEfdf9sHyHgwH8e2v8zSc1vPdc
+          QvQ3X4e17K2vlXa14eY//vnLdePxMHkkEglWkjCWqLSAjT9dqRHoV0ZbN+LQUQsc8sf/6HrXFDnP
+          fyH548u/0wuYcJs/nGrJYcN4UgcbOjum/vEc7zdogg1PMU2mbvSSZt7dzh1chtjf6uM+mk7VZYZB
+          pEkYqlE5LCr3M8DG+6avR6ZszdRUhrjCIwY//xnNeWVxsHIlh9i74uRt84cTNOYhpt5hxzI2yNiG
+          iScGxNP3p4bHyJ8h89r5Hx9eEz8PoaO2DX1ILWumm3la4R9vry51DhanX3sQ8ccPNW3OYuMZzhhh
+          MuI/faCvX6n+gP/HigLxf68o0O5vB++Xo6wvi9/l8GCtC+b4dNufyWETvhN9oErXz95Io8pE9vCM
+          Jl7gNTYDvNbIhSigWvOlgPVpa8K2/MQ0P306fd75jQbXS7xgiDqnET+rssKPR8kElCOOVkE1FMTH
+          8W9iox0AzjCCECZWNxHr+gsBk5fQgJoRCcSu9yNg0SfpoX6oAuoG3y5j13vKQbXb5/Tk+CfGDp7l
+          wqRmKbFvfZd9ulixoTdVJn1U+S9a7mfORdwb5ljcO1+PXR91iMbxlVE1Q3Ykasoow9/RvBB7ei3e
+          8jodO2B0Yk+xf4T6Ys40lLfPU+Mw2MPME82ESyt/qeqWDLBpWXyE6cMjXiI6Gff17ivU8nmil9jU
+          Gi5CoQu7afejVha9Iva7nGwg0MOXXA6wy5YDsCe414hMzVHZgWl25Byt60+jD0tMwHyfLwFSUdxN
+          40Ktht9PmoDAr3RpuNy0jD9Wnx6eR4ejNoyaQbgaVYxuCn3hhfCfZrCNcYUXFWCiMtjr810NDfQd
+          6YkcQ/8IRBseObhqJ50oWjMNs7yYKXIvO5f461n3xKncHMqDM6k1H5WBv7wTGYBf7U7iLxPYArKj
+          C9PfxaJhLcSD8PbPJYq+skWVYd8CMe8dDdnJeSL2A/Bg7aZXgqQYA+pmHovWoHxxMPE4F+95KYr4
+          8yvYIcMfeZJsZwAvL2jNYPmYmKoVSr3JyFEt3799M3HH38vjwRsUcLx+/Ak8em0Q8NufYRALF1oo
+          WNZHeTETuBL1Qa2d5EXjtZoUWGRdR5WUTfrSXcfi4H4oR1SZ8t7y+DoB7Jjv0rP/0fS10NwQKEF7
+          pOd9jTP+sXxchHX8JW4gG7pYAtJCb8wS6j6iqmH5a2kR1ZOR2Mp7n33LK9Tgc7865JbxCpDEfq9B
+          Y8nPtMjgA4jVY9tTfwoxvSYT8+YnBBpMY4hIvrOPOleGdggPon+gd3usdLYTqwJNxRBQRT/2Orup
+          O1suL6+B6OI+HJgfRRpsi/pN8de+sAWnq40q52biMqKvaImt6wVJ36QkSSz+9HW37DpozcGH3oE4
+          6IKcf1pol+ubeHuZDWyghwDChXHEU94OE9QwnhDXSS+azLnacG2xT6C8MzVKFH0/rIX7daHotRax
+          +e8dsGsAILS6n4SXo/f0BKjAHDTqalJ7urBoHiQjh6IkVSQcoyFbU+rX/66nya3c/JTxnSJhVRPi
+          WB+HCSlYBdQ+ipmYt3nvsezCOHS5XwgJKgMNLAmBDH91Mk1ipd0z4Qd8G8BCcfDMJVrDiYofwpDy
+          CXF/WRWtXD678Anq33TonYQtykgTWBM3oafq53kzeXQuPKHDlehSMDezoSgCpOPzOyGRz7O1CuYa
+          Kc1ZotZv5IYZaIYPE7t/Ef3wOmZ94kojFDjHxMzCPJh6V1OQkCpnuuUDoyJNXSg+bwveHVWW/dU7
+          4IZcTpLCZ82kHwMNhbbsUDz5HhB+emNCM7hUWHQt3HBFE6cAmwmexAPHRazFpxgWaU3J5TGabH6/
+          8Apfu3wizsjq5vtXb+ZJfVLTzhLA8yESgFPHD+rdFImNH8GoEa+qKdFzZfF+Zd1fwO66vxFHLTPG
+          ipitkInFgyjo8vDE5NljaOimQpXM8ADzlmcO2jw9nNYyd6KVGLME1Y8aUdzYusdMErno7/6OhaB6
+          bHiHKdJ+3Itol8u34fQ6H+ULexhbPdObT+QqNbLpSLf49HReVWEPDymM6e1vvD+CUcKrhJ/EyFpp
+          GIt0X8KtvxB8HnAzD0Psyh94wsTQL3dv+WhPDP/qk1dfn9nEXe4hjKIabe/rlXHfaO4RTIWWklPJ
+          6ZPxyhN4yJUTufY/Xedz0n6QMRxNcpdsR5+vd8uFdPe+Uf15ug2sdXgJ3N8lT857ZWXLpZA1eNZU
+          hT5q0jUfF3xXcKRcRcPbegKrmqIWLsFwwt+i4Ya+P8qdfO+nFE98FUZsB2YJPrEfETMdumFxpucM
+          ncN5pGd3WYYFctWKdkSxSLSzFMawPX7ghe9WarnJGK3voITodBGS6aWeR7AKJzmGz5Odk3uEsceR
+          aZCgFUL8O5xuCCxS1Rro1R72VK/EKhOOctWhJn696VEpWDOYuVPIlbRgEg3d1Vu8V1miKi8Hcr+E
+          r4wzR1YjQ4EhvmzvY3m0Bw4G50YlPj0+wFxbZwmdlBaSjZbrvLu/lqg1VUqdPo2bN3ksAoLUVCgJ
+          NTosdzhJ0AziilxELtHF12xzyD84MtnqobfGuclB91TZNCzzb8b0VC1R/VIXcjOCXTZHHAslab7O
+          NB6VAszRQRvR0GGDuMa6nWGADgG8CsZCL4+xY8JBuLbwdzQu9Fa1FyCoKd9CCx01erzdd95auJWL
+          mtK+0btkf71VG0AM31Wb0fiEV48dtVGA23hj6TjE0XJ5BxIy1tImZ2VRPFGl1IRvmo8kcOJ7w+rJ
+          lpBmnAWqxqakz7HAfLTuHi9qGeuBDVe/reGIiTtxfOqCFaVnE33793M7hcvV5ztyW5j+YoueXSJ4
+          XX8wMNrqJ3W78sCmlEUQPeVCoNYbKQ1nmYEP/SasKd76Y1+Y5wnl77s+7eTXN2L1q1fQzOcVyd79
+          5U//BLLXn/up0vPBe2PjilFfFV+iF0PuzQ5ztz0mV0Td584fVqx8C9iR8kUtpn6jZX2MPqz493c6
+          jGfL4139p8BQDFKCa8P3WNwHLSrLwCF3231tDvarQDvN8qlTD94gop7Z0ON3Nj5s/x8n2U/gVv//
+          6m2zRlcxh3/9zE3la7PsqryEmWx8p92D4Wjmo3MN0yswqcfWCCy7x8n8ux69nNu3x99+nwB2P1mg
+          x0KoPJZk4YzoJW/oY/c0Pb5dkha09WzQgOmnga0wUOBF2c7IYaqTzS845EDLRpEEz+6d0RJYHSwn
+          NyVGbnIZizvHhpebX5Erv3hgDBu+Rlu+kfO78QdO/tAaCqYwEufgKx6v4h5C26gNanbKyePJeuth
+          E7/fU5EXh6Z/LH4By5jep8MlfEVraM4cmuQDo9ryUbzRuXch8NvLnSg7S8jorIgSaCbtQ23l/YzW
+          c4tCMDk39N/zasooATpTkyrcV8kEN//twH2Yz9v4WDpDhS/A0WivlIDiBRb+dgrASCKVqkdvr6+X
+          Dvl//ZHmr9j1BE9OFWha/GcquOnusVQkI9jqG93y3eOl+8mHU/ENqHdmyjAfo1iGf/FO6FtnswEO
+          F3jLU206lIPG5pci7yA8Nudptf0veJWDitH5iF9YvP5WsMY5FuD7CBBRp1gcRqcfNfj9PBZyVWt3
+          4O7OvINponBTH3y7iKX7qYOK9PJIuNzqaNM3IfjzE9vzeOxiP0wYyK8dPpSOlVHtlPmwOOxE4ljm
+          FK1//dzdGzv6jEyzma93YkMFCm+qxp+oWbp++YAXUTqKp4czcP4MErjbFWQ6SPcOzMNGu62P2E3o
+          kBFv7s7zBPdsmXHzpJHOOfcpgPEg69T7BiZY//TSl/fNSbjNT30+qQuGZQ95GnQ2NywV+5hQbN6Y
+          GvGtzGZLU2Z0EByNKE/xqI+b3jmoA5aI6pYRmP/077FRLOLmpR0Jp/O+A4UZRVt9KIeFtrqL4vB2
+          oU7BhRGXy8CA5iErMLoKjsdDZH9AfdL6KdAaPKzEUS6oRvUXb8/HVuskTiC4TB0lBbKbpazrGDgB
+          phTTQANcLAAMnyc3J1o8CWytDBdDPz/caPTz32CWruccGad5pFn7+A2sqtIOmg8XbPqh8Vj4XGa0
+          hDMgwaY/ePs9a+igV9cJbv5znZI5QKKT3+nl81wYY3meQ/6WtHg37V/6nO/dGprieabnembR1IZS
+          gShxCCXtCIcxBm7/r37lzXyLPq6hTBAqpx0e07DK5uai+KDjApUEO5kOlHTBBK/WVOJ505+/2jv1
+          cL/tvVcrJHsrHnL/cB/WM/Van4Dxo60StHlNJ8pJuuoLlBYf0pNaEec+Z9lMKr4D2UMwqREsJzZH
+          x8WAXyUfiMrSqaF7YimwrcYT0eyHNQjVXh6l9VbHhKiWMPzdH0yOD4UYhjwwqpRlLnuJB6l14C4R
+          299+MjSb2iRqgs2IntHOP2TALempWzSdMzN+BX964uzf2mxs33GAXigRt3x4Z/POjlpoYK4hpMh+
+          3rwokQCrvB4mfv5xYHTtZFvxHSlUl1Zfn4yAYFhjo6LXfqeAtXXvLTgkwof6S/EFTL/KBkCn/Ewy
+          +VWA8XonLvpq7X27f48J3/kXg+HJbb+6GBhseeTpTs5Bv51SbhqNYDWGjOSD0FFFCwMg7sRvDous
+          7UgqIhz9vuPSoX/5SIMaLHvhO0HTEj/4cMd7sH7nVwxESa5O7DK4g7D73ib4LhKektun9tbbkmvw
+          FEOTZCyQoqVp+/Cvnk5or6rDmm6/wbWEK6C6+BD1rd5N8pMELsl4y2azawe2/K77EHP7QfZmZQlL
+          lBzclvh2S3QuACFGFpk/RAsfd2+0zMQH6RgG02HT6+OdayZwFtaMbPHZ/Hu/x2nu/+n/f+OXcSzG
+          K7gxtljqz4SYW58/6cwEj0pQKaC3zg0NJfeXbafg5TLW3JVY5Kd64l9/uL40+YQO48joWTtOEPlJ
+          StzPx9YZOMwQ5mf9gBdlUfT1as8rSjVaESKebH0O9t8eqie5mBq7cZmgPo8a/O4EsPnjVl/Z3lfk
+          fS6UVH03fsOr2TlAX5KwTZ9PzZ8/AZv+m3ZubHviX3yvinUljtQbDd/dgQRlpHgkOuhfRpX8rP3z
+          A25qwWwO50aAb1qM1MTh11vhqV+B83RFYmqoblba0Bxs/Iqa3dkF1CmcFCpidaFmFCv6+uXgDKNL
+          xxFF48aM3Q5PH+6nHZ7QeusimlW9DZ//R9qVNC0LI+EfxEE2STgimwhIEBDhJgjIJgokQH79FO83
+          x7nN0dKyWLqfpTvpMDhA2tnj6HJiLBaC8JB5oxK7znZbxx46pJX3elHuUBpyPTTmiXrM/nlTxe8A
+          3b44e3k0Yfo9HKwecveA/+fXf3J82b/Pz6jY+WLHr+CvnoaQyfRg8Tcay/v/YzZrk5GuotXCN30B
+          5CT9OH63Nx1k/769MDHnr7al6SGC9ccpyWXnT+6gJbp8vYuIXC3ga9SsfEm2preNR+XzpfRyy3Qo
+          /vIHHtEzGukoJpb0I7GEzNNwBRzX0F62F4tHp0PthYL7+BWgGUUJFf5HaPZjWBLocWy8+/kwnGpm
+          g9DhoIUs4yIBsseP/MmQj3Z8cbgnqVxZE8t07wVaI1GqIQedvdr/8med8bGCe3x5jKRd6U+c+Qme
+          FOVKXCatKcEVKGBR8irS3r028u7e0dn9CbIHcxoXG7w3+ZG6LxSeSzqS2cH+8bTULrpUz87BH8s2
+          IQiO7Z+fauYLzw/yFQoF0vVkybD30Tew622P2/0SxZJvwl7b1D3e8oz/DmYBk/fcIGV7b863VDAP
+          y+FuYVrd8P77RIf90DXopKyVs6aGwQLma0OipETQFrPNf0BoOo9YzdvKtlZuPNhsvoH+/BWV2nH5
+          9/6M4NWMOCKR+ceXWGKXM9h4CCsY9+2CymKywdIWwvOvvkKU86drllyiupzenJGgJDw4vx2/QI2s
+          BCU3BVK6PPMKytI2IvUG5IaEQzKAk8fckdlTM2Pz8zuX93ggN6jb2u9bfnpw4I4FuXqbErLvtydB
+          Xp8Y9OorPVzIY6nkt0Ae3tIJMeWi5lLAHAzlnz/Svo995gTMpMyT1DTJ8F25SbJSFLF3BILjbCo8
+          mnCSx4C4qWdnZK+PwkyA+8qu9jkSLPm6PIJT50n+18w423UkmF4TBZ1QeHUwhC0D0i47e+ywkHGv
+          1/nw0hYOfjKeDH5+NfNwrp8YOS12ss09W0/Im+xEsuQ+g+3Zdz+QCQwg1uJodH3kdQ1tBWN0sbIm
+          3DQjUeD3VNXkoXWnjLcI8P7+j9jaC4UbmwwT5GcKPF5MYUjlOg/AL7/ciTs9A0pR+nvC7Khu+6mu
+          S4h3vJHIxrLoAa9KxpHpp4Jdv5KLNZ00/hMtT/nPb2jRc6LTqo8B5OkZoevcztq8LGsOWfun7n4x
+          ceg4vnmgnQ9P8ucfWOY1LHDRYf3v+vlnHbswjl4qPk7KN5tKI2Wg1HU28tlEHfm9fgUpFxOiivYc
+          UiWUbCiLg4yiferUDNnvBo9PJvbK0OzHhYRf/a/ei3Q082M3KucJDqF9wyz7WZtlFN0cSoCJiPar
+          GWfmApkV7WFmkY2iW0jhrajgrs+Q/fwKYL0GWQBfakfwSxzulN35H+x4hxz7k4Ubt5guVOOsJbr3
+          Lce1IskAkqPVEuOzgWbe/SkIOntCqnvQnDEJqAQepzVHrsW5dMsUNQLhVzyTtPyMzSYXQyWZBpGI
+          TbOAUn0ue4GO9LDji0rZpRcssOsxomeq7GyVEPKQrccL0bXoqC0CfYlwr49jkTsXGf4ddRcC3J6w
+          kqcVpUF53CD2lwOy87V36I4/sL4Ihne4ojjEf/F4mgLyjz85KCs/2Zq7EzG31B6FJb6a0uE5bcTs
+          itu4ih2oofxG0AMI+s6y3W69/KnqFtOK2tm+Pyn684fofM/vGr30gw/m9MjtepNQoiqtCFHnF/up
+          0GKDnc7lQdCoDCa/0y0kXfOsYRzcoz1+85C6rb/z+/wm1j43bmo/hf/3/NHVf0fhEh7tCY4iEDzm
+          CzznH1/u/RtvyfSRbt/BKyDVliPJPikfznb0W+D8Mp7eD0VrNm23tIcJCAJktJGQzb/1pAJmOBu7
+          Xpy09cLzv796BjrBfZ8X8y0x2Ov13vF6PGa/ty9W8HIRWmKePSUT9vo3FJPoS853RwgXUc8whE+2
+          JfkJDc0iWqMF/voJqJh+lGQR5SFhugdSdv+0LUajyH9+xEh93Oz8Z0O3jVOPV+1LtnnKO//jB3K+
+          dNQhMR4UGDLzQHQawGZ7cE3xp8dQEOOY0rWmTxmEcuax32DU8C+oJPlm+jkx3Ds37n4ogr/n6YNF
+          GGrj7oc9yIWOhSz3rji7Pp3+9DDuneNLW4LBZeDYuzpKdj79838yiM8XzHEV1lY/qQfprPAV8W5P
+          H1BVNOA/vf1X/5rDwJkAT8DXk1hv37HSRAUEsXFB2gR+2hqNVSQrbjT8PX+HegJmQXh1NqK2TAsw
+          YTsXCpciJefn+gU0c3MGtGVbIIc7W2D3n7oUiPHN4xYmaFbxFf2go0WZJ53x7CzE3BaoH+L6H1+u
+          z/oogeHKyx6712NJ6xkxlD6TSDzyrpolvYgQbo8qxv+e/42jClxvs46bS9w6tFsUHqI+uuM64Ttt
+          75coULtVNimMYHWWe8cVYH9/yNvjc+sF14dB7zbESXRfozf1igHzc+lfvTJcposdwOzS/9Afnk71
+          OZUAMz8TTypnD6z78XdQjdMW12C4Nc0ffgGf8t46iO9mLUUkwsjQKbnFjzto6u+a/+M/Xbq2Dv3r
+          n+39NXJr415b6s39waJkVRIZzndca5NxgfpWHHLavyfffQDH3j9FKDT7hgRZHQP36Ej4x3gvsObO
+          fd9hxtvofGzihv71DxbLOiOr2Uztrx4JrhJZ/uHrdMd5BM9rXCKLfC2N+05rK4fwIxHTa3qwVX2h
+          w0smB//00fZYIwV2jnBC552fJiU3YiAcmQQh/rJqtFySZJ8JFaJwSr1mjc+vCB4q1SSu1RJn8AV1
+          kYl3Tojh1LlDRajk8s63yN71y7KSWIE1LDaklQUXLrn2VKH5sna/cP6BlWStD37cZPx7vsKfXyVX
+          7U0uZTxTmr2eE3S0OCPK+gIaZvPFgtYm3dGVKU1tY5MKy/RthMR73N7a13nfdFmXXRblfF2BNYwz
+          G2bvbkU6a8TOolT5BlXChHj8vttxZlwUwb0ei1CzvAEtMs2DnGrckPP1ezDF56MCd/+013vo2Afq
+          N4F9/OIw752cbKMS7EFbbzoxWj4Ay1m1Nvh82jVSq/wbkuThRDCHU4z7zZrAAsslkFGtH0kspjD7
+          5xfqq/JDRvDSGs5rUgUalZgSR2u2cTBXlZEj7CzEWJit2Y5Sp4JIPRjePj7SWa3PosBLaQnkT39P
+          VaAE8uUYTlhmloauv/Wi/D8zCsT/vaJAaYSamIJjNtQPux42Jtd7C3s8alhxbA8uMXMi1tqH2U8U
+          b7wMtOeEjEtyANNZ5Xm5XJBJzMZUHLZ6uD7srs5EPG7GdDE5g4V2aSgeJ8uPhvM/1VMSG++Gp9g/
+          0DbAqy8/qK4TZZN1wJrNukC7bnwsdXcvXNcvl8Ph+KjQxb+Uzqq95R6eLeVMoh4+MxycRRbe3uDu
+          8QfQOvM14mPw8A8KUuEsjQtdzRjky96RWstZo3r8UOVLbOynOszH8YvVyAPhNY+I8qslbf2IOIAM
+          bL/kCoYM0MgSeEjPPUe0KfMaenrt53YfrAaZQK/C2QCzBKWJ6b31FDVgvXFpBPrDrtAaUjt8cF54
+          6MKwI4pqV81qiuom6wy2vY+dfsO15hodRqzLIqu+2eE66DOGUHDuxGne55B+Xp0rI2nKSXixjIwb
+          n5II95nV3vs7vUKhFbgCFi2aifEgzDjX9FjDM5O73vqdXplw/kk/+clPGHfLC450LhhWCh75gF6R
+          5za8xr4UCP1AwAc51Zyl6vECysM2Ide/dqMwMZYn/07iD/kkfYGtP5iRXFZjiTx+K53VHscCHslP
+          xTSJB21jeCsC+4IUkiogz/gmukC4HrSRmGWdOpx1A7Z8MpjeAyA5ZexVmSPYavNK0O2ljvOlnHQY
+          K0PvHYI6oAKvvkzISkqIEsnSMz7sXB/yk2RjYNJLs6X1tsDCXwuSuaEOOOKF/b74TkFGSxXApd57
+          A1+yduR+n/SM/rgGgurZUHI9KV+6Hkjly6/HoyaK/OUbLF5mCI+NrKOMslK2VrPJQmWrH+SisLHD
+          RVffhtErvhPUZTVYHP5RyCcD9iicapbS7DAokK7qvof10Wmr/6kS2RAtk4S8SsB6Dd0NXtkkRg/g
+          ziP3MR8BjOiSkddTe2bsp9psuU+mieRiIYL1oMBEcvRfhGyhqBraq1oMx3cckFhU7WYjFVAAdcoX
+          BsLMauvpzriwtFGK7FG7U8pvWy3j608nF+biZO1WSoF0L2QJr/f0l01Lem5l4edh5NqzHbJ7vsiq
+          yL6IoRbvhtPzSpQFLfI9MeUPzZJFJ13+3HuBWAol2vo9qb08ce+IlNd9FcP3QxX5vEo5FtKP6wjp
+          5+0DY+hL70vH3lm446+FOWwzdN4aPqNck6jQ+JY3Yu/5LpBPpkPQMRfkqzdzX9SQQwjkOUVPTSEO
+          qdQmkC0jHtHp8NE0/u10WC4vq0KQZmrj+v5slvwKlTdCXaYCtnXjWma+1hMplpyMwjx9NviXf6C9
+          7VNni1MvdyW54OPhzIXTVbZtmMrE8aZi/DXULcJIfvXqg1wXaGcC3qdOExR+8WE1mIbqhg0Bd81r
+          5PqRA9jjmGLot/sKhHq6jQtTSAlIz6T06Os6Unr4xk84OLZNjErTsk0+jjl88Mz0L54nIvoMFFZ5
+          n0qleyPL3/Va3l7BGUXJAzlE8D0V8uxVJa7JLqOQpEsse3jWkDddQ8rhG48hioIHMsRTM84HMgTw
+          UNomut1TO9uaz7mH50wiyOgOQrZKPsNKO34i5VZexv3+Wln1oEWyp7g4WOoQD8tq34Nr8P24frph
+          gK8+X5BGXnpIn9EjgGkADWR04s3huFdswzhJee9QsqeMuk/RgwTdvsSBqMtWRYr8f/ndcuExWxKc
+          xnJUbR3StdoGmOpHD8LoIxJzq1EolNTC8OjL3p5PhkaUMvMgGjuPRBfV1Pi0lhbQ05ISzzhIlBy+
+          xVMK/cJBhkwcjZ6zxIKy7ex3cn5TYkl3S25FfkQuSIfsJwurJA/KnSH2eXEbQTs69fEDjxjpWTU6
+          rPY6m//w1VElRLdo1QZ40B8SulhlEy4qii0ZD6mDLpOQ0821/A1+xHYmRp7wztbcQQ2H/v5Dpfc9
+          astV6WK5zg83ogrXQ0Yd/5eAZqhuxDKHnH4iVumlemMY/NV4h67jcxPBff4RYlpRPK5/+D92roTU
+          9LlmC4NzEfaTBUi6fP1se+xrot/1r0VmEasjq72QDpsB/vDKajDcrp/7Ipv9I8eH+Ko4fHYVXJAJ
+          QPK4/XpxNVU/OOfpRrRTW2T8U/F8uMhPE0OxSACfS8cavmvPIZ7NM+P2OCciFPV3jp50umosU2yJ
+          fM/OAVLX25itimUEMmGfLbF7N3CWs08WIDbuDT2eXawNM2En2fslKtL11KOcqoqtrOcfEx+vSKET
+          iqUaioqXYXaPDz4Gt1w+DUyOEi7pKXs+XX2o0PSC0h1fyPXIuXIgqQYxt2zfM3lXoFxc7x4Kf70y
+          siyILGiGj/6ffpjn7Tj84TOKv9sKKM3nCByOxyuxy6oKeR7rKpDm54GgHpcje4++6h/fEWXnd3qd
+          9EA+0mokyY6P+OMKMeyyxCLF+11ry+2txHIUGAva9YVD+1vHgnfpRih3s8ARnMqVjoS7xSiWGWZc
+          Dr/jIkumbpFkIqSZ/t4fiHmAha8GwrYuqxr87vERXZavH1LrImBozI8jBnepd4jz5nOZa6SBhDve
+          fE6ZaMobna6kZIVAW97sY5LLRPKIu+Pb8lhUUZ6c4UCcjfez7T4PFfC9j/4vH5e8snJ552cMq08B
+          lrxpWqCK/MuT1KMFpjf7wHJ/6Qpv7wOP9NtfCpgcoIeUe6Jla3bsbdjx3UCul+9J4/jvq4L386ih
+          KxgAWPNzacHfKb8RU0iptlX46cIj1QN0xwNu5gvxWjh6G7tXtHXAApMO0GEOJ48VBSVbu5eeQLH/
+          VeRyec3hmrNpDv3rFBBjeeXNVouVCpff0ycapxuUSwJXAvie3onL120zACE1ZTO89//4l5dtoYVa
+          OQFMg6EKt4jtF7l8jQNRsuvN4br064M/PLO7O87m1BU96eyjFSkkeIy/y3KDEN8FjKzokIFv3leV
+          zGgWRqryUMD8d733eLmhh1F3zUxZLZCVhtsdtnly+ON9jiFhedlb/bof6f209iBmspoY9NYBuslV
+          IP/TH+wy0k1fVlEKo9VFBmt9KX2aOIJ/etfcohpsUqB6MOreBwxiKxjXk5lNElazkrgofjfbjn/Q
+          Lt8pukgPAaxqEJqwvmY/fGCFwFmpUSbA2YQfcdixCLfjvYth+TbEf/yNqVlv8HLbXpgZ0bdZtOvU
+          A6m4raTMdA0sneH7QN8r4aq0mqHgh3MLu4x/44o6t4xVeseHWnNxMGje5wwjsinghfobXq5z7Cwe
+          8FvZuaT47/fhOvVSAOtcvnnSaENnUtoUw9EWXfyJvrWzwVGq4Us7huhUqKdxej4mD0hp9UAPIQwB
+          OzVcBKbVC/FBCZ7h/DLmHu76H+V1q4fbGzoWpEUZEu0IvmMncpEEuKtyIWenSMelLocKfjycIC0g
+          fYPRybMgb8czuby+krb83c+ASod4imbSZdcTkjskMjHUyXP4ZblXQHP4kGgJeoPtsqQQ7nxEdn03
+          Lni6tZA3C4NY77AKuQKBBLLjo8DyU5NCUudhC1Pm0mIifruR9ggGMEnrCzHLLABrEu57cLSkIMk7
+          VDKeq93hT1+iB1KVhh6TOgBcRJ87/1kZFxWTCxoAK+Ts+oP/uIdY/CXUInq3NHS/nl7m31f3v/lz
+          IV4v7XqQ6KRbxzX1w1i6FaBA3rV5A8yc8xzu+LBPJf84m42HBNClT4jFt69wWcwTA3Y+ItFbMTOW
+          15ZCTtj5iy7JFocLJfkGEyeciXHOnXFebLEAybKf8tXFXNap/reGu74kepbqdPttHwXK3omiNH9p
+          o7CsgwStIX7jac9vWqmjD9fI70ja6y3d43WB0+qGKNzf38q5tgnl138AAAD//6RdydayPLO9IAbS
+          KEmG9NIHBRFngIiASh8gV38Wz/sN/9kZup5GSKrZe1dS9UuJ+ufv5eJsSKOAmSM1a+oNgZsC7dn1
+          ScIcWnU5dJIPyDuZZ5AeA6dF1bMEsJ4Mn+XCR7ZKx2qEwVGV5tEAPKCxtzDQuAJvPvnR6kxJyBdg
+          /QXZ/O1OiUMtziyg5AkcPq/FV6W595AAsUtM1BOrZvzyknPQnKbnPAW3K121CTIQyk43C1FXqWP1
+          ait498oFq3p8ARsmogLO8/GD1cc+l9HN4uSPL8z91V/pn72A/haddv9411N2mm34rjXk/37nn7N+
+          f1dfTB6wmqEzsNn4MUEK/Oo3Yswyx5DMndYiDGzOR+/fO/vjg1Dystw/fa9rtsbHvATGS2mI5Bz8
+          bC4RO4tCvL32z3M2EsAaSIgjgzy0r1UTl/xSMKaLR5JkxeFSQwrhu2ob7P6y1unOu/ax6r8FO8LX
+          o6znxwEEfWhjZbV1dXO8TwHX3zXz0eArtWB2soGqoTeI5kuBw2ew1iDlbhvOb1oXbn5lGXDHF/4p
+          f9bD8pxOCQjlCGF8zcaBSPg4nvb4TbxEUAH3Fg8b+PC/1l+7z92hYScuUHuqL1/gfXXvoXWN4cnq
+          45lK3yibxu8WwM7MZ5JmHFWp+A1deB3uAT43+poNgzykACf2Ae/6CKAvdgsQ5IMaK5tJ1IVZ/AKa
+          +qqR8/7/ud6zGADuYUTUobEHQWDWIzQ46YIztknUrXoULeACccM2/Rh0+aenWK1JjKtSqOSUZjzc
+          +TCR4ngKt4zJW8jVOSZxA2ZnGDaJR515OmN33296eY0QRNfz4osdo9eT8BFT2GmNgV/l5gxtZL5b
+          KOXaBe/8pt4YXorR4McZdjq5Gsbdn+HBNYNZPHdGyL0ycQPqykpYIg9Ed3yQw3N2JES9EqPmxtPh
+          CI1rwfriufuGf3wMui0XE/Nn+uom5UUAD+L9i33/7VAekGcLr/1RwFH/+9WUl4YKXs17SgzXWzJa
+          J0uOACIPYnhOGtLgwkpw0voKn1tupstUtC182d6DKFsMaqKplxQap/vVFw5wcIbYWyCCL/aJlTdU
+          sn94pX/gN9a/xnPYYJ4oInObHX+qPkO4qNmzhLv94lw6KIB9104PpNy4ED983UBbfr8bhPwIseod
+          q70HlGhCcp1OWFMzrebrH/6CC3tysPRCfr0KvwHCsaoLbL092VkRCyHc/Q3LhnOu14e7+Oj5LRb/
+          d+OSgbLtOEJ5TIeZXfxO7R7+e0Fzo9lEBdKDrsPDiv7jU+WVpztfC9C9yTDxustTHUMm3BDIHDgf
+          Q/VXrzh7S2jHA+RP36GM/UhE9AgeOE0XBWz1fTJgGOTOPzy2NfVYAHXmLfx8ch5ds95pge/+DjON
+          mqjeYC8x0GqWcKZYe2Q7HpPgX77Jf5np8Ctj5GjXGwmenTIj11xmYdCYCb6uFydcJvpioftjGiyl
+          pkRp5a4a2pQnIGf7YWWTc8pTOPvak9hOVTo8V2m9+B0/Ed7t39lO90KE4oNZsGU1Yd2+HPco7nhm
+          BjL8UHpNgytKI2vnI2MebpPQmmg4Dhb2EqZ3tvH0mMH+vsTe9Qeaqo4LvCLtZlHZqoG7xpgRy/O5
+          2fMXn9F/fJW0CrbTpaJrdprNP7w0H94LN6zJ1T2K+QKeWGJPD3WUhFcM8O/qYa+gAqWDPCRA2pgF
+          G3PH1Yt9T3O47w9RP3ztjIv1MeAzLESCwzIHy+v8XtDfesY3dh12vSmFKR/cSAaSd7gCpzLRyTvc
+          Z1g1TbgMd46BRS6p+FrL15ptDwcTttKdwdJTOajTy5d5yCynwacJlsHW1VoJ//hpXF5j8E8fPlTf
+          eOb++LX01jYUAu+CQ7kpQvp8BAlKnMuENd5G6jYC5wrf0vlM8IuVQ3LM9eTU/8QY4/326JAdSgVl
+          +ZxifJpvdPrTw1D5uxBzeasOF+ViC/f9JI4x1uF6G58mxGKg+tutRWC1nluF8unp+Vxpd0O/433Q
+          j/hOnHS1APfJUhbO3QX53x+qwXJfYfH3fljKontGj24lQopsk5ztq+rQL1PaqAfRh8g849B25/dw
+          Mg819qzlHC4gwTPY/Y3gF/gN00nVRah3ypHI27l1hudYKODvffzTqwpXnHXKP72WNyO+ps9+LKDP
+          y2BeLi+rptl1GkEoRB2+d2qWbUp7a0UBuQ0JTHSs18hLTLD0SfCPvxOssz0c2NjEOhvz4cwFSyMu
+          zPVIcH5aKYm9IwPVwLmRv3iz2O1cQuXvxDGb1ZTvGkYD1AE80UynCqfb+LTBkpEM/1vvC/eIwRwx
+          cOf3xTCzNe/D5pPk//DVWkiBCBcVOuS85YFDk8cSQSUOJXzuxJezJiGfi10i3bASftRw/WRXVlSu
+          aeKPuJkdYlqHGa6t42FXJ2X2TcIkEC7TN8T+PhuCthlogfcJTWKOPEf/4Qks8SbW4kpWqa2ccrBK
+          bEziV7aBbTLZAPq8CmaG3w7OKNrxBm9OPWDpOIz1dk4yFw6xPmIX3fI6z25iC7wnDEhwHMaB3GdS
+          wp2fzLnOG8PED1gT77KwEUkzIsoNow9Be7qV8z4dHmznJPShd3IW4v3qW7hcZQZC0xpGIvOCPiw1
+          cSRYpxeO7PFb5ZYnY8DL2faJdQdbNn7jMfqLh/iKKaGklNYZja+zj926Y+vVhocKWn7s+QJWpIGb
+          NG+D3udi+rfzTc+2aHVakNhCtPNDY5hPIdinvF72nsRQDEnRikc4DdoPF357Venj1wX/+IEer1FN
+          d30J+jNR/ZPK7wfSGpMVFfQTfBHpiro93pYPwTepZjGW9HAZYi39xzfdm0OGFsdiKa7gk2O5fZRD
+          q3+GCFakmrG76wMbKKwGgg+05o/2huq88wnoZ8Qj9imNKL1nnSju+YGYjVcMdHp6wXHHp//i52Q9
+          xRJctWLb7VehLDPnIuQ/OosNJx73E79JcvodvxNWLfOTjWLXmn/1F6KcLwyYo3dnwvLkttieIal3
+          PdyH+Bd45PWpOnWJ7NCFavhpsO6pkzr1ngyhfvt2fsUfrmDZXluA9nrYTMIyp8ufnsuXG51/cjcA
+          +hCc5e/3if8m/rDzzxEI3XLDBj6o9ZKa8ghbrjhiu2M+9Y6/UzT4UUZ0tT+q5P1i5r98PB/6qzFw
+          +glGcG4MG9sZPGU7XvThwgRH/OTGH3jv+gLgFIx8MvoppafW3eD7lzHY/rqbM96MxfgXr6TLq6vn
+          zBN8eCmeun84o3MmVIVuQnx48cQXm21YePWYQ8+uZGxpE6Hr/v//9LAZjqAOt8dbdpEAN2cWc4qy
+          xbK6EhbbsSLXoa6HlcUfH7hy9SIae67AYqJeA+AKQyIrj8gZdJxsUL/Msi/s+Xkx1mqE8be9/tW3
+          Mlr8rOgvf+w9IrphtZ5iBb2SORMZfrxw/qJ8AYvKODM9CM96ibl784/vLB6bZ396FLyXmo/dVKHh
+          v/Vao5HxufltZv31vLAQjVdMpCwSsqVd8hbs9TBcmFTLWLr6ERTTh0XkuEjrdb4wM0ylScGFtc7h
+          rielcPgMITaNFoI5Y6IePlCw4fPxwjj0+GQYAGWrwzLeWrCKxW+Be/6ZhR1fEu+EXOif2Q9Jr/6F
+          UuSx5V989MuaKM54i68N6sxiJs7dyOvVPespOoZ9T+SuCAA9uv0R7P4ys/080T8+hMbzdv/T52vW
+          BEsF89iM8a4vq9P7J5rQvyfurtcy4Sy+xxlWG2R2fVJRiXkcvmCvp2J/b5Y041nKoVa+CPaTw6ou
+          knCPkLgImPzpdXt+yiFvr8YMFJGApRP7HCxZqPjH9eU53I4nQayhH7bNh+ZsXBlqcDh2FtaPh2kg
+          bRKUSLS5irgwGMPRqS8s8k7WQp5B/croiPMI7HoPxrpVODO5JCV89OSGFflm0/V0qEbxz79MP+yd
+          7VNaEdhY+zuj6Pzd6wHSjBi496xlNi+kMe5LUC9hNQ8mQYAcujiBxuek4j2fq1vg6RUs8jwnyTo9
+          humPn6+/kcX3R/kBv72+Cc7rMSfndITZdJWD7Z8+Yy1Vr7bv8G0iYa6Jf7hfg3Cl+j2BzAIGf9nr
+          oaurh/xffiUm1JJ6in9m/rc/Mw0BcrYbA79g2Zy3L55SFuz1oBQexNvXF3b9eysXdfnjYziAjRGu
+          356LkVqN286Pu2Fp50GD3SxS7LjpoC7MU2LhXv+axZmzsnG5FwoU1K0jahKbzuiu9gZtfNj8vAti
+          R/g9JxeGQtzt9fnSWRp7ZEBHPY4oG+ay8eVox//PiYLT/z5REApHhaRf6aJOAOEIchfI+3x0Ig5l
+          41cFTfzpsGM8f+EyhvcYul2vYP23/sLtZrA8OhaCPXPHSQb8yAMeCO/aJOdLtdWrMD737l3t3nXi
+          eg75e90r0KPogr1A0NRVVD0JXhLpRPCt/+13HPgSBLU24eAXaHRS1TiBrO+5OKnDYVj78mKjSuOP
+          xPCkvN6UYYWovow6LrzyN6zM9ipBBLU3fiLxCaj38lOYfK8n4pc6CMnJ4CoUtvERu98DCcfxZDJi
+          zd9SIp8Dni6a1BRQVUbPZ5/TZ1gr1C1ip4XZrvDI6lLT74bwLHrzltaAdo876GGQ+wHxz9N14Ljb
+          xMJP8ulwchhYZ92eBQ9CQVRmAUVdvb6lskdaA3Ofu3YrXRXj6QM38FqsD7IYLi8tv0IvZnTiSEQK
+          2SGOTWSc3YFYL0icLbwuGorYIvEFFFk1nyabL6Jzas3UuG1grpgyRjzLbESeysoReth80RKGJlHL
+          45cuXhlBSH/qCTsNK6ksv0Y5Ck06z9F3OAyT02klqqXhjTFtsow7F4mPYBJyWFn7dz1uSGhR86ym
+          OVdEPLDueVPgdaww8XT9oK7mgkoYm75MwhX4A/8yLgnUu73iFVtpJkhsV6EjvznYrF8GWLXaTsBJ
+          9BGx9MJy2jG8R3B505l4w1sPWcV8FrDrDtFuf+eM7SVRgpT/qUSrVV7dsvVpwmXTPuQavjR1zF39
+          imAXPbCXeuxAt8+lAft+kCdjnJxZ7r8VHAlySXCN7Hr59XYCVTIl5Nn9aocc1icLrU18YgdTAWzi
+          WUmQAY8NCVrGddgzEF04TY1BvLrdwCLq+hXNCrKwrx+cmied58IVC09id21ab9qLHyFoUo84z7BR
+          twkXPGSLUsWv6tQ7bKcuPPwaw4NI4/c2CEk22hBzEiLXTCodNvq+NBjhysFyqat0mXjFRsypT8hd
+          O0fZVh+r/QymZviHWcH1Yq92Au7qFWHZyvauipIgIjc43Ih9s5qB/s5RAYVQ2k9QiEZInEFuUcvY
+          A5Z/k6fyyrxsyH3cHeI+kxvlDxd+hOilXLHqmmew3H5ViY7xPnduvdzrJcj6DbCT0hHNXIdwOwhl
+          hEqtOmHLZh8D2wgn5WScmQTbqeAM9HKrJGTwkonNhrsPbKy6MSRSoxJN1H2VY7Z7CUxoHXBi1+4g
+          cPcKIuAxX5+PD+Mwfc3YRpT/qH/xKGtjY+/pkAhnfyvabtjquxrBF5NYBDc4Bdu+3yiIvmcsvS8R
+          EPLufgXXYauI6767cPn9nC90bfaDbYojdeHXqIAueYrk/Mh0h32/5Rw9b9mTKIamA1Zt5RyOVz/y
+          +T3+8CxjJdBum5gozncdluxmzbC+zPrcdfep3hbnaMCU8WSfr6Ip3JT5uEBXnpQZQPAaqCEOLWRO
+          bUIsmz3VpIqZAloeDIgq0Z9DRekgwt4qCxzLQ0ZJZPwYqLhTTpxNrR1+OCxXJOS27i/qfZ96cLVF
+          +PzM0d5jQ3eW2Z9saNW9jfPayMHgvYwUXoRGxbeFJdl6Pq0MPHxhhrXNzQYBoYuGupP3Iw9+Gyi5
+          G6YCL7L9xPI5iAEJoriBF17rsGkUDO3itjNgU3fP+aDkIdiGLDVgF7KfmfZKri7NSlO0DK5JJPer
+          DiuTSRKKbi3ANrMEjpAFMEb9wb37MyffBkqm/Y41mTNyPvgO5QWfpugWTILfT1Mb0nyMDVg1iUIC
+          f5RVDh2fNhhCbiIBe74MK5vVORIv6QV7elWHWzP+csjP2o2E361zpscplFA35DP2xN4GK8MVMaxL
+          f/DD8VHRLXMZCByjN/b9/gCuZ9YRDjQRiMQ/+mwTlusGmpstYOUoRrUATwaLhIVm2DkUcsg5oxPD
+          zSvu/km/emD59UoKi0WucXCN+mHc6neMWk42idGelYy/l4MPkP47EmUQxpq+mSyH5cpDHz4Tjq7m
+          LdygcHw0RBd/msN5ygbRKDvLLDJNBNjF2iBKpeOBRGxdZXTIgx6t6OTtZ0YVwEMlcSGrnzt8fmCL
+          0oxER/TwCpeoFaNlQnB/pqC2qIb1s3mj24ceSxiccE40xJSDoJz9I3xkWJ+5a3cBHAIhg/b8SpS8
+          ch32uqURfGRwwzG9cyo178ERcQ5v4AKKpsPppfOF8DIG5HqcT9nGRKiEHyEzsOd+M3W5Ta2J3IOq
+          4XPIOiG3PWMWjQoj+6gwfs4cv+4bYEqPweY4985m2y8N9gdGwzFXlbWAiqVEi2cm+DUrpCZTGLlo
+          cXpmXuwscTbtxcxgfGgLVg/xCKaTpbSCcfYHkphdn42x4ZioGiXHh+7NdwTBITPQ39VGdCt8h8L7
+          gQKYcoKD8W7f/HPsGWi+bOqDcOyzZbq/v+h1m3JyD9tFXVXp9YUiGsh+5wGHuz0WaBGLbD6i9QKW
+          JBdHeLRy0z9g4epQreSuovfJB/xIyrc6hZdTicJLkc/idXtSavt5DmpSICKr+aa278LuxZF5bVje
+          +Ouw/eEtz6t7/9Q512wTLpuLXrwSkpeWaSqVQsVGcnzhyePHvetlTNsEWdvxSV4NFukYJDVEtiW7
+          2PZyL1yq05jAkXluWO7ww1kDxhyRzf2q+TiKn/A37V3NrGyyyI6fslFt5QI9vTX0T0czHzb/lQWg
+          O+HfTJW3r67ZdN/A7fAOSdwobL0l6fQFhXygGF8ejLN2L31E+/cRezrV9ZaT94LoTz7hV+Ze6eY2
+          Xg658pX7XK4F2XpThEKEimPMNEzEYXkr5RGdj8Qjfgo+lB7muwSb4fLFVuds4TrdughmLvmRfb1q
+          lqOYBySeN2KiTzz82Qe6Px4jiWXh7vDLeWLhjof+3l+lV1ZpkCskgNx2+xeiveKr9NTFhpDv/vX4
+          zeCXmQS7rL/Vo6P+Fsjj6xF7joadLWUzCK3eIz77+p3pgiVLAuZLLLFfnw5gRim1EYgEdcdbIFuE
+          +dnDb8wE2D0QPxuvWxrDe58H/m3H1yyUMuPv+4lyeXzrzYzTSOyZvMc+Tju61cd+BBNkHz4f3mu1
+          m69VBfd8SnRdSUKqS10Eb13SEG0wVcoZ3G+E/d8c7zsfZ6v/ziq4DV4w07t7yCbj+C2BM6omsbra
+          pHRpDHs/0aYTXS3HbPQZVAlWeA9nri6Z8J+9EiMDxDG6atj+8mMWmwq5qtNBXW6uuUAvX/c5cluc
+          bVSWgn/5zSYHWSW/lvKAa4XrzOSXuF79A9TgUrFXYt9xpS5ZIlWQj60btnZ7JunxEUNGZV5EbSPO
+          +fNv4Aop8MuP/cvmZiQF2PEitq22HdayZBuIzfxH1MfdBjQaH1/4xxci3XrT7bOsM3Si5PCf/wJ0
+          joEZ9D5R7bTc85PyhVf3oGO/yt/gm2/tV5zGuiLSmIrhKpFPA5vmYZMUmN+aOO5bg/IvSgl+XINh
+          9fTTBtnxCv/8KRttPEjwmr2GP7ybrSENRIjPs0hcDj5qjkry/C/+imVfDpuvriVUZ8zMdI/fyyFy
+          ePFJX95M937Oy82VNijL7ZukjbHfkRwOJiRnq8XnHS8SXzsa0MDfN5ZsnAMuOeIGGM97Siz2aNVc
+          W7kV+MNPL4Frwm285QUs06H1R62N6fqW2na/M8HOx9egONyB7xbot9aC/55/KYZ9zq2T2dhatQbQ
+          u6Qk8CVrZyJ76kC3Q2wyEArsgZyVdb+TY/MKFIRBmB/XaVRJ7nEVstJDTtz+dwRrh+YAvqQoJc/3
+          vNTLW2lF+NOu4f739TCjwyVC30k74LvxmBzKR+M/vEMkO0vUpbAfR8ipeUDyjLfAvl4sdJWc4PBG
+          3uHSV5cU/fGTEAUkXJ7GuUdOdfpg2cJnZ8u38gvNe9UTrd677pqXIoVesU/90142XfZ4DNGoFj7H
+          z2s9yfjR/vFNH0BwGNYAWCIs2uzhs19BpK3EHTVYl0wyMztfnT3jGMNetlvsD19BXesAm2CPV/4x
+          DmpK0sdxgxdz+848qYNwPVlKD+nblclFrjCdGC6OgdCJDPanVRm4xylT4O8E3bnc1hJsqhYxwFer
+          DOsusKlwuVUKuF0rk/icfKvX9Rte4Z7/yOP6DiiVAicGJ3W9E+mXl84KtaCAx8/0w/79/FaXM8c2
+          yPPe/R8+VreZM/P9IE2A7ZPSqR2ULz3y01QguKTusK2/Bw/vbXfYP4917zZ6ga51MpKwjTh1PblW
+          Dpll70Kv1bpKZE8rYX+AGr6wSTGU+stT0NEcM6wkYlXzI09Z+MdXrHPzHfb17gFHyOwLbUJCioql
+          QlDgDwQLXJMtnJiIgFH9DdsyWMPlu5gKZC5TQvSPoDu83pgQMuu18uHnYmZ0QEcF3uRXQ5REVAZO
+          VYsUbu6DYk+PW2cBzRr8xcv5Lc+Duuz85g/fEem2JJSaK3Hh4M2I+MpdUReCly+8t8MBG+Fddcaf
+          OSjwyC8OTqvkN9A//PiX7xRYFc7UqjUDN2pesASISDfAeEdR8yuOnMshcbaH+PiKL9Tn2GNhTadw
+          yq7gIW7YZ0QCB0pfIUSC+kz8w6ke6uUzs18ohEpFzs2iU4oZMQUv3IzkSl77iS0GlWDfT2wcmXnv
+          In5o+ees2XjXG1Ri3hMRvu8V5zNVLtNliqoNXKwqxH/xf/arq4Fe3PWL5cqRQ2EUthQ9Dcv24+R9
+          rLeAPaRw4vwAS6A+1Gt5eNqA/+buXL7uW7ixQGnQ++wx5Gw8JnWC1aqgItw0gg3tGXYH2kPAtdwV
+          x15CVfrOWe0fn/vD9wuTygUaZfAkZ9etHMo+ywY9dC0mwd9+XAQrFZ3NOfqLX+pUsEck/eEJon9e
+          ft2fPEmBm5vR+RbUUUg4Xi3RBLVx/nquQLs8lHMkVec7NjujHNZ3JUlAA/zRH6/qzZkvp70rfLhg
+          bF4OIqDC9bj9w0dayw/qtNhrCSy34H20CLwzNitIQFk1nf9VUTRU2oufgQOaK744VheOvHA1wD0F
+          BvalSzNsaoVK0Jr5hnE1zXTbimcKx7SxcXyye2e5husCj3dmJm7Wz8MkvqcGYtgZ87J99Jr7088k
+          d4z+6Sl85G45umo/TCTV+agzFzcp0q3Pinf7Upd6+UF4Xm53Iu35cwmyaoH3vgj8BlDOoUyyiShX
+          cx373+A7UD/HDUTj6vzT67bjFEIYj+fOV+/uIaSqmUbinz7jO1uZ0XvTfuG9Ys9YJqmWcatGj6hY
+          1JqYF3qox6XNKtFGhwS7D3geloS5asiEpxFrY8EOv+MQjv+e32YRyEavzjaI2TAgTts2gHWsKf3D
+          1/imjrseyMUxBP1h2/WHoS5+70MPytNL/+Pr9ZqvlYIeuhFjo4q8bMcXJdrzDX6mXjQQsN/g2PWN
+          GXU/1eHQg23Arg+S8O6+Mq41+g38yve287PG2S736gr3/OF/nsy1niQyNX98fAb3iIJFHmkPNP1x
+          JSp3z+k63d4x+OPnxnZg6ikPrRxIonInzmUz1VVVi+Sffrrjs3AJx3wGK1ffCX4YXLYEyQAhZKZs
+          53sVWMVmTCGWjHXXx2JnVK32iBZtqebDfdHqnc/O8HX3eezJz9t/+sUff4p7WRuotW0tRPXQYj9/
+          Xih9Tl0Jr964kWTP9wsSn//58599z0bmVjC1pQpfmLs/kOamaujqIp1okTGq2+8t9P/0nTiZQrD5
+          kjgDXejefrbra1PDLRva8YfP1tlYT07nlvAmPxtiBT9/WPNAadEeb/23lSkhX3ynCE7Dyv3TW4ni
+          PGxY9u6GLfjQh/n0i2OIzeJHFMH1Mm4ZkyO8MfTr872s1SvPPXhIuBJjt/8ldGs7c4a69VuJFd+4
+          emK2VwUf/TMhyRKHw8LfthFsN+M9s9/hVdO/7xtCYSKyPA/O/nw9JNPPJ+rdfYWj6d+/sP6xLH5+
+          Ki5c3DHxwefHQYK/77Beb3uPxfecXYilhqiejjajADybJXbn4xtMn0v6FXNlToiUAU0VDvNdgapj
+          qPjsSl/QpsEhB+a97LFRpo5D/Pzc/PGdv3wXUoQeBsDf8k6euv5yNumqNpBbSwnradWEbO7qAch/
+          94k44diHVO1OI6Ct9sTJ+9Fla1wNBbRO22le2eYBdn2dgTQANna0+uMMt2/GgEqsVh98ZyVb52tf
+          wdVE8h/fcHb95gvq0h2wdFuOdMaPZw9u7y0jhufe6fwWYwNsOuvjJPuy2TzANYc/LQj3GyFD/cfn
+          wAGvZxLs+G0TotN+julz9pmDpA+b+XiY8PB5yuQxPB1ng0rgwml8V8Tb1iQjaxuk6MgkkX98DZVK
+          j/jEgPzRsD4YNeqMNZ0XGCzClyjWiMPZl7ZR7BxHmJvOl+t115Mgz8KN7PijZhtWdKG6KAHenx9s
+          R4f9glrOfGKOs60uv8ly//I7UW5yoM7nm3GE9zWasCmbc7iq0v0Lt0pIiR61ispu39YUYfsdsV/U
+          ONz6O+ZhZCTtvDCGXHMjOfQAgGvnnzynzDjx/flCNrpBoleqlM1/evps/65YhvqNtvUjKNCfXr3j
+          DzCXatpAmYwBfp3uQbh4GnVFfjZu5Fy+crr8nI4Hlv5N/UOTxeqyWCKEv7D6+sddj/2rT/zTv3xJ
+          77IZncIZvV/HgSjq9FLXlD4TqF5cir1Pdct2fdEEDTvKWLl4Q7b+Tf3IHcPHGltXIV1+LgNdmSjE
+          Lgs9WxUrkmDWtC0Jdr65HLjBh1+je2DtrDbOKjk4heFj53tl8Qlnhs9FGDjggeXrA4ZzxbQx+MNH
+          j1XTaOH47xjG4A6JxHtEnf70v2fULX/1gpBt9caHpi/Af/x3y0m3gK6fZGwnnyudgvwdQZhcOOwV
+          xk/9ep/giPb4Nh9G0wvZfGp6KIQb9imoWodu1XWDez3En+RLmm1rn7SwjVFCZOJ3zmCPnATz4HXB
+          5+XyplvHNDziHOPhr52KaLt+8hle04kh7rhp+41LtwTJT/n6x/enGWgPxy/Mf7cJ7/kl3LDIKfBf
+          fHgVYUj7w5DC0nUUvNejnLUOzjYSOuey85GezgwfifBsD0+fg29jYD/ezQWr9HOIj1OLrmud5yAk
+          +E8f/qi7PhIDKvfCf/W04SS2kJyddl6dm1Z/Pn62gD99+vEugLNUmcrC9bcUWOXuENDn0kbw5Iri
+          DE6vSzh8XKP/209foCdHHafLAIFqxzHROv9dr+kSN+If3hdaN3Lmu2L38OQ6nn9iS6qOFT814CS6
+          aGb2et2867dA5puJ/OHHuSxgBB4Zs2EVBTjcYP6t4Ob7mi/s/G996SkLt3zsyflluQ770qIAKguA
+          +Awsxllbo1rQ4T6bRNXU1BnP91yBQ/RJcfScPvUSN8P4Zy/+8a9+F1d1Dq9iaGLbEeVMmFjawNyO
+          3z5rD94gTMEzBZX+/fjfytyc7SC0EXzdSE6UM2M685+9DY148a+r98yoJ31Y1HIbR870NKhLlXgK
+          vOp7DcTOjs5f/v/Lx7veINVrW2klOm0Ni/Hs93QpvGAEez1yPlC4ApKFwIRpYF39o3sJ6co/FgOZ
+          SXDH6q6Pbat5ugJv4t97fYNVp9AXR7gG5g1rkeGqfB1pJdr9ZxYd8Z3RK2s30PL8mvzp45u6n6Ib
+          5dMTywsMMl7C7nba68V+vdvfkl2VIzLinMFWlIXZonx5CQVqXfuFlZ3AaF6KBM5OdCDB+xLt9QsC
+          QeNh3q8yqVT/4cU/fLvHF2cERjBDobpyRNHseiA7n4BX7YOxtQalypWHmw1TszaJPJlDTatQaAEp
+          YkDsqlHoOhyhAaRuKrHOx0Cd//DHLIAbsfyWydqf07Fid3yJWKszd5jVc7bBVtKgD7hhHLpsem1A
+          UF+JfzonQ0gpYkvUYOWMvV+Qh3P7lY/w6nrJTLCw/eNrYqlSl2ClJequJ44ipxYB1p/MdVgOkcrC
+          X2YTXyzCE1hM//WFjefx2F7fk1NH46MBez0HRz006/lDlwqFlzwnj3PiZH96OnzorIqfPP9V6fo+
+          R+BbNHvPUUOuF54Nr4i2vE2wFMfhcrgrPeLWSiJel/0A9xlhDMphdnf9aRlWxU1ZMXOnvYcmLw1r
+          Sm/J/+dEgfi/TxQkS9sTrf109XpbpRLB/PXA6revB9prGQ86z2KJr/UHdUHCckXb6Lj+d5DWeobg
+          5yMeu0dyZeav09dHLgdZIZ3Jc/IvlGXZaoTKV5uwDC8+FWZTHlH9+VRYFRHe70y5LJSfd3E+Dup7
+          mBx/GqGSnSSfdTZP3YL16YN7bZ1xnDqsQ02lmmHjpfwsjFKTtcxHTKDFxRhjNlWdTbsXG+R5D2IZ
+          Vr+BHgYrhvFoq8Tf/Cakt7yz0ShfNJyvHz/jHudTAefzcfEPZd6EI0gG5e/z/nxkoCEKeZi8R2+m
+          i/7LVj1HM4xesoDdIXTA1gzeBmcya8SKzTITlEAcIdMfU6xbjAP4/vu6wk9SyiT/nS6U3IXSR4Jk
+          3sn10J3C1QFVj5zLgnFyeQJ1Pd2vV8RVLUuSgYkALT62hJoHBCRI7uVAf1Mvnk5jcSG4XC81e3bU
+          AmFl5gguWJjNvW8yoHHvoc9q6tdh77NjINOfK2IcwudAM9tm4OkafLDunSd1lfzOhQfy1v0e5e2w
+          3Y12hFfLzXHxjVKHPZ71AiG/SDFWE+JslXw5os9gDVgfUKRSW/3xcMlulLiX8ysk2uXAAuMuhyTb
+          qgAs01mCKJWFlChHZ3A4KbFnBCopw3p1EcBqzmGALInhiT3CJFyQEkPIMmE69/qF1jSBoILbaLk4
+          7OfE4XlLuML38RiQs9a7lCbHKQYu4/v+GkmaI7z1qEDHo3acGenbD2s2pgoUyXImWTZ2wxI8dRZG
+          J9wQIz+YKmd5ZYO6s4HJzTt7zkiPyxFxsUGw3rw+gAZVvsDH2pjkqV/ZbM1vEgt7O9F9imt12JbD
+          UUR49mJckKFT2eq2Bsgm1CMebvhh+j7CEm09eyeOs71CUnwUBaUyl+Kgz4uBjbxl7886eMQS3tUg
+          YO6+wcpyKiL7HbvPMQ9mpOt8TbDxXrKZNTcbqjlbkFdAr2A+tzCGjnMhxNJPLZ3yoopg+Vz1ubQi
+          Th15/OaRIx73ObuWoK4X6TYDM7v+5u+DN8B0s+Ij4h4rxJmX/sJNFrQYBX2ikWecv1QWIbVBvMvu
+          d0DbRW1vkiyht86NBLNkCckd8180KIQhhqcjuhlrMoIMefzMQOw7XKHaDCx8GeNY/5Fwc4utRGyZ
+          vIla5jfAno9SisJPm5P0ijAVpld6hEwvpjjWj61DB1EMoBRHAs7vGhtO1e3Toyw6T774vbX1Ij/C
+          AkWyzfiiJYlgPXaBiaKWufqC/3sDmkBaId10AbbipzewV3nOQfmkOvFJuoT0lU89BN/VwQp3F9Sl
+          XJNeFI4+wtaEy3A5nr0cnvS5JG7zShxOt8QZ+N138H+w6VSKjHYD98PkEHPozoPAHFcFuTdDxx7G
+          XbYZkv6FF9jf5/XybOmSWoqPbkVfkrNAX2D9YVGCtZNx82lqpmGzvIiBkq4o2FJsedjMTOvROZ/e
+          OHgcH4CWoWvvXeO9mYr3buDDs6XAXMzPWE76ol6004dHWIqfWGmqUz0ahV+Awrq+cOFfv2B766WB
+          MmU7YlV4yQ69Hs0K/L2fDKvzILy8IwvJhnviS1IzdH/xyfwUEbbcdc1WrxsXEXUQYoeDNlhisG3w
+          vJkFDuNsVLe41n1xSK8SMTg5C9dVlkcQnOL3fFhBnk2b8ZRgcjNNLBfHMVtpdyxQXr4n7Kooyvrs
+          +fjCLx/wJFTMUuX6p5wiP+19/xAvZ7rEPuJF4Xs0SaiqNFz7xU3hDY8quRxpma34/j3CtMgKjOen
+          B4jEgS9U09wg2tgV6uJEWosGLC1E57pl2NZN9pF703SipJtZs/1BY8VX/8T4IjhJxiF4YMA98nti
+          lSRQt1bmR0BPLcHFXz7UuVcLN/2R4RebquoevxioI2DNydEb965mgQt1q16wsrGA9q98vSI/bX1y
+          te66IwyS14rErTOi1jweJtFbNmQO9Qn7RaDXwlZsDZyOMJwX9xlny31WNcjfFY9I6NU7NOmWEtlT
+          yhH5VWwZaV/UR07zOGHcOM7APdPIRG6jVDOVhXPIHbwkR9OY+PtcOhkIA98YIMriHqsPLDk0e3oQ
+          4CfUiNk2mcPerL6E7M26EJdTc8Bx318ClvwRY31QJMBHzNhA+Bs7HCXxplLrzEh/+RkrQ+xlrHLT
+          r6hYlAve03lNrwZKTgGjX3A6eau6HiwlQPxd8ojj38SaTrc1Qe3t7JN//peU2d7FE/g+K7gnZ26P
+          NxuunG3h4NtVlB+woSFPEk7Yq71WnS7wnMNrYp+IHn3AMBKxiGCA5Bk/4+kOhN7sFfhVDhK5sc9m
+          GBumldBAUxv77QVnPJshG/7FuyddXco3eQIhYJoXid6T7HC89E6hAXNMEm8OAQuwXCELtxJ56kcx
+          nLzmUoD5tvb7+gF1tN9sjILw9MHO5zxlWxwFKbRHTcKp85Iztj+4e1U6zfCNSwSw1bQ30dd3MUmR
+          g+nmq5kJ77f3yz9ZbaDSp/z9wtNBB1hVnlO2vb7pBsXtsfkiH56zZVbEBfKP5EOC7j1k7etgxsj4
+          oH1YAZzB+PI4FyVUVkggB2W9JecPi1Lep7PQuUs46cmvAu65eGJn6M1w+0qnBmWCfSF/9rExJWYB
+          IxQvn831U0hxf7Jhwt4EfHaKctjUbGTgq3uMxL6sNR0e7MuGBsP+iHarGHU5yx2ETJSB+T0w1dC+
+          t9sCzfyTz6KZRnRCs2NCttpOWDUJAXP2cBIIU13GWFVlwOPrGUKZUYz59NZQRmJ/zeE0pr5/SG2Z
+          jtAGEdjtCccGBfudM8hCP0MqcZ+yVQtBlW8wpWxGMuhP2XKVvwWqmfxBFOkLne15TFr4+MZHbFhy
+          F86wTo7oBJKZPGVWB2zq30p4P7efGTjbIVtJ3mniVp4Qdn9NM6xDQl3YVc2KrdiUwlHGJwhV+YOx
+          PrQGWNqX0ECnHEecaoM3rElzY4CDrgt2TvrobHs8F5v7t/YPTXkE1EesDaXDUyJe7ZmOUNOigklj
+          fIjdP5BKveDgw21yCTmPt0s9X+qpOunwGxE1YX4ZSY6fWDxvdkGkLxOoiyBv9t/6kFhkgpCTnrEC
+          /+KbF+cvZ9ntWXxJl3peX9bHobcf3v4+E2XHD/QuyxsCd3bG2LppIXl5Cw8n43Al58/8pnSwfjZ4
+          PjqeJNIpzrZvCBsgOecfOZtmFi5vvoZ/+XHmCgGBpc60Bh5mN5gPhR9nm3exWfHxWCgxP8ZY/9kb
+          lEL4msdf09TjWKQ2jIFR+utl451ZqE8+2PMRsSyEMtJ4SQV/cmngS/S1HeFu3AzI1eMd58ygDUu2
+          LgxSPVEm+ic9OUQ6PA24cHWJ3XsvhvRzP9gwO1Ud0TMzHmhzrVmoBfbNn2rqO+tkcJFofA6Lf3i2
+          S0jXNtdAS64iMc/5Gm6RKjCwuPENsTt6Vxf12I7AWLGEVQsAQJU08qGSAWm+1zhztoK1cxBoB5MY
+          Yqc6HPb1BAjBoyPKIXJV7vkiOVyyO5237Hyjc/e5mhA/GW1mP7kZjm06BvAn+BExhPJNl90+gfu5
+          34kaeEy2gCcfgebe1NhdHkq441sRzm9FxNbUTDUlTGvDq+Xn2Jhql9Im96K/eEuM0uxCqlfJFXr3
+          KCH5E0K6xnzGw6vzZfHf+i1ed0zhSY1NohzLY7Zy/kODIKuBX12yx0BD5ecC/Zq2M8cldzCaKJxB
+          eGBNPxDjFqx+6Vwh5wDFX3Cl0OWRez48is+A7HgvGwXDjI+2LEbE+0IcruacXWEuFud5G2RJ7U4l
+          XWDLeHf8KEngCM+jrsA/fMhkUx+SSmhG9Dv/BvLHD0lp9b2Yl/WEpfutqLc9voOIK+8k7KhGKbck
+          NjJMjcf6eXmC1dDHCqYBFYjbTx9KH7nJgz2e+cuN/YSEnrsEwO9gzKsvTsMyK9sC7dCziBR+hYxw
+          B42BsOBW4k3+Smc2Q+Yf/vazZpOcNS/6CF5qWs5bgKeagueJAYXTCtg/dD+wKl15hWqX1CTszatK
+          /WPJitffOSZq/7DpaPXBCAMwn3zxstY7GPvxoC75bWbtr+ywLNuP//yFfdkful7guYB7fsFery/1
+          +jKLK2TvdefT+q1kU946PLTazSZG2B7BOEbHFk5HJsRyw73r3+ueRtCNUxFrouQ5m2zjBDY318Em
+          pzkZFcWiEPefE4s5XNR//ipk/ZH4fH2pV86/GGg83XsiN5w8LD57l/a5nj7J3WcDiMbABQLm+8JW
+          n43hDAHZp/YcY+ITXVU5yX+74H0xIMarZKnrGrkzyEbY4Pwpd/VW9CWDvEz++Gtn+4D90wfu8UfB
+          WWaoNX+M5QqNQ7H4H9VQM0GIjgtkIfPBym6v6yOer6BYb8bMPWG/z22vfWTlvu4L1yHOdntPwNgZ
+          CZZM01I5LmxSKL8LBeteooPplJU5XFqUkeTwM0L6nIcS7nwIn8VfSMfnC2giaBKKpRF8svX9oCxa
+          OdMitiWJdNFOEw+/U8YQXR27cNz3D1zpkO3x+eCs9pUeYWu+LfzY8cAeT5m/fDUDX5rCEeilhP7e
+          17BkK5y7bOlRoPYh8S/LLVsEQ4rQX3xyAlYOF6C3CjzdDhs2VWafQmKTFg5un8zrvt+LqRADHs7T
+          io3HZ3bG5RfGcHGMt88UwafehjC6IuD+Cl9EeVtvpwuIRc89mzMfni71VihC9Bf/fT48rfW6RtoI
+          fZwwBA/RWG+2+PhPT5AHRqnXd32LIRXHKwlC6ZItUH6wYOeH+PJZIsq7tTlCDfQ8duicU3oMqi/k
+          6+ZK0grMA7WXcYHN9y3NzI6f2wykBdBCrsKqwPoh37AkgH96g2X4vDodu30qBxzKmfPDh9rt096h
+          z5sLkeRAGnjzdoqAt7EixndZqod1RV/xJajRLJ4mORSQ0S5QD6QAX2h+cYS/fDJVL5ao4e+oTofr
+          sIk7HsO6jbVhCfubC33eXrD2h5d18fuFcyx42PD0J10vPDTh/UCc/cSo4yyf5RjBwj2M2Lqr4UC5
+          0WIgjTyKje/QZBvjuSxg9Fcxn0Jxy7rwQgOEFKclmrNN6gqFsQefB8uT4rUWdCN3xfzHn+HZ4OoN
+          pWz1p0f4PbUNVTgMVgTl9FAR0zu4Q0uVaw7BnZ+xZEU3Z1kfdY9eNDpjyXzIO383U/EnVwYuEkWg
+          3+NrkERzeJ+wa6k9pQm9Sugvf5yjU/O3nzzcnw+78ypmVD1fXNQ4SkFMrl/D+fLbMZwxaNg93A7O
+          xD29IzzmTYyjMQrosvyyCHhLURFDKWaV/uFLS735JHnfP9nyiE//R9qV7C8LI8EH4iACkuSIbLJJ
+          EBDxBogoiMiSAHn6+fH/5ji3eQAXSLq6qjrplmT8K1r8x3+X+UF8aFLJwf5OrupVbewEFtu93s2f
+          qVfdenFIU4IMa2NSaewLb+OfH4EDdbmDTY9kaMv/2NqE40L6dIWlk6jk9RqrXHgZnQmuql7gwtYc
+          Nr3NWUB5qZ6wqdoKW+RiauCfH+rnpyYn4koUsErTBZ/xnkbL+dfMMIwym5qazmuTXHxaaID3gnEy
+          f9m8+RUQPRdMdvDMgamTqhUF3OniV+KbZ1PDrxJC4zXG6QO5uRi3ao+GzpLp+ch/QE9DBhE2jPM2
+          xYFFm/7KQJrjCfujktTiOdh5MHW4HPs3rtbm3Wv04CIujGrf4FfT29GeIRQR+PPParbpRfgefZOg
+          FBhMcOSLjBx4C//hDaVcZSG713dkjADTZs8sOPCC3wfVr+eAsZolb6B+rQBbqZJE9BJ/TIQkcMfG
+          Azo14aoTD+5La1Gtt1aXORrlYcqFF8JqgdbL7C8pEo9Pzxf3ozKwL3wSuOEndjf/ipjCqAIGjydq
+          ZMzO+biYKnipzGjjs4Y7O1/cg3AHz/iodG7e1bnXwmt4I9ROHtOw7vWB+6ef8OPwc+ccZAXcyeuV
+          HDKPB6MYtByYb31LtSJ95WPKkgz8+eexxB816gyXAsZ8yWPrTBhb+1tfQrjICbZ2xypil5SFsBLh
+          ii1k7UArnycLVkd8xGpVflz2C90K3D+qRs/1+xoJmx4BwbgW1PTmoF4IHc1//FbhBsFlVAPVn59C
+          PdNVo9kszQKCu1sT9lBzsPz6RUfwXEHCDfDFBPtSW3B/s0J6pWOeTxu/OLjRScfHT/4dlvqEOCgI
+          GFK3bEJ3kshQAHz5hfh29S/DeIYgBSdbiSk+TUU+q+4wgw3v8DFWGnchTA7+8pf/ecC+/u3OaQny
+          IK1ooakvd2zHtwM3fvLnH7t0XB8SjPmC/9Ob+fiZ5wRqT18g+9tuH/3TS9eOP5NEpS3oUpakcPnJ
+          B+qeSA/Wcf1V8rZf/LzEOpif1kuGISf3WOlPS7Rqyb2T64pfsXLljc0/cXWw6SF/3fj1PIw2D3X9
+          YBAmd+9oIxsO2Px3nzuYHhMK/JPAd1K+VFHdCIifm+jAP/30eJYKW/a3tYHQ8040Cs5ltAr2LpAV
+          1/hS/efiWvBK+Q2zh/6jfve2tPXyZQTaChTwGT3siJV5VwGmwRfGVaFHvIRqGdxQzKi1xTOf2Y4P
+          RcW5/b3vaG34bwg++1vqI/32conNVSZqn7Xug00/zmXoptsNNoadHxO1Ndt/JViM/JP6pTDU5JyN
+          Mah+yhXfud3ikljbcWCrn/ifQODzNXw+E/DnT2rJfIiW62K9kRQO8J/fS/izFcPGnwasH/epNtUH
+          W//TC9hSoiAfJy8OoHioV7KPP3nN2FMk0DU8h/7pnQWkzxQ+ovZC//RwZzzsBHaNNpBuAY67rpDy
+          YBEZo7ippM3P7hOolpcUWx6r6j8+DKIWXbCx5x6MvcVxhN7cfqnjd++BkYSr4OYP4mDe3dl8MWAP
+          oeefqMdlA5tfk1MAIZsyAs9pqtE/vXA4adc/PTCw0/PkISyaPnVl5Rktp/VXwMMVrVjzyptL53ku
+          /j2Pc292w/pGbi9/RS+mRgQq8JcfYT4cjvjv83O2q3p01e4XrGYvzNi2v6Dnnc//9OdqW3YHYfG4
+          E1E+TmAdoiKEt9jrt/oAV09/9abNz6NKQtN62fxBOP70FBeV0wP2cugKP6r2pn/+4nzg1QxSS/Cx
+          9hf/1Tv1wW4WAT2LrgmoZwYNwqLu46z2vtH68V8B7NNgR30/bNlMuaZCMvod8cbXB4b3zxkIa0Xp
+          5R2a+SoIvAdyeV/SY6gRwKKD0UP3rb0IF+xVd5W5OTiM2lAQSTkI//gIMN9BTWa18Ib5W67lH5+i
+          ipxYjB4ulPzza05J9qjnIN31YODJjnqdv4vm6WRx8mNnrhQDacnHa+Qp8N7RALu5qQ0zf+BUWcw7
+          yZ/TW1Vv/rgFRmojIgzKMizNcEn//HdstccnYLqmKWirR5IdN36iefiCHjAuf2L/dW/Z/LNICipf
+          6PzPplfZMvYh9F69g43yHdSTkdI3fNv2259HYOT7qOFCKGaJRt6bHv3Daxh1Ck8f9yDS1tMzFEAy
+          WhrZL/IvH23aKGiLtw3f9//yHyyoL1JdkXX258fBjx/Hf/l5mP70tMI5Ljb4h14LzUcuYWeLnA9e
+          nOeuv5cL4Y7Whv/a78ZoUdlxRbz19UhtXI7571WqKjy23orPv2dfs4/txPBzDhd8ml6oXjVvtv75
+          e+c2saMZsbeOjg4JpsR7imDe6j1g9zqUWAP1NqVjOIyw+qnXv3pBTZELdWj97BdOHjKIZsbpHnom
+          s0pT57PW9E8v9spzIKBZK20KlopH1044Y7xsJ27qXulh+3BGrP+Waz3r/UEBcWDk1OburrvutDqB
+          /WzGhKmPKeJzEJZIF2STGucp0sQI5TykPmAEzjsrWmPLtP7VFy5x22tDmVfvP7+EWlL5if7xoe15
+          sTMeDXceY6mHLc8d/bn9qWDdWVwPS7cX//ld1MOyDq/gIWITrXpUg5mHQGXkiv/qqfQ6PCDc/ITt
+          +42Il6SggbPn6NhXFH0QP7tuBiYssR9tfu3MfdbsDz98oSA4WpP67AHU9Sd/55WfmvpSJSDry4+0
+          fH9WNm56DXZQfuMTcqhLUJMocJ/tn3/7fZjPkGVIP4QlPTofJWI382HCprE6bGz6du/nSIeVz3dU
+          3cmhW298/F8++8OvX3ye9P/nRAH43ycK+OtXJzOsv4Cmcjuj3yLC7arEI1+vgtaCFX5+2JC8ji2e
+          fvdhg3cy6YXpWq/h7vJGnumJNFmzmv1k6dZCKxVlalh7vl503+7gsMt5n1eufbR8Ss1CnnBefL4L
+          bbCM6mOU4yjQsYImPh9Kz44BL4o2Pt9vJ01cpEaBWqB12Itm6LK2NBpouh6j6qfdzkh9fxasnzvF
+          Z8r9mPMTZg6yzvfSR0KpD8z8ZR1QdPFN8eGM6lEpKIF+TQpfwBrnzqePUcHdJPfYjW8vbdl9iACF
+          rSJ/ZdrRZedlLpGwv6tYcR/KsOdb1UReV0B847vzsERdTaA4M8dny7PLRb71KvBpe4hPh/k1MCJ1
+          Gbx4C+fzptrUTL8pK1qvZKXuNh11ngYqwN68evhUGZdo+cwkhGaoPOh1p9F6+3yKFAEaNFij3bBu
+          HX1goA0qYeAsaOzu1QLyWqvGR9sqhvb3tnz40+0cn0evccV3LGVowpFHMZqHvL9cIIF9xhd+9aNc
+          NKdZYKJH2Df0dLzc8zmwTR7Wqu/5O30sgNC/ryY6KqaFjwzQnN20oYK3a3zCt+SeA7q2jxUEVsRT
+          Txwtd+Fp1KGPffzRbF/b2jy+jym6ympEvXfzGQQZ3lWUhbLso9w8gGWacQolI8hoSoSunuVUi+Gv
+          vVfYQ5mgMT8Jeuhl4wlfdf7lCir5OvB6W33C3qZUT6phJ/CyFyciDbvjwDP8eSPgIkydcXcCvPY5
+          yjCwLjw9wa9VLw4IePhweZlm59aueaWgIxzNc0mvTHu5s2FzEIzHZsQZu1j5niN1hb64S2h4V8Gw
+          cCLSIZeTA0HyU9JmwxY4lK5cjE+eeWL717XQ4bW4K9QCOcqnYpYCyKmtiI07urm0cmfz33o/v1U8
+          CASFKeyEtKVxJV3d/RDXHLStTqfPvGpd5uiNjOB8N+kpu2n5PkqUBv6eX4Fad2Jqo14YHnxOu4v/
+          cF6GJlzrqoB7GBrUvgyxxo63UEVY6y3q7YZaW+lrbWB2eo0Y66dkWKLEatA6vnNcpDGql8vZCtFe
+          8270chkfTMzKjke5mMz+++3gmljNi0eV2zRUExeLzWv69lFv3jyqBK0V8eGkCLAcWEW9IjRcoQ2K
+          EkrREePsGYXRWo3lDO/3vYu9w/09iC+VZXDKaEdVPmujObNKAqj8anESuGFEJPgO0N0zE5z/qpu2
+          0kF30BStD7L4+4Sxx7wShIvZob4dHQcx144K8i/1B+O8Ml1xfjxKePf0hObX13nY+0++B4cDCej5
+          ZrQRrdW3A6PxA7FKhYYtp1vCQ1l87HCemIxRsN2JRrGcUC+OHFfwmk8Jr729p5a6dZWWWDrCZfxp
+          1E9tbRA4GPfod/YxgQeuAyzTvtYBhoyRXQlZtAZT8Ya4r3bUZvJnWNbHcUbHn+dQN/u9hrUakxUm
+          kZMQ9K78Ycah9kbPr2r6PJ7DfHl9dz48LUgjh8R/AeFxl1Rw/bw0bNUFYyw831q0xTe+Ofdzvjph
+          3kJvaXb4ou148O/377dqxUbA42G0RPUNXeudYD1r+XyuxjhGHa+f6f0JaT3n9SdEjoNuVD1+9W0u
+          bd9J7ZP02JiSFiy5TDlwmIcCG1Gvsnn/2qWQeqJB9Vb5aUve6Tz44n7rknioteW8OhIknuBiK5PW
+          YfkcdgKshmHFxrhUw9oPpIfcIfFouli/SPBzp4fyjzr+VywVxtdBHcKb5+nULYAa8d+hapHa7g7U
+          LQ0fjCwVTZhXrxNO2ygDy1M+l/Cy30/Uxx9VW6w1rZB1kSk57K4T+8VPqsJiAD7ZDz8v4gUGU1CB
+          KsCZ72IgPLlJhX/4Z97jAxiF1z2EL/dT4subedsd6qmAJl0uOH3mVUS29UCKWtsEFYzPV8dWVcit
+          3wWf6qpn9LY5MJqNOepznA9Ew9dNKBb+QLX1F4OFSUECtv+LVe9hMv74iRywZreG2hfzzOb1BWZ4
+          L5+I+r9L6870NVtopxaY3sl5yMlvuKzImm4Otpvq4vKxcLSQVOgaxqkx1vOMfxKy8Rfj4xcoYO9U
+          Nwd+hgvB+vp5gOlZpiu0F/HqC7lSDeJrTj041Ls7YXQ9g/31qciQK7UQn8AzzkV2bFKwU0uM067p
+          tLkOhgDeSGFjJf4CsPSt7EO41yQi3w6Pmr3j7ST9lfNxAOsv67rAmuFDPKt+mzxWtoCjq0N6iRPq
+          29KXjcXlqUJoqBx2brGYk2tdlSjnLgCnUflgwsWqLaTezzrWznURdZ5QKehs6pjG4rQA2j1gDFS3
+          wPihXB6uQNE3RJa1jjSN29BdjPU3I68WHOqAvT3s+8MjgbOT6TSv4JXRcOl54B1bgyrW95wLZ9vo
+          4LWSL1Tx49Mg0Nfaotp9zDSuIsKY8qtCtEbpnbDq9mFre29VmERWQi/TzwdrVRs+nI+3ANtf4Axr
+          kg0rSG+/DF/NOgKdJk4Kep2B+I8frM99XkDB35ovwuBXTzt7ddDffjLUt5Lv5zhJoPH4Vvg8JfPA
+          zOpogr98ey3wjbH3N3+jBR9nagbums+mmM8QO5+IprZrMx4xq0PW7n0kz2/F13PQWJVcKlcDG/PK
+          a0u4VOlfvGBs9U2++OdLBR9B7GJnJzy0tazLEjqX5Yh9/Hm78y3jPPBC45HGTsdF9HjLVIhOPqV2
+          xoN6fu4vCtL5LPjjcxoLunf1F++EY02mke/ptoKlUy74+jQGdw5DbIH1ois42fbj+KqfnbztZxpe
+          W5L/RHoLDkplFf7BRa07/4Jrf1AnEuPT8OI2hWR1f/nIZxZS670Y9e0ff6HxzTm5P7PK38iXK4Ni
+          7noZeLepFWRNVwf7++CojStOO7TtT3qTr18wf4eugXF9DambM+KO18krIHjPB4y94QYWvRpnsOED
+          qfIH1X7DF+vwULU+WRJ+YcRVjBg+gMhjD1jXfGbbEcrc7jC+i8nKxnSXJNB/nl2yCuSZL+3hTuAB
+          Kj9fKhQT8Kq4TV26tg96PqeeJnC+KMCi4TTCnPGab/ljhnG9ObyGnNTrAe4c+ae7OT0OkM/nW4Nn
+          uOExfaSrwsTv7LTQnEqJGu7wBatGbAXS40MiB2M8uB30uzfY8gE5nHul5uNdSJBsWCN2nlJar2YD
+          TPA4+RY967gZ6NpeVyB+fyea6eV+IN4kqkBVTio1mvUzUPpQGlSuX4fct++fCRp5GKZqQ3Tsju4i
+          OpeNH0oyPYWIDExwmhFFWjr63FV8s1FKixbWTXnBrnOSwNq/HzrYs/OFnuEiARYuvQA79zT+y/98
+          KIQpiD79gs+HqzuwXLNV+KpctOWDtzZuzwddR6D4L5+I48F8//FreupvFzY2J9KCo5q+/CY+ZNoC
+          I8sEWrtq2N7Jsdbp9lsHyrvYU639mdHacM4bkFfaU8O4fyO6tIoE//Zv0YXRIHpPUMLnIZ/8Qxhk
+          OXtqei9veECNcVEG/qz4FYBK6uEtHzDhYWEJbPnEh9SWc5ospIAqeNj4hDzCSBsUBfw1xokqD4rB
+          xLeqDnUq76j6FX9/+b6SSfZV8DHB7rA+aruF23phi2+Eel4wN4PDW/5iL3dmtl7HVUE6RpH/W35p
+          zebYleEfH7wuNzmi6bV7w28nfn2eNE4+C69LIAun8YW1P/218I4CFYEz6FnsFne5L98OhiIXUC0z
+          Pu6GFwVAh4LHp7i33V7P5RU+b27nAwnsonlNex8iu7Z8E2ulu0SJ0gLg7jA9lsFTow/tIcPchR01
+          pUlyqeF7OlCaxcdR1ru58H3fS3ipS9Mnk7gDJN/vIfpJzY3q3jVl0/vAe3/8BLvHWqzZSX+F0JVK
+          lYa/polGVzQdNOBUxua3eYFV/TxaGF+bB3Wun2ZYW7SE/9YHj78iGjd9A0+7/ofdDsxgyeTIhOen
+          f8Ge2r5cZv3sFP7lB0yumkZc5ZyAe9XtyaI0IJ/oI1Lha9rX1H7eftG84Q98DGuLFWT5Ll+MKYHc
+          zTxRtSmCiH3bRYFvkL2p85Sk4fOdHwpkX7ne8uMtX8KkHf/xWT8lP/DjpY8HnYVMBAyLrnX1+OTA
+          tcgV7Gz6jLfzN4H3vWVgjF+utuyPAUFez67UeSkjmNY2iuEcVSf8vPO7nJXXUIJbvveb+dKA5baE
+          M4qKN/H31j4epiPzQvnwwQqRdxcBTDS2yz+/gJ7eoKlnYCATNiDgaUpzAsZf8Ohg/UQKTV5Srm38
+          f4QbP8J28DrVWz5XuSC1wj+9z+a0fvX/+If+eOQR80Epw99rJVjLDEMTOKMNwT5TFOqFLR0WeHxC
+          IMxzgHNUblMPf7UMSRp+sTfvj/kf/wH7/uST6XrfDXM0XQkI5eDnSw96H9jiYh/exMbDSnBU3eUm
+          NzG4HL0jfla/M1urU/BGSv5zsTvfSo3Nj2sBd5PU09t2A2aCxb2DBacb2C5eiUu56JmC59BJ+LLX
+          tGHvgs+IllL2qSXs65p93V0Ch8RraGrPs0binRIiL9qN1LCzDLDPnPPgSxsVB/7hDP74+R+/8AVS
+          FPlPjPoG3N/Ix0eHr6N/eBrfZIEgfkVs2h+DEVEJn30OWqAeo0xuYSxbAJvD1tU9L9EI7HMg0Os5
+          fUfrc9x6bPjq4K+CfXCXQbV4aJ9DgSr+4cwWXZdMBBvYE0GapK3nlFfIdZ74GF/9T82ChTfh5XLB
+          2OG5NxuJyRXwqGYvn3D3rqb0YTV//gN2ejvOJ+H4aaF1kSi9rckI1tfpk0L3Rb++uCYjm/MsGmFr
+          th1Wz+qkrf7lYIKHb0c0/sOj71A1iO/1mNrfpIk2fUbgI+YQdjc9tw5ml4JXeLfo1eDOw/IpXQsq
+          6tbr+VJoGqvV3oLyR0+xV0kcWKY4G6UbKW1sNsSvF/98fwMtOHaEHeszEwe3EGQFgQCfepi6fFan
+          Gdz0EbW1TIomWFx6pLvJa+tpc6vnjd/BtGIP7I/WWWPHH3Lgj5YxPtZppDEWnyooiDTHp3MQ5/P+
+          JWbyfsy/+GQixd3XV4cH4rw42JWSg7vymHKgwUgmsJJKxty1bKC5QJs+ULLTyK73s394cVwMqC3C
+          Z+bRk18B4V7hoP35CzB4BVfqNYIZCWXA4j89Rs3o+KnXXW/+wzcfcN1hWK1pmwsNZOCvzuzlLFna
+          EmrnXsGP3Xd0fwpU4D+9pIbOS6OCZFewfz586pqmkQt3Sw+hifcfnJ2PWs57YTFDryshtdX9qs2y
+          UXlI59MAH/vzsV4wcD34GSKC8fN1GH4gIh7MxGdKrUwK68mpno58+1JCdUeTtPV7eq7yq5VSHA9T
+          na8zHmPwvNkd1m/OSWPG7dHD0Jdf1G93RrT++WGHwxhQNSizfKFUsNBLQC02AzfM6V/+c/Hj5aNx
+          cfKVf6Um3BfxHZvjZA/7TM5NMD60E3ZRedaaFac9/DzeB3yegMCW8jA38BR7PnX08jqw8PxswMvk
+          RvL3feO1XmV5eNiU+mXeuP2jthvwWY0UW/lDYmsZgBj6WXunTkk8je3Ih4OJrJ23niHi0PPImyGO
+          pgTf74YF9u8D9KFv3d/0vOHZxBxWITeJXGqw7hQJ+seLwR9ft8l6qQW/KTw49Y2JE92t2ZIkngf7
+          9K3RP304erbmwZIYB/rofFgz2mYO8I6NQY9N9Y2o7Kg+upMkx6buaoyR46qgJrRWqjXpIWKqoc3w
+          3MUxjedLw9jaPmbAyJ7gLR+665P7qMhahBwra/SsZ38aErDxQfqX/+nlrI4oWk5ngtzhUvNwO5Gz
+          OCmHtVav3NkTOhVySlj6nPMg2mT4ZQUycrviI9Jv7t68HSGUPp5JXa6715OdzASuWg3IMCyNNtbB
+          EELgRx3Vs6Pv8lIatxB9I+SzX3XO9zo7SZCvrwfyudyv7M+vBH/+thcef/nM9RoHr7IS0XtTXbTX
+          YzlbsHuENVamyWDr535xUG73GFt30mqr8rEIfNjVa3tfNVuy5VjBPz4RGHObz/lpbpCxm2/YornP
+          +PziWJCPwoGe+bNY9w2nvuHsjQ9qlsUZ8I/MKSHMhS/W3gMHOi+MZ/ANjxyZTTkEU+ZxAdyJykR9
+          hVaRsLxeMRyzkdDUUwMwo9TkACJqSj4FvoFp0yvSTOCbWvJlBr9RvY6ASuczvjBAo95vYh8aSX/1
+          W6yVGklH1MMaNdHmBwOwPd8Icy4CBPIXUk8ft1NBbQocPaHl6i7fxyL/8UNfyqIxZ8dbpgBtSPd/
+          /MOd/ea+ypwOgr+evPUcX5AEn2ElEP7kLS61f2UAtyM89Dwufc5YjN+yKxUq9srjDZDiBBTQoO8Z
+          G8nrx+aDacZArVTyr97ApKMtw6vxuWHtFHTaujfmAAWv8EpYaPbu56++Ua3Pwqc27uuNLzfQeTg+
+          PUXvGPCPYJah/DFTqhqdFC3uT23AUko+jbcGxgzaEge0x+ST2X0otTCnexkufdIT1iUK6Ma3ncHn
+          0Eu+WPVivX62ngtXLeCoT7gqYq/DbwY0vbyIsO4cl/dzxUdvtxSoUVYN+9Pj25S5kx8o3RotGvRa
+          +KNFTI1zqub8iDhddi/dnf57vg5VKoT3t08dk77rRaS3EGz1D3ze6iP85h/J2uo9/WWurmDlmKsD
+          GUhHGk6fG2PXzlKRe+nvdJsrGb3v+58OKyT88Invppp0bxKD2+d+xX5jh+7qCmAE8WcYfDhMk0af
+          1+8KZ488yPwWa7AKfZCiTf9SHQr3gYlR6vzzU43OX8G6x0cBpCuM//s+g2cAYWDscn8vPP2I3eS7
+          BM+J+yDzjntFv8KbVHmrX1Gt1bcpnpkdQO4MdZx3IGCLRA4O5H2WERE8+WhJvLWDG7/DJ2/wtdmy
+          JwJDEQY4dOUUEGYeKxQePfDPv6Q6xAoYcRBiCz/6fFULTwF+Ia3kD3/XqD6MgGt4SI/Tj4B5NtME
+          /vmllhqrbH+sXg5STjbApwtT3T8/BtrHw8+fDE7M5+jyUtBfPIDPsx7WpZdlGGg/Fd8Fr41W5Pcc
+          PJ60iHq1PbBJbM8EcGojbnz8ma+LSXuY3JhF//jNTF+SA6xddaSaM16jsQ2IBB73L6WOugD3nz/Q
+          CJ2KCxjY9Z5IKw/jmyRQx3cpY+fzx0N0XBeqCuQZkRHVPbwVl5BGh7KO2DW5yf/wA+snYfj3Po8C
+          3GFc/SY22+HFgkON7tQcrTpaFWVfwnPfTVRBU5yz+hC85cfhlfs7Iv8iUp3SN4xlB/iLGqjaniuU
+          7M+/pVYYL9F8CvsUwobrqUUEa1jF1LGgJ+AF/+XrWculBOJLinzhKnk1/+MrGVy/YkF4sE1J/xxE
+          AX4bf49x13VgvY7yf+uJ+vpBbPnzO5+iEeOstSq2iK0xwjBLENk1Ru7O733cI67yHKy9Xms0Auz3
+          8JrpX3xaQeSKebnfpmIcthNSPy8SB9UNwRBJLr0a91M0H0w/AZu+wX94+n3OrwJt9TVqC16bL3zr
+          mKBshCtVP0ehXkc0dDAqxROR6iJiU9BVOnq53xIbF67QRJ4mJvzbT9aGnyNgdwk2VEjp6eH3EePI
+          xUQtcE0CQqPKx1MOOCh9fHPj13y0ckYbwPnKV1gPompbH1RAFEsJ/cN3uitUHiXFXcbb/ncXC/T6
+          n9+8+cHKsB8Png59tvT0uNe0eirmOUCb3sLRl61slY2+g2aoPqgP3q969XO1+/Nr/b3kWYB5yUuC
+          mn3mKL7zu//u7xzs3thO/BdjzwPYpoYrLT03559LFBIG0D6CbcrUkmv8bAYJahLWEB4K91q8LVYM
+          J+2wIxzhlOgPz+HuXT8w3uprZPMbJJOyC8WPs1LT7yG20E+WZ6x92QoWSwfkr/5HRFn23bU4MRWC
+          jxFS/9Se3DGtX90fH/al8HAb/vmn71m7+vNWjx2b4cBDu3G3E4ms0+aTnfJAuD4TH5ZHEayb3wxC
+          DuT+/HMkd97qk/B8PmZYh5comi8XOYDG4Qh9aasfDFu9Gd6v3p3mFhqj2k/SDhrnKsQX2yrqlb7k
+          BsLSVvBp8ZRokcjioHvV7ze9o+fr4eWUcNMj/v7CQZdFHbGA+IkJ1TBvRywrKwHy+/dMtfEW50ur
+          DvH/c6IA/u8TBbVncdh6ulM0j6Hwho0Ry1S9337Ruk65Cat6VOiN+9B6dvykhxGeS2xYnMDm2/5X
+          okyfGS2D6DSIMjgRyC3STOM8kiJmSK8AGv2lxkahvdl6FpY3PP2qI/Yu7wNYIjWa0UfZel0cRjPf
+          y0bQyivjB6xPjqnN6fsowXSf9PiUiIPL7nawIjtLSmpsjGodBMLD0C0Kv+TYeRAQPQqQ3y2+//s+
+          8bDS1zZHNei/VAmCt0bCjJcR5YiLlTBIhrntPhWg3OhS71VM9XKTBxU+uCzz1wM6acvynkM4+cGT
+          OknqRXTkRxMexRP2f1V8G0bO0GQYJvmDCGXmusKZ7io4OFcf393QygXPuTtwP4mav98bC1tdtBeg
+          dNy6zt3TP6rMceDAOx49C9KR8dff2iIYqSW9b/9/Nu3fjBwrjWh55w/5WtqeCfu7o2BTuEzR+sqV
+          Fo2v2aDp1WcDNQxfkiAGPx+Q9p0LOLFU1Dl6R2/P1qyX7nDlIEcY8PkJwmiBi96gR2BX1D+5+4hK
+          e1JAMaozfA7JNRfkOuTQKCkhjugnAvSLjhb65bqLPbkshrV5ogLgxBKpTlI9p+7+U0LOCXyaepI/
+          iGBCM7yya0Pvyz11xc+ShUg9iCI2Pt/WneN6hPAZvm7UuTpEY/tHqMBYrU9Ewv2sLfexIejSmxZW
+          3JeZi/EiQ1C8jeFvP2jzNTBSKFHBpse+9Ab+mkQqEnJZomr2FtzFL4MUPs9CTYvrNi+s+70a5LwW
+          mZYGumld6KYVhPjwo/euh8P6c1YegLN9xPi5HsHeNtcO/fy0p0mR7cDM4zmDO/NbU2Me83y8A4cg
+          9+DN2JADrp7RxlD17xvQY/LtIxKANkWPzjlT7cz/8lVoOBU+5ZjD0eR1YK84kwPNapsrylnRwGfH
+          IEOztxuoPUity3JrzKAkw4DsR+qxX7bKsbxLLUSLJXlF7HB9JXANyxtpz1enFuNl5eAWT1TVww6s
+          drNX0TnzIlrU2jsisrBWsEw8lWLj/WDrRN8VWq+fGd/yc6AJsv1ZUTpd4dbV/FXvafJQoFh5ADsS
+          aNzR6TQeiT7fUg97iibu2rMJhakMqRMvt5qX+EpHt7VRqfqWcpfPvrWDju+s8MdXcR4EmByVbUTE
+          ppiMl8s/zHMIu4sUYHW9S9F8dBcPnp+mh5N2Dty5fJo6FDdjvexOUr5gRxLQqRAq8vwkn2iRpH2I
+          yldi+/s/hfXB4wjhe59idd8EA39gjKDzdnObNtbX5TtirvDrPO/UXvaKNodu8EbnT5JhM9NbjcVG
+          1oLf7zpgb7dV8A7pbKGA2iPFR84AwrZ+8mWOE+xUYcz4Bh4K+WhIR6z2Gsvni+HI8Pe7DdS57Y4u
+          H9cNhIfnExHe9EU2N8HFREQafhQnXy7f8CVApy97+fzQvusZ7voWntZDR7j58ciX36iraMMn/Di1
+          ZJgNZJawA5HgH+6fby4EVBFQJzkPH+yJre1TRU9QpugKVvTEz9egMwVISyDg9Ct3YOF6K4R7r6T+
+          mny+wzxfohK1Etj5jdLHA7/hLZoDs6KurvWANvoio0VTfbK8re8w7X65BJWi2mH8SYx8tl/CdjvJ
+          u+Or0l3A4jxZAU+cz1McXHQwnop465L3bch85z/5VAlSBa+XXYBNC8bahMW7j9T0UeN/8eFjOYPL
+          4XSl+BdFTPw8VgI/ih7Qizp32tyfcA9bO3Cpz4aWLcu+LtGDSzOa3etmIN9Un0HrWzrWh1YdFuHe
+          8nApnZj6B+xEe3mRIMTPqcZbX+p8VvWdD1vf0am+rQ8RGkGR5xje8a1aokF4HfMM3j7ya8sfqra+
+          cquRPYBSHLivNlpu1Z5H9U4f8e3egmh+sEyF2H4k2FdXDkwLTgncSzsHK/3hw6g2miqCiStS99LA
+          gV3LPIN/8XQprCJnZolHKAfdF0evYRj6HF0dODxKlao+fg/z4VxBNPG3LzYWq87X/OakMDSONr3f
+          7V2+bngP2fD0/q2P6NhnHfp5/iJcHCQaG16ShC7pm2DrY33zpeBePpz46xd7thczAmG+Au0LRmqm
+          wgn8y2e3Pc2paj+ug1inlxJt+dRnra66QkUqCx7faYGf+vyt57K1dDTfBw+f95fJXS/WkaBLaOT0
+          ZL0yRud7LoDvdD/h53XfuSv4VQ16XCSFSI7psb2cajPysq9OVXOI2Co0ggofIjrh4w7Ww5p+lBG5
+          b67GqRE/2f5bPSx48/UDjqJ0HObYFmekT9CnN/9lumsFziP8e17vzL2AOH0zB5rpLaVPjcvyFb8/
+          DaofY4qz2bMGPo+1Bmo76U6TN3vVgjl+iOxwj4zqtL64Yn+9dhC0lUaLWT5sc+HMEgHrYtHyodYa
+          O2xzorf4JLDxcTTPg9Ij70NdUt8WzRUOTmZBhpQWhyoxcjEw95184jwe3/TGi2imJjPib+hLFvep
+          1fuWwRGCKMdE4HupnrsjyWBg6x9sFaua85cbpwOX26aQQCi5i+tFEpJ2w6bl95a2dmWWoUXuKL0N
+          pxos7vOk//0f6uymT72urBvhGHgDvf6auCaec3GgER8TrKsXO9/irYMEvd7+ASQXNkW0eoPFdL+E
+          L0TACFa7Er330uJTf3rlS4kgD9ynSKh5C/VoTcKbAIMUPKn1lS3GH1cXwhcXYJwg5VHTVmAVUi7Z
+          gXr5k9XsXswxYld/h1WJdmCFia38xQ/VHuHVXQQWK9Cy3T2RffWUM9J+Zmhk5Q+bv/fCVjW/NrKZ
+          LJy/lwJHEyWcj3K7204ACPAUCVn/gvD4Da/0LPaSS8g7bdGp4Ct6v39OuejibIS12obkeiitiJ04
+          tfrDP3rSkQoEhf4q6LYnSMa7VIKlPb8gDN2ywP61jPPlFgcB3K8vkUhbfFGQqw2sDTrQc9E92BiM
+          jgLpWxnwzaVN9BvK1UQceBZYv0caEG+bnb07qwJWcMkD4iUCganFY+qYInWZTN4OVEIuovo9qtki
+          mKyHzjn0sCXt7IG1AnuDR9IyetIbL5+rC2/JUmT9yMf5DAMrQJbJf3wmP15sba9faQ9eQvT2gbMf
+          8+k597J8gw7BtoRMsNRgygB+Cym1wwxr+6gjb8gFVeHDsnfZrN86E2z5g0y/mbJ1/3lnkPf2Fb36
+          b9ldzcqyQL2vFOpaxStaqy6roBfbiy9/BMml713awKeyQxsfLfJ/fNj5zV9fQqniTp8MSLDpgo4a
+          JHkN5CEEAnglRYdte4gG2ud1CPe3RsAbH3H57thmMBPJC59h8cknFbYrZErl4dtw0sA2yVYBy+VC
+          qDr5br0E9dIjxCtHHPQHg82X4dEB/OZTWkhl53ZvMfDgFv8k3fTRG7zbFrR26GKtPPzYXzzBzu1V
+          qjGujyY6qz48FpZPveF+itiSfkvY9v0Tn3oy5XMv96ocreRMMTEqMJvJKYZU6W/Ua74fwMRAMeHz
+          lAcENZ5XC9dxSZFiBivG0+hE8+M+8kDhky92XV9l/wEAAP//pF3Jtqo8lH4gB9InDOmkl6Ag4gwQ
+          EVBRIAnk6Wtx7j+sGtXQdVweCNlft0lCfw0ngJIqd4JOZmkSjQkYogfpSID9olydR0nhrcsfyD1X
+          J0ZTb09BSt8Q+SAlW4d01CCabhbmDtWzo9Ucbruajg8SGve+W05Qqf/qi0TvjmOL5r98OE1SQNw2
+          O4zcjVNt2O4f94jHmT6Kg4BcMPftHO1LbXtDLt12zVW1NwluzydjXS8a8PWse3LgDydA9TXYQbFs
+          HCz43QDWP771k6YgVZqC8V+9gx5+yQ2hjK38inYwUbkEw/H2KemawQmy8R6i4o/Powde//gz4ite
+          NkdVOWnq7ckfSKJKZ8Ds2plglhIpkqFySZajhBQ4efBOrgcim9M457aaH7GLvOjedOwsiWfoNe+J
+          WH3pdwy/5xUcLm+KzMbKwGJAvCqu5/FRWokArOen0Sitl2dY3PCWmY+mUmDmiUQ71U0y325LC7/v
+          8EzM7qyVyy+fatAuhhQtwM+6KSwODQxa2CHtg+dxveavGqahw/B+l6fJoqulD5VxnbFUZUqy6cFK
+          2fRStAaWXzL6+NZQEHgZWcEKAhbwcwU2PIkk+u07dj73DfQ5BWO1tJSEbX5RMUM4Eo9f8bhmt5MA
+          18MqRE0ct8GfX4bBzv6i6Iu3c2fzvlLXMPJwRZVbyXi6vGUqKs7m198Bzd0KQ9VC0x9emSQEzIbE
+          RyYJyqeQfK8axaqT3Y6k+mgUTKR+5RB5j4wYrW2x1ZPSN0g+W4freu7L5eKohnqJbwKJNv0i5Ms7
+          hvJ+S3hNWpV40w9gvGoEXb8NMtkda7W6NyufmNxVD/7xzYZ/KKfffsyezteA+xu+/vNvbPg+38Dw
+          IzNq5ec+WISjfgbPrB5IeG0PAac/cgqvX1Mhjh08g3WEXAW/D2ihx2PVGfenR1lztSOy6V3K7kYK
+          9MqPiFutbcKs/F2DPb75yF7spluO4y+DFxoF6BC1RUAvWaKpxTtwIjaGf3mAsFNd625E/KLcusXy
+          fPsP37D88H4jm5RlgtfIlpFbZUo5325yC3hJ9VF0PXjBYpGUU4V4XZFpxW9zlve+AoXmMxGURklJ
+          Jf04Qb+7piQoSGOybpQ10OXCY+tYiyWje/wG/NePo/1d08oviVEBDJoECFW9lDAm8hLwv+sHhWla
+          dmwn9z/1c/UVPAf7g7nunSyD97tmo4OpQ7am2s2AgvS+IZd7J+aSa1YKZ3y4ETMwPh1NqCCptuEE
+          kfKUv926anb2z59W9emSMG4XNPCWTTuUcNuKDxI7OVTmbUWD76Ju6XFZw9MrRJte00fx/lE4GCU5
+          ItX6xAHRtbMB0Y//oXTIZZO+vidJ3fQ7QR9CE8rMi/vnb3FrnV1A530twYPxgBF9b2+sDfU5h9f0
+          6iDn9gblFBbHBiTwLEfKrNNkOaPXBG3qaXi9Xb2SF+XvDx7B+0NQHcwBiVlmQPlrHIjnW35Ag8+j
+          BuVxh5GVZB6jeiBH8HfnGIqK+JksTZdDWPePFmlmoG/+LRKU0ytCyNpbUUmta2Or9ldakZOXeFzO
+          puZDPAgEHTf+pdHjTdUdvWYkMjR3pI7guwDchHbbswCY690QKnXzE5ESxK9u+dhkB6WrWG16/WgK
+          LacXwP0UFIvMOpfT45VMyj5JLPTnZ0WfXAVozbsIuaGEx7mSkjd4Mt9D5UllJtv0LNSeiEXvy4i6
+          dXP+MrlcfUw2v7vtkWRDGqojOkwk3E6VE0MVVb1PIizf2JowdfrjB2SkuzP4KYLSQBymabSLSG8u
+          wtGL5b/85/hiz3/5F3RffEaC3zEA4uGlY/gw8B1pctgDprKrphxcgKLlk6RJpzajALxfkBMtf38C
+          vOljqBXq608/gplFfarexb2DPJmoAUPFPYJ8f75Gr/TBBZPL36l84W2LOO9LBoRND8CnFpYo1zwW
+          rEVfZYrSRnokCD+TURDCHp7ldtuVup9KdjNNH27PI9qfHp1Jm13aQHSqLGScw1MwtSbPwfb6FZBt
+          lq35V+/qn787PgsfLLXKCRCf3y3xi4vCZugvnJqGB4bSP/241S+063Uk0TZe85/eT0UN4WUPzU7Y
+          4UsNb0EvotvwqzqOBEcO1gohyLl8Q0Y2voJ9/NAj/qAP5eqf7hUs3p5D6ge37bHR3zK5QfqCYXVq
+          2JbPKn/5GKrS+Wxu/Evh71o2mO0fTicI0uUNtSsXkhz6TzZN510L/vy/6wIt2PSpr8onPUeB94kC
+          lvlMgDspNcl5q0dRO3ZQ/d0FRuxu63C2wS6E1hXrWN3yydW2ix2M7Akjc6WhOUWP9wq4UGzw+ueX
+          96IsQaTuahTGtBvp7Xlv//AeBVVidhOp5wI63M8lf3i86aUV+Mc4JDeQLYACuYKgTEOIrDBYGFvW
+          6gfNujmQ4t6wYMFXvYB/9eZ9HqRj5gGcwb2sJmRv/gZjLPjwzA05Xrd8a6u/GqqXxSbaPH46eqol
+          S749xQNyY/MNmNPfU/CyoEduBjQTbi8uCqQ9S4i3J27C9uFDAlXCBcjf2RRQbaUp/PPXBt/TjvRj
+          +INX6OKo4QoRLDa3syB/CrqIy9yxE75iuSkf7kDu79d1nHgYr9C7P0QU4LdR8hb/q0Dor/4//bjY
+          0Zgpl675oausTwm5+qhWXu/sjpeqt0rOJrAFH6N2UDQ1EWAVOOeQ+EcT8zh7duzTlRD85YNoOlzG
+          ke7fb3g0oY1O7qkIVpV4HPjL/3TdY8m08Sfk20klWttpjDNPRQVTJRiw9D3wjMoMTDCLHj+i0/I9
+          sueqpnCUbib661fMG78qm35GocV2ybKgeJI5IWyRR/h0XO6ntIL6I/mQgMffYCkOcw/T+3iMhior
+          yn/5+MDi/p/e47a8E7IvWLFanK1AwLxJYTd35z/8MGnIlAluejEqjOcvwHCAISSsPyCD3kzG3UpH
+          geOiUnJ877BJ7npjwVPAH5Fz3y+MDbBaoWsbFXLa3E3EpvF9+JdfHeBX+5cHK53Rn4nFoDOyzZ+C
+          Yz7ISDeoa/7Tu1zKRShCSgzWBg8u3MvdCSHtuphrEXMUGpfUJ84x+QD6OA8N+HveIdplybLhN3Rt
+          rULoZIKA/em1DV+w2r+6YMNPQ+0zJSX+KJjJovmzC8cdl6H7aqfgTz/APr7r0dI8V4CDO51gEUR2
+          tINF263Z+SFAd21FYl1JU+KW83JA4seJaG4CAT7pnK3yhmKiTd8FLP/5Z2A09WlbU9uwf3qu/dEf
+          CqnbmbguxRhedj+K3M/znJAt/wZDMBikLssoWPiiDCEOsxRpB9U0xe36geGHJrkgKiWbXqjgT/cS
+          5Fqt11Hwfq7wWWZP9JfvTaWSpWqDBRvvvhdhJO41KMA693cSBFEL/vAKULPYEyetm4D+jpcYJJk2
+          4TUcjoDgwg3lLa9Egc9PyWJHXaZufjqSYl41R7t2MIyN8RwteusB7s+/nD+7rf58ZTvF5lfDe9Yz
+          dOvap4kTxmO49YciOb9uK7ReTaH+5UXplWgl+xptrW7+ltjPMGT/8OYJMY/HL8eXdNOvigiuOkFB
+          +w1o31kxVPeug2Jx0BNuN9Y/0BqqiQ7dp+jWul1i9ZbhHTlydpNgTtZ3MDv1D3I6PKRk2o3ZALb+
+          GN5TsR3XsvtAaKnMjdaEvdhXuL0F6PrKJbpGh9ok3G+EShf6u2i8H2TAjvjpQn1UXeRePr9g/Szy
+          DuoHRUfuJUpGckpyBconM8fSb/I7sS4a6R8/6vK2h/UFzhRArliRvu9m8y9PBm/S5khj6MdYsrck
+          actfiTmGaslU5aZBw1wp0pTC/NcPUv/4z76e+2T4clUB81/YoYsbNh3lUWhA7+fl//QOp+ZjpVyf
+          2vDnB8e1YxcDZNH9h/ywXMdVLX9nIFXkEv0i7t6toMW9EpNgIuGL783le4s5ZdRoS6IoqYIhngxN
+          3fgZL9nHT2bxwdVqbEUTBlXSjbS9dxqMjq6OpSANgj+9q2551l8/oxOxcszA1q9Cf/Wx9DqzwQBO
+          ArLjxmD8MTdWeDWWmejN8wymx95VYCT/KhSIWO3oKKsDHG6gRfrughnd/JeaaeaBWIHWg3kWowai
+          xZEi6bUd3jq65gT7gB6RhfO+XMBvt1M2Px/t9RcqWQ7VCmpIrJEfcp5JeWRpKv91Y+Iogx6wn0st
+          9avIB3KCuDDXufIh5Ea7jOh+906oj081bC6nPXI2vUnEWLMA9bgzcXLuELAl9BT4G6CCxfjw7fhB
+          iw31w8YLfsdOmFB9FFJoXdRnNHOS0BGfPARwsyULPcSvFTD+K77hIzx5pLj4UbC2nJ6r8euloVoe
+          NLb1n1ew5XvRe6+5bNoFt2Y7s/VEnK7jzb9+Koy4tfs3votpwAps/h4lR85LFt5pNChf65HYvInM
+          v/40+EjPCAVFIgXrYM2RMpQVRF5B04BjX6LAgysjZE0L6fBTL3MYgn2OtLZrwD/8tbvHg+hK3pRr
+          8MtiiAKNoujdpYxK3GBBWeL2f/0Fht/77Y3X6YmI8Tltpy5u+QBrLjbyHp4/8nY0prDpsEb0c8TY
+          9Oc/vo+dRSI/O46LtJLh//NGgfp/nHrwmQoS6KDt6OEYTXCn0T1Wzuq3m+xt18buKE7E/KynbriM
+          xxpKyruKxGR6m8ugft/qe7fGxH+KAWB2/IXAwC4hh6bAbJ1pbEG7Gs8RHSKdiYzcCmhvu4o578cE
+          5mKQcyj78jp7Ux8EvLZOA5j9i4cvzlkLCD//GniokxXppkwC2l/oWx26U0DMqHNG/teyXvXqdSD+
+          u3uOk3zp3vA8H0G0q0ajW0b3+IZA3pUIvSvDFD3DXlXsTl9koL0w0uX0/cGBUBLxLliTpb6edpBW
+          jMfKIH1Han0gBg/ei0j5o1a3NM63APenZWAR56dySVbZh2v8nLAqnE0g8o+Bg8d9z6FblxRATMLE
+          UGd+f/4bD/B7uaceomvMEUOa1QSXrSfBS069SIhhnyzYPNcQB7VGzsgh7DvP10Y9E/VMCnqQSirj
+          doKnYMH4ja0z46VyKNQ17iZie4iyKTiLrdLedxbSStoCjl/dWr0fjIBo7eHHcHLQBNVCb4jMz1vr
+          hCYCPUQn00CeLbbBim5TAzX3EKNEzL1R2B/5As7e+4C02ETj4ucwhc/K8pA17rmR+lK7U4OjTolx
+          1dKECkWNobNpz2J7HtxlPFSQTtpCynEYSy6mha0erBChB5HNgE8XI4ZG1N9J7lmwY4+U5qq8Tj5x
+          SJybwg71Z9VxdgdUDumzFLkr+UFoswy3ed6Yy6Vze5gVPwsLAl8H4he4BlTOcUWKQfJGcv54GiwD
+          oyV5090CJpkBhPFlSMltParBHFwkSx299EcSVEYl2R+kBq7ClSKzrz8jwzzIoalVN3IlcAYUno8c
+          XIULJdEhfJvk+oSFmpTxAV3A8ZJwPVxyVSdaQK4vdjVXMdxP8Pw+HvBy41CwjBk6gwfuD6ho9q+A
+          /3wwhmbRQRI8gAjWA6sE6I3xkbgkP4zMvUktsG8AYGqvj2T52VII4Gp5pPrEXED3eeyr7n1conWb
+          z1Oma4XqcJczOsTDAbAhtA01m/FC3FAax6l6SCn4vgIboUEFJhPwXoJL7z5QPXt6wrcKb6g/pybk
+          KApNwhXZl8Jf17xR/MEQzDIPJYj2HMIUuB0b27RWYKmfL8S9Z35Cv/EIofC8SMSNKjcQmtNegNPT
+          +KHDr3uN1KixC5/w9kCH4nQFIho5X4WNI5CDi2BHvtToYTHWMSrCQQfCIKwWIOyRIVN55eUSnc+Z
+          GkhZh6XLN2Pscvm+Ib50PbL3hVQyejQnNSyGkJiGX5d04hxDNWqZRcuNIwF3bp4NKBlII+Xwscy1
+          4WEBL0NWRld3HDpmnWNBBf19RonTeoyBXU4hrRaeHNbKDzjakl6Z3EeEjsr1lvDYLGqY704tcs6u
+          N3JLOitQ5CObVM6maPRdSEHeq+Y2nloiluJlVQ03HSPW10636EHHQRZwNQk2fBF+v+MKl+tPQciM
+          VxM/jGFV/XjoUXpfhoTkteTDS2t/if8MsoBb5tmASwo6EmVt0Am3Za3VTP8BZJTLvmNBn9bwqLEP
+          yuTFLmm+P9VqX1YuqgRfKJdCp2/1911zorXqaeTF+Achal4pud/RD9AnawV19WCHItkpy+UIDA3O
+          DAEUnG96sqxQKqC41gT5THaCFfvTAP/4AHXNDZAHWEJIImtPfEXpAfvka6oW57tDELy2JpGqwP6b
+          P+hgvP2Sy+7mqsbLFJDidDuanJJADMfK1okTpcVI7p85gi6s9ngRaNhxn/G0U/mD4JALkMdgynQ3
+          h7aZMeQp6zVZ5EqlQCo1P7pF9N0t7jBJ0G5dnRy27y+X9fH7N3+9gUJziY3vCh4u1VFyTc9AFCxF
+          gKE7ylhhxyWhnGnX4HecniiV7sm4zS9N3T8eHPI1JHXrqLmTPEVEI7qvTGBRDIdCa3BzhL7fmOHd
+          0BXqrL0hMSWPDyjPWxhYr5tOEp9I45TW6wCjOjdQ7pS/ZPl83tO/+zezKgzWqbiHsFXWfaS+XJSs
+          qjBbsN81PLm4qBrpbX35ajIcNCwH3K5kI04mKO1scetQVIye/PegfsPXD6EEU8CS3e0MvzsfIeMS
+          v9lPv7MalvUBEaSmu3LxWJvCe/++Y3WevVHg3nGmvjuTbr8HwXKOuP5v/qLU/SSMvjwrVgNjd4yA
+          555Kpi3LTtW9fYBc3W7H8Zl4DcSHMEJZe/bYTIY+Vz9jcMe7o7ok9NWauSqJlk7OEfIC3O79HVR2
+          2Tm6rNUvYL8ui7ezdlf0GA5pIqjCy1I3vEEaK+OSjtWOwiAzFHKsmVMKZaqF/+63tmk7iofkHcLq
+          sffJORynkf2WK1QPghihoPSnkXZHsYBRiziife2ho4U72PAVOQ6WnNIvuW5hEPpCNRBNO9kjn899
+          qzpUEIltGG1ArgpoYdbUXsSyKjTX/VEtoCPaX+St+mT+BotxKiTpgpKhU4Pl6n128MicGKEC7EdG
+          v46lNqLpYfkmyIxeuUqDeb83MeQKl2G0jj8o7rZdqF/vPJiMb/eGTt0dkXuk53INRhPCqPyNf/xg
+          TuK1qFXd2F1I6LuHTng940nNZxiSw+12Mtderaw/fUSQq6sdsc65AOP2jpAmTU23PlfqwuDWXYgx
+          f3xz+Lv/Qdu3yJyB2y0HUbNV6eVgdCT3a8AOyTsCVmg9MT19woDTnLsCt3rG3OoGnbA4kQR3kf5B
+          xz4eRurPrwZqsxtv/IsT+hUnCi3pgZGt1YeOPqqihxSvLnIf6zwy6RdSWJYXiWjoKncEVzCGEuXQ
+          33iYtLLlUDn4lY4cd3Q7MZKaGs578sCCuZJuPeyKM9j0EfECri63+ftWD8ddRfyEj0tRKaoQVlVN
+          UFAKRsJduVRTx1f+JLHuHUtRjvRcTVd4IIZlryWb686F4rYuTBCfCVsyJefAmzdMZJ1h2U38x9Sg
+          eJbuUej2nvlLpbQCk17lqCYvq+NuH4LB9jzQaUnGkZnwrkCz0Uak3Ro6rvod1EC5UBGLRqaNvJ49
+          z2C+lwKJLj+LUXFXnoHGCpMc8qefsGFnxcAR+ZnUvDKNuH0HHPzTy743ydvvsQqsycmM1v33kHDa
+          WNRQlolGzOPYALqOgg+dmXpIK61jQIl6ecPpc64IsvMg4aoHzVSBZQGKhHebbHxSQeHe2MQwLzbj
+          Yu8UQyHXc3IJiV4KnJGscMND5JvXpZvPvJrD8R28SGg6DljAolO1X+Ij0rlGSTDGzwjeZpvi7Uji
+          kqa1MsBUWASk55berVFEbCjVzQuZe7UDCw2fGjxPlRiBNvBL5vyEGtRyd0D+bfcpaS26bwhoGKET
+          LS3AN7iNgHr6lf/4cuXiewU2vEL32jlt8+3cQ3tP9uRwwN24VDdNgLvmctr0sRHMx1lL4Xa9kaS+
+          V5M9UimX392mf0NR7X6KYnBw09dYtMRbN12NywpJ52sIFQSPS26PE5TSNUdGLUt/epyDett8/vzF
+          KAa3h6E8+CDC8Kz7Jte+ni5o5LVBx/cUAGp33huctVInx2V9mEt0yXrIQUElmjF7YKUHuQWXfPVI
+          oLTzuKTvkIPlEqjIuXmfZNpzSQNA9B1Q9hP7gN4n04WteO4jVU2e3b/5E11ymSS7m5NwURQWYOXf
+          E9E87HTC5ZT8lA3/8e4a6aX43fa4SJ5ihEID6QEf7qIY4lb3kPb7vYL1eVsMONqfJJKXS1ayW3Ns
+          wGFucST7UBwpGxZFLQOtJRt/jhNidQY1682TqyXeRvZM9AYKKl8R0ydSt6oWFmBzrURkLq5YLkP/
+          q2CUu/HmD/SSE72Ygj98tmhPgoHdueoPb0lydXW2VJo3QVjFVsTu9y9jhwRH//DnrlzlZPXqxVdN
+          GwfI4eZb8u0FXVEjEKnEi6yYLcwzz6C+PyIUyWwJqGG0K0zvvoZc4cWV7Ky5IRy83RgtZR6MPN8k
+          rWI8x4BYmiGPuLpc3nDvKhOJbNp2VHae20Q6T8isXp1J82MSQf6S35CeTTxYR03D8ILWPfHSo8Mo
+          ydIVtukOkKihubmczb2vcA5TI7x7vMFieNwOJjhbtyVEQ7n+IiOHWhmo+EcPebJOxSUET0m2iPUU
+          VEaPP7QCCxf+durGHFD9Hq8wU7opgvH92G1+tAIPafwikzxeJnNvtNkOwTOQQ2IpYDc5/Cm7fR8i
+          FPBTuQ73FcL9uL8RX/qtf3rfh9eX8CDuTytMEqjJW70ftIBku/Y0Lqu/uiBxG0ycnUDN9xSWDTio
+          Bw1pwdkKVsUN3+qfP5/JyTbFdb3/FPA5Wv/4Y512pxX052qPzhueibtpydT9qN4Q0jgMFj/nUijL
+          s4bOWbrrpiThI3j69RCFhI9KfgjGFFzMpEThNF9K+jDLSukvYRaJOF9K1pn57t/9OHflPI7c9TPA
+          zU8h91nuGSM3aqj0lhTI8dEhoLcb1kBWDBa5L+veXEMxW0EtPw/I2bcvRjXnokDzHd/RfU8s8Icf
+          //zMRJ9WIlIubP/hZ7SuSUlfL2eAtm35JLg/uHE97M4xfKWpEoFv6AZf79UMoGuCDPnLlwP4Tw8K
+          nIOQ0wOafMXkE/35o4jtVROwRo16eAoYJtrmJ5lUmbZaYDii86W6AsGBYw57ksh43fCMDtbJVz+f
+          r4GOByvvFh5mEhT5vY+QPl069ue3I0zySAo16+/+JJgKTMAy+NXmNC/+Cp8SsJD/pXEyu2im0OiL
+          FnN+QBOiM5RCf/ZdhJR9lrDfJX1Dt/AyoknSMs7b/FDUA1CQE/azSat51f7+Hwqk3clcH99nD69m
+          K2CRxHnAIHxKatNU+l89l/TWdzV8DGmCqpPXJQxZLFW5pK2Je9XbBGP4UxQuaWoUJq/3uH5OgwuJ
+          xl1RuP9cy5UQof+bv8j9aUrw56fhlkcQZ/P3i3HdDcA7nyxkrlod8KMSCLBazxzyuj3tqG82NQzT
+          bYVKoDwD9uffpPzMEHKgA6bHoxbgcU3Dzb/myXp85xx43tIVoV56jeuAlhjuIvMTtaerkUxxmxdQ
+          KChB22b7bK3r3peD/eUW7Q+N1f1dvzpUk0H+/DN7G2IF/UzKSF4rb7ZwrZGrgwdHhKQuCeYtj4BN
+          Aqvo4pwbc62nmQP5LmkjtvmTDV9cJet2PXI2PKaX8VjBS2t9kVXJz45dlgkqr6t9QuVJvzDmvYMM
+          nJ2biEJV9JhwoHkP5iaIkEGJWS7+AmrA5wqHbDKQcl0tY1Cz5/FItEVMGP3TD5t+RkbaL+YU3rlt
+          xYn4iihPpeSvXkCWzgMW4HmfEOSIb0hO3kxC+Ngl+C4XLnjFwgNZ3+5UsqoiK6xVs0GulSzJ8vBf
+          P5hjEuO9Ax02mwBE8J3DEBmb3pg2/aiev3NDjHJ5jIt9qn7wphsKch+2CJbD5P/g+Y0OeGlvXsmJ
+          4d3YVuUcUS28jZKLbp4ANnxHG953uMLVBO0dBAQNyy8gmAcF2D4jbUquCTNoYIM/vJXjWw7G212P
+          4Sd9X5BbG5o5+VIL1dfP3yEPu59ufYNXBEkHLLwbD3u2nMpGUrvGy5Brd+Yoivdlpwa7/kK2/ALQ
+          5ClI6s2Jp80frAkry6IAU0d1VG/8Ie6eQIL7cjDR/WBJHb0FUwrt6nvGViS+g3/8iqznneRheAHE
+          qeAK0o54mAdyYAppelVUVgUc0tbu21FNrly45VF4qkeZ/ekR+LTjAzk1fMjmPzyqbcWOZDuYykmG
+          faFu9Y62vKvDh+uthvsxupDjp1/HjZ9DOKyNQ65RqnRMaIQQCmmjEk/KuIAa5VNRa1uyEXqUVrnU
+          Xh4D4vc9OX7y57ieZtqonLkLI0E6emzhnSoHxuGVRgzyPliKQS5AALgdKcY0ZKvUKm/g3r8LCaTd
+          EuBr0YZQEIQ3cSfYJkt/N1aYH4qFOJmOg80/2WDIqphktxl17H7rWvh53mJ0fGtC8BvQcoaF538w
+          t/HtlgdiWM5pgML5Fmz49X3DjX+RY0af5NvgXwjcIsgIOvtpwiu3pVVlWF3JlgcFtHAb+18+5tlB
+          WPJ/eXR6dzW8wOMRkJ0LV7jlHeTgfYbgLw/78xObH/4komdEVHl9yg85+A/Mhv0+38H5619RKBpq
+          wO7qYIFxkWxyep+PgJ1Mq/7Tw8iSvW+CD8VlAqES1STyTmCkO/FKQccPFvG70k1I8ipcCD7IigTl
+          ezS56GM2suyUV+KtR9WkmWAoID/kC4oXMQGLXPEUbP0gZFwn31ycKC7g6fliGL7zOWFPN7TBT4/U
+          CN4ZM+dCpz0sVvO2+e/3uADO1uCp1H7IFvjaXPnPPIEJJ+d/+eZ6AmsENz8ULb9dNy6f26zB+OYT
+          cuTwp1zdTISwwLuReMZVLfHxe7Gg+Mh14gTbioLnKrkwWiqKjMw9jvzX4ypI2D37p99Wof9WYMtr
+          yF9etCqKz4HWQD3RxJRjS6bebdhYhY93eXD/L0/c8joUXiKJ/fUX4FZPCGXbK0jrehlAonBXopkv
+          OrK/vLu7W8PGt3yH5d7ZwUv1UpA+tFrJiddzrTZ9smIwNOdyfT+enLrpPRQeL2nCnz9gAlf7fMQL
+          AQ9Aq4eUwW2+oNR7FN3afBQME/Tso7U5WSNlhRRBtkQ88a6XhjEO9JG68RFJzxbHfhj+pH9+LrKt
+          ia0Tq1L4KQTnT98ArEyeC7XZj4k953eTdL/bCsW1IqSQWzmZpO4V/+WReGVxFRD4JgbEz/MhWtz6
+          O063GzYgkt93Yie507FkdzqrTsq06FUc3gBrp2GAwvMqIb8rh6RnBQ3huVJexA21HtAM+ZYi8qH9
+          l8d2gnP5/dPb+NFLh46vX8kAt/mEd7wSbu8Pd/jf9R/VRB/XoRFDFV+ePfnLy9bO1yr499l48s+g
+          +9N/dTLrxDevp3G192SQYlDl0QscLyVR1f0Z3Ex6Qldt7YNJ3CVn9YLfR6wXjcX+5cub/9nymNe4
+          Ho0qhsaC7hGn6DRgf/nLVi9IM3QumDUNF/KhWK54r+yFZHlziQEVuwNIOyc64Cl7CupWn5HclpzJ
+          ju3cg60fghCfWSV3LSsFzJZzJeaq7YLfLhw0uPE7so2ImX/5rlSZbkwiaHYdVe9BDPkXIVhs30O5
+          vu0Ww8vICaRsxcakrnLjYFDFE3KSyTan7G5SWPv5E+na6xWwJCwNCIrcQ2FZByZtd8MZ3pzzFEE1
+          4Mwp1OVeudfZb8sHl3IJW0MBz+xTRnK1GubffAf7x50jkeyAhCGZ52DH/yziXBfQ9Q1/mtTH2Xoi
+          Pzu0254MRQjr5gmRwQ0fRuXmqwH/dBuI04M4YaN8clWiCddI/eT6+NefhNv4I9swDJOr0lVR4YM+
+          UH7PfgmDXbvCzd9E0nb6+pLuAwiu8XVB9kW/AWqUXwW+SZdGDStpSa1AzhTRHJ9YFNSxWx9xWymP
+          s/0kmx8wWfc7Ubj1u4ixzfepO+5zqKmPN3HD8MLYC40r3PzbpifFgAwFmgBdXwrmt34h5j+vCbIa
+          BMg6hZHJpbXyg8TKVmTU+jmgRtzXMHFbjOs25TqGTdWHhiXE//HNtWijPz2DdEIwW8DxuJ3KWRJk
+          Gv5u69d0UH0x3yUHs15Msvnzf/zlpccPWOhrcOW/521yMR0X63mIIBH2SyTtzh9AdgrQ/vqnqGxF
+          zRRCMaNqeFV4cviKiDHD82OIwXXAglYfRqaERfOnnyLZbbiE0PKQQ4qggu6W5HS0ShUFvIShQ4f3
+          Tw+ESa8tGIm4JXrlXU1huCs7GJ8VlXj0jjrhPuYV3PQN5jZ+ol23nv/GF/31n5rnMBdQrqTDlid/
+          g+XhzwNkS8ijSyucA9ammQQE2nMkjO9zx76et0KP62eCEhwz9nyaoTqC/RcP12NSLvXuTIGNuCsq
+          mgQl88bXsEl2FUGXXZf809vEf/f4GbGlXObHXVB0A142f2GP/DT5LoCP9YG85LmAOegWqvry3kT2
+          T7SCBbzD7K9fQFKMPgHd72OodvHHQn446Ex8oZH+9ZuJFZYGEC6/86r+wyNY692wSNspS4DLUJDV
+          GeC98NvA6CTV0eCUfjLtvpkNj/LFIObu9ik3/J/+P28U8Nz//krBU/icI7JOBZtGcvqBkD6PkVrf
+          o0RA/d6HUnA9EWfpaLckF0kBHxR/kObyAeMSKvmqRiKfBL/roZwvu3oA1j31EfJTjq3V6ZLDn3w2
+          kRth1WTl+qqUo2Tb6OivBCz2U29U6XjriB1/nuNyD3IX7sQFIe3WJuXEDIuDYE1PKAtCKaDLU/Ih
+          F54ccpLVO1jStZIgmc6MaPIoAKadHoqiO6cLCpLjWE7pMY1U73oHGHB2Zs5GtMfwO/1KpKn+Wi6f
+          3VVRr5OZYtU1QbdmxllRwSl3yPEOv2D1LvAHVCocSZ7Lc7fmqpLB8PS9ILcXomDgntpZvX5VA/P3
+          1AoYXQQOEjVU0f1UW4CDrp2pa7wMxIgPmFHpcHwrkmC/iHm9TN1KpC5Ugxlj5KjNYlKF+hQA+X4i
+          iZ+ZwVIOk6Zql5tOHhV2AKd5bQafNouRze2djpnjKqjx1Cuk1s2+nDSvTYGgfA44b2/lKLYDfMOq
+          n37kMlAtoLG8+DDQgIfC+89gojEYITw2V4SQlrQJ3dXSANk7RugOHKGkq2Aaqnv+BMg6Z4O5avuY
+          U2EavlAk2wNgKYpt9cpzNfLHzANvWSjfagn1gUSIuwMBqpCDpidvLdnlPPLS4earZe4YEaXBANjt
+          nAug8fSGBOkejGs3apVaySzHsnPUE9HSDqsKzbhEl88lG3l6LX04zZAnubfLTEE+7wc4ju4V2cYx
+          GoX5UEmgJ9xz23bZ/+/vjtJqpH6GHzB37UhhF/kBieZfCIRGM2PVtWdCTpfXFSzpU/mB8MBkTNOT
+          A/jrxdNgIPc+8az90mGfk2zlgK0MaWutlutdA74a228DBQX4JsL7ty26M6yaZKyvzTUeCIXfJ/Gj
+          VQy/5sy3swGL1fJRfv+mI80ZVVQEqUTuau8C7o6HFgqf/kvuL89MWIpyCy4YJNG0XJ7JyrfSGz65
+          p0birtZNKjZ5oXKnq4B3iFPZN6l9A16yIiehwpOEvOhFUF+w9jCvPOhIAHtwCve1Z+Ro4dNcTyxP
+          Ycc7EXIb2zF5TYAuuHqaRAponjru1hMBfmOmIUs5eiOdPm4Ddxl/2Op/BOz6SXroGeW2LXKUBiyF
+          dQot+SkTB5j1yH0/tFCb/YBQfau/YBmrrlELcKqQa82/UbzjcAf5EQTEH0Wc0Cm1Ocic4opS7gQA
+          FS+Eg9uWNegAtEvA36W6gMtr0iPhfrA7Thy+LRSn+w15hq6M9HV592olzjY61C80CifEFaqzjDoy
+          h7Us+VUkFGrCJSDoA54moRbt4esgFsj4tvdkDS/hCp/Ppkf18emVzN+9ILwF3xwvFyspxcOnyIDg
+          TgsKdkbHRJQ2tcp9tBLFGCls9nO3hQ+smOTiuXogKK8lU9WHYmBpw79VFpK36mkZiPDh/AbrsyKx
+          guxVQ/pP1Tv+fHbP4FyFu62+XEaumZypizTdUap/tW71f04DtfqsRbs6ooBzxeas2u3eibjgl5gi
+          qzhb/fs9w7vJDL/XJ4bBU0xQuq8v5iqu7Q4qx/2CkDnsweJzUwt5frqQ9CCZgRhgc1LJLb6SsnVb
+          sM6qiVXXJuTf+E6yK4QqkVyEbiZ+ATKQGYKzAM8oqfgCsPKXpvCVLT3xc3gNVkORUuiY4ES85V0E
+          NN8ke/mUQhInT61bGy04Q9FtGZaQYgKeXhNf/cODe65VpuCcQwOCd8XjpNb1kS8o76tvnN6JHqdh
+          InyjiySn49cimT75AQUMFEqZ7GqEXvOS0L/521fcgdRc9mTrpYxiSD3rhkI28GD5PS0D2jdPxZgT
+          7YTfUfEHm+/+h2yKmlHcThuAYvlE5D6JejI051kADRKOyPtKVYB/fFtB/6FxqEgql/HXZdakSJRc
+          khC6BovrkRTgb30h5oq/Jvltr+RdPUMi5qgFActEngN9ZubEfC9Dt34VA6rcHZ2Qec8eyVrIJ06t
+          vH1Gjqe6BywrT5x6/36KiHb1M2CBkAqQ72lAKjZcgDiQeQd5cf3g18NOA7FWogmGRZPi+23+JIwb
+          VA5u8wVzdeKOa6osbzUw9mUEQnoz6R03jbJMvE4iVUBsjOZSUNYnWYmZcW7A5+p7VWVJNpEXVB1Y
+          pNukAaWWQTRP4jMZSqZrqtrvlmh/6I+Au4ZTBsO8NpAbwqO5fDy3hfe2KlChaXMw7ZnZqMhXeixL
+          LOmE6rkfoPyBMinBa0nWcb5lcGh2O3L4rt9gCWmoQWZ8ISrxwUzEv/uZR/xCx1vYAqrhaw0egr5F
+          +p8h+TZdMqilqaXowkevQBTNnIMcdLx/9SnGi1WpX8TxqJg+l1H8PkwMs5e1HTTJyk7c7h98BPSL
+          5OXAj4v0VH05e9nn6P1MvUDA3qGFXGadSNxHQ0Kw8Mj+6RPjhy1zYST7wT/+NEM3KRn9vibVyM0n
+          iroHLFd1Lt7wOukpOn32/kgBYznUORugP75hyWBW4O96Qixh9o+/F+EXIA1+j0Cc3tMOXNz6h4Lh
+          pwT9S/Yb6Aa1T1JbCU3+sHSr2khiS3RBac2l2e8HKPjM3Oab21Gnf/cA46dJzIf5G+tI8w25XtyN
+          17dtS9/XtoKCv5jRnBGaTEr45qDA6gMysPrufhJyDAjmF49uQO9KGn5ICqTBFSNWt3YgiAq0YS79
+          IDHulpSw0+FRKNzOK/EeimNJnbicIMdBj1zOAwnYx9FCdVVfx+jDaV1ASXDgYKLKKTkenlYpLs+q
+          hyHMB2Rp9xubHtcohRp3lpEp7rluTmpDg42tRsi62sicElk/q0wMb8jGnGiyuGlDwPkRT3xff4zM
+          rW+Z6ovOnhiBfQiEkzPGILOjArkXIHRLNnM5nIDUkXSrF9bEi6Um3g7j+fDsSzqS06Du96mLd363
+          JpN+mSzYnfwhAto5DUjD4xDM+eSRlDzTkTa7QgJM+/yQtuGfqJBPBC/j3UZuVoEA/+nLdo1pdBdv
+          eUKvqVrD4h1i4vArC2j4yCIwj9Prb7zM9Sy6EKof3dzmRwRmZuQFsNpcJp7R7k1arzcXZJbv4l0z
+          muWqesIPfkGTIEueDibXzdqqakbxQmG2mzc9l9uwy5MheqL+FVDN+2XQhjQm9fwLGU8/agFVcX9E
+          weBtLaVs7f/wjxh5uiTrvmmgWjmyRcqQ3gI+/Zkr9C7lF3njy0nW/SFz4cT0njgPTx2Z3+pQ1Vfj
+          i+UmxwGD9u0NoR55xFk5GKyBZxpQ9hwdVxdosS6aS06ZT6CKVlMeR/xevxiuQ/PAfFo8SpoeqxAM
+          14CL0ufaB+tetXbwV+QByqS62/CyEsB1b1ISJJgAMh+aGvKzekene8Kz6WqVIQwfH4M4vdmOFPll
+          9R9+pd67G8+aIMGoC2YsPrx7t3jNbYDBMGgkHuHbpPxJDOGOx+q/+hLj5hcB3UkuyPjQY7Ak78SF
+          yrM/EO34OoP1xj8jiO+9TG5vvinZ9WHU6t/9h31YmMuCDEPZPSYLeRryRyp42AKZHRbEuV/nkQzk
+          tYP7c9ORh8/fkp/xmBtw+3UxOdhjNNKfe7SgwvVPFNPAZeRAxgqul8ZEZzhdS/4j2BRSZTfjVbqn
+          bP2b//OD98gtGd/JKu96DIrieo721dFMRPEBBahcVxDxnQCCJa5lBd590iJthHYgLjezUg/nCKDo
+          cH6zyc57AaZ9Y6FgzzEwh9QywEfZHTDYPi/DxYjBt55mcrgc9HHpivNPtfIPIfGiSmCN4imG+i9Z
+          MDcZt2SpH52hxudlQX5SDYAYj+2gYN4l6MRjUG56NoPSstM3vOjL7fua2py5H7mNr0+5vH9HDbpq
+          WkeKi8iIJVtVIDe6e3K2vitbK27JVOWoLsjwX01HT229g8l7TYnPrA9b8fStYa1nZ7zb/AQV9BSC
+          aDKmSDUMF6yK/27g5fH6YTn3gpFzJhFDwfyZWOovXreC73MHsZI/yOP0jBLGlghCZduC473XFHMK
+          gyCD4VjVyL27NFi1PjLAn54OaJ6bNKHUh0FFe3Jbnbqj4vBswS6Xr5EyZl/wz09Yp4qg+sDGcrl+
+          Ewlyp4uAM+OIu+XG3weoeOKR6KX2NFeUVgrI3POE//Tj0GZZLcfH2YjYerS6lR91DDUS+qiMdi7j
+          Vc2l8Fof3ugoCz+wOHNwhnE8P7DQs777Nvv9D0xA6Yhu2lrAn4tToZ5lrSchihGgq3/ZgVfTY3JZ
+          YxzQoBMmsN9nblQlj7Vk3jpFgPGuE8mnuzBOAIatGp1RRzb/0Yl/eBotTolc776MC3geY2iFhYay
+          bfxmn+tbVdqfROSEPzsQ3w/Ww81PbtfrAxoqT0OVFqgj+/pTg83fD1DVtQBdEo915ONoEewiN0CR
+          /DmWFMbXGn47VhFT3KfdspYBheInH1HwCupuwdNdgcD7Iizd3ThYmr04wGUv3onlD79g3g3vFoLh
+          6SD/4zjlP3zjdkGJ9L5B5TLJTqNCPfTwd0eeyUqtl6Eqa4uQW4pWybJEw8DtukvEHJl2M3YGCi3x
+          /on2yCrB2rhpAXeQV5H97YJxlZCjweIQ9yi+JJRNO5D+4JCYMQlat2VzoT8qxdTwdsyG0yasOg07
+          CJl1R+cRuaNQ3d4t0KxaJIblRglz2kUA0sUtkVnxBVt91YuUXCEROhwl12S+Lfqwn1dEzFrXO/5o
+          1TtgqJNNaoPnkj+8/cuPoi1/CX6Rm/jwZ81v/Nu3E6OWFLvqVn/IPn5/3doh9QfJL3pG9BkvyXI5
+          xYWC2++A7qs7MUYvBQf9yP5t4/lJmDH4EXSnIo3A4F1Gnn0Ojeq35gnvNe4NqOWutfIUXmdUfUY+
+          WdazMvzDd/8+7czBctcK/tbHduzby+9+RBpDuPFRxCH5Mk735DL98Ss6NooG+B7PjewBKyKWGjQd
+          q49aBa+18474G++O3Faf6qv9asidD3G3FFV7hvhbXYg5rKD8h9f5B6/I5btLh01b8WFFe3vjw/uI
+          C/m08ZNOiR8c3+X/kHYla8oCSfCBPMiiVHJkExCQQkDEG+AGiMhSBdTTz0f/c5zbHPvSai2REZFZ
+          mf2q56RUGn0fIcNDUxckPkzfzdmHEpHoH79wH4JA9nIl9bSSrULuM/9EvUC56ozcUAx1sy2pfR5i
+          /Y/PydZ9yLA1hB99tI80gcNsMqx+ZolN1nafgaSRE+Hfldlz54MEyNs8JOwmrpZzaUoMWPUAtrKb
+          UDFZ3RsoUPkvvtZ6uZaYfAla8RfbYvBis/pQpj9/y9e3nq+P93xJ0Z9f5J+nHk2WrdVycSY+Ke/G
+          Lv9g8RNAH+Uv7MzXoie89Zngte0wPq78oCuG1v/z73yhqEt9eQWfEu7wEdZ4R6rpHt0HaTa5Iz3d
+          RitfOr57/OE39ufUYfyJQQdvrR2ovYmKqipbrpH386TjY/2dKuIedUXmrc0HW3RWojl9awtU/MHH
+          p/lr51wbCwSAdyh15iv0S7w+En4HTYZP1ekSDcnunEGTxyrVgoblg9i+XzKvcyotJB7nPwfkVmry
+          RPV3obyt6KIEBmDjIJL9Pk71GUFgoL6M0n/r+xmmTfrPnzAkQvTp075CWPUxEe+7RR81qZ5kFBU3
+          HHbEcAV7HzTyXS0D6o9HO19aLbZl4T4///hfxJ3l9yL/8TfvkypMPOhCCsdH2GLjEbUVq4aPB9+t
+          +SMT2jf9DCNWoHopL3r+PhqXPma5lr4SHGieoWO0bPZxIr93fY11qfjq49tpg7/7Q9Um3P7Ftw3g
+          7+DQ8Kem7LdXdh0oQRL5eLqZOfc53R8ASShi35jLfFzjBbAmxFTH5jpm8uJNoHG89W+9qzmQYmju
+          WUS2B/2Yz2dmLLLe8VuMg0BDS6Q7C+q98EnEITy4MwS9Am+X36z8OneXYR8ZSD2uTf5wl+rLYl86
+          eFZKSyPFDdn0NGtTOjlJQi2T7PTRLt4KNPc0ovo8vPTlOq9P2gbhRvV7ss2XqbU3++ppXqjt5Xa1
+          iI/URKt/jJP8mFbz8eAL0qo3yf2ef6LZVhMfnq9NQpBUtdHsRq0GE7XO2JfsM1rj+QuhJTnT05ge
+          GD+zrfPP31Lae8d+57Ir0cAfBOwMX75fnA6//vwc+uevD+lr68ORO7+pc/b37mI3iSTi31T4iCWE
+          iU4Xmejzagh1n47L/uIbMg7fCFtGXvXs5/8eSBTCinr7WkdTuI4FoTsH48N2KtzldgQbLU1xWvn9
+          3p32tuBLe2l7Idsvertz3OkTrHqarngRceYPm1Lt7670LuwiNq/+G/qLxyo3hPkoXlAKKPvesD3z
+          12pd3xrd3D71v2G/q6Ytc0vYxg1PXsXLiTpHPnrgICGjhm99+tGttQB65baWSC61vuy6MydN2nnA
+          dnlD1fiq8g6cG5jUOg7UXWoyG7D6u9Se48hd3rQFdC4yIKv/1nPF1lmk1e8nbDgnjNf87QC34Vv6
+          8qaZ0XTPGw6qPV7IrudqtgiiW8CRegu9rfi05Kd6QapiX/DpcfdzqtS3BoLdIyFcz5JqHOIsRWu8
+          wJZfCPqSqb0GLZMzbL0udTX9fX5oOQFB5FDlq19aSpqwv5ElKd8uU+pbjWy3cPDaLcedwT5L8rM7
+          ttQ+u7hfyI1O6Hl1PYzVixuJ1TD66HnDD8KF4tDPQdIuf34Cvr6bSV/y0zChKj23BCFjQJMtvgJ4
+          5E9CT4nfo3F3ODty7bhATVY/3N/pmrbQce/Wh1v+ipbN7HrQkl1PeN4/6KwoXe8vfvibbrigiZSt
+          BpoGjCYrXk/OVn3BUYkRdclBz6dt+Qv/6ZG45p+Mmawwod84zJ+s+qVPf/gw78jdF4PUcKnWaj68
+          1N0Ln4tljua/83MTsoEepq3W//mzu5Wf+GCpYzSv+R70Ntam6MkrZPPKT2BW9IKefvsup/vNFKJN
+          PDG/0jui//MT6oXb+JtV74rI7wjcDy1H5L3QMXZM7LW4JatXfOrcJZWXBKbC4LEK2ZTPF+H0WtuM
+          3/DTPITR+HZeofznB4iK93b76TeSPbIamYBSI3eo5F4BnicXbJ2zkzuahw2BbDEdrHzHqmerP73z
+          c7Nc8zUlW36Ss4E1n4W9TewgzvvSBB5xQvAhtLqqGw/xDj7PriLIfgv9oklTjMhzLWnODFOftFO2
+          QWkuJlgfN5o+102cwXNnJf6ff/G3/3KHkovPmTern/lQW2Sv/Dxp9m17tMwBmcBaB5DufHLXp0Ef
+          Gkk0MvyPT8/UvWUI/GqLDfdIquHxfBaopOsA46Bbv4WdSGhQjwH981dXP7xGf/uvXRIpp1d92wH5
+          pjpVshqiJdGyHcoan1CDv5fVPD3rENnhx8WnQFaiUTdJjVa9jL3uN0STCcoCkAQiddfzzJzyCHsQ
+          seoDmMidle0pgCvID2x5R7US7gpz5DTnE5zAeKi4bqW43XJP8CET555a24uwLw7Cgv3wPOWrnjZB
+          eiDkc1eTuout/hZwfKMjNVMCfSrnX4g0bcPomq/QF+CrAdS9+yO17L76Nd/FQUeLPfX33zGa/vIX
+          GxDl9fdpPbevVUW+hoKGzcvu5K7+57/zuD75tCPxHf8IKFr6wWpc+YgtSmrC9+E8/Ynm72pytscS
+          PlAcqbvRdNR8ierLN60wsTbOW7aE+cYGK9mP2D+9jxGHnq/XXz6PMIrH6ofDyZYdxZaxLS87l5os
+          NiCYfz9suF3ksmKzlig7zf3PT3Pp6qfJnfLRcLgUml6zPWvlXzAr9LTj3mx6aKhAKZYl7K14wYT6
+          niDzvkvX/InO2F14GJLanWestBucL3QKTXhWWrvqnzqf3Hys/5uvq54QzcMel3CF7YOqYt5U7BXs
+          DdiHO5dm5VWJeJVUGnhpoeHzn776zU0I6R219ODfhn7RsR7IAisOuAAT6V0oKhvgx+3dZw/1XU2z
+          v3nArX1syOaqitG8aZuX/BPmnc/9xZc/v/978i74rmktmx7Pa/HnV5H+Ky3sT//t/vD34ILkEvcc
+          26BuNwlW9n2CRLy/NOhePjLs5I6JZs7PbAh2RYLPqdbl//QMJykivTTc4k5Rg2tY803/+C2DjZLK
+          4D1HAvkpyOfL7V3KD+Y11GBK4C7fTVUgk5crvxneWs8bI7PlGxdnVKsXIfqnZ22jv1Mn+2nRPzz4
+          f0oK+P9dUiDW0YGqhveqWDtoJTqLkkO9bDjo4pnuffCrPqRKEqU6w3skwfhGF3+7uesRd7k2jWy5
+          MaEHo9+jzrFMBW2u27dvdqLKuCrMORROZo+tabzqwvzZBeBTzsbm8CP94I75BEd+n5Nt/xsQuabq
+          CypYjmQie6QPZwFeYJiiiY8/fuwnmtgLdK+iodFmP/VTfpYmCFioU+VwTPK5lMkGVc3ti9WXQPul
+          3t0aOIs7h+p+WbM5EWYNyvdrwsn7NEZ07ttGdr5479/ur0afjSJeZNd4PLDZmKyf3DFawGnkI72e
+          Xyd98a/tA9pA5H34bh5o3lu+BO/JC7E2420/h1dHQ5DffvipyrHLX7rPQ+ZPikpvAqqiZREsAYIh
+          XSXnOYyY+40KSPkpxKa7UMbu7b3ZC9tWphdefKK5YK4nj48Y0eu34t2Z/5w51N3zAfuXbpXsTfCQ
+          k6dGqPFAMWre+jVGWfrb+NvxeupF+yJocrJ/OtQ8FXU/Hwy7hNEafxjvqMZEbpYf0mcfq/5ecHh3
+          /MWPGCntccEH56zlnHTqHJDgamJTpkY03+9bAByPCpFyNeqZkFMf0Lj2DT+NX31C2Y7I2imxqFuz
+          Agn1vqgB93FIVTtOe+HH9Yv8pVXic7ejnU9LcW/hWJoh2W7eOOfuNjKAkhHI/JSPuiCW2kO+i9cQ
+          +8Rzcy7Tfz64ZLen4XVK9NkGd4Jt5SF/eyMF4q9R7MO1Yy01NvaDTfWzS2HP3RJabL+0H0AIOJAs
+          mVLXPqhI0MuTAa+7cKDhAx/0Gb+OA0z7ocFOc7OREJmDAds9dzoR6b0gepwfNrI+r4wqZNZdEarj
+          RrbxcMBZ9G4i4a0/YxhG5U4LFtq6kIvTS75cjY2/z6V7zm6lTaAzAwmn1l3NhWEndWA4NqXpgkwk
+          NFMhIdEMA4qXIGI0iG8mrJ+H3YfHs/nRmyXQ/FjRy7Gw3aVTl0zuOG1LxLsi5fRtex0s9utG9t6j
+          iObf6y3J1+61o/fG1lCPEiEF15A32FCLvh/ypxrLuFV+OB4OCRMlKfKBKUSijpYHFSeC1wFTBglb
+          qRW7XaBdPHA6RcP5zDdVfessDX6+caPJo1sidtx+B+Dv1ZvssouOxNyTH5AXywcb6/9ntZe/YGM4
+          KRmj4JWLt8odJOWKNLriQcU+X9Yix5vvOJRfucuQXXNg1CLCRliqOWdc3g/ZC04pPmzvdiQoBfLh
+          zqk/fPwoLGfuTd3J1wg15KV4hiuezicfinN/95d6ef3bL+l7MJD/3LdHt/aXsYPmlEvYT56kmg6z
+          2siH8Bpgl4Z9Nc4RpIA3Uk84Gaf9v/URtp2MjeszqkR2yRL0np4amc7JB029sE6+Ht8RPfvMysXi
+          mzXAS4tJAIpOX+ZbTeSI25ZkOJZVPoUH00aCL27J9nRZ58yo6QOt99+X1/WZHnXuy2oR3HCWebd8
+          Uh9vGzakYb6k5VPP3/0qkzND4PHpmqYRt3mfGoC9S6hXlGtdd7NkQK2vg++P6dS3no02YOzcGnv5
+          T8yZp+46OEzrnIHTjc+nW4cV8I7Gh4byKllKh29kfHa+/nzJ/JwIl80O7qJZYqP3pmheyPmFTI4W
+          +BTmjcveceABip57wk3S2R03vFWA3TlvIoQXB4mRe05lQ+wMmhW9HQ1bEjvyx4COesFpxW/02sjF
+          9vamD095MuETc5O0u5g6Wdzr2mdLnEq5eoYcNe1eWPuiHjWxiGWTXo6vWJ9b01vgtHxrrFXyoeID
+          EoXSJeF86lP1UD0A2ya63z4nv3eDtUp9TY1mBsfTpyjZkTg97xr6FFaBFeFU52K9zsUQ705IbweR
+          j8gNqwIkZZKukzSlnib3mZOXJ0JYVce5mrrOt8FqOZ1m71fZL4WpObAfq8hfeHizmeazKVutoFNz
+          t0nc6aiXJuLrvKbYLip9yQcK0CB36z/FA9GnTMhqAKt5UfO0vlr+NZkH/P39plbwNdHUJN2A9i9O
+          pskzV3qufpaZ/LdeYnKLcu4Dhxdo6DsQgVN1xremN8HbrR3qHNkHLdebV8L1+tySmauGnHC9YwOR
+          HUI+7W9x59NPTf/is988n14vWCrnyEt1+mDf36r5nBKIwdYrTLG4ifPW8ukE5k49+CAtl5zH58mR
+          l+JSErEzJ30ZzWMIx3ykOFfmdO17f/Dkev6p2L7VfLWs+CI/nH1PQ2Ef6//wS003Ez3Sp+by8vUd
+          y7fTkcPJowsj/tsNNcz6LsRP38f91BViDWKtXGj8ENt1fcpBPju/D47P/NMVijgaQBW2Hxq72yaf
+          LP87/X0evkBYolm8X2x0jVqBFscuj7hjsFpgyhIQKbNPOWf/jB1y+ZdHFWQ6vbgLaQl/+2fEgabT
+          yL1l8Jw+b4Jocq1I1g8biEPfwueoffeElx4DUEOZsTvBK5+My+8Bfahy+B6fN2gxaNTK7yZKiBh8
+          TcZd7EwBt/QrXz41PVu0eFPCo+a+hNov7IoxZ5twUiMTO+yb9NMjxS3M3Eah9yhQIvH5vEiyjMuA
+          TMLJyFubT4t/++Ubr6Wa6FH29lVYP+hph6/u1BXbBs734YGTDp0Y66dzKm+IWPllLXgui3ndAF0F
+          jB/YuuW/e77fwN1UOPxI8rKfMvHeQGLfCuwYh0Mu6ry7gbpbNuu8WqrPp37tWxy6OT18J7efzG1u
+          w4p/dOUjiAuKZwr560t9zjx/0NyNUyHXyjciyMxHlwnfZYK9buywng9NNYSGa4N+KVus+iLLx2S3
+          VdDzVhB8Um9FxMxeDWVt032wGet9NTFhmZAvnnpqyOur6K61HXk1r2l+0riVX92EfeVh5It1gar5
+          D/9vDafQ21UVXSZktSc/cbWQJREpWm7KFMjcIiRYjVq1Wpz9NkUb3jex5jRxTrbxGf74IfW1ZUaL
+          WJ0K5C7PCCuOBtU8h3Mmp7Z2wl5wEnKWousG2u/axeUjvnI6bMoN7CZ9S/be36Q0T/Rh3M1XmkZL
+          y+ZYYRNEHFkofp1FRM4BmXasdy9UZ3caDd+PGAAlfkGez/ScT2F0JSA3Px1fD3iu5tdr7Vp1f6TY
+          4V53d1HFNIHg9hGpzzZXND1SqwMxO9xoCN+YzRseP0C4+CU2VczrY4la2HPsKlClfFnVvFUCQ15y
+          QaP35e7k3OnzNFDRNynWl/aiD5n+9mQ0kQfFqr1H7KKdB5kPeBfbn/ngcvTqTqjkhBs1erteOUbM
+          QfP9NORTpEnFkvssoNMSc/hY66q+xDHfgdWokr/dfPKc9fVJALnZM+pp2zafnVoqof16V3wWuwMS
+          Dmtf3d9w3VAzvCVs4e6oAaPmEU667Y3NT+M1QOIWF3o9uJ+eoeOXA/V2evlisSvz6W4zQ4bHGPrS
+          uBf0eTZHTrptcEn9t1ix2StkDg2Y7cjEp47OmUrio+/CK1h/Tks/3+8iwGy4GvU5cZNTaR94aFt8
+          WvoUj5k7S1H2gOpZ1dTcDEnF/GeyICEtXd/qtFqfOYcjcMmy9YKIj6h9Z0EJJBwi+iBeH7GV7yPp
+          48dUJcqcE8ybNTwZcvDKx1lJ3r/0X3x0tKvR8xbhPThau4bMRJmj5ZoeX2jrGTbNMnuMpvzlJWh5
+          Zjn1meQy8bq/bcAvyieZFK92J1N5+BBxw4L1d2blon8PJtnzMh8fUyqyIfVaCYRtesJK6Cru5LwO
+          PngyGahS+22+RGZtgNdwnb9TxRQte2f25ef0fdNEZkE+JbM9wFfIb9Q7NjQf3yTMZJX9APuHs4Km
+          u/Qu5FYqFVz8xRvlqklgBJNGTzdtdNf7lcqmOXzp471F7IviXQpokn3sXCKmMxRPmbzXBJsmpsxX
+          LMo+G9mTbjpWT5Kq8/zyykC4PENq2l8np9fBzf70LPnCYchn/DYG+F2wh62ro/YCLyUDxJ/lRdAE
+          Ss6Dky1IH7LO3/qOXjHjTB3ghP5JT/3h4o42sifQh1+G1/OIRnDvCxzztc/5yjeX8s4V4GAhocbr
+          d6mYyB8TgJ20odZ9bqtpz7wNmOJywSsfyuePKktwGUmAj2fs9HPYPGqYokNM/XJaG3E/74pURzug
+          R1r5+kKpBFIcjl+qXzaNPhjlpAAYmkW9yfWrWVBlDQUs0PHzg4ZoefOnARah8rH2iGZ38d6Vg4xg
+          0Yiw6uXP8nYAev+ZUe3XNzlrtECQ//Tf01z8iFjK7PzpDRqs50FMzbaG8znYkUXgbXfedRUBeC8e
+          9SF69nMJ/UaOdlpO/V8u6wxqA8DUbR9fNH4dkrbet1Xf4eg5LVW3v2mLnPJLSOaVf//hD2goBpyH
+          Q9tTdgkT+LUv3pfdXe8ugG1DYt8GsPZjj2gqxkz6wzeyD+0A0SMNnH945U04q8jKl9CFnCZsqu5P
+          n80w/BPGb1924epOLTtwcD4rNn2apMxZlI0blECpEyl6Nzmb+1cDhW/pa3zg8n9871heNJ/wtVLx
+          2lJpf/oYH4LdMZ9PnZHBqo/87+/Ms/HSjQWyO/uNLb48/+Nb0sz7E2mt+zsaBLIxd+daPK/+RdTT
+          wD6VMGi1RtPP/NGZ/8gn1H+ud2rVRV4t5P3O5OG2rbH+tS3EGddvATu0OVDHCjrGzP4YAkpCbY0P
+          svvjv1cBVrzDONJ/jFzgl0DzTQQiU4lGf/iJrEd3pYZ5mdmkGIaGSCjr2JDdoWq39GoDmk8bf3wM
+          aiWccTXJK97iuLsKOgu3pwVcfgG/X/khQ/HaBe1+03Dsbs2Is3/eDr7MmvDq0Fd83dyHHUb8Qjbn
+          xs7FP75yH4+Nv3lvczSY28gB2fO++Frsymjluy+Yni+JOgWN87GUG4DUGGp/9+HP+pzsRAXekx+S
+          icyVyyq6i0Gnso7N3TFHy7BbWlg/jx7hZ+hCMz93cFUuJzJrzRUJJM7WrgMJR5XDJqhoX2kCbJ/d
+          m0zV/Ya+xvnroDd2Kqxn7BWxBL/CP//D/+rvgHWr3gFzrH9Uod0vb//45kXxAWuHn+ZOf3o1hrbE
+          Kx5GYxumMVw4USdzf3Ny5ulCK/NBk5O9+7kwtv0lHBSbnUpvXlJHy4+9Jllr9chf/SR9hg12QBWK
+          A1WziKtm9RM/YBgr2SdYP+XCj70WxB0uKj6rs9wvQrTbga75W+xL7xAxwleB9HcfouKt6rPDDAO4
+          FGX46L5/bB4+aIIebai/QdleH/JQnpBahDfs2Ju+Z5IUeXKz8XZ+uxujalauzu4fnh3INaym5hjb
+          EBwm4odyWumTgNQO8Kk54dXPqhhKNpk07sov2RhFU3V0Y5T7NR5Q92Ef0fSt5lC+Sf0Re+9rqvOv
+          UAqhEz8pPV7Vq8uKm6r9xQ9/7dkSTUFxTWUt2L2xj5LRHYmWbECtbh5WW99AYu1FJdxvvEwtZ2+6
+          /FZ7mIi+FwGfVv5OVz2ATr2nYts32mjeNPxO/lvfNNxr/fw0mwHOZ81eS9gf0Z9fJt2LziRUPEr6
+          osXCC8S64vGfXiXOQVFkZTEIVjNs96TeFw0kez8jWz/M+2W0jEAWj9zvj78g/i8+/tqSp396nIT7
+          ewHVUzPWFJHM2BTYD6iJ8cX3m2hFvBndDTB3+gH76/pNno0A1XN6xt5Yf/tpv6gA+WFX0efQbPN2
+          fOkF+qmoou7gLjrbtpEgN9F0x7cq2VSU97IWVj+CbPDOqoT5M4XyVrqOazw+6TPA0UZuGn19xWEd
+          YpEXZzB5+Iyx79Pqnz9aZuhO/b/7HHOKIdv6vKWOv3blscbFgybWT2SihM/Z8xFKf/cdu2emRT22
+          tBL5VVpS91A2aLFY94K0HG++T4Jrzooca6juSt//6h9AVMi/3p/+pWo03frpUQXeP37ypx9mfN7Z
+          sFluMv33ij5iHwc++0SlqnRS0Lr/BnpZ94ra4m7vjvzSppAa2xgfH/ddNOK3QeAm7fd+OeNttSj5
+          zoe0LH6+uN531ui5gA4/HGLbur/z35tkGTKzaqLWneeqf/j7ESKbetbOrxjvhe3uLyXoY33M2eeL
+          WtjkSkavdYr7pT68fLkm5pee5gHcYfXHkYZMz5e+IdX5venu0OpX0tDNSUWuTWTs7Vmr8ekmWjn/
+          UZYdWvGDHi5jVU1JLTt/643X79OTqrM16U/vHi6j3nN2e5Sk7m4D9fHu289NkSuw4hVB4dBW07T0
+          Euo/xkivK18ha7xDf3z0Dz+WTpVSsEJvxn/3m14E00BvV5D8l57z1fwVkQOnrepQ5SDVFeGKXQ1g
+          hOK6H507f4WdDbu2Laiz1a+MmfuQg/2NONiIg1Kfa46GiK8kYz3fZj9V54ADYTBrapUdQYyb5QKt
+          fJ6qy17sF0MePZSW9EZEV33ogiquT67sX+l/p1F02ZEGa/zCm//q5zTniz89T3PzuXdnbTvXsN4H
+          HyZLq+ZYFRYgSTBgbC5+Ph0Wr/7z7/DxN6sVn477BJ6Tqf7lD6pZ9JUGJGtLCRU3XL7MXyFEz1G3
+          fOHPvz/9jinyvPS//Ee0YqQB+nE5TYYx66neSCn0DlKxgod57XpVSaDSWvN/Xjblg7nNHZhceOPb
+          4zZES2GdVr9K0Kma4bZni2Uv6M9fUw6b6Z/ekjc8bejhl72ibztoLxCPwg9rCmzReA1vHlyy3xX7
+          4FrVt71PIBf+hZFP0vFoIZeLAKogf+hTtfds5ae7/d/5w2t+YnlzzuqHKw4+v+mFkV8T+v/umzmM
+          WcX2D2bC+G4/2Lbep2jOjrcSzjtHwkacSjoL4rMhc5QVVEkXyxW/5zFFmVp5/tYudJ3/y8fs9WtL
+          zTXfs3iTSMB8uzrV/s6bvDxKtFT441ftIcuXOFpaGfvW6AO5qYwr29IEC9LC58R7j74x75po9a+p
+          K5ABscN9IUhRpuAfnyML73HobF0BG8IgV78YHxKZffnBF6/+aY1nhb17h/JI8evkIrKoQSFnSyNh
+          rBonNEWPNgSadTo2XSFHM6J8LRXxkBCwQKqWm4pC0E6xhdP+eej5Pz2++hN49bsY6y4QiLeIWYQv
+          4q5nlYIB/eW/7qbTVZ3U3BKkLNcHPmCfq6ZvtQ/hcWYE/+VL2N/5WfHjL3+QL4JZJmtXFIveErl3
+          m7d6i2FdX+rORhqt+bgJ/cXP7Rip/fpS3wfPQwQ7knmofqs/gNTifcRY3bfu9C6SEsXQldSMwrs7
+          0zWz+7HDyBca59lPsaRk8uNjXclWVVQmVIOpoYs+X6gh3nu26uNC+uMj6r4I+4EPsw1aSnLAx/mS
+          sUU7/DZQtAElXByU7nD7eQq0SH/5uyP2XS5/4xiaoOBWPQxsqgwtlPnzAeGzdTV1oT60PnBCSrHT
+          QsWGix1qcLhkR3ya69qde/fRwV175FjdOsxdJid14DdYvl/+aBhNkhI1//yeSFK66BeMaQCUv2JC
+          2EZEo0VkDzhNTwnapWsJbm9lIHvPK1XV36HnYPfgYO267i8v9cyGxfcMqMnliiMuuTAuNFwHbAz8
+          +qpdRVPLTgIiIYnIr+u+bEhz/gHUOL/9eeVzZLRx+JePwF42fHRWbb3Vv7+E2N5cbZ1b8Usus26k
+          vlE/KjY9MgVKferJ+015NNvOPUWrH4wdv1XcFR99SD7SjP/00LjGC3l8GO4/v5Bk9yiVA8270HMb
+          Bmz5Qlv886/YyCRE7uUYosVmR3/L1NIdzrrtoD/+HwXxuZrk8175p68USi75lD+P8V/8pO4ESsT/
+          8dWiDSk1zPOBCc5eTOEPTzLuJeuTlQsxWva1he+mn1RkY/upxH/pifrpsconVAiDvOpNnCvjUBGx
+          OjzguL9zPgSnJGL+87HAT+0YtbNLxZbf9+LJ9F2a/ka6m2z2colDsGt9fE0tTmf2ZaPtz2iMKGaq
+          pot/8Wf9fVhtwwCx5Wgu6C+/q+PcdfnMKV5o2Qui/6cHWDVvYpiepYRP8bFB84Yd1/zy50Bvj8sr
+          n17KJkX75HcmbFy7TEivtyI/L9KFagr3RZM2yuQv3lKLoC5fHpQJ4HTvEh9XvJ9tpEyy+dpFRCTb
+          JFqezlmD7n4b6FEunahvn5uN9P+UFAj/u6RgPGcS9WTJZNzhIXrSyXwZ+CA2WiVIdkrAfIUVVUdd
+          rYahkx2QTFmijsq20XycziFcatyQrRjFjFl9F8MQEZt0b4fq7HDJY/CyMcNH9Sbos8J5L/S87nzq
+          ztTrGXxZLb+xEVHvOZYui85tDcKGa6m3f+n9uG9UD7JLneL0uDn0PDl5nUQYSenpbRo557QHAzpi
+          YxoXj4e+1K0UgjS8O2zYRor6UF0yeG5Nl57416+aPOtdyjr+qj4f8bhnfLZOJ1tMhl0DlXoXe8pG
+          rnvXxg5fczkjYhsi7fpQaNItmj4Fcwgoi/fY//aem49y1GxQUEuYyAxyd/l6DoBwuVxwXnyePb9l
+          bgJTcp5pKpyyatr7PrffBM+G8IBFfbq/SQDf04NgVe/UnB19pEFSBR3N7hc150Z9N8nit2X03HeH
+          nCvuiwLXPO+JLD3LaBI4vZRptrvQvIfRHSqpS8D/TQ1drVyXK3xXADU5NNjhPqgndXIt4XZY2xA4
+          8U6frehXwH2kPx+Nl2PU8Ye+ReM+UPAxq3SX+/ieALv92Sa/w8btp/d9ecl7+7vFbvHZ9tNcxwIo
+          RXzF7s5Sq1kZKyIParrQx3YMXU5J7zuouDbGB3Jqek52clO+Hnj0b/0mao0KtMcqItJGDvLlRbMa
+          kE5b7IN1qcT7+d3Ib+f+xef7XUO8whklIFPZ0LR1nWgxzSYGcr8/sGNB5YondZbgehAR2YnXTz/K
+          z9dOVsZzTh/r6LVhE7Ad0O0ERCoSu2IjCXfy6XJcaJT1abW0xyQFuhgRLgLkuNwnmFNZIqJCaB26
+          bHLag4mWRDOp8iTgzvZv95DbSvniR7P/VXxwOrZwW0e7R0TwGKfO50LujBbW9U71JXlWGZTFfMB5
+          imqXw+2pgSRQAlq8rEMvtHjro3V96A22O0S/vqCBqFk6PjVPJeJaqRNAfQU6LTyr6il/HkwIzVeN
+          j3JNItLRqIHikXak42su+u2er0xeCTbZjXUTsb3rTmCaDJMm4EhF41xuQbi2PU5l56ALRzgT2TTu
+          e5oVbhKJU7bbwHXaFmT6vfN+vpk/QWqaYMZK/nTdpd+qi7wIskPvuR1WLFSXFGozeNPH1i766dGh
+          DoRuf8DBR3zpY52cFZluDwwf9C7PhfhYZMg7Tgm9ZbIZCQZzBeS+zjrOJqS4fCZRDvWH+IRDx8zd
+          +VOFHVRLvMHaTfR0UXh8iGy4feOjJn2gaf961/IibB3syc7WXfAxkuS9/dkS6Vq+kWjYtwG0m3Gi
+          plPcqim8z5pcWfiOre2sMTbFqQPnKnDwMx5GfamT5wvERGv8cxS8XF6C1EO/abxhvx8nfdnTuwHE
+          bkTylnrSLx4NGpk7Bi299N02n5LN2vjsu0uwZ8aGzmmbyAQ/eo9YP98JarlA2u14q6v8LlskfXnk
+          OweibSLgPzxciuFTyu9geON0yPKIzb6o/Vuva/e9IBFmewO2NGrk5yWvSogOL0f+wzs/XfvKMFEM
+          YXN2KM6/V7GaMkIzOCohw4qKJDSn30MK+zr7Uf0iLu6c7tfGn4sZUfd6vfTzYax9+b6WhHiu3fWM
+          qnIKx+9jwup3lWz7zArhCROP9TzwXUEMPwG6dQ+ROm8H6wNxzrX0GKaRYhwFaNz+eID5F6b0aFaq
+          +63JWZLP3MnHKuqsnHNVZa0HVgQa8f6bCXDzG/GiKDZVh+iG5nL/c+Dy9Q70fL+XjMdmK0B33Sf0
+          8uXnfg50LkSHN37ik/f2GPdTdYKOSsCwtjMCxNSwCeD+Go44TvpNv7i3YwM94lWqLYLJROGdhfCS
+          C3M9fy9XlK1Ak9XEaqiOn3d3aq6ZD6BZFdYPg8xo46qTHDqFha9X7xj9uMxc0EEsUprR67nnlbEa
+          QNBXZoe7XzSJ82MD0t1jtFjPJxNnu4DSFkuqHQcRzZqT12A385O6x1qIJlEaYhgr3aKq/qT6EitY
+          gclqdXo6mWXU2u49AENgFvW3NvQLlNFGVjJdxf7BbPM5fNUGFNUmp66xjlrVDQRw2vQHn3e0KWco
+          PWWwaxNj/TvI+WN9jYFPv19syu3NFco9bcE5KS9qJienF7yCF+RdoxEyGxGPvpzRFShXlAlbUmqi
+          5ZFPtnw+TgI+ttsEcesTMVlalhofnFPZE+v5lpBk/XQceT87mlzVHv7OF3mk75gJ50KIZevARTRU
+          IsTGX3Iu5bd4upDJ3Xu5wNupBybpLjgGRdGFu6eCzBPhhY+vROvFheQpJKSX6aEUdmj5XoJYTtlG
+          x7Gvdbk4nasGZOX5oNd+fqG51oVA3peHCIdVYLu8q7NFEkYvp3k6nZl49JkCrDR6jB+x5C7bw9DB
+          TfvO635glztbpQatsHGovaQDmpEgmqjOfIMebkzNyQXLPjI+nI7PsWGy5Xcz6r/4hzEzKjQDloZd
+          n5zPuPAsvV+sgo/lHx8f6Gkb2YwT5weAYo09VdxYWBtzRg+A7euGzeBX5ALrX4ncEQdjT724K195
+          BfLP2vs012eBMd5OiHybtJa62Daq+RzIO/ie05Dat7qO5tq5GkD4+EOuRcbyOZqyEDkfrscJa+xK
+          EG++AVcWLlgjhd7/gpPaQr01e3y6RgabEh4TkEOrx+r9XqI53b8c+fZjxId4EtGCj7kEyI5j7KnS
+          Cw30mnuQGlggjAaBO1m86UGaujeyZ6C6SwPFAIqdcVj1z998VuXzTlZul50/ebbcsykObPgNjCey
+          hyZ3XOMDtK/x7IuDa+S09OUSKXbKYYUPvYpwtGgA3aLUR20SVAuEnwV+c1ZS4+4fGNGHXpO1Tj3T
+          08rfRFudU7TGf+pQ+Opz4lz+7RctlnVwkre4LeTPa0NVqfd71tv3AcblsiGzPd7c1rJ+CqjbZo+P
+          H+eoz+eoe8DP8WrScNsa9W55KeAGloF9XlrYhMbeh1YABxfu+HJnVZcWcGrvsd5XjPpCZRlqXe63
+          xrOny8wy3sBffLnvL2K0mDfVlFZ+RWPrsGXU2BqB/I2054q/1J3CuGph9O0nNj8+Q1Mgsgl+dIrx
+          7Xk9VnymyyGkxXSgATdeXUIXpYNbKnT01B0RWlyzEYAf3g7Nv+WPsb1RbuTXWqI3dX1TDX98domG
+          yd+qpp8TyJkN/OZ4pSF6Inep2yWAlS+SiyBIaCz3P/s/AAAA//+kXcm2sjyzviAH0kmKIb20CQJ2
+          M0BFQVS6ALn6s9jvN/xnZ+jaeymQqnqaChWQn86ZRrm1R0JZbDm4U9vAnJHX+fwa6Pra9m6gek2e
+          /mTo5ALUjZ2V7+rJQqXnHR6bXCfh43VEvfesY0Cz98Aiv1mPIsytDVxMeUu0SyKiCZRrBhf7GRAj
+          MGPE7dFHR4kpPyg5dhZbXtbzC9fOU0lxlF9s3nuFiUI4U/qHZ/OWGSdYDnk2Tvbx7k+bsjXRIEBL
+          3CtDyazND4DDlHSrnukrah2TVPmLn9CukDH3U7GgkfUXqqmmufKTX72+hXMeK6tX0OxkXrRrDP1M
+          cTzy+WxgIoAneRr1Py+Xia9nFIHmmhYlVjlXw9/31VO/o5Fzf/ns9BVjOGrfI73MG7eaF2O7oP3S
+          aCNEp7Rjysar4fVKBXL5qpJPk0PZgLJLvsTWMPnLTwk6QxeJbsVpNY+nTEV5Agm9telozIshLtDI
+          OKX2ke0ToembEh3G3w2Xp87shG7rToA2ooN54PqOHeijBowyhuX35ljddA0HyPgUDtVcsTaGdX3l
+          j1QTen17P3/+TEWM1oYcFmz9lMx6Um4U23zs6EU8W1UXvb0Skou3o9qzZ4gZ2iQrK3/Ak/Y8d/2U
+          jl8QG6UlhTodkaARFqBBODypXkVff9pl6qQYzbSlD80e8++qX+BihsKffqkmq509RcLcRC8waWyO
+          z6dRIeR0Grdq3xnj+HjLkEXbnpoXEvrL3T1mSDiej8RwiOVzbu4X6MagIod1fdnKl2WQ0fBP77B8
+          CT30Zdtu5LaR5I/vJhzRil9Ut8TWH1ua17DGByHXklXt4xVt0Efc2XTVU91SFqKAlM8vp7h/nhjz
+          7utUi2eur/H7ZtMmc+N10JC+8qdPzgJ01eF39/ZYuuEtmomqqqCdB0bddJuywRrqANxwW+CPk61b
+          SPpBh6tVRyP/u6p+d9ZkSS6Vu03t3fVrzJ4gm+hPn1pJlPlfpgU9mEFKx13pT/m8qW4OXFWlpnaZ
+          Dfmyk9oCGu3iUXPF23nVa//0s38G359STtWVJ4e2NFjxjk9VtUR/8RsGzx5N8k3RZfnpnbFSnqAa
+          vspNVpTdZsYKh1JDCOilgQ9TKuJ6IXQ04+YR1u8jGD/MhE8FZis+fe8wTzTX582boIJ5kEqSX4/7
+          vGdebCvek2z/4ed4XrcAfJqNTla8RTMLQ0dm3A+Ik6qC3891IcCWux4wv3xcRClKS7Tqd+Jy/pIs
+          7aJ6Cu34K3GS2UTTlXNKFDvOTG/Bs2e91c4OaPZCiVvNPaL98XxCp9u4I5aohf5fvCoKRHu8FcZt
+          MptTtCjx3FqjcnKviG2z5Q7KDmYS+q+EzcX5PAJvbzPcR6g1+jM2Utgkd4tqSVQaS3+8cLBU2mek
+          D1NjLF8sD461RPG0sZnByDFNYcUTWlh1kf8IZ+vKH96aN/xGi0Mj8w+/qfur+4Q1W7IgL9qYeIoP
+          WjII7vUCdmb8iPOZGzRbXKgDss48sf18Sej9eBSQrovpuFg3m/3xPwjhSInZmwUbau9hAncLfOqW
+          fpTwX7tLYfB5l/i4IsafvoZHiiiGVf/MT6amYIWqSg58f2Sc9awmKJEzkag3X2w5PFoZRffuSL1D
+          HeXi8rZVpaggp3/6dPVXYiiiXiPnfdNWvZnrG7QcrtnIbOGLxlUvIeXT5eOiJjmaUH3j0FoPRuHR
+          NB17L9gBJZ64kV/0NJ9TJZdgkaVhFGqCjCkbPxcIyk36T8+K1v07Isd4WtRPT7jqhaEd4aF1x3H5
+          3VrErHs5gvv4/IgdcbhaLH2xYY1X4gb+mHdb59H/e94Smh3Uvqxn+49/7Ztw6f7Vv7UeEDOQtjmt
+          7l0vw+VU4GmUxa5f/TvlxX+iceOWXTUZ2E2R3XUC9XvUGCx63QCmVrpR3Pc/Y7rU8UmpgR2p2eVO
+          snyJiFFUJhW11nogpG56gSr3VWK8jbDj3INgwy4MOhKdyiVhVtwsyM76hdxOO63iJ+ZykOJuGLe/
+          Q13NtsGif3iubmzmj+f+rUIjByk5zrcNWjbrYP/H1vRpprFtPu5PRSnbmfYb7/HIJ5P2PpewyPKA
+          4Uqjiv7xrxXfqOZIfr5spHYDerMf8dOwOrZ0PdXR/SJ6xBE+Cpqz0GvhKHExOZ3CtprN6bKA+3j/
+          CKl76vfXL26gl2pK3NbV2B9eKormfcZX1Xb52GzdGPZT3I1bvE7BOfNyBl4E5j9/Uhi/KIbAXU7E
+          b5Q0GSEeFqQ+ZYda8MvyKfRvE/z5Yb5Lv91CG3tEof0yqXO7xgbr8aZEf+thJTytZifTI3BMjElQ
+          n07VlEfSBVjVS+P2Zy/JHMmjIHPO5z9+N6x8XCFRkhPD57VOLM3vHSyvb2kujO98OBVTqizvk0Kd
+          ryoZ01SW9T/9GjycoPrHv64XriW6J+jsD9/B0SSJBO9ynQq48uu/60vPeZPQEiUqIFvfUNsTgDHt
+          ureVs+OvWwLGPJkXu0+BzxubqvLBSv70O9z7fU1Nx5TYl2nmiOice+NvR8tu2iOqw3o/xJzxns0D
+          mTFolvOi93i377jW3AL85eurNjGijtQBouvenfddX3IWbq1YYZiLsOi8q275HC8p3OZfSNWVH6z+
+          1KRErFLGx/LAyUSe9YJseyYk1FPPYM42mOQ3zDeada+KjSseorpzHSypb7ebmx8r0Jme8YjQeVf1
+          r+Hbw/DYVyPvhXrHrfwBkouzG6uTbibT5pz2sH2UP+L8vMpYxscgg95YI260LzMm7f14gT8oPdGQ
+          NbM+/WIbrXz4H/9njrbL0KpX/uH16sfjPz+CYiSZ+Vfysga2VRES7c9vcnhJksHuCY2/98Xv01Zu
+          4RCYAT2sfvLk1hcBfRrQ6f557Q1uUS4XFE8ABJt7Ug2pksgQTlCQ3PTlfO5+CEBJ14O16OAZc79r
+          PbCGBBFNsHf+9BM3Mdw3F5m63rL4S2R/TjBwXYpF/jr67Nh1HNyvr4TiS783OMVLTHTZnk/UiyvX
+          WC6QY7Q7OISk63r/PC58ybB9XbFyoKduqa+BjW5jtoxvov0MNuOtDt/xKo8z3/Pstc5sQntd5Kj9
+          xgliq96BzhhVinN4GPT6kUyoz9JIDrp8rbopvTiIz2t79V/cfJQWrf3nHyWBGTN6HC8XOCwfbZz2
+          1hsxaj4BkqwgNLAorsR4G07QnSAbG9Y4HW+1Youc4/omXSKVyRx4aQ0fMpk00Qurm9sqcZBw2r5w
+          mrdi/tTm8waS7poQff39P3yB7rO7E+Pw1QxGUfGCLFrjAX5qIoT8nKGAnmzqVGej42+7sgXr3obU
+          4955Nz00fUTmp4jIbZ/oiNWvx+YfP9cX++G3PZZU4JfFJGpzMBJu4PQXnN61So0Hb1RzHfqt9H1M
+          ATmbKjbmP340TQUmPh61/Pdk6gnStjjTcD80Bl0+cgNFNGp03yxffwjLroaVv5PosRuTue2/JfTh
+          A+OeJSXqb3XMKdGr3lFdlQ75ctHBBryzbvQuPL7J1J6zHppKyVe+m3RzW8gn+d0GAcnDzbVa/bA7
+          7E/sPMJAl4rNz8MCaz0l6k58GX3KqSo8jORFrC85Gv/8Hc55Pyg2hsBf+0NfWPUDLtf7F55bdYE8
+          sAKqj/r6isV4vaAVb6nmxrijf36vFeoqXuMLrXhsKqdDe8PdXMvJP/yIrxZP3VFxkn6/yWvIYDlS
+          Eu3FjlnH5ASS+EV0nzdBNfzhyxCbBkkORtexDTrpSovcae2XPZLJSV8tvOo2wOKT0Xx8ECEA2zXf
+          IxdIj3ya8/0dltiuiP557PyF3+oL5PXuTF3uVFfj6rfDk9tt//FVXk++G9DMx5XoonVmUzSOGfJq
+          fCf7PJfzNmiM4E//0ZutbPIxVYmOyG7/Jo57jQzx7CQjnCQuI4ndCdXKNwvFH7b9qKx8/E9/QIOv
+          LT1vPiIadhmJkPI91f/wb1S+MMmXYrEwK39tPl3yYIP8CAJyy52lGjuey9B+UCOiW6Lnc/LHu6xT
+          OboRahQY4/F0Wg+SKkq8KIZbMb17CspR1R2633xENoXYS9EmnoDuH/ytW26Q9IjLwp4S79HmlHUN
+          h7Tudxpnwd4Zsy0P3G7ZxSXdO9sGzc5vKuS//oGnK17O31W3hDNfO1gp9l012eM7U4xDSEjCY43x
+          Khe8ZDXvMF7Xs/vTN7sVP4irZrd8wRBj5TJQl2A1fqJZVB4XSMzwNL6yQqt4Iiw1PM6htfL/gE33
+          795Eqx4b2eGG0eRraq+8f0+DWinnVDxcVVuR951BcRpXaEkDB0Db7CZqSdkvmfmS1mCqUULiyAwY
+          3/bf11+/jLjLIvis3qgmFM+fhXeP4JwsvdW2IL7Pt5FD2sdgO22TASR0ImE+OP/xVVOY98TQxMjg
+          fXlMoZTEjFpfwvtjAln/1y8YxYDjurlSPRXqXDgQs8+e+TR9tAxuvnP513+ZBM4v4fErc3IcmJHM
+          B+9mArs9NuPmlabGXEbbC9yX9D5Ki7KeJeWdaiRF9w1x+izP++J8wOD7l5wc5pvJZr781Cj9uhoJ
+          xPO2W5QHvcCqx+nq9zBB3voxJKb0oKeKFytWv87wz7+/1u+kmsF+p+Cm+ETMCxmMadV7uwV7Wxrq
+          aevPkdxwyl88noJtiaZI/p5A6BqM1/5mxf76EW0W9CN/+QlJGY3NRVGL0/lv/ZMZ38T1oJ3c+Os3
+          Gl2fSQXalQ+bqruerv5YC1DqaYOl1P8m7Ufna7hlD4LF5zXwea5xejmqiU1w7D7QfJuSCHoZA574
+          uO/6v36U7B4ieneQ4LNfN3Cgstdx3KZPma1+YgHLSbVpfN5l+ayFqqe8SqOk7tovHF+7T4vs7icQ
+          fzsXaJn1TQ3DY+nH558/0WqcDOjc+RTPx1syn5LXZvf/2VIg/u8tBe/usaPB0eMSliIpkysvpON1
+          ubvJooV3DDR5OtQ6XwdjuoVWrNTtlsfQtnzF6IYIwLmmTvFg3vPZt7Mv/NBlpu6+sBOuTde3cEnh
+          kZsnnyuxNnYFkiwhofaBHHJGcGQqyTnoCfl8/Hz+bKEA0ToRvGzcyl/Uq1PC8j6cyC3eSP7M5coJ
+          +Ie5pxet1n1B4E1VcYtGo5bYkbynorTAK/hG5PpCG9TT12WBjyu/x4OhH5L5cGe60gWqQOzsM1df
+          Pjpg6Jd4M7Ll3VTLWdY2CryEZmyVacyXHF4RHOMNT9OBehXtnUyHbV0w4gTpC43qqN7hEfOPcdEj
+          Hs2G8Jlg71xjcuKfSSd+rHlUrqZi0fNwKas+zEwMb0e6EfVXEraEyVYF7mk8iFkfW39wTPkOsMCH
+          3ti26qZCtBxlcg8+vfsFQlMuDCa6HfgzMc/3bbV89MpW2CZQ6Y0/1/7YZ12GnkUcjqNt5h3/UlJP
+          KcPta4T4zvlL+HiewP/efaKZ2YFxsLnW4D60gpBB6316WXCKJCE4kf2BHnOBpFsPuLExiEWOfcWy
+          XzQp4a9ISOQ1IRLUq/pS7PCAx+vz8zDGealGJV16RuN3cPHFfD0oIX01C97I5cMX4Y10Zfc8ZeNu
+          dn753J2LCV7ny+EvXnI+LZMA1vUk6txEleA5Pw84nGrkai52x/HmKMHO78/0rB7SZLFsh4PjBXM0
+          GPsjEriHv4HTzkREP+gnNF4bz4F76ek0eY+Qj+bkjDLH2Rb1vtFgTF3aBIqs1QW9sUhOllP9AygN
+          8iIG29m5UKmtgKKvj6gxC0pO7SR04LWPIxq2b8MXCsU7KRV5jMRMl0fFwUIjeBXTgd7AMiuuF6Ja
+          8QfOx4vQXo0ZnVQPnPv1TO51/snF00F6geF3Gg1VpHaclnQZqHIp0UuTy4hJ1TqF41g/CB74D1u2
+          pB//4o1e1cn+u94NzJl2IMZOvrEl+Aob4DaCRfzWmXP2HNazqC7PZCQXvq2WMBFV4JtdNa7X4y9m
+          rk6Kd+ZscqOKkXAnzlwUfpu6NG7rS8JdHr0MVXubyV62JH/0PgIHKK0QVryj5s9tx7VwfQwWjf7i
+          WfL7AEDpPzTMme6LXJCVMt8n1bqeU8WaHQnAVn6H8QPBE3GV/G4g7NBIA9ipyXytkhpM+LkkNxzc
+          cYMwfaH+lhm5DcmaD8EbA677CzHv6Jew93X3gq6RJLqaYWyyu6pXwhqn42d313KOvzaCIuMnj2vR
+          9X2xnXecTPjTjdreY6rYvCUbcDZzRgKcJdWMOGUBSb0nJHxzu4rld1zC80o+1HT2Ts4bvWxDKHSU
+          GKDtK7FfdgBKgn/EjO1tvgSKIMMh8TeU1IrfiZ4a3oE3sU2c7RaqefM8qcBVG5NiMXQMWld7B44K
+          J4y95Aisr+dJV6RvNRFrSymalqFpFSt2fXJ3fCufjK1RoEd4u1NHVt8dv3NvI5iH3576tDxUHA2k
+          BYKdHWGm8kYinj9SDc319SKJKVZVfS2DDRxvNUdSm34QS6Iohet2u6dOdKp8PijiXnkGqk8jLp7Q
+          5F9NW8n9Y0g8UQ/RvBVPJti3KyFHsaPJQsrfXfYk2yE6hUsn6Ph0Qn/r5Vmph2hxvACwqSsoPn4a
+          n1Xj4aKIyI4pzrZNvqS9eFecsbpgvuRHf+pLqYHP9vul57YvEyF4uDE8UGbT8BnqiOW/SZAbPD1p
+          3nOHSrALR5LMeefSRER2xyQsc8Cqu0q0xU4r4SK3OlifX4B5V+kTFp6GGujm6WHU3ed83h1VXUl1
+          TDFqnUPCrc8fttReiK5kVyQe4VJAFoUeLWCn5n1zkzDEh/RLrtb5ZrC0iS6KGFcdISXykyUJnwtw
+          EbpTQznuc57cZBldflud6GKuVzPeL4KiOMWXRqVDO/bUDzGU+dehD97qq8EcMgeC8UQwh/wuYZVx
+          vYAUVgM1+IONZpm79dD6rwVTLXqwrr/JKjzxejZX3sk+o48YK8apMrBiS+ec54vbBR6+boy8nwQJ
+          x8lXE95NKNMgWRha56rWqEyOFVUtPHQT/K7fv/gkJKA/tFyQZ0KrTgvdQ6Ah8QTTSznk94o4l6zq
+          lnffAuzmIidnbjiz6cumSBl1ZJBwPhqIPz3zE1RxeyKa4B2NefsUTIghbclxPIxsznf5Vyn9137k
+          2Qezf/X9UX5tepvqLJ9fH/DgFesHGrgm69j5M9XKrmmuJGNqWPHQOwFsuTWew9s3YdNvL4EsBnea
+          BuPWn80n5eAVtBG5bea6E66odQC8ZEPv5fXjz1vDKRTteJHJuh5Vd/levkgjkkRjYJUhLkJ8gfe0
+          RXiSNxES+9uig1vUGtXUy9wtu8CMgU9Pe+oy7pd3ehMIwPtolTTno7EcjtsYvBQ9ibrm03KNWgfa
+          buHwRmcmEl+vvQqC+PySq3+3O27SmlIJ1PxLvZfUV1NwrR1lnyTpmj+20W0+OYYX3rYkPM8a4vwx
+          0GGUOpu4T3U9duJdT0pQNYTGQ3xhrK72nvL3/AzjWSUsQVGqrPk0bgZzkyzY1wpFP8ob/No3PKNq
+          ZC5y1ioeKSRRNdh+s+Kdoidja0Cbs8GRTUSye07Mk/hlw97Msv/w9uyW+STQUYbh1Wl4cqMQiQWH
+          G+gCXRi38b41ehHtIjC1ysaosWe/Ma3vBM9E+dKwC65oOerbGtxrVxCXGg9Wk8mMFP9y8ohXBkG3
+          4kUN3L3QSdiuZ4//HpcLZO3WG8XhonbL2diOqFJeZ1xFW7MbeS/QkTzDg4TOVjO45jZhKJ6PlJrX
+          t5ssKAwyhbRNgYtndPS5zz2K4OfrI+a/u7GaX3HqKHKqibQI7QRNXD2byjPQ/ZE334E/UXGalLUe
+          EV3evxgN+7CRjW5nELKcA6P7ZHO7S+N0wqya3/mKl1+o9lP6L/4nWTncFfblVPLQogdi0S6yFSUU
+          vljAzwn1Xzc6AVqKatxcjk0341smwLDMOo0idawWc3J6eEi3J7VOxmD0G/a5wLGyBbrW44pRV8bI
+          VrrDuDGnHE1st1OhukVHEheVZoj4bLXwsFqDGsmvTlpcapli9k+Hhhs192c/X1r440ePNumqWVZL
+          T5Em4UTUHkf5tIsPBZJ8nOCt+nMNlkSXFA633Yau3+//rT8675otZlS0fVpLxITkIPTUVSc7FxJJ
+          6xWBDzD183TyV/xvZLk4FeNG5O5odrOlBPkw3vC5Ck95/5CGExhNZxL3+dn6PSyf6K8+jpcZ3v4E
+          t88JBRcBE8MjFWJRHXngbpM70aVZR8JB1Gp4cmFLnVOyY7T/kBZR7O5Jxl1GY7r6oQrviO7oH54P
+          24uTwUtRE7xEtzKfRedW//FJ4t2FIKEQIRMCwe+oq0RewnW/0YGos0uKI2IxsbrMGeDTbqb+6M3d
+          KPK9KUssmcchSqV8TvjcAT36bmkxmUI+vd5WAzI0d4rTxE7EgWUq4Na+jEl/V4zhnRYZkibuRM9K
+          +exmb8pesMPaTPVkrxrLqX7CH18ad8vdzbldTk3Y3mEmf3qOmbGlA/1pLjlurgRN37uUQrOb6pW/
+          bisGKFHR2eCAHkDiq0V9nTwo5s9x5PzTppp/KteC2AoR2W/2Uj7z+1uBusITSejJYjcTJ5Mgje8z
+          cVQmdaNY9g68o2GH+VrEySxzxxEsxQ+pHYazz375bvnDLzzNTdR1t3baKGu9oPcClHywLnEA93bS
+          6J1O92Sqi0hX/EvqjaLkadWiVH0JztW9/Kuf4uzJJ1RKypvq50n0h9s754B5uwu99uctGtBD4UDZ
+          4I6ESl5XEyoohjn9HehFSaBixqHeKJwx6kTdS0My11+IIfD9L/XNkhjT/tRN6H1UpHH1FtEU780S
+          ziO3Idl0Gli/dWkpi+/msOpDHXEhDTw4XgIOi1+97L4pC1pwjN/5L1/9eboqJmKs1vAw1XIy2GMc
+          Q57fMHWzeI+4wZe+MnqdXtT/LodkKuvwDqJk9NQO/YhNHNl5cOWPzh9f7RgT5Du8f86DuJm4ydnI
+          Fls+OjZb861JZln9eqgUWp+q+Dsaa/1a/uPr+Pz8w08ZBXeM6f6dqv6iU7WBXfJkZE8M3v86fceB
+          H94dLNpbrps4o5lADnCOmSfI3dIMhgq7pr5SN8xv3fSQ3ido6rha7+fDPms9VJQqbenVtu11y8Hn
+          9Yd/dF9Wp45xz6WUV/+AWCU/GnRksq28lijBcuEFybpjNQKwjC1WXs9XNzlnY92C7K9b5i/Y4B9N
+          slHsW05G5LxILj5Ke4LfcaIked3e1Ri/fiNU+yUlYeXs8iXLRAmMg28QY5fsjT5l5vdPn1EvMrtu
+          6dfJVoPk69RpH3k+FzvUy+lHmAneSttkORF3A91pEqmBtnt/6jypRPSeONS06Yf17WMqkGv6Oa6Y
+          d/Nn9FCE3UK6hKgEBjRFeX1RfF1w8bTGG0syrIM4HDpcbtXZoGNChF0aF/O4Xc6936/8AFjnZiS4
+          BlzX90c7QkF3bYmVWWHOmppE8Ey2X6IWNkl+i6XEIHoNpth6vI0lkdwRSXma0qQ4Ff5MbousYI67
+          ktP9N6Np0sYSdkKfUfUbRMb0548UQFJiba6Udb98nmD7ExjmxwNGs17SEUjsCsS1zjd/0c3y+8f/
+          aVLpYLDneY7h2367URzHspp+j+iiZMe3R7Fcbg32Q6SAZ79ONSnInU3YCmJoTP05Mk/Iun5oxwil
+          N3Mid+ruKxGHk6pEVnEnK78z2F2dbVjvn4a3dRB3A29OmbYGHpmDamPxQFigvO8rEo4fI5lGMerR
+          DsyMWpd56Ma/v6/6kOrbR+XTWF6nDu0Ki+o359z90wtR3Ml4uz/+/FkRjgBYdBn1JVgqet1EvdLM
+          ty1Z6z3irM5vwXgmJvW337ha1nhDu/HAiH0gczJeo9ZDR6kQCKmVrpr7MwgIp82Z4ICDpGPPuIDq
+          LB1J+OD1fLHykvvzSwg5uV7S0sJv0J9fY0b+F8105FWo6DBR/bbT/PkSJqVy50udmtRo83G0uxrC
+          OZ+o48SvimlRzSkvSnr8czibzVo3bxR4zFuqSunBYOT8HuFfvZGjOOf//Kxqu12ImXlbxJgvF4gm
+          lUNwemgMJuFFAG8ZHZKSbGuwwVlMqF7TmahxOyYsoRaAzCW7ceM54NPxSya0tL8Tse5uiPiYY6ay
+          1lP6l38LVEEB7dPhict7bjW3WtHDoFQ60ZPxxKZdfC1gvOQmtVI6JGPw3Wxgrde4XH9vJiyWQTtm
+          8h8fQ/Nk3mywpcPhn7/BTiC94JTcd2OmTDiZ/TlrISi7lpirv7fkkwhKFLoS8cKLn78VXT39fT8J
+          jbeai6K09GAtFabexQl8UTkkHnweWUOCyTzly774blAU+hKG5ssndSGGDqzxQfFbVRC7fqX+j/+O
+          LNvkVZtoYgHLY05XvjwwdiCXAFY/j4TQo4QercaB9qlyNDzlGluHNhWAtk+FekwdOtbuD5Jyvp4U
+          4j9OUc6tfFHx3sZrXc/Z//nTUiv93pGpz0mLwQ61hGH8Pi2Sj96h4i58IP3zI6+rX8tr3Qxwf2ca
+          ZjfrnbNzHNXK4YY2VEc3HU1zOEnw5w8dvmOa/PlNyurnkGJbqcYfHwAkCtWKv2djcZlmK6tfSm+h
+          9swn0fqo6K9e6sm+9Jd56dYtYaQlhOQuo3ywNND4FFE12WCfE9AhUDjbKEbldzCS2TLiCNW35Tpy
+          x8r1ObOtmr/8G1soxmo0c2dB4a5JqS/gxZ+MtCqVCXc+nvhvhP75w6vfiuczjhmj7hLA5heM9NhV
+          YzJUMeD//JPV36GRYXFo1UcjPT/aZPGcpyf/5ZsfyqSbBl9qIUjumOLp+TVaIytHOOoCR0zT+hrs
+          mx9G2OXcZYQNUROxv8k6rNc3ik2r+p+/9fjjp2bMOx27JtPyzw+N8BcbnNUZLfD13SHh71Alk9N3
+          Aoih9iM2ESdE3+auhnuka9RI47wae6u7gyvWJo1zv+xmavAqfNu2G1OedQZHT00B32PcUQOauzE1
+          y+TBTRtOJE5GAU2j2lzgO3nDuPPoLaF/eJTxfECw47/zwTbkE3zq6TPucPDx2YrP//xGy0+CnMuD
+          0IFKKc8U02mTzO8Nd0L8w95TXbBeFftMh+96UGZL7m/68ceDuj0hzWt3eHmEJZuUza5FUqJR6q18
+          atXndzhEXE3sXZUwWgqYQ6t+oTdnq/nLfr8OVQoSm5Ji5xusdzIV/flnq//Llvu3LVDIpUeisYLr
+          ek6+2mjlDzT84w9zdZHAPHR7au9ZlM/WRY7gF2V01QdXxtHTWAA3uyVxfpLZCZfrdkHKpfmN4kbQ
+          KvYS8Ub+W6+jQTrEcBWpMO1+eFQq8VXNEZMBrfyJOs7umk/sMvbgy6gh+pGVyWxdYqyUduQQ43Q0
+          K3H1Y+ARi49ROL0P+bIl9Qg/lM2YzcItYWK4jAgF/YNEPZ7yReR7G+R586Cmtrnni0G/X7iL2xvB
+          1sNa9e7FQ45rnKlOeqH606ewy4ULwd3+k0z0FU1gFvkeS31ds+keDqkcvceSOmH7Yv17P3lKZN3v
+          1Mntqpv/+HR8OH3X9e6T6eNMkXL7nC9YOatSMiyZEsDB2rzG3bXadDRLLwKQ2Beog8kumabXxIHK
+          O9IoFevBmZpHW9Q2GUd0Sb9WTK4XGXxlE1A9upXJPI7FBgJnDEfU2Ad/tgYZo+YuvUjivyvGutsV
+          gGRFTj1H9X3BtMoJrXqb7JtWNYQL8mx4flaDzxOthOPl2YEL2avjdDzhrn+DtexWP4T4QU26Zb+o
+          oLjbw53se1NlorN54z8/euQuxE1YLEcB+NahpqQ5dPlo5aUAHGIPjMySGiMRtAKZYuISrdVTY36n
+          aabsHLaMst54yXR7hRiy9+BTfTNeu3lzLlN4vJRy5W9DvuLDCOKZtcR/mRMaFVw3oGyCjq5+HHvd
+          v6/7v/gJb66e/+svCRn26X7lV7N+WnSAW/EmWlTG3fxOiwta+To1O83zF3VUC6gHKSOpdf2wJbe2
+          G/TnH+ldolXchtELnL4bGfdflUdL/HqOIEpa/9cfyZcx70qQ++L2Lz/7B7Yu0AelTF3TUPK/9UJ/
+          9WDV12j0XQ3/+aF4Mz2//hw5MoCvQIDl1R9eRrtr4ONK73FLuw41wUOLoS9aRPXLs2RL9HRfaNgv
+          W+pPwgGNi1NxO9b5GWbnrq7m7xHqXb/3ZPw4uW3CzvGl+evfjIsJgk+L7STBOdkoxORZ57PudgAl
+          sx7uuPJXxkmHOVPC+TqR6/OzNZbLVVzgdaUd/ovfWnFoBmoWC3/+pT95l16HyBPuJNQPaicMuVFA
+          3NtfYhrx3X/9PV9C6iu9oNO7YvwrqxHx0ow6a79y3o9Lqqz9wREZDq641ovuit6o23GpBBGNR/yT
+          oJ4eIr69iWH8+fN/fI5EYYwN/hqyRnn1ckx8wxm7+SjkESis6alBxcaYHjjMYO0PYPF3KpOeq2f7
+          D/+I8wnM5I9foLWfQVzbthO+hFRVSr/c02LfJ8ayL0pQCIWCatPCoRkvMgcPQbeJ+zo5OecOviSv
+          fg/ROO5o9N7FgD88Iab08PPhJLBUWevdOE9Liib3XElQG+qVmtr5041xJ6Zo9UfW57UxvjfxHYBi
+          mF9y94ucUfMzqEh4GR9iG/Pkz/Y1S6E4HBMavJqjIR52kgf4Cmd66WuTiUdrdNC9dHS66ik2bba7
+          BR26piYeMrHxz599ih6h4do/+iUeCxSj+ZkkGuop6ZGye0G1yJfx/efHdWX0UoRH41NDXzJjWfNZ
+          Ud+6OK5+HFvXXwUJ1ScS2/HJmFIn+KI/fx5nW3vdEqalUECYrv7Uw5+hPEpyp2UicWTV6rjj3ezl
+          x/dcEqt6hOsrJVkBc9odiGm292qZFqNVKnIb6eFzYoj6x6GA+NR/Vz+oSeYngAmhivqRxXJutNlB
+          x8rKzynOIsWng7If//q5VL1eMOK/alrDS5Mi6qb+PWfh6d3Ai6MjNQKlTubF4uP/18EH0v/eUvDS
+          2Uz1ym/ZNB0jUDJXOhB3cJJkcm/GBsgzascf7lXG3e/spTxiUyUFWIGx7Fs7U/DmvqeWTko0WltO
+          UrRJGEh48fVEDM9uCQ/XrQie9oUh0FdygR0ucqqx2PGFq5R/0eNbxsTNzl9/XiKqIvOFp1EMlwqN
+          0TDECLKLQB6eS7vJp0MDPLcOst4qn469eBKBfnzsqYayIZkLTY2Vwd8NxDyhZ87u2ltXVMkOCaav
+          S8VNptQqr7wP6UMAnI/e53lSNvxpfSvznLLP3//rri3iDSdWiNO3kQPa/opo1l30pBuUSwNs1F9E
+          PQgbtjRb3oPZlFt8rTZ9x9TSSYEk9ouo0av2xeMG27Dndyq9njjbX1z7O4JxDxs8cW9kdNoglugT
+          zGfi5sI+WZCUy0AvvxO9uWTpJtP5qQq7yjd61J9yzob4nKLZVSnde0Rj/NV4nxRP//7oUQrsXNCH
+          JIbNNvuNb+/36DjOYbGiiGyizqlM8+HmHWwlPT12xE3GlzGB4abgjN2D+Bo5dGK75VsQXVCINVwQ
+          Ytl7ecGJWgW1/gZWj9x6EMVotiR8Lg1iQzhtlGOTKtTJLbFicO9iCNSnRE9TiHPOqZNCqTnOJnc+
+          1ZB4/j4aOHJPixivWEDTTVk4mJWfSx+Vq6FFPB5AkTmrJM532xhCszdHcDdlRc6PZ9hxkPy+KPYu
+          L5r9dpWx3OqmgF8nhlh+uyif8uAgA+lO/cjo58neC/pGIJObRZNXfELzLZsvYLRvgepvElZLefiZ
+          itHFDi0+r3M16pbvIdSLWxIezmfE29vLFxzgP3jqpCVvyuYzAVeK1iiVPeRMNmVPkYYDR4y5qyvx
+          rg0qcEW/pdfs/DWmWMq+kKebDSEHwqqp5UBFFrVSkiqPq8/V2+gCV+zH1H3eAmPmjodCIdtipif6
+          9RhbDqhBgYY9at/RK6fn5CLD8AVCr/gJ+VxLbw+2495eW9Bff7gbWoPscjqQwxg1aMlfgwfuNRTH
+          qZPiZN6X1xRp99djHKrqUfWHSC+V/hZG5PApIJmLnTQqYnuLaYS1KuElrY7hmq2DVNb8pBupyJB+
+          jrcj0znD5xVNfSkvwC/qfcUnWt5NoKNU/JY0jAXRX4zGn+QvM0dyF06VMbVBJSlpfObGbj/FvjB1
+          PYYft7/Qv3oxyZqjI8e0HFLoqpaLT1JG8HD9imBmydWw+UUt5KqwG7lHaVb8MTJ0hX/ylPgyldDw
+          EipO+fmphEX4vfOh2nKt4srZnVoNPSJunzAPNakqUdNWonx4a4KuTI5qkbsVyRU7NvcYOmHakGy3
+          DxL2eA1fqB/VlbpGkvi8nB8zpCT+iRjLYDHRyBYH2Po2sntcTCSu+QDFnpU0uCHms+1sykjgOYmS
+          QpbZbOWjDH/xQj7j25/X9YLZ1Sl1k0qpRrU7bBQvGDvikCzK2Rk9dcUt04LEJ61OqD0kE3yGpP8X
+          D2L0qzxFWV4W9SrrUvEP7vVSNk/8wptpDwa7HrMAguf4IMcd3nd8cfg2YF9KmxwvFfYnfN3a8KAf
+          nvrUPHT8s/FLCMxOoW703vqTe+TuShjEAtE3Zp4v4St35CS/BuQ07NcW42vbAkruhPjUnLsl+0EL
+          nykOqGPOXje1QSfB5UPVUf7tDF8Yrl4N6QFjapHrJ5+cOr/DZJCRaMVJ6mZJ62P4kbtCVbG6VuLz
+          c7HBZvad2B/j07Gvwp2gi52aunvplHB6b/fIdSWXZolI8rE8PG14nAILN1F9MvjUXrCy8LFBPEFz
+          knkfEgnKM6ajQL8tmq4vNVPqx/OK0XbODG44CzqwsAzJcbjkSOQSkCRt3SlUZDzxZz9BX/TKDUT0
+          3LAqZln+FxyHhMStq8kvrN+zhU3Q7Ed+tvWcj1CpKpItvkiASqeacrR7wYpHdF/a8zoIXLNh7lKg
+          Zwnd2XyW0zsYu3uE1/wyxEu1CyAjAqZBNPfdGi81KNJdp+seiXyUz3mExnyaqaE6G3+tnxtlzXeM
+          LpPZ8Tx2Ad76GTCc49DgbkNc/uUfnqUCo0lB8/KvfvaG/upYcgvvcnTcmfj73um+kHRqoCi795ea
+          ++3czZvN4wSCoZ+wsv3x6KdMpaN0p/hHsH0rk/HzDVMY9uw7vo2x9HmoRwzHKwwkZlKSsEnyLuiu
+          Xk1SZJfZmIxwf1Fs99djmMdttWiFDDBSWaB7Cd3Rso9PDbj+qI2KaMYdy6lxQq9D65NCuO8TQfGU
+          GmTj4xOrjcX1/ZWnDr77iai1iRzEWbIGys2PU3Leem/0V68A7eQnvcpqmXN0/jjK3/WlopTmS5Jc
+          MnjhrlvrmcaEnb38x4f070K6JbumJ9hFyw+DgKxc+LvetT6P0lE7GitmRajhniHF0TqFAbrDXblR
+          eSSuUum5eJG2EuzGzUTsT2l1i8a1y84ulwOJNS32ufTVO0obnF/U2Ip9wpRzMv3Vs5VfqehbRsUL
+          dJGlxGPlp1vWfEIoKQi5rvk60tyQlM98XWgyd2a3tC/RU37WBVFbOwWMbaTiAkZ87LEy3LDP3HJj
+          Qv3kP7iq5pMx1WcnRqEecCQdCK5+m0PJ/fGF8Z2buBvsz5eDneJJxJ2wmLDwrJVo4NMHOfPVM1/Q
+          HGSwxhchpV/l02LvNkDU5UjX+sj653z0ZGyEdxpQXsunepRsdDpqGQ22RttNrlU4YJh+Tey2casx
+          RWWprPgw8i8v7sThtf1Cmuj2Wp8exty8Nl/ICnwnxvE95qw6/BoYJV8j+0SwkyWtphE1MVFp2Oc9
+          Wx6aswDL+yONLu0vH96t7SmBHwUY0hR3gmt/e1AlMyQkHhc2OUH1VfwDn9NzF3CIZu+lhCl/SCNf
+          2JM/OWBEwFu3Bw34ojGWbYovqOkT/S/fu9m4DjX6i7+/es8KKXXkzV1OiS5sxIR9hlRHoDv9v3o6
+          fV5ZBrsRJkLu+7Kb8KC+FFM48Zhf3lonOP7eBu31fdP7Wt8X+IEkc2bB01sqX9nc/toLTEL/onFE
+          UM7o9iBD1sRP6vfVwedw8Q3+Pv+7P3YdPB3y5xlT+2v9/KneRpmy/4wSVbfdBlFy9QTIr5ZBHWo4
+          FYsO7KQscjUS93yo/fZxkWpA/FTQW8gdGYX3MikS/9zSWFk0xCEsYhBNTcWbwLQMdjVnSWmd8Eaj
+          436fi/5RDSB+fQeqHRyaT1+vbf7ymzpL43UT+8g1/NMTp+vO6Lpyc4LfrbPIHn0MNJ9/nAPLpmOj
+          XNaSvxwHrYTTkNTEOW9ZN9v9Hv/hKzFmbHdcPPseOn4dnmqfAvJP9KscWD+TuEsiY/4Y2xSOm8+e
+          BkOqIt7zWgFdeOTgyhDMvD/VQwGfl22OEEc+W1DZ2yjg9J7ihWnGdIJbBGxfRyMjG7cSNkZj7tCF
+          yphuTdYNj9oLQNx8ziMvFvd8MoCqqDOgoHeRf+VT+MQNtNHtS729ZzNuGtQF2ZvNhezj88egf/hX
+          rQfdXEcvrWqxWzd3z19Cr3tJSObdq5bRyv9psMYzv6AyUh5t6pHb+/hlw6zaI/zVI3svDsmQvHcl
+          5J+DT4gAY7I8vWuNPnKKabzG5xJc9dPOP25L6vnUM8StAwtYmULI45s4HX+UEg+5Ly8l6/37XGZo
+          pnLUxmrFx3dHw3GXwptebuRya6Sc+Up0gpc+z1Tzzb6bHS3WwfscDKpJxd+2blNX/vjDbVYn//cS
+          Og6Q+pgoZlb2Hz+rtXM8DudcTsbHiRRw3Lz35J58bLQ+r3UL7s6k7gv8XGReb0Kucjsab0yUDKm9
+          BMp88zhiEouvFq327gBjiWicbg1DOBZPrAyVHuPd5m1UQ78gDJ/h0I9SKZbsD7+gZ12FBeYHaN6H
+          ewmdy+W68s8LWmw3nFDOPxLcnrM5X1JUvhQv4mZqXLo4p09xypTO2BTU07OHv2jca1I2l80Zy3r2
+          MOZY+Z2Q5fz0Ef3xqVfSZqhNPwPV8vloTPc7esGyuzjUmsynsSTnzSJfh5dGzM1GytkhDzzYX8uJ
+          uAIV0bLiI9KPtz3db703Y+Sqc7Dy13ELVWKs9SJD1py/qJ+Z24Su+S0fWLKhxP9ixlb+A/OvMok9
+          hWMyEVnsYTbMhZjTMnVLT+ZJCdOnNgqxwedLVbBA+YoBW1/xm3NmH7sNpO9TQi0zqbqhveq2Mr2a
+          G907k96Jx+YeySs/Gb+ZcVot9K8Op/nakkDjC9aufFKRXmYwblf9NHw/Pxut8YS3Nyj8WYz8SL5v
+          rZKcnodNNReHslHqz1zjX5F7iMfNMfp3vVch3rHR8YkNpyp8kcM14v7qQwCFxlfEwI82n37gl4DL
+          WMX8qn/pVMaNQlgjUkdw+2r+bi8StD0vkv3nFzP+T09bbyPGCDaqP1rf0ATX7zV6NRJmLO49KlFj
+          cROxuln028s511FwLVJiho2Q0x16RsoIiY3fUf5lbKfuR8gu9pWqu0+RTxLcyr96Qjz07dnQCKoN
+          1+KYYQW6c/KLJEGS32L7Xbe43YylM7YcqPH8JKpgZgl7prj94we4+tnP/IvwNpCLkN6ovlxTn433
+          Q6zYce3+Pb9ueWY76Z++CA64reahLjC41mKS4KVwyZQHVwmYjTYkbDjH+LdeeeXaWPCUQ/Ld3U4Y
+          inkd/Hr60G7Y/bgMPgJ2SdiePeOnaE4JrtQ0//TJ8n+kfcuOg8yS5v48RevfWkcGjMmkd9zMHRID
+          trE0GgHGGDDG3DIhpXn3Ea5WL1qzm2WpVGVMBvFdIogQq3ci9mQvIqSWVj3l57IRfIrpvN6bbVHE
+          t07BQt8GCl+jU3NM2plQ2kk2epjeEi3idFCEi7zNh1nTh7P5UYnYUX3GbsoilXqfL/fT1/Ni1qNK
+          j4MJIfGggp39kjqDW5yl42OnczPbUb7+SI+nCygXTJvf8c0ovm1TWDZ9Yz0E2yGxkEHw6rw71o7w
+          oG56jofdVxkQEt9xdpAOZFuk5B6wb1539dxzZvvzG7bzc8H4EWz/aNQtRb4TP1TKPrxZCKt++vGd
+          YbZGYxTs5aJjC17zYcVVlkAr0c/zkbosIK502oGf3jXPYHHmBCsMTJniO4vD3cnWq2+7Ijc1Ez7f
+          g3hYjd5Pjr6KCmxK7ieirv514Yx5DqmdVqqUICGEpcZXWM2+hH67z0uHPjISbJKT7TChCSRorvSN
+          5NAmA/mqkX6U0yRFudgMWf/zm+5Lb6Ds+5hUchngDkpvIqGTJ3Tq6pR3Bk6THvtNIZ8yOnk8hH38
+          npCEVuiQkRNjyGVhiZHbB9GcqxMDjL2oIKnlFpV41pXAt484JFm5pDLNTQoh3/pnrFCeRtPD+Ozg
+          5o/5y0myVc4TbAgDY4Qot812WMvj4oszPOvoBr/vTX8oqTgF4wkZB0Wquccg7mAiBzMynq+ppvX5
+          1YgpYnxsfL4rpfGF3UH3zTX+cdN3a++RAkQfmmPnDAXnL5++z/4VoYjEzpDURx827H5EP/1Bdvma
+          iyQIg5kJu8Owhsa1gFpnTtjr+J6SoPiWP/6KpYthRIe3qjTwWGsc2vScOokJicXX6egiewZaxNwr
+          KYF+k5yQUvZ2RK7xyRS9+a0iTS/DjJPAGopXtsb4VHzIsK7TuxSDVLX875XRVSbkwx6QreX/PmhB
+          veUvTTQuyYBQ3skRN+gdhEgiF+yhiHVIyIedqPdMhDO4k1RyjT1bCJmdMAvxMaBLK7SFKFFoogdT
+          rw7NdOrCwRLumx4t1CWFlxUG5JFjSeku9Vo126Bz91b54o3ns7WKqhRueIEcbaX1+DkLOnzw9tFn
+          g8wEXOxKvKgXuYqj/i3VC3xcRyFs19Vfk+rt/PnPG17NYYCyaH2SLoFHt6ux9ja2V5aqkBd/egdY
+          TExx1MS7n38444gw6nCslxx8mVOCn9L+q+aSzIawHw8HP9ZFEs27YefC785KsXsQ23qdzD4ArQVZ
+          7HvJng7mXkzBT2+ZymSAxZsask2Rkv1FFnR1UfhRgIxfHXx2iO8qeTeHHFRqoSKjUPxoFQKHCEif
+          Q+Ry4BQtU0SFH/5gtN4GlbByvUJhOFz8XWgadafdrSscn6rp7y3Dz9aNX8KoHm4YaSfPmbT+JMAD
+          EgTsf1Ia/fxGuF0PvoXMW2XDyZKgRlIey1Km0zHQlxl+nkaz8UvDYRMbMvB7f0lY2t/fGd3iGXyU
+          6YK8D2dFBNgJB7rrXMy29tUzchhEVzCe+YR0hR2dRSoAgemJuDP32gbPD3atw2xCJkalo0b0k4NR
+          8ExzQY97Lw/UR4ICgGwryJAePMUJvxfgnYgsds3vJ1onIqRwead7n9vdr87Kc2EDPdNesINfeMDO
+          xXThpPfKfNAkgU71HnawUnMVXY/MCiYdfko4W0WD/MtgDdN+fRXQ4aNp4zvPiKzFl0CrUHyMUi5x
+          FnnPJKCvCwOrzGLWB14hjOiNi7M14mQDoyK+ArJiBz/+ps7DTQ4AfTIpvkf1Y1h++Lz5/75YjBWd
+          P/zUQ71V4cxV0ROsHCgJpBeS/OWjv/O7CE/Fp7L9jCiflh1kGc6feXLqHeLexBWWEnnMe/A612tr
+          uLNQ7osjMmXfHDiTM69w87ewjruedqlqaVB8P2r/eL/olKH7KgVy1b/9XYZHp2cDXAI75Sna/IVs
+          PH1fnZDdDXVeb5en84cHz7W54FQccb1YVdiLm983H+EVDqOhpgxcgStt5/eqN74yio9H+/Lp5T1H
+          a18EBA59OmNtC87N3+fAZQwEXGzxe1jkLIBaZ0/IneISrHl26Y7s2Kr+7rH1v8b8IYXIZm74bOy6
+          mijZnP78CaynjlBPe/seQ+KaNoovM6hpHmgSFE53dW7H3s2Yi9pw8J7WlS/2t37zl1AicFlQzmzx
+          Cer1+l4V6NxogtUCM4CIjQyFZ40B8uBwyJYievHi8Tq3M/+exfpX34ASjTj8w4elaw5QMK/3i99b
+          vjUcNn54fAzrd+ZqUR36tx/7UI7zB7pW0R6MTFnrYunvGmx53qsmKi91wIH3FctmPTrfT5EEx03v
+          +vzml1L5e7fhTUsthNqxBlMrzAX4+f36ZGwlc/CS4LVGFfJkw8sYnJeFaLXhHaNq94jISbB2sLo1
+          Jgpq4jm0sUMNdnmfbXpBVpczk5TC7DVfFNxHu15/8dA493h+h9CK2M1fFFEBc+RcvGpYe3or4MHa
+          iVg+PUxnOobb4o50RVi+B0w9iWYgwVETdSy7uu9QeXfj4KeTZGRt+EKaT9lC3Xk087vFLCC3rUUO
+          PK+Wz/DChU5fXk4h0vh03vyXmlr8W4A3P3ew9GYwXXbnjgEmjyt8yrR5mEDs8YAzzAy5weLW2OVM
+          RmjILPtxPKYR8ZhA+d1/bB+NMVooExBxXj3NP8TZFxAivjTxpOCTX2aORZnhlKfwmT0IuuqJrzIq
+          xIpQU8f2/66HT7seqpL6RD4bvwB2OYkTf/WoJ5NNgJQlw4ChvI7oZDONM/zqMzEtM5xU11Ld/FUC
+          ZdcNsWY2F2d++7Er9h1esfz5SJSdH+4MUwvt/e9k0mi1btb2Fvx88n/+dH/Qlh5OM++ifKaSenin
+          ayoe/JOCQuQdwdJ9XprIVpWMHW2N6pm2hfLz73/4SSmY7BC4Us0jJ4/aelVF0PzxFZk/uvRg4osm
+          inyu4GTTX4vWezw88/rFZ1zj4KxslHRQpVWE0NtrNn2w8BDR5oAcZrhmU9jfeDg5YEKqymnRYSEr
+          LxasWmDJJildRDORoK2T2edmyxnId6jdX75Cp4m9063+IgHROkXovnurw7K6Uid2fhHNzNuwMq69
+          vSWRf90FbC6GEzEkxBzUxmDFAacJ0Va/0EBnVlf8w9PDxqePqSWckSKuL7r6mdMB1wldLA38unXN
+          bYu/wPGOrnuezX7PJ2yi7xOdprqMKDfNCjx9nRt2LIYBDaikTozF9jGLhoPAEnFnBg58w/35rzPz
+          9EtA5vyGb4UyR8vO348weS2qT62H5ZA3LwnwPn8/s+Auw0AqYYKw0G/yVn9VnDW6cQRqXMxip77b
+          Ef3Vc+JjnuCLWmdDv/nvkL1/Gr8TZkyXR3pMoHCbgo3fWBn+6bft+Z6Xzm02vV3voMcd/PkIYuAs
+          xH+PUJKe8bxu/jJBYscJecMoOJZyDKau2cP/r8UHx/93S8FotCHW5Pc7IqcLm4iP9UmQY7IDpXko
+          rbDoA4iMUVkprvqghIsoP+YWXyWVk+SoFacmPWFTApM6IYUKR4eG/Ly6ehgdsmDfwECPLyhzP2fK
+          KZcLFAPmfUJmP70GNuZUCZrmB/q7rnEdUmSiDo0Vzz7fT696WVY+hYbBHNHNO9sOczVUBlQ1FeeG
+          O32c9fSQOWDtoIODXTRkJM62wSDwKvv7CO8zMg2vFN6aTkVuUNkRg/f1KuKrW2B71+8G4lhuA/30
+          dEDqm9RgfThlL552dxm77IfNSHvlbbgkygErp+TkUOuCd7AoyicKjkNbL7xY5uKnWxOfo6bucPpH
+          lSB/PQrIr+MqO0hnKECjFc/zhw3zjBy/XQcO/NeeE7gXKK2Z+wj9x3pCxvk401r5mDk4Px9HnPC3
+          q0PDTuXF6JMz+BbiaVhurzYGQxXd59kUm5qmOpeKu117xXZiBOqg2AcBHoLwi9z9WmXscH8W8LwT
+          Gmy4mZNR5l3aYjwIJySZF11dkuO2YSDOd9iR1dPWXzYr8FbuWaQ8cQVIawWFGIisPzeXTqIMK4Uj
+          rHenHLnHe5bR49vjIIvkxn887xYls2pfwXMePfxovnSg+qGJISK5gdxalABHYTyLq2ux887+Ks7h
+          1q0EOjTgcdyRhK7cU5LExOq/CJ0OZ5WbLa+F1XFXoK0zKuMSLMTQmZIvPsMB1MtuutvwRSUBKdP9
+          lh0qeSkgc1ISbMPspU5etGqweoDZ3+2tmmKohZ1Y622NnSRQao5jb6N4Wb4YeyfeBixJ/StkYH1G
+          XnmYhoUnEhHXJcdIMcZPRvbUKKG8VDr2HtMYLdfb2xbvbMOga7yvIjYZyQzdGNRYq0usYtwsIVyF
+          kUfamzkPhBakgcfwpKBHjkzAkk+rwDl5qThP4MWhdyiaR1sunvjmnXuVyBNpIe32nc8fvRgQgX4Z
+          XtkBDytjnQz0euQhRJDySA0OVs0Njz4Ufs+bxJhFNP/FM3MO5lXVlJpmz7IAr1v4QL6L+xq/p3yE
+          td7UKJSIFLHTS72KEdJDfB8PN8rqe5PALR6Rk6XqsODHyAlcxt6QX913YMYnrMPn4r7wnW3kiHgB
+          X8BvPNfYdiQwzO8+0+H5uu1a/ORyxALfTeDnWfnIi8jdOVTysYDDXUXzwXBeNaFjEsLzjm9Q8dHN
+          jJFMXgFkshX05FVS0znlNVDcyBerPA/oOMVSJXKMd8TOGXEDwZ3Ci96Cr0h6cSeHOX7LHupBLOF8
+          fcTbVJPnCNvmZsz7Xn9ERJWBAm7mJ0LeHbTR9xbrPbxlpxadb0wfUYUpTVF6XROsLM4j43Ckr9Ap
+          wvO8Hp7P+n18KtsUiuSLnPUjZQcNKBx8H609Pht1DBaBGwp4CIIv0rjTR93i9/qXj5zatWoqZzsN
+          Fn0Isf3e7yLctbdRKHefzt+JDz1jGkbjRBxcA58xX+Mw/r6fwZ2+WA/bzOHOXbmKlXi9Y/t5ZCk5
+          drcCnqVLhbRMb6KVidge5ss4oTuR784hTEcfBrlbo+TeehG3v0Sp6BufxqcgTRzaEODCUCMUoza+
+          ZV/vdV9F/rFTfYFvrhmJPnsFPIc4RIWy3jJifapCrMT4jh7TTgaHUjiV4uRMAb5VtpYtxvU7AyEy
+          k7lsC4uScwQ76JiJitB2nv2YOius+hdFV05wHbaeTiFkk1bGBU9yykSfvXQMv8sde/FDoUzXCyVU
+          vtIembshiTjZyVd4AbTE/ngNnZVjnzPUv42FjNstjA7b4D9xfOQllu57IVuXLoxFvECAnLsuZwc4
+          HkfgmKmKtd3zTOn00RvoXCQTJZ6YDxzqOV2s3WJBnpdzdA7U0hcFL7BQKpNVxWdVjaH0cduZxrIN
+          aGzNCsye0uAvzJkAal0XU4zfJkLFLno7HT0nAvTTxMbpWniA+X5yBYBapr/PByv88trvfiErasKa
+          UVE6wniXJhi9xnWgAmxHWBTVc7sfFcXWBUN4dVwHF9HRGQ4CnGfoj/wD5V/PHpikm3gYRkcHOY7l
+          q+xwv+WiDxUW2Qe4ZGtd88kfnpbC65RxUexqkBXc0T9seMstrKAIcjE9sWV2zsB8yuMVCEtmIgct
+          IR3ZdxTAb80BZKTRsV724X0Fp6tU4oAxd9nCrYX+ux8oZ5ZrfbiO9/B4yPkJFaHnO+tX2EHYVOWM
+          CiPN1SU+2zlMaYiwVdyaenYXoEP74QKEfnjugIaDC3f+oDQxiLp0IOLgfHFM5AEucKaPCRrozs6I
+          laR8OKzms4x4dF2KUjZlHe7ljxyU9esOXw3eA6zeyIkovu3nfKhsLZrx6aOL/tmWcdjs3WEB2ajB
+          c22bSObVK2BTCdhw7dw3tisvyegdP3L4KssPvsZ7JWKMqNTE55wJ/m5h2Jp2qeXDziM7JCc5yhZn
+          8QlYleM4iyz5ABLcIlOcXdIiJ+c/Gz53psiLeoaNptApu+G9+NKIjXye5IAy10Mo/vDELdJkYE7m
+          ZYbacU+QVvB9tkRFZYotupY4Ok9mRvtrP4uhac5YPmQLWAvTNiG/wv6HZ9FbPcS8uOE1MvKtZTSl
+          fAAeRgPQwy5RzZ7EfQO9YFt0UF0Wh/jsokDcjQuyP6Ss14f9LACh0hkVe6sG6+Fr+XB8FCWS8eEN
+          1v0lS8E09T52iLTPMDyvsYhHbcTmrWwogVXdQ9SC3l8+NsxWKBkMNCee/eOHRDmdS/FcmyY+jcM5
+          o9yoMLB/Sttb+jigpJ0MBnZs6yAUXqWBGUzOhpfrKmBLcHV1fYZqDIclW+byWSn18HVflbjEHcT3
+          9ZJFy5t+duL9dtXQNqUFMIL/CsQyZF288R+6SnkKf3//l39WPnEkeHP2Pta+7WEgnB3x8E0bDbkn
+          nGTL7kJCEaxFMR8+3YcSVgpnYR2aEltBeq4J7mwBsJ/5jtAzCVQiyFYLstlSUVb2XnRwlbz48Ye5
+          3eJzNa8uB69nLfU3fBwWfwh3wpZ/cfR8X2tGORxXsEzHE3KyS7HFxz6AU+HVeMP/bMGPhoHvx73y
+          L48gHdihM6/AN94NUhprouQT3XvIElhj5w4lh8lHEAA9/y7Y7VUrI0x3XOFE7GoG1dWmXM+/erGa
+          XRnd8+NLJapMpT/8kvXTOcPgkjBg40v4OawpXe7n4xWaQin4xcaHF/M0lLCLBg/rgSAOSyJsU174
+          wfZ7H70GcpcDRewVY5jXYD44FN1KSVzzwkSGb5iA+teCgwJv7zY+32eEg9CEQp68fLhflWx9yZ4N
+          sDo66BoIj4EeitoWr5Dj/LVQTxk5R0wP+i5wcRKc9+qcFtAE33UfIR+XW5d+3ZTQs90cP8Q4AgsO
+          PFuAjWEh7ZHqYG1lRhFmpWd9YNw6lX731gpArdL5U4Ai265fguT7YJByblhKQ3ZoQIviEp3KpzSw
+          0WgxYL+/3LGvdpyzuu5aweK2frGFVndggFhLR0nxJWTfno7KaWUZQ61Akc+t4Sda3o8kgEVmNjhm
+          7tawbPkcqiXnzju2zWqS3vAIxbB3N/21RtP5ygWw+uQOun8+szqSmzYLht9ipItSXHNie27EiZgV
+          yg/gUFNeOoUgvEU6MqXTHK3XtCVATMoUX+oGUsK+o1C0ePeCNGfwhpWJxA4YZ0NFjoLVjFvXqBR5
+          x3r4hze0s1GQrQY+8D6fd0l0HWgM/QB8J0Obd0YKHSydGQGot1bBavLFwzgN3xRK8nHCSbwz6/EX
+          35se+eVzdTkDjhOHJz0howtGSvB+IMDDnoKimzOApWTsAE6fVUKIxb7DNPdTLrbl3sWeojdgJsek
+          hIGnO9jJLjtK7TlVIBjPcMvnfESrnHAQsNzdJ2svDOPGp8DOtk4zPN5BtrI85/7po59e4IxFc+Hj
+          iiMkN6dSxbdOWCExDhZSduUwrAV5uVBDGphXy3ZUMhfWDHejrmIpOi5grZdEhwVxHIyq6hyNn3K5
+          wh12Boymi5odrsWxhWJSpVj3dtdoRf1OAy7QNWzWI1svz97SYcpdDsifFVQv9fPdHEfNvaFkN/DZ
+          7/xFABuEL/uiVtfi8mJg68A3uvRvk7IHR/IhdxEtf88ccUT2FJXHQRgEfIKx4izTy4nBfLHMTe+B
+          rMuqVQNb/PhHvaqdReDqXCxQaqATCwSwyk5OIFZnxz8sJZct6lf0YauMOXbaHa4JSPoCOp8yw4Gf
+          +RkV/YiB9wha+JEub5Vyx76CXj5jf71n5Y8f6yJJbzlWN/1Apm3R1Y8/SLtoiEa0ptWf3k4FPqnZ
+          3fHTCeljf0GaMA2ADqexhGEEHGRu+mmx0DiCb80AdA74OlvMSPWhOhUB0tPHCZBTfdHheUzPSPaf
+          uTqK56mAybM64I3fR1/j+hqhVuIPku4WqOfRbXTxdE+zP/6BDfndgE2vYdcJy5ryXROD+5e7IaXZ
+          j/Ui12EpxOHt+4tfQOgYBKL+BmAW9flbr2WuVeLmB8wtqAiYEJvbYNP3/rEI93S1kcHBZHFqn32L
+          mvO7n0CQ+pP/Zb4iXapLk4L9/nafufvno07JyM+gvbjbW5e9rtKhrHl4x+Mb+aUtZ2uvW6mYOg9l
+          e/5vA5kLeRQvEj9u+eVbkxeQOfG64hNSp8oCK8yDRLTx5COUtbZDpWfegu1n7PbHTp0kJpihUIVk
+          Pn5PZ7qw3qMSyvftueFvNdDmQhkYzW7vkzyRa+JYWiOKL9KgS6HfI3zRjQoKFu/ji3I+APo0phDq
+          bNyiBzLjaJGYZIRaf0j8/utrETtHQy8QLnFxIBVZRC7v1yii6VtiE+5T8H05gw8lUhC05ZeatuEj
+          Bm/jFeOTJ5j1appvGzBy//AZzzpHZHyVrvjDP+3SSYBBjcDAM6i+/mHzT8h9xRrc+D+23tFA6aT4
+          krA3mBKZPs0yLsJOAW6Z0WLZb0q6nNpihqIuGf5i1AwlyLvv+B0beyhjTo06r3vqw0/4OSBVCc81
+          ibOyFPffh4PllU7ODL9EF/ZgXyCVCW4OW13GBDZXckbOtiP9gGDXwuG5nPzlW6s199iZvJj6jIiv
+          0/tB53oJdDFe1hhLL6mNOrQXXLD5B0iXdFMlfne+gs3vw/7kBnR57ZwCJNdgxo8ZZ4A837sAXryS
+          oNCM3Wjl3/38x/cSQY2GqQhLX9Q+jwe6dKpbM6uuVwCW2PcP6/XiEBWlM9QDdZphNDoZRVjigUmp
+          hjb+B1bo8yMMcr/ezksAWN19Y3iH197ffdOt5D7DbdEba6HT9CIqXR5MCDhBxDPsj6ZDyrrbQdse
+          uzkdqrZmb685/uOPN4Ft6rm+JjPY4gmZHXxGJHXOvnjAd2muklJU6cOvQogopyCHrY3he7aujbhe
+          5vDH38AoPk/2zy9Erna8D1xceCF4jO8Xtr6ADOTxzdc/f0cp1Hc0b+cPz2tQImvD227zU6F3jxnk
+          m8Yb/D2/nKQDdLJNpabKex8A59x5+J4fZZX5xdfYjj2Kl4cS0TAdXbD5F/6OMXmVSEbLQfdVd7MQ
+          1I0zh7eqEVmyq5HGgRNg/G0KzJbf8KVTxz+/UKx6Tsd2Im4jaz6qBFzL0pEri2HEtjshhnk+68iN
+          xiFa70YWgtQ8GT7vPo8ORbYZg8nBAXbV7AKI5kAfPIwW+OmdTvUyNi9OrD6Fg/w2+A5fqgsMEBiq
+          YpS1vboYs9KA61lPsSM4YzasCiPAa91OSJXzkNI5JbroMlcfKUkpOj1PlVRctwnm7Ns4ZYdX9ulg
+          PXysmXZaG/3wErgERj/+7Bw+8tOGsDlZm39SZLgQRwkG+vWCpOPQDqS9ElPc8AghJN6j1R6CRKyb
+          QzR/YGRGP70ttEnbY1M4rtnSfGcCZGWvIkMPduq0u/AB/CpugD0vv9IVcHoDf/n0Iq6nYWG9SwW/
+          3M3HNphvG7+tK5A1vIBkLBsO/QSSIu52zfXPb2XIzR3BZ7Sj2ag5GRD2nYXC3Vs+yJhmkq3rmlVQ
+          Op07f19yEWWBNI6AuZUDkja8xOmx6OHikwn99MeSvekMqyMskLeXlgiPDy+FWdYT7N5XShf1y/pC
+          NfsyNhzaRqzudiXk3HmP9bl8ZdOnFiBwLUf3y0Gfh/Xnj+3Hg/7nF5P18UmFg2yf/fWUvFXafnf+
+          j2/4YNISZ3aU4xVq5fTZ8tcybM9nALXzB8077/2NyMv8tmDj51gd/E/Us8HTFvrb/ojMVkrUdcNP
+          eImSvQ+yS0HJmWddyNdc6LPls6zX/mUS6OncaxYE5U6Xsq0a2CZNj/SbvwfrCfXmsRfWJ1JJcXa6
+          fKSBuPkhKFpiRJdolBnIG7RC+pJ3w5+fmtZsg7f8SclXyXJYuIqL3a/XD/Qc1Apgz5yDXNmg9bLa
+          51jcvv/MN3E6HMb+rcPNP56Pki9Rjqd2ArCkitiqWdNZ7SFJQON5DNbMlztg2b+mIOq5bfERazoH
+          /dBcoZmllc/t4ys9nINBgpv+m9nm3NaLe0s2/0kxkXa6DdkKfTL+/APswOdM+0swtSBWvbN/LG7N
+          QIFwTmHFaAGONj902uIBnNewxNIAp2hl9McObn4D9i2qRnNT3gXY7O0VWXxpDmtnXkeoCaE684T2
+          6subjznsWqvEZ0tOhs1PsEV0CvYIVdWSzUoldiATzxm6iztXXXWmGkXyJcbMeTsuIpv/Aap4eGPV
+          4sShu+JeANGnYLClce9hLkzFFEXCQ4Ta+JCNsW8y0HVf8txu/HYp2779+Wn4ueUPugwZBDEOJKz4
+          nV8TctNG+FX8AMnW4RD96d1L/jX9eou/5ds/euh57Q05B/RWye04an9+7/lb7OjyzF8pTEKB/vQ8
+          WLqQWaHcfGYknRwNkO4RMiJn3bQZuqT5/f8QwsIX/EPfoB9exlB8BgH29Pk7rIfsW0JxTjv/8Irf
+          A7mISiuYqiLPnGfeavzLT28bIOy+r/vtXVSxB2t4vfmsT4xo8wsluOkhbG75dl29dAer08tDms3O
+          dPM7OOjdrwxWzHjM5ovq8b94Q65YSA4n5ekODqojz8xWP6GsXRKhMPszkrV7RDf9TuCR3NZtcYWl
+          zpdOz8GGfygNak1lT+KhFXb8bPrCxenoYqFmFL9OccfK1RXqPj3HhdieOQbL0meia30NZkhgmGLX
+          HJ71GLn7/rg9r/NqSxYlJ4GF8HQYHORlH5WSWNoW1X4jz1+1Ss8OrcZ2IBAP/szvBj6iXPzdQWmC
+          +eY/lZT47FGByK9N7O+it0o0whBwZCQJ2+UrBbTZpQJcdzpFEu/V6vo+CET4dCTBz3MpUy68VS30
+          R+Hhs1aoUprGVw0C9ZghA0ZdRmQQBL/zQq457Guaf1+9mLNah5L5adTsuN+cBIMp//THbA9BKk5b
+          C9/zXL4oEc/vQrxvU1aj9vUAVHUtDRLt4SNv2r0oYbw1he+js8fyx84jvNXb4BQbq8+kwbueqlrT
+          xM0vxVZ+fDm0Y7eWU19+Y11J+4h+3Vcpbv4r9tG5BctNnyG8trsOeXFSqEz+zFo4Gk2IYvn9zjCT
+          mPlfvvbQjqrLeX7acH4tHnbQstKxCvsVcmrOIfk8V+qm7wPoJucCaaTaUfL95BLos87AOdLaYb1X
+          YSdOt8eENDcKB1K6pgZ/frl/HMpstg6tAHXpeEH2xj+WyzqWsPFKA4U/PqV8pFz81d+2+lS9BurG
+          p0eoIe9+MobRfLwC+Kx0awb7L6ZrBAcbertLiu7y51l/1TCPYXHms7m5HbZXLmQqia+Ir/DmL9L1
+          V+/44a0y3Q8RzZ5dIQRDDHAe+gZlskrQoXuYLaTUry2e+1GDyl34YNl0jgBv+kcU3t2AzuTZA9op
+          5fWnL7DsP6FKsOdUkB20+KcnsnnnhK6IdUr91YzdbLpI7uanFA4yKrvJpgnaI3jq8n0uHkuqrncj
+          CuDJJVesSKTMZq3srgCd1dwff378B5Dxl6+Q3HCpuoz7rws7+65i9501w8Ko+RXGQHlg51W66qqQ
+          Swwjy9eQfEJrvSA2N+FNKM9YUXyVrm0O0l99YD7uGBtwm/8K5MvjitxHcwfrVnaBm770hx++/uon
+          p7B74zu+M84symcJ/ldLwb/+4z/+19Yg8E/bPYr31hgwFcv07/9uFfj34d9jm77fv8aCf+YxLYt/
+          /vO/WhD++Q5d+53+99Q1xWfceg1EgWf/2g3+mbopff+PX/1r+8D/86//CwAA//8DAGZcw6K6BQIA
       headers:
         CF-RAY:
-          - 96665c82dd8acfe5-SJC
+          - 96a9b549acfd7ae2-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3002,14 +3001,14 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:21 GMT
+          - Tue, 05 Aug 2025 22:25:04 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=3JUChou_j71oqT.eFAUEI.fy44NOfmyXp6zqhHXRrQI-1753726521-1.0.1.1-wgl04pfyWkQqatEBB7WMBsJNB4PcOG8s0bV7lGUXj6XkyphhxtSaXp4qwSgFfoWAjHAwDmZ1HCKDGEHWLbEZUCG3qQ3uAN5qbZZRz9iLWtc;
-            path=/; expires=Mon, 28-Jul-25 18:45:21 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=3oHajtAXtPGAPMX1lFbG3r4TVKH678Do_cGiU8HRHnA-1754432704-1.0.1.1-XNS9Y6eHLS3hcu17z7nIn3o8vXVSXNPAWhecxSEmzrUOYI23Z110.jad9Ik4I1s_15yYzNP6_ilRX5x8HliohaUxpBHV9IivkhA8Xr1wVEI;
+            path=/; expires=Tue, 05-Aug-25 22:55:04 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=Iz7PZK7VRZcOT.9zgsctYqFDIEH8voMPmjfoBzqIlEM-1753726521306-0.0.1.1-604800000;
+          - _cfuvid=KGGnFxPtFH8_4mtbwcvqQS48y.vrg3E8eghbLIaJuAY-1754432704162-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Transfer-Encoding:
           - chunked
@@ -3028,7 +3027,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "403"
+          - "1112"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -3036,23 +3035,23 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         via:
-          - envoy-router-86d465658-xks7r
+          - envoy-router-canary-558449d794-9bxtw
         x-envoy-upstream-service-time:
-          - "429"
+          - "1115"
         x-ratelimit-limit-requests:
           - "200000"
         x-ratelimit-limit-tokens:
           - "200000000"
         x-ratelimit-remaining-requests:
-          - "199998"
+          - "199999"
         x-ratelimit-remaining-tokens:
-          - "199979817"
+          - "199979820"
         x-ratelimit-reset-requests:
           - 0s
         x-ratelimit-reset-tokens:
           - 6ms
         x-request-id:
-          - req_9556291208b21712e27cc5dc3faed125
+          - req_b735777139b80cb557a9fbe79d107233
       status:
         code: 200
         message: OK
@@ -3076,7 +3075,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3086,7 +3085,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3213,7 +3212,7 @@ interactions:
           98//e/xf61/917/+FwAAAP//AwBe8oQ14CAAAA==
       headers:
         CF-RAY:
-          - 96665c86a8e9cfe5-SJC
+          - 96a9b5524ee87ae2-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3221,7 +3220,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:21 GMT
+          - Tue, 05 Aug 2025 22:25:04 GMT
         Server:
           - cloudflare
         Transfer-Encoding:
@@ -3241,7 +3240,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "138"
+          - "91"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -3249,15 +3248,15 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         via:
-          - envoy-router-5fc5c4954b-ssdrl
+          - envoy-router-dd5dc9fdf-5ctsf
         x-envoy-upstream-service-time:
-          - "140"
+          - "105"
         x-ratelimit-limit-requests:
           - "200000"
         x-ratelimit-limit-tokens:
           - "200000000"
         x-ratelimit-remaining-requests:
-          - "199998"
+          - "199999"
         x-ratelimit-remaining-tokens:
           - "199999930"
         x-ratelimit-reset-requests:
@@ -3265,7 +3264,7 @@ interactions:
         x-ratelimit-reset-tokens:
           - 0s
         x-request-id:
-          - req_977c47c171fb453c1df8c5a9aaa59338
+          - req_edf59a69e2f54609110c9a2d68252003
       status:
         code: 200
         message: OK
@@ -3285,7 +3284,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3295,7 +3294,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3311,118 +3310,118 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//VJpZD7pKu+Xv30+xs2/pNyIgVbXvmETmYhY7nQ4gDigyF1An57t39H9y
-          uvvGRMSINaxnrd9T//Gvv/76uy3qqpz+/uevv9/Pcfr7f3yvXfMp//ufv/7nv/7666+//uP3+v/d
-          WTVFdb0+P/ff7b8Pn59rtf79z1/sf1/5vzf989ffXjSU8/40+kXPy+YdVg/T8J6mtYDWNIsQ5lJt
-          z/v7aa+OjNfHEEx0JceGlHQd5rOC8mgvk1hUkmINdNVC3PtyxJampvbSelUKdl1OyGmUHsVkesiD
-          UyoAfJ1tD+yvnLhAwZg1fFT4EOzZfGKgDoE1Nx55RJuWoxy23tPDiuCSYfPTlwZDRbKwUm89WD/A
-          14BrTgfvMMbTsD1h4KNMiy/k5iVStOckdYFv9lPP9fCs6GrvLgwkxDt4E9h3gF4j/wlT/S4Sg9Gd
-          iNtNfo5g7RfYPyU9GKnql0gpDIvEQTLRTQ6WEnbJKOMMVK9oc5XWAKVCa6JrF6vgXupRhFv54PFV
-          sdRiKnZdDs6rUs+M73v1SrSgResSY3Ll4qVYhZs+ogvUAhJJaaAunXqO4eMTlCQ+DeGwPy9yiw6y
-          JJKjrLXDOlWAg58xKXD6TIyCC2paIavMAK64DwPo/lAq8KLyAtbvRgj4xTqkUHqsb1LKJI/Ya88p
-          sJ7EgNjd/FA3k3/d/3wfMyujzm/atxByrwBH07UGVOAlD4XlvcO3upQoG1yLBWYn3SRGWe7UWZci
-          D+htFczz8NIGvtBlC0XeKpBwC9SB38CHheHz5ZMgulk2W77kCvWipRIPneWIf22+gmxT98jxNC4R
-          Dz7cAvG1YYhWs01B+61JEWv0Llbsc1MsrHDJkXAILaLv7JvKZa6oAOKfEdZOwzYsFr16QHyXkMTd
-          EkaUUa4veOD7hkRHvh+2eZ0MqIfZgBPot3R5hqMDDhfWJ9kG+IHWudyjttgqYuf1GG1BDSrwcvY9
-          PsmbpK7UEDOw3M86sVFs2Jy6gAzs6AETyeo8Ol3lAwszz9+I/A51wC7+4EP3vdeJQT5RTbngWSH1
-          Tnz8+/+rxEo6ajtpJrFz96OtPTU92vMNIeZUFHTJi9QH246fZmRFL8B26zADzpHbGdyyDmw35q1D
-          veFf8+7DRyrvBycFCvH5Q06kScDyEJMNqQZzJ1q5RcXEZ7khrinccDxVcs0bQyggfBf2uBgeQ7Tf
-          ax8INoudsTS4MV1CV3gB4V6x+DceVJHjDOG6fROvOxN1OvCJDp1z2eOkYGixaQ/fQac0/HgjYB2b
-          q9cwRLoeVLg4dY+CN5FeooXLRqweeWugqjds4HIBFba9sRhG9GpTAE+mh8uPcQVbctQ50LFDjo/m
-          QVf3+CNZSKidFz7rbhjNn3DqoXViFGLK+VLw8Sa9YABgjf1FCQc+ZX0G1X6vzCCZZ7C6XJ3C8yrV
-          c+udg6EzPeTAEyOn3jI/BNAURzuH7tyWOBGel4E/5J9MTDt2wXjHH4uND/caZDHHYbNtqoLQiAmh
-          37k9LveOVlD/eefQgZMPJLSCnU2PBTVQ095uxDOoaW8e3+nQDypAHLexbe6gBiJg5tTEphvI6sTN
-          mQG3IWexZD6lYZ9/IgmUvkqJOjwZumJ1W5DHus+Z/Rh9sWKBjsDy+f3cg80fyPIZRORUXYQ9wH2G
-          FQHEQdE8E4yHtxLtP0uggzgjBbHwsbO/69+DfkUsYsPWsdmjJt5h90hexJnUJ6VHYZoh/9Qzj494
-          NxrR5RHDyM8bLCl3t1hufsT+9B0X3HhX+R1aU6gt84VorN7VSxmSFlaLcvSYIpTr5dozEowWOOL4
-          8fHp/j0bs7CfdDzTLFWLLZAaHSEjyEnBHm/RYjGlBy+TG2O/k27FJD3nEOSPY0BsHb2L+bv+kc0c
-          XVISVBS/eoCIDVRsA+oVe5w0Kfz+XyzjVaTL3Dx8pLKyTkyGNnYHa1JBeLI9bDCbUXz1Z4Md55+w
-          /XJPA3/f9Rx80duZuFakUf4qMyN4PpwdxpCJAP/CVgtHLt7jOHrWEWWMJYeTUioeHc9wmE5E9OD+
-          LHUk3rs82IQTp8Pjyr7JFexNwPFXqIn3KTmQ+FLhguafs4CCIr9i6YHe9jaGvQeXSOe8oVOO6nIZ
-          BQ0+1bzD0uvt2tSaBh+yiXglSds/ABtIDwcNIMRYRUSnZIcOMayYyxkHO1Giy054+ch4DSp2GhnX
-          Xa8l+sF57CxseOe13j4Xo4K8tzywNAPN5hFoIdQv73hGDxcDHup3BV6N0JlfxfAC+1JPFXh0YYtj
-          lF4ianr3GMG4qLAC9Ic6i0LOikshOETVNKlY6httkFqGKT5pUglYK+1b2PBpSbzLcVT78CQJ6Fwv
-          EVFC1qZ7S2UceAzCG9YE8LGnLFdbmErX2CvkcqDrUNYlSo7ZTOT5mdZbcx17mBRlTLLkogFWgKIB
-          g3mMcHkbsoKzWzlGSGUfJMZsrP4ZL+Z2c7xVOd3AQoSSg5ekensj7DpKjz60wO0ta+TEEz6ix530
-          hOEd1cQ5fNKBqkOaopY+PHJ+po26XMznjLaoPeKcC3iV2s8xBDC+VN7hq09UiU4QjlVzxFfhoavL
-          vLYh+gAn84S7NUf0dWccoDClhC8bTKL1XBEOZEHTEanCe0qji+gB0/YJOQ+2OlAxtR24+WKCzSRX
-          avZ5enuoPnMDOcY9qSmNOB9ek3tJ8tVq6bY6KkS/9Wy9d7G9JtMaos58sSQE+sNeMA02lKfNSrzy
-          sqjLObzECDO2RmRwrKNtFf0UYFW74bLwdXu9YK2HXSn2WJKm2qavO+eg0m4OxM3fHzqqSm/BEROD
-          GDbphq0vfA518SnG2guwtHO5OkZ16jLklLuvgZJkiRGpB4qVd6TVW3iSRNE9GBD/9Kq9Ma0D81t2
-          IpKYmcO+MpoMWWH5xBd+3ShtgkiCO3AUsHRKLLqdw9QBZ7+6fuunBrjuhiSoKYI672XNGHhs5RVM
-          n+2JVF78HJZsjX2o3iR9XqzOA/ODmAL07UtMHEylgcv6sQF2mlZYzohoz6SZDdCkW+rBnbHYsyiE
-          LMRplmBz8y/RuH9GOeBLZSWG+WTttShjBQoJjHFVNe9hi2I/RNe7BLFXy2dA23cFIX9ENk5tebO3
-          ewqe8BTdeaJ/OBms94NtwftecrH91cc/35erip/Hj4HA/FJdEQqn45kc0+uB0qcMQrGvFI6osjLZ
-          76/eAT2xinlBUjDMwonRoKD0FcZKRerZPkAB1japsP5MjGiD4uDA87nT8ZGEZs2S5PTHbxK7kVZK
-          z/I9Q/wlkL1DHBcRu2uEEaZs4s/E7bSC3q1OhAmUco9lSmmgQ20bUHnJiLgz31MiO0UG6fb64JOc
-          sXQ0bkEIry0XerRt9YKA+dKIMfdYyXF45RHdXHlESMEsMS/7GGwv8z7DB0oUchSeh2Fm7qYCXXZU
-          sTtpUr0EBKdA3b0c7FedrP78F4SmwJDSOK7D4kIpQ+hsnedduGz1Kry7JyysaiJKne2G1eDmBfzG
-          41tPIi5J2x4+yg3Obn+91TN6TR4w6F0il55GgD9O8yiGH9kk2Edt1F8KM4Ps59N44svpvutJFtB0
-          mwPi+B5VqV96iqhjjSXOKkgRhXqrgOHDXIipgw6sXMmlsHIMA0eCaBZcmMESPlrBxAHDXuqFl6sM
-          jjAriLVyGuBHcLmDDlULOR31l01/43tFke3dL8apHj/hfUF8Ka04uYq3ghS6bCAUaCuxHIHWC/8a
-          FghKvSXY+gTqIofHGN7cMCbmg+GKTVU+CmiNqp0XozwU9FsfwTdPYHl339ejADcDJW+GxXr1uajL
-          YaUWEuLk4yUFE0UrszAsvC/dQjRrlw4d2fgN4nhtidulQbQMLdLhi0hXEnuyCVYVuRI8XDgf26Kk
-          qUu1+hyyW/VDDJoogDskbiV6l0nB52YVCqKtq4Ue4jskLkrVgaV3VUIq5Mp58XeNuj70/AXt6+Pk
-          gW6WbRbKIYOY1q8xPvSEkq1CITSDNCVWMAX2bOYHBp665krk1WLqaad1OtgyncHWyr3ANhxzDna9
-          IZLjtC42jVXWQYbvWV4LtIu9LM+UhaUXd+T2qnExDYdKgvsl00lShOfoTz2t81cx74N0GH7+HzjT
-          08cWkwv1ON/1CuaPU4CtCxfWaxEkCoz2xdsjDzxEk727QPDNS9iUoqc6t55cwqmED3we7Hr4o4ed
-          2bDe1uQb/ek5jORHR9Qz5OlyPoUpmm948WZT5ItN8HwRfv0kVlB8pasfWJZAQKwTK5ETdcNjr/+p
-          J9fIUCLylIEPxp1lY1Uxi2Id9XeFtvfZm1kSDGDpn9IT7Zk1w+FU6vZqILOB88uKyekWh4ATp4KF
-          dB+f521SRLBZ1IghlmKCzflRq0Q8ZA2SxUYj1r3K6+lRHw0Yv7gWW6Qj6mLRxIELXnY4o8mTLr/5
-          bNTyPv/0Y51x1sK3vrpENe152Jx9NkNvz2HiQihEQ/2WWuQWaYOtpJAK7p0KGnTnvsQ2s3/UnK8t
-          d+TIF83b3x9SwZ/B8Y6q4vTApws9RGvo8DMIqO2Srx4XdP9pS6S8VITlUmXU6eS8ONihcsFfP0JZ
-          3SY6rG9qTPTkbBac+oYZVMhHxcrVqtWBMZYM1MB4eB8YvqIV7s0Qvhy+n3nI3Au61woNvs07IrbJ
-          M/YaRuETXoqnPe9Okasuv/r49Z/4bCuEzhcFKKJ1GWxvb9CKLj0DdGCcc4qVx06q2VsRsjBVNBnn
-          xlUu2Og2WjB3aT7zcbgD255Zdagt4wXHkZGpe4XzK8Rejzq2z7iNlvHKaJDreIbgUEtrnuHTJ/zV
-          d/vQCOBdsM8cnUzl7DF00u21HoQUZiJzxFqdA5WMmWPAt5LG33x9t2d02lr4HJcjviQSH1FppzLg
-          q5ceU8s8IPSdabAegxdOvryIsqPTQrCvjjM6tGxBLeHpQ3iIZVKqkADasPccyCrMcJSldbSZ8c2A
-          Irg/SZjIck0PBmXBoKqcx3z5xV6a3j78zu8sfv0GVV7dApJjPs+CY0F7qZqAQ3RBosdWu5ZOH5Bp
-          wGLN4Tv/u2gs2qkHYF8ecdWotsp3aEqh+K6gx1zFrv6Tx7JPj7GVP+dhnbHf/+FPckg9lT70WwUK
-          OMGZLQYN7PVnz0F6wZa31AsppvtBNcTmWRyIdqlw9PUTEtSqxxubsFm+9SDN4M048ER6oKPKs6PW
-          /+op/vo9lYy6/0Kz2ng4Vv17QX/52B9Qjm9Jltc0GqYNJuTgfv3OvV7YRJN+PMYTdM9Xt68/AAxo
-          LvNvPoYK5SW0yLxi+S6A+psHOVh3NevNZMfR9YKdHnzzlSe8hxPYmBU2wtcv4LDeLEDnPL7/8gjB
-          m2tFS89QDR7fU4a9IpSHDYaHFnbzm87TC6vqyruRcTirnos1qbvVC6zcJzwqlyf++mNAu/u2IADb
-          GWtMeR9GY9VY9F3PP75IaZJ8Nii6o+PxDyiCV87UPcKqfiN2lnJg+eXbI9skc9WdiU0dZw7/7N+z
-          +rYicsO6DvfqTLHcn7SCK0QthSxmOZxWn4O9DWc9gxfDSbCERnfYOlWpkH9+Pua1vHYR/eVFzbju
-          8BHvVzrHZhCCOZNPWB3PsJ6uy5kRPwEHPVbqdgM9ILgcWhi/cAyxbbO2PbTwfvxYRBEO9rB8XnGM
-          ksm+Yid/WfYvfyJ+n3vYDEZabG+maxGcAcLHhkDQOycnBQILJqzqdVMvHxLFf3iG8+UtVD8iA07s
-          FHjsj0ctZwvCWXi7RC2Ph2Frrq8ebhY3e6JxfUT7gu0z2ObmSkwdu5SuwSSAL2/xmG+eoyzONfDN
-          D95QMVW9wvpTwXWSMpIxWxstwmn1kL0PeI+mpjvQjOtSOJntgI3BeRQ0fB0cyJzmq7dbH3O0YF/L
-          0Dc/EQXFiM65+WkOh8nQfzw0mllzsdArgzz2/V1j0/xzEwDf1AE+fa5ssX3zKMxOmklwr7WAyqTd
-          4G/+fzxu8+JoRHA+IPLNU8OafNISAtjPWOJ6MaLeZd//Wa/fPFRsnWDnsLk8daKTXR3RNLBfAL2y
-          D3a9+ayu9vmeQwNeti+vOEaNZF97kLeBSOTDDtpfPyjCbx6daRUSe12DToJhEks4Mm2v5l9Y6dHK
-          3wVipeytWIMnmkGcxwa5fPnT13+xoAz6+4wIc1GplcQi/NZLfMpdbeDanlVAPUYvckqyj8qHZPHg
-          kM/mn/lfE6sWobskDZaC4WlT9wVbWIHhQ7TqjeqhBIoOSb5IxJz2H3URLbkBuZk+iONNl2LjHOMO
-          045biHusPtGqRLb+x4/qjLYAYhjuHZZx7hMjGJJ6+/kpC+5kbw3pbA/2cR2RpHAOPu69/UDaJYG/
-          8Z65IkrU+dRfDTjdxgA7hd+oVD7nOawnIcBXa3cHW5EyEIKnhbGpfo70y0d0qOtRNS/88hjGg/de
-          wMF5SeTqAwdwYnH3oRLGLCm//ns6decMGv3Y4firF8tZ4p8wvaU+wbj0ag4/Bg1+/SyRXLTR1RvG
-          FDw46+ItX/61grkpIbc+Q4/NzrJNjCEXwM7gOu+SzfpA8OYo4CV7xrc/4db7Ccw5DPWbOsPh2alL
-          vRIDPoxU9bZEftSLU4sSHAKBYnVSs4Feo+wOO32vEb3XmHp4Olt4iA/Rg3x5qL0YUOVgsWo5Dqai
-          LVZ7FzBQzYfuW1+8+o+fEbfJ+er3Yq87Rncg2zkRTj6mof7080BAqnto0qSBn9yjBavi+CB2tjzt
-          dYogB7/1DWsEnej2KkofOtUQzQeu14bptb0EEIgJR5TUIGCydwEE7YM7YnmDD7BVt6eFjtnVmGm5
-          JgNFbvkCjGOn3/7BzabM+8nAZg05rNPoVKzc7Fso74TPDKbr+cs/PhxEzEHHyqLN6vabv/qtA4zL
-          3AWz219meJlwTJSpbFQK9bsE4jw1iJesrbpNp8EBv3z80w9ac6wH9y3ZvGXnftQxuxwF+Dp2MlG2
-          oB428Iw5oDuJhh2CimjbZA3CuHbBfGx7GdBzlUOYZWv4zbP3euNl8wnuRbsQGU12sX75JHQcaOHT
-          KMnFuPh1iMyja+Lw1J3Vj7nhCgZLcf2vfkDQXbPf/sexN73BWmHZQ/gw29gwuMzefn4IH0Ybh6mB
-          6SZOEYd+/v6rX7T/5bFvvwyb/LqBUbiMC+fYyZVcx0wbWMDad1j6MsWy8hjsyXEeGZI/vkds9+1G
-          bOulKXAPFiTeKcTR6NSbhOZG0Mjtge2Ijz3giSl/kmbh+7w8sDUP3gzA41M82t/nE3vw8/cOU7Lq
-          GBAci+ginLw1jj+AuIT2wFSul/kd1Q918J93Fr6DYCL2t38x3HdPDqDk0WO5ofdoYvigRzfr/iLG
-          7sDV9F3mivit119+4aqLdGBEcOX05yzIl62maRcw0M/DYF7IR1LZ+gYa+OPtLi85BacuNIM8XR/Y
-          /vJpgtVdD5r2esOnb17cPP6hIZPtHY+uAqeSmoMe/L4n+mAdwT7QVQOVLJNi2U7bYXu0RAddKfQk
-          2yMLbEr2VhA9PeWZOd7e6rjQpf3xtVnIJoMSa6rDX/+IGO3OpnNzrHzAvO+UZKwl1V99lg7f/EQU
-          PgsAyztJBg3LqomZKdZA0ydvCLuJb7G1LNdhfSdbDP3kwBLLb55g/fnrndxLHlyrvbocvGmDHlla
-          rEVGZs963TbiI9o18wU2fr2lbez94XWGY0F16reu+uVbD1RMNazefYr/6M+RUSN1/e5v+NV77xSH
-          N/qHN8Crx87iNy/tz4vZQjOIU1y5Hii27bq7Q7OWbzMUZfkPj0O//oK+fYbo57/APOsT0Ut2plvM
-          hA6YemfFbtsnxdb2UAKczzi//VeM8tVJwbf/6B3ufKuytWpKgOXKGtuCCOnPX/zpb/z8N5263Qw/
-          wMs84bt/OArvM7q0W0HUbx5kDw00fvz6uz7GaIoFIRTBemznfV0IYElTdAf+a1vxKeKnaFnFyYGT
-          2Q/Y/lBoL5/wvqHDZOlYE+e1mH987Ld/tep9rTc1NjgIMfOa0d5V7B+fg8RPENYgHuzt+bJE+Osn
-          eg05qZvgZQL88v8vD40Bn3GPGB0YHhO3v+6GxX/wzE9/PPic7Joz47MFv3kOq7tnYK8BcHRR2QQR
-          +45lDMsoq0+08k8Ba3RkbDpYcIHehSgzqBim3lK762HqknpmjscDWOFe9tHfv1MB//mvv/76X78T
-          Bk17rd7fgwFTtU7//u+jAv/m/z02+fv95xjCPOb36u9//usEwt/d0Dbd9L+n9lV9xr//+evw56jB
-          31M75e//5/K/vj/0n//6PwAAAP//AwCh+t4M3iAAAA==
+          H4sIAAAAAAAAA1R6WxOyPJfl/fcr3npvma/kJEm+O04iAhIOitjVNQWCIKicA6Sr/3sXPj09MzdW
+          ibGSHbLXXmvt/Mc//vrr7yat8sf497/++vv9Gsa//9f2LEvG5O9//fVv//jrr7/++o/f5/83Mv+k
+          eZa9vsVv+O/H1zfLl7//9Rf7P0/+76DtHz1NJ9byrnb73rUPaMmt6Ra1u9jtR7YDOLRHe2LLBxcO
+          FybxoIDshWj5+0GXthdUFGtXhWTAu6ZrzlUGWk5Ix6aWXQHtkzwCXLcSgrFU2sROzy5c9BrgMChd
+          wIlax0LX+Oj4nE8BYNs4Y6Am3k7Txw/KcL3HYwK5J+NiTC2Sru+GlSGyLics230HlmqPWWlfvwV3
+          LyljvyzS4iE5hHcSM54cCp4YzpAq1WcqhuCprRVqGaihD+/2gt8CahL59VsPObgPJ2Tfg5ygNSpS
+          nFv33h5nqYiRn8wWCUN+rBabejH0D7GC86yqw/npxjoYmOpFjiKxUiGo3jxcdqWA78mqgqlNTzmQ
+          79FrYtzEreajujSIe0KX+PQ4p1TsjQFF9uCTR/D0tdXNhAuM+PFB/DkLgNBMfoPkad6T8yI0/fyS
+          bB4253eKU9U1U8FKqhxFvQhwYhg7m87CQ4XLrhIwfu6CXii9NoL5cnoTb/wkIV+UhgrBiwmIcnZK
+          bSZcXaDcgwDb3o6pJnuRGjjXsY8v11sFFsTNLiqfQ4sfha9o7PViz5A2rxNxqbzTSKtrFtBiw59I
+          CPRemOzFQnA6iCRkbK1n4+7LwsdF9kj+kixbYNsyR88s14gmckrInxZZRUYkucR4KHPF1u/PDMPM
+          YMjRiD7p+mX4CLk2c8auPX5SOvFtgrhDYG3zPzXWLVoGFNUNYTNp1n6Gy9kFsdlAkrffIFxUF9XQ
+          vBsfkmp119NDczYhs4g9TtlPQ+fHAnWggItH0vokgDV7Lh26EDUnDqZDuB52dg4Ou32HHXGVNfow
+          uwdA79AgrjibNsuiPgYHsrhEtRsckqt8YqFbPig5ipMBhOgKPMiGmUEMnQ/p4uSvHHnTDeNf/LMs
+          yQa6ifVE7rsbpmvJfz7oJViEmIX80NbjjtdBzlfTJDHnGrAqsD+g+t6/01w4nb1I97cFz134ngRM
+          Qo1Fy4eBR3/3JQciXwFdBG5FB10qiELNEIzBR3IkyZdXnF5jpeIP7UtE4xxz+LZL+pCfxFgGZRJP
+          WEm+F7qSs9gA7p2z+ASiIVx7o45Ra+k1Oe7SiQ6axxnQM9kO+/JM++U2FQ6KfKt128JyUoHwa4AG
+          yuU4avKyF+6D8UBvDg74LHNWT6tP3wFzf8sxtuwUEGlpItB2votDQ8/A8nKNCdTsMcGq6hkaD3LP
+          QvPOrPClTH1K8hx1UOkDlVgjmVNWPMg1zO2iwsltH6RCus4MEk+JOtGgnMAS+jSBX7X2p5GtPdDJ
+          zujA9Pu+urzgsva3AiCCzMJmOFzme7/tJ5S+Y7zgg5Uc0rmNDzqMSlfA2C5ze0piJoBF/O6wxyt6
+          ulZBwaPX4b4nOZV39nJPNBOFAngSWRxPNlX91oC5xwBic9C2eb0NVHDH/Amf7g+5IsW7MWHqrOzv
+          e8+6dmUCFdCVHO7OTptVaR1QYu1f0xLUXboUr2oGt0gQp/bDe2BQ30BCK7+EWBXCb78er2ce3pQn
+          wUaI1FDAQmmB632XEu1stfbiPh0X9s7NIiedcWwesckDvk9ZRdR7+qK019EK+dW4u1wTnumUccoF
+          eq71xmqWn1M6O9UMhpghOJSsQmNp4CewdvOY6NqrDVcW7xoId53uMgdFobMvuDI8zcWA7x/eo7x9
+          tBIxWF1v4gpPS2eeiVQ0XfcJiU3uGa73cXDhMGUXHIxVDqZd4Qbg6JSY6MR926SGVwN9UXkm926f
+          pls+6eiTphpWi5Pbs3xrRHB5pinW8VeiNM1KDx26+4HYWvW1m6XEOYz43sWYmGbKS+qHhzORj1iN
+          j8eer5HEQ4mvbuQcdToV+oPbALuRd3jLZyBwl66Bcq1zOHztqnDF7pzAQnJ0d54g7IkGOxeuUd2S
+          ezgI/S8+eKPOm3iFY9l8HaYPid0rexLdI5wuei2IyH+/MvyLbx6fnQvzKODdceQP2oprU4dr9Gkx
+          jhoHrIGUenAbT9KhLgF7N3wH+dcc/9ZfTRl3usCWtDecs5Oi0Y7oHprkp4bt8oGrLpBqa7+9f2zS
+          01LNw9PMoV84LyznnG6zshRDuMcHf1olGQO+OXgqvPKROb20dw24a/BRoTYWDU5kcg/nuyFf0O77
+          zPE5aMtwWioJSntQOMSORjml8Eg/qJHUCB9S/tEL3jlp4GtIHuSHL20uyCyiSAyIbWCbCrGSO3Bm
+          8yd23s0HkIxWNVQj/+xmydBTaiXVA2mNM5Lzi0ThanROBwMoXkgIeR3w6NQ50PfMC07YR9xzwrhc
+          kKc9ShIJ7kX7s1+cBGx3vapPsD55h4djx2TuOL9aOjc36AKpX3RisoVQLagoXtBklYpgwYnSVRE+
+          EYr41iXPdv1o67xXOwR3jY7Duyto1E5gBKpAyV1kOHO6YnoT4ZEJDvgqrgdKXSEO0KnXL+6qoymk
+          jDk54PXwZHx3o2u4liVZwYanREsqjq4Zl7hg59SEPESq9ZTrbAfaR/6KLSioFZceri4SHL4nBvsi
+          FX1/Iw9S6j1IwDENpZUfQnR9hhE54f5iLzYtA5Q6M0seICvt5d35K7K/zEKOjDFrKzzdL0hDvU5M
+          vK/C9fWdXyCX2Se+a1/DXhLCfiBXGD2WY6uy1/1oOEh4SHvinI9fOvamZUE1/ZpEzbIW0LGUedQr
+          5ILV2uaqDrHVBfmnM0PMoK77pb4UFwSMnmLrq+gVTTIvkVyehfg8v1rQvZDpwP3MmkQ1/FMvOIYR
+          o2SVXziw25XO6ruSIfaX/cYPLUoDNzLBPfk8tvqpA+4rjzJ8Zg9tAvfE7NnvLcnhtHpH4jP41S/J
+          Qb9A8aIfp13KYHsY2b0Ig8NyJYfjJPdcQuAHKCDJsSG+JXs83XIHICu5udR3ZpsQJxDhN/OuWMnC
+          ezVMVpWA334aW32ZrwyrQl6IL/haft79LCZegNbyAvGpE29g/V5yCO9cZuM8SFf7xz+glA48UReo
+          AEoE24K6ZJ6xgetnOIehHKC1Z9bpuzIIDNmabXhV3ohem5K2FGfblaR04omqww58pBgnAD3XYppr
+          17fJ8pp0GKpBjm2xIXSIdo4I3y7Nf3gYLrIIHNgtrYH1ZDpVHCTfHH1fMyB6mS50OchejMLvormg
+          sdKQ52RxgCs8uNMkEr2fFanlYaToqbsSX+7XuulNSCOEyJmeOzrFY3qBdV58sPlkWDpFqhLBVbEC
+          VzAiIx2+Q2dJSYJmYh2TJFxi3R8Q9b8sUbT00tMSyBPcHa8qccd2n07R7a5Cp2U1fO4tOVw0/IzA
+          M2DP2K9jJVyGdU2gcKsZ8rxfln4FJ/mBMGPcpt1zt1ZLv7Yv+ETdSMzdedfPsz2J4A+eCa4bCtIj
+          7iBKXnRS3mOhTZ8QuQCyUCbPvAsB6wu5KNnobhI5KxvaN1obw1LqPy6bTG1PXh9FRFt8xGArqq1F
+          6aqSEDgcUeBeDjd8kMBWb8kpZ1uwPOYogqs9m/gK6SnlxWF4wC4aTtjfzteqCNMFHiwnJdZ2vnle
+          OdUgzNSZqLFR28v+bsewZjXH7erbURt9X57RvHMWnICysMlkLyba+CDRrmdaLcWrn+G85xtyfL99
+          bbamdwzfxeu64SufLo8Dz4Ct/kzw7uzTtZslCUT25OMDkTk63dTAQUPYsfg4BPdwlg7URWqBAteX
+          57BakylnIWxGQuz6dAONwAgrtPqsIRrR/JDWMDMgOQ45uZ4Ny6YoG2WogMjDx4Osa0suyDzyKP0S
+          m2AVCIcA5dL2/nH83Ikp2fgN0s9jQJwNH4VHrsmo8/PHJDbpR5td3ipg3V6PLscbii14oiohsXYq
+          rEoyodNxN7rQYT8RsUHoAaL5ewbycM2IUR+YinzjvQGmw8pgs51rsOaWxMNNP/3w0qbZSXeQfVyP
+          7vdq3O01mSIW3pDYknAcsT3wTC7DpBQN4gunW8gyr0KFWvZIpsXHfb8qr6gDv3j1dybSaX/jX3CS
+          Kx8rdz+oqNDVDIwnu3TfeOm1SW8TGVh2wOATDUs66eHygKYblzi77aqebvFD826xrriwK52P8ekC
+          P8W+JXLjidp8YYIAnaUbdUvhJKSUvcgSZGSbw8eVyegctUMkyifHILrlXTV67SQDGh/4IMm7V8Oh
+          rHsPiPlqY/dG03SR9GuOvqg6TygZevCrFwjNyh1n16thz0d138FkYS5EtdsAsOnRZqExeLeJN0UJ
+          zM+9eYFDDAm2CX7RgRHND7rFif4n/weP58zf+cRnrSTaUopXBy7Y2+HLKL7onBorgw66WEyzvjPS
+          9WQ0zY8vELU4Tf0KqDjB+mRgYgAsVL308Bq06UOs+5Gc8j9+ZWpThtVoLivhq3nFjy+6KGRkWwhE
+          rkAHqSqxltT7kK7i7QX2enompsk908X/xA/006PYLhltYqwLD0/AmXHczjoVdubOgLvgeCWHLZ85
+          pYIxtF9Ex9rGHxunkB2w1TO3nds6XBZpH8DDDnQTCpnCXlaQyvD+jhFRTMrYa16oL7jt/8RT56xR
+          U489GKymivPGn7VxV6aWRF7fs7u/TXk1Swfggg2/sD3u5Yrb8S/2j5+x4WPKjfRhwUB+JhNqjjsw
+          r9/SgLX7iPEjKWONE8YiR9x8PmJV+zThDJdchxUUGGK/86jiOBi94PGIL/i8B3z/NvpX8sNXd//j
+          Cwc5TiAzJAdsNxYIyS1yzN95wNaGR4M0BzXkrPqA4+wihPMppgwQ77nnzo0n2tPv/ezvXI1Tyw/A
+          HF0ef/TFtL4LNl3u99WDY8Iq5PaRCNjwPQeDVsQ4tGmlzZ32NGHowBdJGEcJ56nQIFDuhHfhLtd7
+          ft1zHmzMsJikoGzteT+fBmAsE5mWPQftle8XHl3rBbrMLm21ietj/ecPEAPXu/CnR8ErggecctDW
+          WB2METxbBnDRPWqrRd2PPDS1DuNjjScwP6u5g5ZcS+Qs8662/upzUV3RtGpvHQjWwZJgcaYnl2df
+          pB8TnT4kOB1FgqmFw2VX1jI8cOcP1o/1HC6vKorh094LRDYuB03wxbqDpRUE2LyHIR1zx6sRrVwX
+          p9FYpKtXewzKsJLgfLgk1aq8zivc6/fzxneKasY2K0NO4pELe9nT5sfqQPAG7n2S7p6Ydj+9uPEb
+          rMkTqNal1njYDxXrdtqdp8uPLx0IdV2mrI9gw3NVZONFwbeTatu0/rAFkllbI867tcKZDagJk0N5
+          x/Y1Vnqqfk8N1KuzOLULo2nLaLLzPlLU/8N/YIpekK3eL2yjHdfT1yeYkbEMBLtqVPSjNLEsImuf
+          YiO7BXSJwuMKlaJxXEkm+/S91+iENn1B3OuRBys8+Zcf/5/8q0fs9afvw0yecSZzVjXcOsOAh2e3
+          YnnT+7yxu0QwKh0B51G2t+fbhY/hTPQrlovzuacP85Wj19coJ9H32nA9MFINK8gx+PzQFjrcDd8F
+          k306YjtsoTa8vS8juTwPXR7Xu3SR7qO5B2iocfAwbZs7n/oGrktqkWPh2imdX/oFHR/fDB/qwEp/
+          +hNhcXKx6xs0nSk6NShmQoRxoaK0M3aPCExFP2LHenyqZf1oF7jxQ6yYNLd/88HzswxcaSReOKeG
+          BP/gp/UE+3SOLpcO5rI0ulTkypAt6ySG2SVbiRK9z3Q5DaMINj/G3evMoC3yWdIBGbHmErTPqyVL
+          by+YwCYmwQ8/dufSRRtf//kLYM24UwTNcO6xxThlutSXvQNPRZ67/CJM/+0/+Z5hkNOAEZ2q5pTs
+          A8ga+Lnx1dE2Cws9qSzgVBY/Nt1HWATC6+ZjFQ5sv9ZIYiFtihM5CdcGzCsUV/gNuicx/EAJqR6E
+          A4oZHxE9GtR+zo78A3aDO2HrmEghzXOu+3NeD+BRpbSdQALlk2sQq46qag3edg1OHPvFpmHetD9+
+          gAqWlWju80g/5yz7gJ3wlgi2bACGkV0keL965SQRndgrzFsZGhEr49tZcyvBOwcdqpxCJMqre6Y/
+          PwY0LTRJXEZ5Sq2DJQJQqcXGx+7hIk0s/8dPP0hHvecoqCQQrt+aYN//avwLye7Pv8WycPtU9NaH
+          PCyy/RsfefVlb/q5gcnh+/2zv02xfxkQN43880u0VQ+XHAymVJIDVe4pVaH5gI+Jn4l+j7/hPD2B
+          AUvcpOTEqIs9yUL2gI+L6hG5k67VEpMTA8u3rbizPUxp++Pztpo7+OgfuH66gIMIATWmCdjfqzb+
+          zuMzK3x80r4fbd38Irj55Tix2QLQ9cVAyJQrxsaNOVCarI0B2QZk06yGZT+y8NAADosyiWTWAbzW
+          zB58LSxL0jKj2iTFxxgCMW6xl7BFT4/MMYeXifGICq5uJVS+rcPXwrMEC/ZKl0J0oh8/d4EQnNO5
+          lvgHdKMucJetfg6XUBLBTmH2bnRjDoDw3cMAz9gyf+NDTsNuDjf/ZRIkvdVmm2ATaqaqudt6q+X5
+          smSIkoJi87mL0wUVTQHFnaIRdawY2mcmjfYw7kuyxQsWxQ15KE7sHWdR36T0pK8qVPm0xfJHdCtO
+          vzY63D1G2xV+/PTnB3BJE+JcnE1tbuOzsRcvxtEV5J3cCx25WvAglSWR5/5lr3fN4SFjLT126tDU
+          5pM2bH52dZn2c6unpNNrEXxmJJDT7kbAZD0CHWz1GB9zvgT0ka0WYvOTNYEPvabzaXoUwEbPiMhD
+          8LQX2VUZqMo8/8dvm41RttCFjZtJzC63cP7pg8p5G1gZl2njN54HF/0DsLy7n/vh0uwneJeECzlf
+          rx9tmfFsApB9TKL4bqMtH9n2gDtYwsRmyRTOzY114ZWkq7vb8mGcKCdCdykVYvaHqqcHdOFBHl11
+          fBybtFqNCwsh+fLSJN+fCljtRwfhkiiBy231kbJtmwOR9RZy+CA7XTd/EtZubGGn8BWbRFcaoO/K
+          nfDls96q7yd45hBmuwf51eN190QxRM+5wM/g8wZ/8PBxj+w/fsAiV0EDH/eLjX/9gV++osEUS3x4
+          wS9tT7evAeKYWFv/bAVE9Vtd8KZ7RvJNfwvwCGo4m4cVH/DS28PZVuI//aejFZ9DfvOvQGx2kFgb
+          Bx2fL1VGrlHrJH+/7ZDbOXYuaetOnpYUhhX72z8nDgX8868pKbsPwAaK8TFI2HBAywQl4+GcXKE4
+          f8FYvKoVbPx/Kla5oL15n1mYImXY6n2Xtjv+xYOff677UREOZ27p0JqKNTk1Z57SzT+VNv1Ozvfb
+          Wdv8EQa850818Ya9VrNblMwv/6ZdncohN2TgA63o+CTH195Juf1di2ERnktsBN9CI7v82YFQ2D+x
+          kjdiuiBe0dEEGdulC+E1UkHHhRse/cmvn/+OJFmKsE5x0y9tv1MBV+g9yWLF2vjMe+tf8fKEoq6m
+          46GXG7Quz3gS28UMJ8RWEZAWvyNaO9jVoAPGAR0vUhJcDnIlzEIt7n/znQXsA/4ev2Oo8EZFzBuw
+          /ugXMURhg8+fJUvnZAkusB9Klsgy8wI/PwPhppPdXXbltNlzshW2+txi97bewFhfmkg6cPgzXY61
+          F64My7rQM/lu67cAOiTLKf/pHXeH9nm/hu/xAmPNo/h0lkNtjna6CHk5vLpHBT4pdQw3huOccBPf
+          UjnlYv0+QC2YI5w2FkjXb7sr4La/k5DzCl19XMnoqAj2lh/9n3wGP3/F8U8TXeTzqoOfPsJseU3n
+          qbAh+NWDg7JL0/ESOAEwBLlzGXFuNM429zL49X+0/A3pj1/AljsS/OtHzrZOVmie1NjlrlZsC5s/
+          hEATpUTe9CD7qqEJ53Zytn7vEA7beqTNf5jWxtnb1D6gAnzkbsGq2I3VUoqZA+Vp6jGGFUwX35dX
+          NK3BEWs3YbGHdOkd2A1mgeUozaq1Choecu/kM9GTqdobXgUQOArEOvvu7fXQdhKkzrnDRtQetbU9
+          NCIcNSRhU14v4I8/fXQqTPQg3PU0UQQGzlkGXKQVdsVxvmDBza/GSvLx0/XJAUYKg1j60w9eCoHm
+          iNUS8ac/7WU04Qy3fsjEdixT0Re4f+DGB6ZdoezB5v966O/frYD//Mdff/3774bBp8ny93YxYMyX
+          8Z//c1Xgn8I/h0/yfv+5hjANSZH//a//voHwd9s3n3b832NT59/h73/9tf9z1eDvsRmT9//z+B/b
+          RP/5j/8CAAD//wMAwm/oZN4gAAA=
       headers:
         CF-RAY:
-          - 96665c88bce3cfbc-SJC
+          - 96a9b55609f1251d-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3430,14 +3429,14 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:21 GMT
+          - Tue, 05 Aug 2025 22:25:05 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=.1I2A6ZESGbSTwcKtLr1Noe0PRcViHJiRIiLOqFhZQ0-1753726521-1.0.1.1-StmF3KW14wOM5kp5c9pJ7iEf..2RC6sg4urVAG83EMvaLj31ldYJP54PjwLxNVLwen3RskrG.UZ6Gpgpg5wNGTguf6qlnOE1FKCHnTDJ1aI;
-            path=/; expires=Mon, 28-Jul-25 18:45:21 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=FNIynyT7Lu2qhXwRuug94XnE9O1QDLh0.XBKTtfybmc-1754432705-1.0.1.1-zCR_UL4LmLNkbhGXMWEFDvQ5V46ypY_jpi4.ryFmrJaYSYuU4BNsNaJsV_5Wmu06nv3zFMhh1Lz2tvaLvdg.22EFquYIjFFakVk2fxco6K4;
+            path=/; expires=Tue, 05-Aug-25 22:55:05 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=P1b_lJTcCyeN.QnADurbWUDM0r7.s6ZEMt8VVNuHWLA-1753726521863-0.0.1.1-604800000;
+          - _cfuvid=j6dQmcueMTfNkJTnMHueZKCRYW71VAyJ39uz_ebq69Y-1754432705178-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Transfer-Encoding:
           - chunked
@@ -3456,7 +3455,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "81"
+          - "125"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -3464,15 +3463,15 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         via:
-          - envoy-router-f55cd75d4-7zhl8
+          - envoy-router-7c6b8c8c54-q8qjt
         x-envoy-upstream-service-time:
-          - "88"
+          - "129"
         x-ratelimit-limit-requests:
           - "200000"
         x-ratelimit-limit-tokens:
           - "200000000"
         x-ratelimit-remaining-requests:
-          - "199998"
+          - "199999"
         x-ratelimit-remaining-tokens:
           - "199999996"
         x-ratelimit-reset-requests:
@@ -3480,7 +3479,7 @@ interactions:
         x-ratelimit-reset-tokens:
           - 0s
         x-request-id:
-          - req_ff646ed5ce13c08176b0b12798e52128
+          - req_010b82f8cea15a81c62fb1f769b5d2f0
       status:
         code: 200
         message: OK
@@ -3488,387 +3487,14 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 25-28: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\n2021,
-        25, 1315\u20131360.\\n\\n\\n (9) Wellawatte, G. P.; Seshadri, A.; White, A.
-        D. Model agnostic generation of counter-\\n\\n     factual explanations for
-        molecules. Chemical Science 2022, 13, 3697\u20133705.\\n\\n\\n(10) Gandhi, H.
-        A.; White, A. D. Explaining structure-activity relationships using locally\\n\\n
-        \     faithful surrogate models. chemrxiv 2022,\\n\\n\\n(11) Gormley, A. J.;
-        Webb, M. A. Machine learning in combinatorial polymer chemistry.\\n\\n     Nature
-        Reviews Materials 2021,\\n\\n\\n(12) Gomes, C. P.; Fink, D.; Dover, R. B. V.;
-        Gregoire, J. M. Computational sustainability\\n\\n     meets materials science.
-        Nature Reviews Materials 2021,\\n\\n\\n(13) On scientific understanding with
-        artificial intelligence. Nature Reviews Physics 2022\\n\\n     4:12 2022, 4,
-        761\u2013769.\\n\\n\\n(14) Arrieta, A. B.; D\xB4\u0131az-Rodr\xB4\u0131guez,
-        N.; Ser, J. D.; Bennetot, A.; Tabik, S.; Barbado, A.;\\n\\n     Garcia, S.;
-        Gil-Lopez, S.; Molina, D.; Benjamins, R.; Chatila, R.; Herrera, F. Explain-\\n\\n
-        \    able Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities
-        and Chal-\\n\\n     lenges toward Responsible AI. Information Fusion 2019, 58,
-        82\u2013115.\\n\\n\\n(15) Murdoch, W. J.; Singh, C.; Kumbier, K.; Abbasi-Asl,
-        R.; Yu, B. Interpretable machine\\n\\n     learning: definitions, methods, and
-        applications. ArXiv 2019, abs/1901.04592.\\n\\n\\n                                      25(16)
-        Boobier, S.; Osbourn, A.; Mitchell, J. B. Can human experts predict solubility
-        better\\n\\n     than computers? Journal of cheminformatics 2017, 9, 1\u201314.\\n\\n\\n(17)
-        Lee, J. D.; See, K. A. Trust in automation: Designing for appropriate reliance.
-        Human\\n\\n     Factors 2004, 46, 50\u201380.\\n\\n\\n(18) Bolukbasi, T.; Chang,
-        K.-W.; Zou, J. Y.; Saligrama, V.; Kalai, A. T. Man is to com-\\n\\n     puter
-        programmer as woman is to homemaker? debiasing word embeddings. Advances\\n\\n
-        \    in neural information processing systems 2016, 29.\\n\\n\\n(19) Buolamwini,
-        J.; Gebru, T. Gender Shades:  Intersectional Accuracy Disparities in\\n\\n    Commercial
-        Gender Classification. Proceedings of the 1st Conference on Fairness,\\n\\n
-        \    Accountability and Transparency. 2018; pp 77\u201391.\\n\\n\\n(20) Lapuschkin,
-        S.; W\xA8aldchen, S.; Binder, A.; Montavon, G.; Samek, W.; M\xA8uller, K.-R.\\n\\n
-        \   Unmasking Clever Hans predictors and assessing what machines really learn.
-        Nature\\n\\n     communications 2019, 10, 1\u20138.\\n\\n\\n(21) DeGrave, A.
-        J.; Janizek, J. D.; Lee, S.-I. AI for radiographic COVID-19 detection\\n\\n
-        \     selects shortcuts over signal. Nature Machine Intelligence 2021, 3, 610\u2013619.\\n\\n\\n(22)
-        Goodman, B.; Flaxman, S. European Union regulations on algorithmic decision-\\n\\n
-        \   making and a \u201Cright to explanation\u201D. AI Magazine 2017, 38, 50\u201357.\\n\\n\\n(23)
-        ACT, A. I. European Commission. On Artificial Intelligence: A European Approach\\n\\n
-        \    to Excellence and Trust. 2021, COM/2021/206.\\n\\n\\n(24) Blueprint for
-        an AI Bill of Rights, The White House. 2022; https://www.whitehouse.\\n\\n    gov/ostp/ai-bill-of-rights/.\\n\\n\\n(25)
-        Miller, T. Explanation in artificial intelligence: Insights from the social
-        sciences. Ar-\\n\\n       tificial intelligence 2019, 267, 1\u201338.\\n\\n\\n\\n
-        \                                     26(26) Murdoch, W. J.; Singh, C.; Kumbier,
-        K.; Abbasi-Asl, R.; Yu, B. Definitions, meth-\\n\\n     ods, and applications
-        in interpretable machine learning. Proceedings of the National\\n\\n    Academy
-        of Sciences of the United States of America 2019, 116, 22071\u201322080.\\n\\n\\n(27)
-        Gunning, D.; Aha, D. DARPA\u2019s Explainable Artificial Intelligence (XAI)
-        Program.\\n\\n    AI Magazine 2019, 40, 44\u201358.\\n\\n\\n(28) Biran, O.;
-        Cotton, C. Explanation and justification in machine learning: A survey.\\n\\n
-        \    IJCAI-17 workshop on explainable AI (XAI). 2017; pp 8\u201313.\\n\\n\\n(29)
-        Palacio, S.; Lucieri, A.; Munir, M.; Ahmed, S.; Hees, J.; Dengel, A. Xai handbook:\\n\\n
-        \   Towards a unified framework for explainable ai. Proceedings of the IEEE/CVF
-        Inter-\\n\\n     national Conference on Computer Vision. 2021; pp 3766\u20133775.\\n\\n\\n(30)
-        Kuhn, D. R.; Kacker, R. N.; Lei, Y.; Simos, D. E. Combinatorial Methods for
-        Ex-\\n\\n     plainable AI. 2020 IEEE International Conference on Software Testing,
-        Verification\\n\\n    and Validation Workshops (ICSTW) 2020, 167\u2013170.\\n\\n\\n(31)
-        Seshadri, A.; Gandhi, H. A.; Wellawatte, G. P.; White, A. D. Why does that molecule\\n\\n
-        \    smell? ChemRxiv 2022,\\n\\n\\n(32) Das, A.; Rad, P. Opportunities and challenges
-        in explainable artificial intelligence\\n\\n      (xai): A survey. arXiv preprint
-        arXiv:2006.11371 2020,\\n\\n\\n(33) Machlev, R.; Heistrene, L.; Perl, M.; Levy,
-        K. Y.; Belikov, J.; Mannor, S.; Levron, Y.\\n\\n     Explainable Artificial
-        Intelligence (XAI) techniques for energy and power systems:\\n\\n     Review,
-        challenges and opportunities. Energy and AI 2022, 9, 100169.\\n\\n\\n(34) Koh,
-        P. W.; Liang, P. Understanding black-box predictions via influence functions.\\n\\n
-        \    International Conference on Machine Learning. 2017; pp 1885\u20131894.\\n\\n\\n(35)
-        Ribeiro, M. T.; Singh, S.; Guestrin, C. \u201D Why should i trust you?\u201D
-        Explaining the\\n\\n     predictions of any classifier. Proceedings of the 22nd
-        ACM SIGKDD international\\n\\n\\n                                      27     conference
-        on knowledge discovery and data \\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
-      headers:
-        accept:
-          - application/json
-        accept-encoding:
-          - gzip, deflate
-        connection:
-          - keep-alive
-        content-length:
-          - "5833"
-        content-type:
-          - application/json
-        host:
-          - api.openai.com
-        user-agent:
-          - AsyncOpenAI/Python 1.97.1
-        x-stainless-arch:
-          - arm64
-        x-stainless-async:
-          - async:asyncio
-        x-stainless-lang:
-          - python
-        x-stainless-os:
-          - MacOS
-        x-stainless-package-version:
-          - 1.97.1
-        x-stainless-raw-response:
-          - "true"
-        x-stainless-read-timeout:
-          - "60.0"
-        x-stainless-retry-count:
-          - "0"
-        x-stainless-runtime:
-          - CPython
-        x-stainless-runtime-version:
-          - 3.13.5
-      method: POST
-      uri: https://api.openai.com/v1/chat/completions
-    response:
-      body:
-        string: !!binary |
-          H4sIAAAAAAAAAwAAAP//dFRNb+NGDL37VxC6tAVkI3GcGMnN/UARLBboFrtA2nph0CNK4mY0VMmR
-          GyPIfy9mlMTaNnsRhHnDx8c3JB9nAAVXxQ0UrsXout7Pfzy+r7FpPv6h8u7T9cp/6D7cvb/6s/6p
-          /nW5LsoUIfsv5OJL1MJJ13uKLGGEnRJGSqzn68uL9fLqcrnMQCcV+RTW9HG+kvnybLman5/Pl2fP
-          ga2wIytu4K8ZAMBj/iaJoaKH4gbOypeTjsywoeLm9RJAoeLTSYFmbBFDLMoT6CRECln14zYAbAsb
-          ug71uC1uYFv88tB75IB7T7DRyDU7Rg+3IZL33FBwBN/fbW5/AKWa1CAKdBRbqQwwVBDJtYH/HsiA
-          A+CJgacMscUIHd6nP4KKHBtLmHd4z6GBXsWRGRlIDZtbsKNF6gyiYrAelUIsM5/2SjFJLXPuIVSk
-          qeAqy48C7dBhsAXcbW4BuctqsaqUzMC16D2FhgxscC1gSjBYLAGdkyFE3LPneBy5a2QNKYzDVNL+
-          mNQeuEq6KXkXMHWAQS2aimOFXqliN56KAo6/C/jYEtCDI+3jaGayxuCAyjIY/CN6b6NTiViUchlO
-          gqM+WgkRHyRIx2SjROx7z25Mn/xxfsiyOBoMRkl5zeQrA8/3BK6lji3qsQQKpM3xpaiRTakZPEbR
-          I9SKHWU5C3hHxyQhKu+HsaQxEcHPm99/23xnWWSv0ih2JdigBzoaSMjnp+54dvWVOfulZL0E49x8
-          t1DRgbz0HYWY3TICqmvRaNBy03pu2pgbiLteNGLqLKmnnZEfMGf6MlhuxWzP14+42BblOAlKng6J
-          ZmdOlNJEXG/D03R8lOrBME1vGLyfABiCxNH8NLifn5Gn11H10vQqe/tPaFFzYGt3SmgS0lhalL7I
-          6NMM4HNeCcNXU170Kl0fd1HuKac7v7xej4TFaQtN4NXVMxolop8A69VF+QblrqKI7G2yVwqHrqVq
-          mvPi6rUIHCqWE3Y2m9T+f0lv0Y/1c2gmLN+kPwEuDQNVu9OQvXVNKW3qb1179ToLLoz0wI52kUnT
-          e1RU4+DHJVqMPbOrOTSpyThv0vSes6fZvwAAAP//AwCBXCXuRwYAAA==
-      headers:
-        CF-RAY:
-          - 96665c8a1902d03d-SJC
-        Connection:
-          - keep-alive
-        Content-Encoding:
-          - gzip
-        Content-Type:
-          - application/json
-        Date:
-          - Mon, 28 Jul 2025 18:15:24 GMT
-        Server:
-          - cloudflare
-        Set-Cookie:
-          - __cf_bm=VMc2K8osm8kO1ZIe0s1SjsitnUjVPsWldkrtEZxrglQ-1753726524-1.0.1.1-RZ0c4um3RJbMVJR17FWel9eIUEcVVl_DwcVpcMYk2Q4UCrhnh5WLenZ.q2F2FfboI0wxFhnrZMnoxRK3W71Xqm_.Wn9zcZAC6ZQEt9bHtno;
-            path=/; expires=Mon, 28-Jul-25 18:45:24 GMT; domain=.api.openai.com; HttpOnly;
-            Secure; SameSite=None
-          - _cfuvid=2ORelj029GiOqkqqnvcbHbIslaeSNDJn9tHCSxIvYtA-1753726524028-0.0.1.1-604800000;
-            path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
-        Transfer-Encoding:
-          - chunked
-        X-Content-Type-Options:
-          - nosniff
-        access-control-expose-headers:
-          - X-Request-ID
-        alt-svc:
-          - h3=":443"; ma=86400
-        cf-cache-status:
-          - DYNAMIC
-        openai-organization:
-          - future-house-xr4tdh
-        openai-processing-ms:
-          - "2022"
-        openai-project:
-          - proj_RpeV6PrPclPHBb5GlExPXSBj
-        openai-version:
-          - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
-        x-envoy-upstream-service-time:
-          - "2026"
-        x-ratelimit-limit-requests:
-          - "10000"
-        x-ratelimit-limit-tokens:
-          - "30000000"
-        x-ratelimit-remaining-requests:
-          - "9999"
-        x-ratelimit-remaining-tokens:
-          - "29998614"
-        x-ratelimit-reset-requests:
-          - 6ms
-        x-ratelimit-reset-tokens:
-          - 2ms
-        x-request-id:
-          - req_7ba2b5d36f0ed551aeed2f4045837905
-      status:
-        code: 200
-        message: OK
-  - request:
-      body:
-        "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
-        the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 3-5: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\n
-        a passive characteristic of a model, whereas explainability\\n\\nis an active
-        characteristic which is used to clarify the internal decision-making process.\\n\\nNamely,
-        an explanation is extra information that gives the context and a cause for one
-        or\\n\\nmore predictions.29 We adopt the same nomenclature in this perspective.\\n\\n
-        \  Accuracy and interpretability are two attractive characteristics of DL models.
-        However,\\n\\nDL models are often highly accurate and less interpretable.28,30
-        XAI provides a way to avoid\\n\\nthat trade-off in chemical property prediction.
-        XAI can be viewed as a two-step process.\\n\\nFirst, we develop an accurate
-        but uninterpretable DL model. Next, we add explanations to\\n\\npredictions.
-        Ideally, if the DL model has correctly learned the input-output relations, then\\n\\nthe
-        explanations should give insight into the underlying mechanism.\\n\\n   In the
-        remainder of this article, we review recent approaches for XAI of chemical property\\n\\nprediction
-        while drawing specific examples from our recent XAI work.9,10,31 We show how\\n\\nin
-        various systems these methods yield explanations that are consistent with known
-        and\\n\\nmechanisms in structure-property relationships.\\n\\n\\n\\n\\n\\n                                       3Theory\\n\\n\\nIn
-        this work, we aim to assemble a common taxonomy for the landscape of XAI while\\n\\nproviding
-        our perspectives. We utilized the vocabulary proposed by Das and Rad 32 to classify\\n\\nXAI.
-        According to their classification, interpretations can be categorized as global
-        or local\\n\\ninterpretations on the basis of \u201Cwhat is being explained?\u201D.
-        For example, counterfactuals are\\n\\nlocal interpretations, as these can explain
-        only a given instance. The second classification is\\n\\nbased on the relation
-        between the model and the interpretation \u2013 is interpretability post-hoc\\n\\n(extrinsic)
-        or intrinsic to the model?.32,33 An intrinsic XAI method is part of the model\\n\\nand
-        is self-explanatory32 These are also referred to as white-box models to contrast
-        them\\n\\nwith non-interpretable black box models.28 An extrinsic method is
-        one that can be applied\\n\\npost-training to any model.33 Post-hoc methods
-        found in the literature focus on interpreting\\n\\nmodels through 1) training
-        data34 and feature attribution,35 2) surrogate models10 and, 3)\\n\\ncounterfactual9
-        or contrastive explanations.36\\n\\n   Often, what is a \u201Cgood\u201D explanation
-        and what are the required components of an ex-\\n\\nplanation are debated.32,37,38
-        Palacio et al. 29 state that the lack of a standard framework\\n\\nhas caused
-        the inability to evaluate the interpretability of a model.  In physical sciences,\\n\\nwe
-        may instead consider if the explanations somehow reflect and expand our understanding\\n\\nof
-        physical phenomena.  For example, Oviedo et al. 39 propose that a model explanation\\n\\ncan
-        be evaluated by considering its agreement with physical observations, which
-        they term\\n\\n\u201Ccorrectness.\u201D For example, if an explanation suggests
-        that polarity affects solubility of a\\n\\nmolecule, and the experimental evidence
-        strengthen the hypothesis, then the explanation\\n\\nis assumed \u201Ccorrect\u201D.
-        In instances where such mechanistic knowledge is sparse, expert bi-\\n\\nases
-        and subjectivity can be used to measure the correctness.40 Other similar metrics
-        of\\n\\ncorrectness such as \u201Cexplanation satisfaction scale\u201D can be
-        found in the literature.41,42 In a\\n\\nrecent study, Humer et al. 43 introduced
-        CIME an interactive web-based tool that allows the\\n\\nusers to inspect model
-        explanations. The aim of this study is to bridge the gap between\\n\\nanalysis
-        of XAI methods. Based on the above discussion, we identify that an agreed upon\\n\\n\\n
-        \                                      4evaluation metric is necessary in XAI.
-        We suggest the following attributes can be used to\\n\\nevaluate explanations.
-        However, the relative importance of each attribute may depend on\\n\\nthe application
-        - actionability may not be as important as faithfulness when evaluating the\\n\\ninterpretability
-        of a static physics based model. Therefore, one can select relative importance\\n\\nof
-        each attribute based on the application.\\n\\n\\n   \u2022 Actionable. Is it
-        clear how we could change the input features to modify the output?\\n\\n\\n
-        \  \u2022 Complete. Does the explanation completely account for the prediction?
-        Did features\\n\\n     not included in the explanation really contribute zero
-        effect to the prediction?44\\n\\n\\n   \u2022 Correct. Does the explanation
-        agree with hypothesized or known underlying physical\\n\\n     mechanism?39\\n\\n\\n
-        \  \u2022 Domain Applicable. Does the explanation use language and concepts
-        of domain ex-\\n\\n      perts?\\n\\n\\n   \u2022 Fidelity/Faithful. Does the
-        explanation agree with the black box model?\\n\\n\\n   \u2022 Robust. Does the
-        explanation change significantly with small changes to the model or\\n\\n      instance
-        being explained?\\n\\n\\n   \u2022 Sparse/Succinct. Is the explanation succinct?\\n\\n\\n
-        \ We present an example evaluation of the SHAP explanation method based on the
-        above\\n\\nattributes.44 Shapley values were proposed as a local explanation
-        method based on feature\\n\\nattribution, as they offer a complete explanation
-        - each feature i\\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
-      headers:
-        accept:
-          - application/json
-        accept-encoding:
-          - gzip, deflate
-        connection:
-          - keep-alive
-        content-length:
-          - "5792"
-        content-type:
-          - application/json
-        host:
-          - api.openai.com
-        user-agent:
-          - AsyncOpenAI/Python 1.97.1
-        x-stainless-arch:
-          - arm64
-        x-stainless-async:
-          - async:asyncio
-        x-stainless-lang:
-          - python
-        x-stainless-os:
-          - MacOS
-        x-stainless-package-version:
-          - 1.97.1
-        x-stainless-raw-response:
-          - "true"
-        x-stainless-read-timeout:
-          - "60.0"
-        x-stainless-retry-count:
-          - "0"
-        x-stainless-runtime:
-          - CPython
-        x-stainless-runtime-version:
-          - 3.13.5
-      method: POST
-      uri: https://api.openai.com/v1/chat/completions
-    response:
-      body:
-        string: !!binary |
-          H4sIAAAAAAAAA3RUTW/bOBC9+1cMeGoA2bDVJO76lqB78CIL7KHABlgXxpgcSdNQJEuOEqtB/ntB
-          ybGdbnoRRL55w3nz9TwBUGzUCpRuUHQb7PS2/7vCf77fVt0XbG/b6tvd/P7j3tfr/Y8//lJFZvjd
-          N9Lyyppp3wZLwt6NsI6EQtnrYnn1cVleX5XlALTekM20Osj00k/LeXk5XSym5fxAbDxrSmoF/00A
-          AJ6Hbw7RGdqrFcyL15uWUsKa1OpoBKCit/lGYUqcBJ2o4gRq74TcEPXzxgFsVOraFmO/USvYqPub
-          dQE+wp/7YJEd7izBTRSuWDNaWDsha7kmp6mASBXFBOKhJWm8SYDOgJBuHH/vKIE0KBCif2RDQNml
-          w5ygBJWPIA2BIc2JvZu2+MCuzsaaUqIEvoKbNQy5SgUEjMK6sxhtD4YogCWMLlM+fL67ONjN4N+G
-          LcHnu8MFYCTwlZCDhuvG9oBadxGFivx8P+CWUgJ2QjFEkix5Bvc3a0BuB3G7yKYmkIYT1Bhg14Ov
-          Kor5cXaJ60YGvh8Udc5QtH0GW9INOk7toCZEMqwH+TNYC2h0sCPQKFT7yD/IjE7YScxuNXxIZKvp
-          a9587C+GBNP+aBB8kmnj9cVrBcbI36T68A49ou1yQ8IOExnwDlAk8q4TSmD5gQCH6HDHlqUv4NDQ
-          5CilfIqRtIwH41tkBxiCZX0kVGxo/It+16WDbQ45dVqzG9kz+NJQ+qUfch3e1HiINncfO6iYrDnE
-          qBtqWaPNnRIoSn+W1wKeGoo0VjbLRsu1gyeWBh6cf3IQmj4N7FNpZhtVjIMQydIjOk3bpH2kcSAW
-          843auJfzEYpUdQnzBLvO2jMAnfMyCsrD+/WAvBzH1fo6RL9Lv1BVxY5Ts42Eybs8mkl8UAP6MgH4
-          OqyF7s2kqxB9G2Qr/oGG5xaL8mp0qE6b6Ay+Kg+oeEF7BpTLZfGOy60hQbbpbLcojbohc8adl5dH
-          EdgZ9idsPjnT/v+Q3nM/6mdXn3n5rfsToDUFIbM9NcJ7ZpHytv6d2THXQ8AqUXxkTVthirkehirs
-          7LhIVeqTULut2NV5ZfC4Tauw1dfL5fWn6vrTXE1eJj8BAAD//wMA8Ae31VYGAAA=
-      headers:
-        CF-RAY:
-          - 96665c8a0cf7cf26-SJC
-        Connection:
-          - keep-alive
-        Content-Encoding:
-          - gzip
-        Content-Type:
-          - application/json
-        Date:
-          - Mon, 28 Jul 2025 18:15:25 GMT
-        Server:
-          - cloudflare
-        Set-Cookie:
-          - __cf_bm=VJDkENpRsapGMYClJsuPE5e3jQN7J_EYU4BnZiN7RzE-1753726525-1.0.1.1-_L0DFcMIk.ktRKII82zNvKFz1Z3NfIRSzrqwzLzFe7B9ds.5jOf1SZARYbOHjsICq25iSh0MAo3vzgSa9_Azds4VCdcCEElPXjULgOAyojc;
-            path=/; expires=Mon, 28-Jul-25 18:45:25 GMT; domain=.api.openai.com; HttpOnly;
-            Secure; SameSite=None
-          - _cfuvid=CAWISLg9jFPC0E1CcuMRur81pIZxXa2MqXKDCYNJcdU-1753726525235-0.0.1.1-604800000;
-            path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
-        Strict-Transport-Security:
-          - max-age=31536000; includeSubDomains; preload
-        Transfer-Encoding:
-          - chunked
-        X-Content-Type-Options:
-          - nosniff
-        access-control-expose-headers:
-          - X-Request-ID
-        alt-svc:
-          - h3=":443"; ma=86400
-        cf-cache-status:
-          - DYNAMIC
-        openai-organization:
-          - future-house-xr4tdh
-        openai-processing-ms:
-          - "2745"
-        openai-project:
-          - proj_RpeV6PrPclPHBb5GlExPXSBj
-        openai-version:
-          - "2020-10-01"
-        x-envoy-upstream-service-time:
-          - "3227"
-        x-ratelimit-limit-requests:
-          - "10000"
-        x-ratelimit-limit-tokens:
-          - "30000000"
-        x-ratelimit-remaining-requests:
-          - "9999"
-        x-ratelimit-remaining-tokens:
-          - "29998619"
-        x-ratelimit-reset-requests:
-          - 6ms
-        x-ratelimit-reset-tokens:
-          - 2ms
-        x-request-id:
-          - a3932dc2-0608-48ec-91a1-09deb5569ac7
-      status:
-        code: 200
-        message: OK
-  - request:
-      body:
-        "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
-        the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 1-3: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\n
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 1-3: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\n
         A Perspective on Explanations of Molecular\\n\\n              Prediction Models\\n\\n\\nGeemi
         P. Wellawatte,\u2020   Heta A. Gandhi,\u2021   Aditi Seshadri,\u2021 and  Andrew\\n\\n
         \                          D. White\u2217,\u2021\\n\\n\\n     \u2020Department
@@ -3935,7 +3561,7 @@ interactions:
         a passive characteristic of a model, whereas explainability\\n\\nis an active
         characteristic which is used to clarify the internal decision-making process.\\n\\nNamely,
         an explanation is extra information that gives the context and a cause for one
-        or\\n\\nmore \\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        or\\n\\nmore \\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -3944,13 +3570,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5817"
+          - "6093"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3960,7 +3586,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3976,25 +3602,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//dJVNbxs5DIbv/hWELm0BO7DdxAlyy24LNGgL7O5lA6wLm5Y4HiYaaZbU
-          JBkE+e8LaZxk0k0vPsxLUuTDDz9MAAw7cw7G1phs0/rZb/33Ck9iQ3/8fnbtrnZfVl/7xd9fLv/C
-          P7/fmWn2iLtrsunJ68jGpvWUOIZBtkKYKEddnJ58PF2uTpbLIjTRkc9u+zbNjuNsOV8ezxaL2XJ+
-          cKwjW1JzDv9MAAAeym9OMTi6N+cwnz59aUgV92TOn40AjESfvxhUZU0Ykpm+iDaGRKFkvd1urzWG
-          dXhYB4C10a5pUPq1OYe1+XzfeuSAO09wIYkrtoweLkMi73lPwRK8v7q4/ACsgFAxeQdVtJ2Sgxig
-          lXjLjsMeOCSSVihhZqMQK3BELXhCCdng/advH6BAgVbIsS12U0DnhFSzSaoJ3u082pvZLt6/g4Cp
-          E8qhUk1Kg7cewdXFJSA3CilCgzcEn74dNGiiEHTBkWQmrhSGwUGSTtNdlFT3sOshVhVJfpIygHDI
-          uYqSX2IZZ3gEX6mHRNIooGq0nPsNd5zqkggH6ztHIwA79pz6KVx3WojiU6XBDe9l4MXmqJAee4FQ
-          RVIqyzQc7YUKgbprMLyqbLDnkISDss0uOGCYwl3Nnn5KAFAI/u0wJM5NuiVoKAlbhVRjAkcVZVA1
-          jWBxIC29fAXk8xga2hzL92A9Cld9iVBgBPTgyLJyDLMGbzLvVqIl1elocsqs3qeCx2KnNDRCW7I5
-          +ddPF+IKLUpi23kU3wM3bZS8AcBhGFEFzzcEtqaGNUmfgdCrwRhyeQ4MFgPsO3YEdd/GMm46NCxo
-          nsHDeOFtZAfadsKxU7BRhPwA4mhtpsOGCXm6xWBpozYK5U1bzNfhcR222+14S4WqTjEfidB5PxIw
-          hHhYpHwffhyUx+eL4OO+lbjTn1xNxYG13gihxpC3X1NsTVEfJwA/yuXpXh0T00ps2rRJ8YbKc4vF
-          8XIIaF6O3UhePakpJvQj4eP8ePpGyI2jhOx1dL6MRVuTG/nOl8fPRWDnOL5o88mo9v+n9Fb4oX4O
-          +1GUX4Z/EaylNpHbvAzGW2ZC+Q/hV2bPrEvCRklu2dImMUnuh6MKOz/caqO9Jmo2FYd9PgI8HOyq
-          3djV6enqrFqdzc3kcfIfAAAA//8DAGWWD4W5BgAA
+          H4sIAAAAAAAAA3RUwW7bOBC9+ysGvNQG7MB2nQTwLdgsiiDZPexhN9h1YYzJkTQNRaoc0okQ5N8L
+          SoqtbtuLAPHNDN+8mcfXCYBio7agdIVR141d/La6/9O0V4frr/dXn4o/Hut//nr59/7m9qG4+/uT
+          mucMf/hCOr5nXWhfN5Yie9fDOhBGylVX15ebzcf19fKyA2pvyOa0somLjV+sl+vNYrVarJdDYuVZ
+          k6gt/DcBAHjtvpmiM/SitrCcv5/UJIIlqe0pCEAFb/OJQhGWiC6q+RnU3kVyHevXnQPYKUl1jaHd
+          qS3s1O8vjUV2eLAENyFywZrRwp2LZC2X5DTB9PHmbgYsgFAwWQOF10nIgHfQBH9kw64EdpFCEyhi
+          lkQAnQHK1d1wUPgAhqgBSxhcTpnePsygUweaQIZ1FzgHNCaQSA6JFcGHg0X9tDj4lw/gMKZA4IuM
+          CPXZcgGPN3eAXAtED+QqzLxjSBI7HknwwJZjC4cWfFFQ6BkLl1WUTN3Dc9UCDmxqfCIBaUhnQcbk
+          LuCeWtDeaWq6zO5mdtomQyMNhuummb+hMlDHuUo1OkjOUMiDMu9hvni/ejaHL0m6OQyyTb8mdJGz
+          rEeCmmJgLRArjHBEywbjoELf77MPsWJHIrP5jzOYGhIduOn/fDF0fW4QnlGgRkOzXlMWaDBE1sli
+          sC0EsnREF3PnuqKaJYZ2Ds8VBRo1luW9fRgLBxodlIkNQdU2vpvesCRO8kj7SQIGAufjeUmkSYF9
+          EtA+BLJ9Ixc7Ne+3eSCkaS/aB8pbvVoOWF7SPddYkuTzAq3Qzr2N7RGoSILZnS5ZOwLQOT+scjbm
+          5wF5O1nR+rIJ/iD/S1UFO5ZqHwjFu2w7ib5RHfo2AfjcWT5952LVBF83cR/9E3XXrdbLwfPq/MqM
+          4M31gEYf0Y6Ajyfku5J7QxHZyujdUBp1RWaUu7pcn5rAZNifseVk1PuPlH5Wvu+fXTmq8svyZ0Bn
+          X5HZnzfnZ2GB8kv8q7CT1h1hJRSOrGkfmUKeh6ECk+0fSSWtRKr3BbsyO5f7l7Jo9uvVFV1dbvTG
+          qMnb5BsAAAD//wMAQQBhgDIGAAA=
       headers:
         CF-RAY:
-          - 96665c8a1e611694-SJC
+          - 96a9b558fc2ccfd5-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4002,15 +3627,17 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:26 GMT
+          - Tue, 05 Aug 2025 22:25:06 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=uaI1haIReSoI5NAQdH7qPi_Wqbvr9X5QMTe.4wL6JFY-1753726526-1.0.1.1-p8XpvwTTGynN1bQ0p0q2qHhiN2yXD1GnWq5FO6jZfb6XaPmfyVijaWCSkePa62wcoiZ8zDJJUTGmqP1whb4_e9BfEpoGtH3i1wJ3dlJ7KpQ;
-            path=/; expires=Mon, 28-Jul-25 18:45:26 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=l02P9XPxObedtKE6petfM3etonagiVazC2KDO8A4SpE-1754432706-1.0.1.1-lIs0xM24m5tCX9r7NXklzcS0K1VJemBqn22K5EJ1ckEC.ZHOavJnefKxjd..HLMuSOJz1j0GhiHQaBwV82DznCguy6jQhmoZDp74UHH1.Ec;
+            path=/; expires=Tue, 05-Aug-25 22:55:06 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=XQ4x91DIH4.3Dwlui7Sh9loWCKIJkq6Dp2hhgIpqT7o-1753726526694-0.0.1.1-604800000;
+          - _cfuvid=Wq4bgGSK9gp6_AmZv9wFZgy2uPZSLH6fbF32sj3ATjc-1754432706983-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
+        Strict-Transport-Security:
+          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -4024,15 +3651,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "4700"
+          - "1469"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "4703"
+          - "1474"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4040,13 +3665,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998613"
+          - "29998544"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - req_6770feb1f047612edbc33b560ee58f93
+          - req_060908554aab4f9faef9874be645429b
       status:
         code: 200
         message: OK
@@ -4054,11 +3679,14 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 20-22: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nnal
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 20-22: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nnal
         molecule.  The counterfactual indicates\\nstructural changes to ethyl benzoate
         that would result in the model predicting the molecule\\nto not contain the
         \u2018fruity\u2019 scent. The Tanimoto96 similarity between the counterfactual
@@ -4123,7 +3751,7 @@ interactions:
         the input to models informs the XAI method.\\n\\n\\nConclusion and outlook\\n\\n\\nWe
         should seek to explain molecular property prediction models because users are
         more\\n\\nlikely to trust explained predictions, and explanations can help assess
-        if the model is learning\\n\\nt\\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        if the model is learning\\n\\nt\\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4132,13 +3760,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5811"
+          - "6087"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4148,7 +3776,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4164,24 +3792,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//fFTBbttGEL3rKwZ7pgRJlutUNydoAadNL+khaBUIw+WQnHq5y+zMqlYN
-          /3uwpCyxiZ0LgeXbefPezM48zgAMV2YLxraotuvd/O3xQ43vV/TX+98+lh/fvP31vz8+rN80v8sX
-          bt+ZIkeE8h+y+hy1sKHrHSkHP8I2Eipl1tXN9dXN+qfr9XoAulCRy2FNr/NNmK+X6818tZqvl6fA
-          NrAlMVv4ewYA8Dh8s0Rf0YPZwrJ4/tORCDZktudLACYGl/8YFGFR9GqKC2iDV/KD6sedB9gZSV2H
-          8bgzW9iZT7d3BYQIvzz0Dtlj6Qhuo3LNltHBnVdyjhvylgqIVFMU0AAdaRsqAfQVKNnW85dEAkmo
-          GmC8J9CWoI9Usc01Egg13N7BUAwB9kqxj6RDxkyTfEUxy6+GXxqgTR16WcCdH7gGJw+aebrgyCaH
-          EfoYeop6nGQq4FPOc1Lo+J4m921IOXONVhM6oGzb4ygwq6hIbOReQ/wGi3R2R2OtoHRo7+dleDiZ
-          WsC7H7CzPwR3IOjYc4cORGOymiI6sC36hsbCjkrHA9qW6TA6jyjKvnk2zCQF/Nuyo1clJyGQFGNo
-          UOm57plVNXKZlMC21LFFNyGdaMAINaGmSLKAP1sSOheVfIveEmhMogWgtSTCJTvWYzF2U4fDqens
-          oQodsj/1Y0gsGo9Q5s6FA1fZGw79O7+IMWQeydEBvQJ74abVUQ2MU5MTXNoNFfXkK4EwPhlMFeen
-          O77T/B5T7IMMUfk4qddiZ4pxPk75LO3Fhkh5Tn7e+afpUEWqk2CeaZ+cmwDofdCx/nmcP5+Qp/MA
-          u9D0MZTyTaip2bO0+0gowedhFQ29GdCnGcDnYVGk/82+6WPoet1ruKch3Wq93IyE5rKbJvDm6oRq
-          UHQT4GpzU7xAua9IkZ1Mto2xaFuqJrGr6/XZRC53uGDL2cT795Jeoh/9s28mLK/SXwBrqVeq9pcl
-          8NK1SHl/v3btXOtBsBGKB7a0V6aY+1FRjcmNq9XIUZS6fc2+yVuMh/2a+zl7mn0FAAD//wMA1tlO
-          6V0GAAA=
+          H4sIAAAAAAAAA3RUTW/jNhC9+1cMeJYN2+skRW7ZoiiMbYseCjRAvbDH5EhiQg1VznDXQZD/XlBy
+          HO02e9GBbz7ee6OZ5xmA8c7cgrEtqu36MP959ekP1938qdv09e/ffv/VPRzvP/10+nii5uMHU5WM
+          eHwgq69ZCxu7PpD6yCNsE6FSqbq6udpsPqxvllcD0EVHoaQ1vc43cb5erjfz1Wq+Xp4T2+gtibmF
+          f2YAAM/Dt1BkRydzC8vq9aUjEWzI3F6CAEyKobwYFPGiyGqqN9BGVuKB9eFweJDIO37eMcDOSO46
+          TE87cws7c3+3rSAm+OXUB/SMx0Bwl9TX3noMsGWlEHxDbKmCRDUlAY3QkbbRCSA7ULIt+38zCWQh
+          N8D4SKAtQZ/IeVvMEog1HAPax/kxnmAwR8CzUuoT6dC4VMvsKBU5bnjSCG3ukGUBWx5KDspOWsp1
+          MZDNARP0KfaU9GnSsIL7u+2FaPCPNIm3MZfONVrNGICKesaRZ2HhSGzyvcb0HZboIpJGy6CNX0E0
+          Zas5YYCaUHMimRCkIrQOubhYNPgLYU9SgWTbAgqIJdYF/NWS0LdtLTIcCRL1iYRYyZV421LnLYZL
+          99I0AeNIJCA3GRuqyjg8N6VzB2gtifizty52RQKdChlZDJYRt8iWBDRlUfB8HlYF6F1RAihSanAz
+          IhAIE3tuqsG7QtZzHVMHDhVBKNB5ImP4MfvgLuFnvwpW/rpBcaGGdU1WIWa1sSNZ7Ew1/r+JAn0p
+          DPdiY6LxP14tL3iZz9532JAUrMYgtOOXHR8Oh+mKJKqzYNlQziFMAGSOOjpflvPzGXm5rGOITZ/i
+          Ub5LNbVnL+0+EUrksnqisTcD+jID+Dysff5mk02fYtfrXuMjDe1W6+v1WNC8XZoJvHlFNSqGKbDc
+          VO+U3DtS9EEmt8NYtC25Se7qan0Rgdn5+IYtZxPt/6f0XvlRv+dmUuWH5d8Aa6lXcvu3FX4vLFG5
+          xj8Ku3g9EDZC6Yu3tFdPqczDUY05jIfSyJModfvac1NukB+vZd3v16trur7a2I0zs5fZfwAAAP//
+          AwBllygLNgYAAA==
       headers:
         CF-RAY:
-          - 96665c8a1e12232c-SJC
+          - 96a9b558fa9e17dc-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4189,14 +3817,14 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:26 GMT
+          - Tue, 05 Aug 2025 22:25:07 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=b8qPaCTsVt55Gr4Uc4YiI4rsq9f3vo5dJdtbkd14VYI-1753726526-1.0.1.1-veWOv3YwWRsr8.xv0OcLEQV58ZLFG6KEyafaNVw0ztAPr_RkWeJjXecKEbWVFN7fXm2gj7lg41Fb4Fd5jvaGxj76skJo0igZ69b5JrsyMTA;
-            path=/; expires=Mon, 28-Jul-25 18:45:26 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=kuZeJUMxUHXVXKwvp5YmANBAVDGTZgNYcmABmPG4h_g-1754432707-1.0.1.1-zxfN7ez4Ic0VdoG0Vwk8Rxq5uYaZEyd5vR00VudDNHF7SPArWZxd7672uZGDksU0_iMTgWAiVOHKeYqGzqzUtglL0D3ACicUu4oKic06cQk;
+            path=/; expires=Tue, 05-Aug-25 22:55:07 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=igfPQsNHuIPnKm452NzcJEaUZpdg.ze81.NnVUx4KPw-1753726526949-0.0.1.1-604800000;
+          - _cfuvid=n6rJGumqcZVrdTjJxWDsqjNW3LLPIU6b9WkLkYM_gwI-1754432707131-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Transfer-Encoding:
           - chunked
@@ -4211,7 +3839,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "4907"
+          - "1540"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -4219,7 +3847,7 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "4918"
+          - "1549"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4227,13 +3855,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998612"
+          - "29998544"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - req_50ce1e008403a3aca09f5d2a6f2f0425
+          - req_a8bd6b680f0285e4f5e0b95558a242f7
       status:
         code: 200
         message: OK
@@ -4241,78 +3869,79 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 22-25: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nut
-        to models informs the XAI method.\\n\\n\\nConclusion and outlook\\n\\n\\nWe
-        should seek to explain molecular property prediction models because users are
-        more\\n\\nlikely to trust explained predictions, and explanations can help assess
-        if the model is learning\\n\\nthe correct underlying chemical principles. We
-        also showed that black-box modeling first,\\n\\nfollowed by XAI, is a path to
-        structure-property relationships without needing to trade\\n\\nbetween accuracy
-        and interpretability. However, XAI in chemistry has some major open\\n\\nquestions,
-        that are also related to the black-box nature of the deep learning. Some are\\n\\n\\n\\n
-        \                                      22highlighted below:\\n\\n\\n   \u2022
-        Explanation representation: How is an explanation presented \u2013 text, a molecule,
-        attri-\\n\\n      butions, a concept, etc?\\n\\n\\n   \u2022 Molecular distance:
-        \ in XAI approaches such as counterfactual generation, the \u201Cdis-\\n\\n
-        \    tance\u201D between two molecules is minimized. Molecular distance is subjective.
-        Possibil-\\n\\n       ities are distance based on molecular properties, synthesis
-        routes, and direct structure\\n\\n     comparisons.\\n\\n\\n   \u2022 Regulations:
-        As black-box models move from research to industry, healthcare, and\\n\\n     environmental
-        settings, we expect XAI to become more important to explain decisions\\n\\n
-        \     to chemists or non-experts and possibly be legally required. Explanations
-        may need\\n\\n      to be tuned for be for doctors instead of chemists or to
-        satisfy a legal requirement.\\n\\n\\n   \u2022 Chemical space: Chemical space
-        is the set of molecules that are realizable; \u201Crealiz-\\n\\n     able\u201D
-        can be defined from purchasable to synthesizable to satisfied valences. What
-        is\\n\\n     most useful? Can an explanation consider nearby impossible molecules?
-        How can we\\n\\n     generate local chemical spaces centered around a specific
-        molecule for finding counter-\\n\\n      factuals or other instance explanations?
-        \ Similarly, can \u201Cactivity cliffs\u201D be connected\\n\\n      to explanations
-        and the local chemical space.149\\n\\n\\n   \u2022 Evaluating XAI : there is
-        a lack of a systematic framework (quantitative or qualitative)\\n\\n      to
-        evaluate correctness and applicability of an explanation. Can there be a universal\\n\\n
-        \    framework, or should explanations be chosen and evaluated based on the
-        audience and\\n\\n     domain? For example, work by Rasmussen et al. 58 attempts
-        to focus on comparing\\n\\n      feature attribution XAI methods via Crippen\u2019s
-        logP scores.\\n\\n\\n\\n\\n\\n                                       23Acknowledgements\\n\\n\\nResearch
-        reported in this work was supported by the National Institute of General Medical\\n\\nSciences
-        of the National Institutes of Health under award number R35GM137966. This work\\n\\nwas
-        supported by the NSF under awards 1751471 and 1764415. We thank the Center for\\n\\nIntegrated
-        Research Computing at the University of Rochester for providing computational\\n\\nresources.\\n\\n\\nReferences\\n\\n\\n
-        \ (1) Choudhary, K.; DeCost, B.; Chen, C.; Jain, A.; Tavazza, F.; Cohn, R.;
-        Park, C. W.;\\n\\n     Choudhary, A.; Agrawal, A.; Billinge, S. J.; Holm, E.;
-        Ong, S. P.; Wolverton, C.\\n\\n      Recent advances and applications of deep
-        learning methods in materials science. npj\\n\\n      Computational Materials
-        2022, 8.\\n\\n\\n  (2) Keith, J. A.; Vassilev-Galindo, V.; Cheng, B.; Chmiela,
-        S.; Gastegger, M.; M\xA8uller, K.-\\n\\n      R.; Tkatchenko, A. Combining Machine
-        Learning and Computational Chemistry for\\n\\n      Predictive Insights Into
-        Chemical Systems. Chemical Reviews 2021, 121, 9816\u20139872,\\n\\n     PMID:
-        34232033.\\n\\n\\n  (3) Goh, G. B.; Hodas, N. O.; Vishnu, A. Deep learning for
-        computational chemistry.\\n\\n      Journal of Computational Chemistry 2017,
-        38, 1291\u20131307.\\n\\n\\n  (4) Deringer, V. L.; Caro, M. A.; Cs\xB4anyi,
-        G. Machine Learning Interatomic Potentials as\\n\\n     Emerging Tools for Materials
-        Science. Advanced Materials 2019, 31, 1902765.\\n\\n\\n  (5) Faber, F. A.; Hutchison,
-        L.; Huang, B.; Gilmer, J.; Schoenholz, S. S.; Dahl, G. E.;\\n\\n      Vinyals,
-        O.; Kearnes, S.; Riley, P. F.; von Lilienfeld, O. A. Prediction Errors of Molec-\\n\\n
-        \     ular Machine Learning Models Lower than Hybrid DFT Error. Journal of Chemical\\n\\n
-        \    Theory and Computation 2017, 13, 5255\u20135264, PMID: 28926232.\\n\\n\\n\\n
-        \                                      24 (6) Duch, W.; Swaminathan, K.; Meller,
-        J. Artificial Intelligence Approaches for Rational\\n\\n    Drug Design and
-        Discovery. Current Pharmaceutical Design 2007, 13, 1497\u20131508.\\n\\n\\n
-        (7) Dara, S.; Dhamercherla, S.; Jadav, S. S.; Babu, C. M.; Ahsan, M. J.; darasuresh,
-        S. D.;\\n\\n     Dara, S. Machine Learning in Drug Discovery: A Review. Artificial
-        Intelligence Review\\n\\n     123, 55, 1947\u20131999.\\n\\n\\n (8) Gupta, R.;
-        Srivastava, D.; Sahu, M.; Tiwari, S.; Ambasta, R. K.; Kumar, P. Artifi-\\n\\n
-        \     cial intelligence to deep learning: machine intelligence approach for
-        drug discovery.\\n\\n     Molecular diversity 2021, 25, 1315\u20131360.\\n\\n\\n
-        (9) Wellawatte, G. P.; Seshadri, A.; White, A. D. Model agnostic generation
-        of counter-\\n\\n     factual explanations for molecules. Chemical Science 2022,
-        13, 3697\u20133705.\\n\\n\\n(10) Gandhi, H. A.; White, A. D. Explaining structure-ac\\n\\n----\\n\\nQuestion:
-        What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 3-5: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\n
+        a passive characteristic of a model, whereas explainability\\n\\nis an active
+        characteristic which is used to clarify the internal decision-making process.\\n\\nNamely,
+        an explanation is extra information that gives the context and a cause for one
+        or\\n\\nmore predictions.29 We adopt the same nomenclature in this perspective.\\n\\n
+        \  Accuracy and interpretability are two attractive characteristics of DL models.
+        However,\\n\\nDL models are often highly accurate and less interpretable.28,30
+        XAI provides a way to avoid\\n\\nthat trade-off in chemical property prediction.
+        XAI can be viewed as a two-step process.\\n\\nFirst, we develop an accurate
+        but uninterpretable DL model. Next, we add explanations to\\n\\npredictions.
+        Ideally, if the DL model has correctly learned the input-output relations, then\\n\\nthe
+        explanations should give insight into the underlying mechanism.\\n\\n   In the
+        remainder of this article, we review recent approaches for XAI of chemical property\\n\\nprediction
+        while drawing specific examples from our recent XAI work.9,10,31 We show how\\n\\nin
+        various systems these methods yield explanations that are consistent with known
+        and\\n\\nmechanisms in structure-property relationships.\\n\\n\\n\\n\\n\\n                                       3Theory\\n\\n\\nIn
+        this work, we aim to assemble a common taxonomy for the landscape of XAI while\\n\\nproviding
+        our perspectives. We utilized the vocabulary proposed by Das and Rad 32 to classify\\n\\nXAI.
+        According to their classification, interpretations can be categorized as global
+        or local\\n\\ninterpretations on the basis of \u201Cwhat is being explained?\u201D.
+        For example, counterfactuals are\\n\\nlocal interpretations, as these can explain
+        only a given instance. The second classification is\\n\\nbased on the relation
+        between the model and the interpretation \u2013 is interpretability post-hoc\\n\\n(extrinsic)
+        or intrinsic to the model?.32,33 An intrinsic XAI method is part of the model\\n\\nand
+        is self-explanatory32 These are also referred to as white-box models to contrast
+        them\\n\\nwith non-interpretable black box models.28 An extrinsic method is
+        one that can be applied\\n\\npost-training to any model.33 Post-hoc methods
+        found in the literature focus on interpreting\\n\\nmodels through 1) training
+        data34 and feature attribution,35 2) surrogate models10 and, 3)\\n\\ncounterfactual9
+        or contrastive explanations.36\\n\\n   Often, what is a \u201Cgood\u201D explanation
+        and what are the required components of an ex-\\n\\nplanation are debated.32,37,38
+        Palacio et al. 29 state that the lack of a standard framework\\n\\nhas caused
+        the inability to evaluate the interpretability of a model.  In physical sciences,\\n\\nwe
+        may instead consider if the explanations somehow reflect and expand our understanding\\n\\nof
+        physical phenomena.  For example, Oviedo et al. 39 propose that a model explanation\\n\\ncan
+        be evaluated by considering its agreement with physical observations, which
+        they term\\n\\n\u201Ccorrectness.\u201D For example, if an explanation suggests
+        that polarity affects solubility of a\\n\\nmolecule, and the experimental evidence
+        strengthen the hypothesis, then the explanation\\n\\nis assumed \u201Ccorrect\u201D.
+        In instances where such mechanistic knowledge is sparse, expert bi-\\n\\nases
+        and subjectivity can be used to measure the correctness.40 Other similar metrics
+        of\\n\\ncorrectness such as \u201Cexplanation satisfaction scale\u201D can be
+        found in the literature.41,42 In a\\n\\nrecent study, Humer et al. 43 introduced
+        CIME an interactive web-based tool that allows the\\n\\nusers to inspect model
+        explanations. The aim of this study is to bridge the gap between\\n\\nanalysis
+        of XAI methods. Based on the above discussion, we identify that an agreed upon\\n\\n\\n
+        \                                      4evaluation metric is necessary in XAI.
+        We suggest the following attributes can be used to\\n\\nevaluate explanations.
+        However, the relative importance of each attribute may depend on\\n\\nthe application
+        - actionability may not be as important as faithfulness when evaluating the\\n\\ninterpretability
+        of a static physics based model. Therefore, one can select relative importance\\n\\nof
+        each attribute based on the application.\\n\\n\\n   \u2022 Actionable. Is it
+        clear how we could change the input features to modify the output?\\n\\n\\n
+        \  \u2022 Complete. Does the explanation completely account for the prediction?
+        Did features\\n\\n     not included in the explanation really contribute zero
+        effect to the prediction?44\\n\\n\\n   \u2022 Correct. Does the explanation
+        agree with hypothesized or known underlying physical\\n\\n     mechanism?39\\n\\n\\n
+        \  \u2022 Domain Applicable. Does the explanation use language and concepts
+        of domain ex-\\n\\n      perts?\\n\\n\\n   \u2022 Fidelity/Faithful. Does the
+        explanation agree with the black box model?\\n\\n\\n   \u2022 Robust. Does the
+        explanation change significantly with small changes to the model or\\n\\n      instance
+        being explained?\\n\\n\\n   \u2022 Sparse/Succinct. Is the explanation succinct?\\n\\n\\n
+        \ We present an example evaluation of the SHAP explanation method based on the
+        above\\n\\nattributes.44 Shapley values were proposed as a local explanation
+        method based on feature\\n\\nattribution, as they offer a complete explanation
+        - each feature i\\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4321,13 +3950,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5820"
+          - "6068"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4337,7 +3966,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4353,24 +3982,25 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA3RUTW/jNhC9+1cMeGkLyIbtOEnrW4oWhVFsD+0laL0wGHIksaGG6swoiTfIf1+Q
-          cmLtNnsRIL75eO+RM88zABO82YJxrVXX9XH+8/FD7a4vhw+bXx7/uvoj3v/5CX/7tPu7ueDblaly
-          Rrr7F52+Zi1c6vqIGhKNsGO0irnq6vry4np9dbneFKBLHmNOa3qdb9J8vVxv5qvVfL08JbYpOBSz
-          hX9mAADP5Zspkscns4Vl9XrSoYht0GzfggAMp5hPjBUJopbUVGfQJVKkwvp5TwB7I0PXWT7uzRb2
-          5vZmV0Fi+PWpjzaQvYsIN6yhDi7YCDtSjDE0SA4rYKyRBTRBh9omL2DJg6JrKfw3oMAg6Ats7xG0
-          RegZfXDZozHWowtS/lINNzso1ggM5JEzdV8IaIJ26CzJAnZU6hQVT5qzuhTRDdEy9Jx6ZD1OulRw
-          e7MDG7rCEkdR0KbH104RLRO4FrvgbISeA7nQR5QKkFpLLlADyoNo4YskA5ejFscS38lEhI2hIXgM
-          2oK4gFR8m6gJ1CzgdzyCa22MSA0KBCoc68QjDVE+QiAXB4/A2DMKktqsJsstGsiOFn6Pi2ZRQbai
-          mhghyoPTgVF+qMBjHShTPuO+PAuHMnZNAylybZ0ONkoF1ttec8YXvXKoD3WNjKRgBx+wlDhxOHGX
-          Cnxymji3Lo492DhYffXMJWZ0SijjA7B9H4OzdyEGPX6tb1GcCdmjPEwSqIlHCF2fOD9rsDJ5M8qW
-          JBSbNEHP1mm50lOHUUQgCOSH7HEFLdqorbOMJ6r0EDhRl+2OIKiZtSz2phoHhTHiQ/btIC4x5oFZ
-          Lff0Mh0vxnoQm6ebhhgngCVK4y2Wwf54Ql7eRjmmpud0J1+lmnx70h6y/kR5bEVTbwr6MgP4WFbG
-          8MUWMD2nrteDpnss7VYXV5uxoDlvqQm8+emEalIbJ8Dl6qJ6p+TBo9oQZbJ3jLOuRT/JXf943lP5
-          uaQztpxNtP+f0nvlR/2BmkmVb5Y/A85hr+gP55XwXhhj3uTfCnvzuhA2gvwQHB40IOf78FjbIY5L
-          1shRFLtDHahBzttET/c5e5l9BgAA//8DALl/gOtnBgAA
+          H4sIAAAAAAAAA3RUTW8jNwy9+1cQOiWAbdiOkxS+BbsLNNgivbRA0HphcCTODGuNpEocJ26Q/76Q
+          xrGdbvYyGOnx4z1S5MsIQLFRK1C6RdFdsJNP868P5j+Wh8dt/fTn17/6cH2L+4crt/z1989qnD18
+          9Q9pefOaat8FS8LeDbCOhEI56vz2erm8WtzOrgvQeUM2uzVBJks/WcwWy8l8PlnMDo6tZ01JreDv
+          EQDAS/lmis7Qs1rBbPx201FK2JBaHY0AVPQ23yhMiZOgEzU+gdo7IVdYv6wdwFqlvusw7tdqBWv1
+          5TlYZIeVJbiLwjVrRgv3TshabshpgovHu/tLiFRTTCAeOpLWmwToDAjp1vG/PSWQFgU63BJIS2BI
+          c2LvJh1u2TUQoteUEiXwNdzdQylKGkPAKKx7i9HuwRAFsITRZZeLz79dHu3YCcUQSQrVnLp3hmLW
+          a/LVFB7v7oHdztsdpZxuxyZHQWM4NwktsKt97DCfsg5tMXK9L3RLmZ6lBNbYZ54VtewMhEiGdfZJ
+          U7gX4KxAyAHn9nfkhAxgAgR58pMkFN60rqDmmGQMhnZkfShsHKDWfUQhqHoB593kvbQieDwUtyVX
+          +LsGKDfKFe6lCSzpPbUs/60zGh1UlGsW2SXWcFH1bCVf+CK3JLkEH4GejzYYgmUyEHySiUTk3IXL
+          KXzZoe1RMot3NUaRyFUvlMDylgALFazYsuzHcJgPcpRSPsVIWoaD8R2yg5JQHx1qNjT8RV/16WCb
+          C5F6rdkN3lP4o6VEp1fIXa5HFdk0w9NrMEBF8kTkBqGHkut9CXZW7pJ4ulbjYTIiWdqh07RJ2kca
+          JmQ+O+J9IrPhDhtKGavRJlq71/Nxi1T3CfO0u97aMwCd8zK0Lw/6twPyehxt65sQfZX+56pqdpza
+          TSRM3uUxTuKDKujrCOBbWSH9u62gQvRdkI34LZV08/kvV0NAddpaZ/DN8oCKF7RnwNXydvxByI0h
+          QbbpbA8pjbolc57zenEUgb1hf8JmozPtP1L6KPygn11zFuWn4U+A1hSEzOY0Kh+ZRcqb/Wdmx1oX
+          wipR3LGmjTDF3A9DNfZ2WLoq7ZNQt6nZNfmR8bB567BZzG/o5nqpl0aNXkffAQAA//8DAD1ViMGC
+          BgAA
       headers:
         CF-RAY:
-          - 96665c97d85dd03d-SJC
+          - 96a9b5586c211698-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4378,9 +4008,17 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:27 GMT
+          - Tue, 05 Aug 2025 22:25:07 GMT
         Server:
           - cloudflare
+        Set-Cookie:
+          - __cf_bm=e97wPPw42xtrZ2GV3XGKWIGENTOYiGBrXLYvIRL.ur0-1754432707-1.0.1.1-Bi7AFtiSDhJBkZebTqC6JYrOkcaaf1FCVdh7jJmJj.zCoASlG0nrLgfEFoaYA8m4F_ELk9bNyJAkxrnlguCkA2M.HMd6C__Ms7izTAgJSM8;
+            path=/; expires=Tue, 05-Aug-25 22:55:07 GMT; domain=.api.openai.com; HttpOnly;
+            Secure; SameSite=None
+          - _cfuvid=EbhutmVn9GEDLhcd.0zFr3o7z2ahefB0D23utTXcj20-1754432707182-0.0.1.1-604800000;
+            path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
+        Strict-Transport-Security:
+          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -4394,15 +4032,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "2823"
+          - "1739"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "2829"
+          - "1743"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4410,13 +4046,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998616"
+          - "29998551"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - req_58610991a84fbf4b3861dc046b3b0e6e
+          - req_bd80ec1fdec34e3f8b47e877880c1eaa
       status:
         code: 200
         message: OK
@@ -4424,76 +4060,81 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 14-16: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nsame
-        optimization problem.100 Grabocka\\n\\net al. 111 have developed a method named
-        Adversarial Training on EXplanations (ATEX)\\n\\nwhich improves model robustness
-        via exposure to adversarial examples.  While there are\\n\\nconceptual disparities,
-        we note that the counterfactual and adversarial explanations are\\n\\nequivalent
-        mathematical objects.\\n\\n   Matched molecular pairs (MMPs) are pairs of molecules
-        that differ structurally at only\\n\\none site by a known transformation.112,113
-        MMPs are widely used in drug discovery and\\n\\nmedicinal chemistry as these
-        facilitate fast and easy understanding of structure-activity re-\\n\\nlationships.114\u2013116
-        Counterfactuals and MMP examples intersect if the structural change is\\n\\nassociated
-        with a significant change in the properties. In the case the associated changes
-        in\\n\\nthe properties are non-significant, the two molecules are known as bioisosteres.117,118
-        The con-\\n\\nnection between MMPs and adversarial training examples has been
-        explored by van Tilborg\\n\\net al. 119. MMPs which belong to the counterfactual
-        category are commonly used in outlier\\n\\nand activity cliff detection.113
-        This approach is analogous to counterfactual explanations,\\n\\nas the common
-        objective is to uncover learned knowledge pertaining to structure-property\\n\\nrelationships.70\\n\\n\\nApplications\\n\\n\\nModel
-        interpretation is certainly not new and a common step in ML in chemistry, but
-        XAI for\\n\\nDL models is becoming more important60,66\u201369,73,88,104,105
-        Here we illustrate some practical\\n\\nexamples drawn from our published work
-        on how model-agnostic XAI can be utilized to\\n\\n\\n\\n                                       14interpret
-        black-box models and connect the explanations to structure-property relationships.\\n\\nThe
-        methods are \u201CMolecular Model Agnostic Counterfactual Explanations\u201D
-        (MMACE)9\\n\\nand \u201CExplaining molecular properties with natural language\u201D.10
-        Then we demonstrate how\\n\\ncounterfactuals and descriptor explanations can
-        propose structure-property relationships in\\n\\nthe domain of molecular scent.31\\n\\n\\nBlood-brain
-        barrier permeation prediction\\n\\n\\nThe passive diffusion of drugs from the
-        blood stream to the brain is a critical aspect in drug\\n\\ndevelopment and
-        discovery.120 Small molecule blood-brain barrier (BBB) permeation is a\\n\\nclassification
-        problem routinely assessed with DL models.121,122 To explain why DL models\\n\\nwork,
-        we trained two models a random forest (RF) model123 and a Gated Recurrent Unit\\n\\nRecurrent
-        Neural Network (GRU-RNN). Then we explained the RF model with generated\\n\\ncounterfactuals
-        explanations using the MMACE9 and the GRU-RNN with descriptor expla-\\n\\nnations.10
-        Both the models were trained on the dataset developed by Martins et al. 124.
-        The\\n\\nRF model was implemented in Scikit-learn125 using Mordred molecular
-        descriptors126 as the\\n\\ninput features. The GRU-RNN model was implemented
-        in Keras.127 See Wellawatte et al. 9\\n\\nand Gandhi and White 10 for more details.\\n\\n
-        \  According to the counterfactuals of the instance molecule in figure 1, we
-        observe that the\\n\\nmodifications to the carboxylic acid group enable the
-        negative example molecule to permeate\\n\\nthe BBB. Experimental findings by
-        Fischer et al. 120 show that the BBB permeation of\\n\\nmolecules are governed
-        by hydrophobic interactions and surface area. The carboxylic group is\\n\\na
-        hydrophilic functional group which hinders hydrophobic interactions and addition
-        of atoms\\n\\nenhances the surface area. This proves the advantage of using
-        counterfactual explanations,\\n\\nas they suggest actionable modification to
-        the molecule to make it cross the BBB.\\n\\n   In Figure 2 we show descriptor
-        explanations generated for Alprozolam, a molecule that\\n\\npermeates the BBB,
-        using the method described by Gandhi and White 10. We see that\\n\\npredicted
-        permeability is positively correlated with the aromaticity of the molecule,
-        while\\n\\n\\n                                       15negatively correlated
-        with the number of hydrogen bonds donors and acceptors. A similar\\n\\nstructure-property
-        relationship for BBB permeability is proposed in more mechanistic stud-\\n\\nies.128\u2013130
-        The substructure attributions indicates a reduction in hydrogen bond donors
-        and\\n\\nacceptors.  These descriptor explanations are quantitative and interpretable
-        by chemists.\\n\\nFinally, we can use a natural language model to summarize
-        the findings into a written\\n\\nexplanation, as shown in the printed text in
-        Figure 2.\\n\\n\\n\\n\\n\\nFigure 1: Counterfactuals of a molecule which cannot
-        permeate the blood-brain barrier.\\nSimilarity is the Tanimoto similarity of
-        ECFP4 fingerprints.131 Red indicates deletions and\\ngreen indicates substitutions
-        and addition of atoms. Republished from Ref.9 with permission\\nfrom the Royal
-        Society of Chemistry.\\n\\n\\n\\nSolubility prediction\\n\\n\\nSmall molecule
-        solubility prediction is a classic cheminformatics regression challenge and
-        is\\n\\nimportant for chemical process design, drug design and crystallization.133\u2013136
-        In our previous\\n\\nworks,9,10 we implemented and trained an RNN model in Keras
-        to predict solubilities (log\\n\\nmolarity) of small molecules.127 The AqS\\n\\n----\\n\\nQuestion:
-        What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 25-28: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\n2021,
+        25, 1315\u20131360.\\n\\n\\n (9) Wellawatte, G. P.; Seshadri, A.; White, A.
+        D. Model agnostic generation of counter-\\n\\n     factual explanations for
+        molecules. Chemical Science 2022, 13, 3697\u20133705.\\n\\n\\n(10) Gandhi, H.
+        A.; White, A. D. Explaining structure-activity relationships using locally\\n\\n
+        \     faithful surrogate models. chemrxiv 2022,\\n\\n\\n(11) Gormley, A. J.;
+        Webb, M. A. Machine learning in combinatorial polymer chemistry.\\n\\n     Nature
+        Reviews Materials 2021,\\n\\n\\n(12) Gomes, C. P.; Fink, D.; Dover, R. B. V.;
+        Gregoire, J. M. Computational sustainability\\n\\n     meets materials science.
+        Nature Reviews Materials 2021,\\n\\n\\n(13) On scientific understanding with
+        artificial intelligence. Nature Reviews Physics 2022\\n\\n     4:12 2022, 4,
+        761\u2013769.\\n\\n\\n(14) Arrieta, A. B.; D\xB4\u0131az-Rodr\xB4\u0131guez,
+        N.; Ser, J. D.; Bennetot, A.; Tabik, S.; Barbado, A.;\\n\\n     Garcia, S.;
+        Gil-Lopez, S.; Molina, D.; Benjamins, R.; Chatila, R.; Herrera, F. Explain-\\n\\n
+        \    able Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities
+        and Chal-\\n\\n     lenges toward Responsible AI. Information Fusion 2019, 58,
+        82\u2013115.\\n\\n\\n(15) Murdoch, W. J.; Singh, C.; Kumbier, K.; Abbasi-Asl,
+        R.; Yu, B. Interpretable machine\\n\\n     learning: definitions, methods, and
+        applications. ArXiv 2019, abs/1901.04592.\\n\\n\\n                                      25(16)
+        Boobier, S.; Osbourn, A.; Mitchell, J. B. Can human experts predict solubility
+        better\\n\\n     than computers? Journal of cheminformatics 2017, 9, 1\u201314.\\n\\n\\n(17)
+        Lee, J. D.; See, K. A. Trust in automation: Designing for appropriate reliance.
+        Human\\n\\n     Factors 2004, 46, 50\u201380.\\n\\n\\n(18) Bolukbasi, T.; Chang,
+        K.-W.; Zou, J. Y.; Saligrama, V.; Kalai, A. T. Man is to com-\\n\\n     puter
+        programmer as woman is to homemaker? debiasing word embeddings. Advances\\n\\n
+        \    in neural information processing systems 2016, 29.\\n\\n\\n(19) Buolamwini,
+        J.; Gebru, T. Gender Shades:  Intersectional Accuracy Disparities in\\n\\n    Commercial
+        Gender Classification. Proceedings of the 1st Conference on Fairness,\\n\\n
+        \    Accountability and Transparency. 2018; pp 77\u201391.\\n\\n\\n(20) Lapuschkin,
+        S.; W\xA8aldchen, S.; Binder, A.; Montavon, G.; Samek, W.; M\xA8uller, K.-R.\\n\\n
+        \   Unmasking Clever Hans predictors and assessing what machines really learn.
+        Nature\\n\\n     communications 2019, 10, 1\u20138.\\n\\n\\n(21) DeGrave, A.
+        J.; Janizek, J. D.; Lee, S.-I. AI for radiographic COVID-19 detection\\n\\n
+        \     selects shortcuts over signal. Nature Machine Intelligence 2021, 3, 610\u2013619.\\n\\n\\n(22)
+        Goodman, B.; Flaxman, S. European Union regulations on algorithmic decision-\\n\\n
+        \   making and a \u201Cright to explanation\u201D. AI Magazine 2017, 38, 50\u201357.\\n\\n\\n(23)
+        ACT, A. I. European Commission. On Artificial Intelligence: A European Approach\\n\\n
+        \    to Excellence and Trust. 2021, COM/2021/206.\\n\\n\\n(24) Blueprint for
+        an AI Bill of Rights, The White House. 2022; https://www.whitehouse.\\n\\n    gov/ostp/ai-bill-of-rights/.\\n\\n\\n(25)
+        Miller, T. Explanation in artificial intelligence: Insights from the social
+        sciences. Ar-\\n\\n       tificial intelligence 2019, 267, 1\u201338.\\n\\n\\n\\n
+        \                                     26(26) Murdoch, W. J.; Singh, C.; Kumbier,
+        K.; Abbasi-Asl, R.; Yu, B. Definitions, meth-\\n\\n     ods, and applications
+        in interpretable machine learning. Proceedings of the National\\n\\n    Academy
+        of Sciences of the United States of America 2019, 116, 22071\u201322080.\\n\\n\\n(27)
+        Gunning, D.; Aha, D. DARPA\u2019s Explainable Artificial Intelligence (XAI)
+        Program.\\n\\n    AI Magazine 2019, 40, 44\u201358.\\n\\n\\n(28) Biran, O.;
+        Cotton, C. Explanation and justification in machine learning: A survey.\\n\\n
+        \    IJCAI-17 workshop on explainable AI (XAI). 2017; pp 8\u201313.\\n\\n\\n(29)
+        Palacio, S.; Lucieri, A.; Munir, M.; Ahmed, S.; Hees, J.; Dengel, A. Xai handbook:\\n\\n
+        \   Towards a unified framework for explainable ai. Proceedings of the IEEE/CVF
+        Inter-\\n\\n     national Conference on Computer Vision. 2021; pp 3766\u20133775.\\n\\n\\n(30)
+        Kuhn, D. R.; Kacker, R. N.; Lei, Y.; Simos, D. E. Combinatorial Methods for
+        Ex-\\n\\n     plainable AI. 2020 IEEE International Conference on Software Testing,
+        Verification\\n\\n    and Validation Workshops (ICSTW) 2020, 167\u2013170.\\n\\n\\n(31)
+        Seshadri, A.; Gandhi, H. A.; Wellawatte, G. P.; White, A. D. Why does that molecule\\n\\n
+        \    smell? ChemRxiv 2022,\\n\\n\\n(32) Das, A.; Rad, P. Opportunities and challenges
+        in explainable artificial intelligence\\n\\n      (xai): A survey. arXiv preprint
+        arXiv:2006.11371 2020,\\n\\n\\n(33) Machlev, R.; Heistrene, L.; Perl, M.; Levy,
+        K. Y.; Belikov, J.; Mannor, S.; Levron, Y.\\n\\n     Explainable Artificial
+        Intelligence (XAI) techniques for energy and power systems:\\n\\n     Review,
+        challenges and opportunities. Energy and AI 2022, 9, 100169.\\n\\n\\n(34) Koh,
+        P. W.; Liang, P. Understanding black-box predictions via influence functions.\\n\\n
+        \    International Conference on Machine Learning. 2017; pp 1885\u20131894.\\n\\n\\n(35)
+        Ribeiro, M. T.; Singh, S.; Guestrin, C. \u201D Why should i trust you?\u201D
+        Explaining the\\n\\n     predictions of any classifier. Proceedings of the 22nd
+        ACM SIGKDD international\\n\\n\\n                                      27     conference
+        on knowledge discovery and data \\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4502,13 +4143,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5773"
+          - "6109"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4518,7 +4159,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4534,24 +4175,25 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//fFRLbxs3EL7rVwx4aoCVIcmW5frmIn0YcJBLCgSoAmGWnN2dmksyQ1KV
-          avi/F9yVJTlxeuGB37y+bx5PEwDFRt2C0h0m3Qc7/WX/oTHXH83v/37sPu13vf3kHy5v/pz3zR/L
-          raqKh6//Jp1evC6074OlxN6NsBbCRCXqfLW8XC2ul4vlAPTekC1ubUjTKz9dzBZX0/l8upgdHDvP
-          mqK6hb8mAABPw1tKdIZ26hZm1ctPTzFiS+r2aASgxNvyozBGjgldUtUJ1N4lckPVT2sHsFYx9z3K
-          fq1uYa0+391X4AV+3QWL7LC2BHeSuGHNaOHeJbKWW3KaKhBqSCIkDz2lzpsI6Awk0p3jr5ki5Eim
-          wOwSSRBKg0F2hqQUZiB1BIY0R/YuQo+GoN5Dj7pjR2AJxbFr4acPD+8GV0MUzr7fP7yDQc14Afdu
-          iDbw2yXwDeiOeo5J9hV8vrsHjoAhWB5Lyk77LQnEJFmnLDQN4gNJ2oOQxdLG2HGIFcSsO8BCpuTU
-          PhcyDeqU0QIVndxofqgwauGQvExrLPRfpHklQ21RP05rvzuW/5sXoB2WEar+N4lGBzG3LcUEqMvf
-          0KTem9Kkg1Fpibeks6Uxcx/EbwkOHJkiWH4kqK33ZloLsoMaRZgEAklPWLPltL+A90c+r8so8dgQ
-          fM3oEidMvCVgF7ntUixMjxWgQENYJI4VIJtBxbE1xfBsHAoy6AFByPBALl4MzauFTVu4dAQtBqgp
-          /UPkYFy6FxmHDnS5R3fS+sCkgh4fSwJOoCUP09x4AeN7ZHeQw0huwXAcRmN/sVbVuCJClrboNG2i
-          9kJlVX5eu+fzvRJqcsSy1i5bewagcz6NopWN/nJAno87bH0bxNfxG1fVsOPYbYQwelf2NSYf1IA+
-          TwC+DLciv1p/FcT3IW2Sf6Qh3Xy+mo8B1ek8ncFXqwOafEJ7BlzOb6o3Qm4MJWQbzw6O0qg7Muc5
-          l4sjCcyG/QmbTc64f1/SW+FH/uzasyg/DH8CtKaQyGxOk/SWmVA54T8yO2o9FKwiyZY1bRKTlH4Y
-          ajDb8bqquI+J+k3Dri1zx+OJbcJGX69W1zfN9c1MTZ4n/wEAAP//AwDP9zAVawYAAA==
+          H4sIAAAAAAAAAwAAAP//dFRNbxtHDL3rVxBzigFJkGRJSXwziiBQCqQ5pECAKpDoGe4urdmZyZDr
+          WDX834vZtWW1dS4LLB/5+Dj8eBgBGHbmCoxtUG2b/OS3+e+f3R+3C/z851q/fDrMPn14v/m4/rri
+          +kdnxiUi3tyS1eeoqY1t8qQcwwDbTKhUWOdvV8vl5eLtbNUDbXTkS1iddLKMk8VssZzM55PF7Cmw
+          iWxJzBX8NQIAeOi/RWJwdG+uYDZ+trQkgjWZq5MTgMnRF4tBERbFoGb8AtoYlEKver/f30oM2/Cw
+          DQBbI13bYj5uzRVszYf75JED3niC66xcsWX0sAlK3nNNwRK8+Xa9uYBMFWUBjdCSNtEJYHCgZJvA
+          PzoS4ADXG9AGFVo8EGhD4MiycAyTFg8cakg5WhIhgVgVbzmKUiugGYMkzBR0DByUcsqkRdW4T9MF
+          R7kU6XqlGqHpWgwyhY0CctvrSjnesSPgIFw3WhQVx/izZOq7IYOylMmxLS0UiPkkUsZAQbpchKK1
+          sQuKN+xZj2PQ3IkOWirkHEhkCt+uN8ACjsV2IuTKE9xh5tgJ9A24VynlWN+5QsoqgCl5tqhDPtug
+          9xRqkoE7phSzdoGVe5OAEIXC+zPmg4DnQ+lTZlIEUkA/hTeL2fz9xRg+diH00oOD6wZP9vL/BT1a
+          jmchi/nFFL42JASinWMSoPvkYyZQvI8htr2CKmNLfe6BqDSV2yISy2jECjJJikG4n6DNGKhNDQr/
+          XaQU70DkoIq5p+9HrX/SUpPjO8pCUDF5JyCdbUrJtqGWRfOx9INyfXwek0FCip7t8WmiplszHsY6
+          k6e7ImonNmYaxns+O+GdkNtxizVJwSr0QtvwuA37/f58czJVnWBZ3NB5fwZgCFGHxpWd/f6EPJ62
+          1Mc65Xgj/wk1FQeWZpcJJYaykaIxmR59HAF8769B968FNynHNulO44H6dPP1ajUQmpcDdA6/e0I1
+          Kvoz4N3icvwK5c6RIns5OynGom3IncWuLtenIrBzHF+w2eis9v9Leo1+qJ9DfcbyS/oXwFpKSm73
+          srSvuWUqR/pXbqe37gUboXzHlnbKlEs/HFXY+eF+mmHWdhWHulwhHo5olXaL+ZrWq6VdOjN6HP0D
+          AAD//wMA66T/jk0GAAA=
       headers:
         CF-RAY:
-          - 96665c9eea89cf26-SJC
+          - 96a9b5590d75239e-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4559,9 +4201,17 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:28 GMT
+          - Tue, 05 Aug 2025 22:25:07 GMT
         Server:
           - cloudflare
+        Set-Cookie:
+          - __cf_bm=Ljf1s6bIokPWYhBPDoPcflsCv5XtgyhDuo1T2c8YtFI-1754432707-1.0.1.1-Jwe_zyuPlihH31zs1xAD6Ke8jZY5qXGOFMglDZ4ywIuOVQm3ADYBd_CFV7IxcbyUchRmjHN8oboD93Qee.FoRxup0uJQcsbWtb4R_lfOlmQ;
+            path=/; expires=Tue, 05-Aug-25 22:55:07 GMT; domain=.api.openai.com; HttpOnly;
+            Secure; SameSite=None
+          - _cfuvid=zYurzQZkj_TJdTjTJGiiLPrhBzVcv0nxDmfHz8dqT_E-1754432707603-0.0.1.1-604800000;
+            path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
+        Strict-Transport-Security:
+          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -4575,15 +4225,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "2672"
+          - "2022"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "2675"
+          - "2055"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4591,13 +4239,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998624"
+          - "29998547"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - req_d631be915450db0ac9d2df7318fdf6ea
+          - req_69fd42b573c6449cb76030117cae8fdb
       status:
         code: 200
         message: OK
@@ -4605,77 +4253,79 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 16-20: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nssion
-        challenge and is\\n\\nimportant for chemical process design, drug design and
-        crystallization.133\u2013136 In our previous\\n\\nworks,9,10 we implemented
-        and trained an RNN model in Keras to predict solubilities (log\\n\\nmolarity)
-        of small molecules.127 The AqSolDB curated database137 was used to train the\\n\\nRNN
-        model.\\n\\n   In this task, counterfactuals are based on equation 6. Figure
-        3 illustrates the generated\\n\\nlocal chemical space and the top four counterfactuals.
-        Based on the counterfactuals, we ob-\\n\\nserve that the modifications to the
-        ester group and other heteroatoms play an important role\\n\\nin solubility.
-        These findings align with known experimental and basic chemical intuition.134\\n\\nFigure
-        4 shows a quantitative measurement of how substructures are contributing to
-        the pre-\\n\\n\\n\\n                                       16Figure 2: Descriptor
-        explanations along with natural language explanation obtained for BBB\\npermeability
-        of Alprozolam molecule. The green and red bars show descriptors that influ-\\nence
-        predictions positively and negatively, respectively. Dotted yellow lines show
-        significance\\nthreshold (\u03B1 = 0.05) for the t-statistic. Molecular descriptors
-        show molecule-level proper-\\nties that are important for the prediction. ECFP
-        and MACCS descriptors indicate which\\nsubstructures influence model predictions.
-        MACCS explanations lead to text explanations\\nas shown. Republished from Ref.10
-        with permission from authors. SMARTS annotations for\\nMACCS descriptors were
-        created using SMARTSviewer (smartsview.zbh.uni-hamburg.de,\\nCopyright: ZBH,
-        Center for Bioinformatics Hamburg) developed by Schomburg et al. 132.\\n\\n\\n\\n\\n\\n
-        \                                      17diction. For example, we see that adding
-        acidic and basic groups as well as hydrogen bond\\n\\nacceptors, increases solubility.
-        Substructure importance from ECFP97 and MACCS138 de-\\n\\nscriptors indicate
-        that adding heteroatoms increases solubility, while adding rings structures\\n\\nmakes
-        the molecule less soluble. Although these are established hypotheses, it is
-        interesting\\n\\nto see they can be derived purely from the data via DL and
-        XAI.\\n\\n\\n\\n\\n\\nFigure 3: Generated chemical space for solubility prediction
-        using the RNN model. The\\nchemical space is a 2D projection of the pairwise
-        Tanimoto similarities of the local coun-\\nterfactuals. Each data point is colored
-        by solubility. Top 4 counterfactuals are shown here.\\nRepublished from Ref.9
-        with permission from the Royal Society of Chemistry.\\n\\n\\n\\nGeneralizing
-        XAI \u2013 interpreting scent-structure relationships\\n\\n\\nIn this example,
-        we show how non-local structure-property relationships can be learned with\\n\\nXAI
-        across multiple molecules. Molecular scent prediction is a multi-label classification
-        task\\n\\nbecause a molecule can be described by more than one scent. For example,
-        the molecule\\n\\njasmone can be described as having \u2018jasmine,\u2019 \u2018woody,\u2019
-        \u2018floral,\u2019 and \u2019herbal\u2019 scents.139 The\\n\\nscent-structure
-        relationship is not very well understood,140 although some relationships are\\n\\nknown.
-        \ For example, molecules with an ester functional group are often associated
-        with\\n\\n\\n                                       18Figure 4: Descriptor explanations
-        for solubility prediction model. The green and red bars\\nshow descriptors that
-        influence predictions positively and negatively, respectively. Dotted\\nyellow
-        lines show significance threshold (\u03B1 = 0.05) for the t-statistic. The MACCS
-        and\\nECFP descriptors indicate which substructures influence model predictions.
-        MACCS sub-\\nstructures may either be present in the molecule as is or may represent
-        a modification. ECFP\\nfingerprints are substructures in the molecule that affect
-        the prediction. MACCS descriptor\\nare used to obtain text explanations as shown.
-        Republished from Ref.10 with permission from\\nauthors. SMARTS annotations for
-        MACCS descriptors were created using SMARTSviewer\\n(smartsview.zbh.uni-hamburg.de,
-        Copyright: ZBH, Center for Bioinformatics Hamburg) de-\\nveloped by Schomburg
-        et al. 132.\\n\\n\\n\\n\\n\\n                                       19the \u2018fruity\u2019
-        scent. There are some exceptions though, like tert-amyl acetate which has a\\n\\n\u2018camphoraceous\u2019
-        rather than \u2018fruity\u2019 scent.140,141\\n\\n   In Seshadri et al. 31,
-        we trained a GNN model to predict the scent of molecules and utilized\\n\\ncounterfactuals9
-        and descriptor explanations10 to quantify scent-structure relationships. The\\n\\nMMACE
-        method was modified to account for the multi-label aspect of scent prediction.
-        This\\n\\nmodification defines molecules that differed from the instance molecule
-        by only the selected\\n\\nscent as counterfactuals. For instance, counterfactuals
-        of the jasmone molecule would be false\\n\\nfor the \u2018jasmine\u2019 scent
-        but would still be positive for \u2018woody,\u2019 \u2018floral\u2019 and \u2018herbal\u2019
-        scents.\\n\\n\\n\\n\\n\\nFigure 5:  Counterfactual for the 2,4 decadienal molecule.
-        \ The counterfactual indicates\\nstructural changes to ethyl benzoate that would
-        result in the model predicting the molecule\\nto not contain the \u2018fruity\u2019
-        scent. The Tanimoto96 similarity between the counterfactual and\\n2,4 decadienal
-        is also\\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 14-16: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nsame
+        optimization problem.100 Grabocka\\n\\net al. 111 have developed a method named
+        Adversarial Training on EXplanations (ATEX)\\n\\nwhich improves model robustness
+        via exposure to adversarial examples.  While there are\\n\\nconceptual disparities,
+        we note that the counterfactual and adversarial explanations are\\n\\nequivalent
+        mathematical objects.\\n\\n   Matched molecular pairs (MMPs) are pairs of molecules
+        that differ structurally at only\\n\\none site by a known transformation.112,113
+        MMPs are widely used in drug discovery and\\n\\nmedicinal chemistry as these
+        facilitate fast and easy understanding of structure-activity re-\\n\\nlationships.114\u2013116
+        Counterfactuals and MMP examples intersect if the structural change is\\n\\nassociated
+        with a significant change in the properties. In the case the associated changes
+        in\\n\\nthe properties are non-significant, the two molecules are known as bioisosteres.117,118
+        The con-\\n\\nnection between MMPs and adversarial training examples has been
+        explored by van Tilborg\\n\\net al. 119. MMPs which belong to the counterfactual
+        category are commonly used in outlier\\n\\nand activity cliff detection.113
+        This approach is analogous to counterfactual explanations,\\n\\nas the common
+        objective is to uncover learned knowledge pertaining to structure-property\\n\\nrelationships.70\\n\\n\\nApplications\\n\\n\\nModel
+        interpretation is certainly not new and a common step in ML in chemistry, but
+        XAI for\\n\\nDL models is becoming more important60,66\u201369,73,88,104,105
+        Here we illustrate some practical\\n\\nexamples drawn from our published work
+        on how model-agnostic XAI can be utilized to\\n\\n\\n\\n                                       14interpret
+        black-box models and connect the explanations to structure-property relationships.\\n\\nThe
+        methods are \u201CMolecular Model Agnostic Counterfactual Explanations\u201D
+        (MMACE)9\\n\\nand \u201CExplaining molecular properties with natural language\u201D.10
+        Then we demonstrate how\\n\\ncounterfactuals and descriptor explanations can
+        propose structure-property relationships in\\n\\nthe domain of molecular scent.31\\n\\n\\nBlood-brain
+        barrier permeation prediction\\n\\n\\nThe passive diffusion of drugs from the
+        blood stream to the brain is a critical aspect in drug\\n\\ndevelopment and
+        discovery.120 Small molecule blood-brain barrier (BBB) permeation is a\\n\\nclassification
+        problem routinely assessed with DL models.121,122 To explain why DL models\\n\\nwork,
+        we trained two models a random forest (RF) model123 and a Gated Recurrent Unit\\n\\nRecurrent
+        Neural Network (GRU-RNN). Then we explained the RF model with generated\\n\\ncounterfactuals
+        explanations using the MMACE9 and the GRU-RNN with descriptor expla-\\n\\nnations.10
+        Both the models were trained on the dataset developed by Martins et al. 124.
+        The\\n\\nRF model was implemented in Scikit-learn125 using Mordred molecular
+        descriptors126 as the\\n\\ninput features. The GRU-RNN model was implemented
+        in Keras.127 See Wellawatte et al. 9\\n\\nand Gandhi and White 10 for more details.\\n\\n
+        \  According to the counterfactuals of the instance molecule in figure 1, we
+        observe that the\\n\\nmodifications to the carboxylic acid group enable the
+        negative example molecule to permeate\\n\\nthe BBB. Experimental findings by
+        Fischer et al. 120 show that the BBB permeation of\\n\\nmolecules are governed
+        by hydrophobic interactions and surface area. The carboxylic group is\\n\\na
+        hydrophilic functional group which hinders hydrophobic interactions and addition
+        of atoms\\n\\nenhances the surface area. This proves the advantage of using
+        counterfactual explanations,\\n\\nas they suggest actionable modification to
+        the molecule to make it cross the BBB.\\n\\n   In Figure 2 we show descriptor
+        explanations generated for Alprozolam, a molecule that\\n\\npermeates the BBB,
+        using the method described by Gandhi and White 10. We see that\\n\\npredicted
+        permeability is positively correlated with the aromaticity of the molecule,
+        while\\n\\n\\n                                       15negatively correlated
+        with the number of hydrogen bonds donors and acceptors. A similar\\n\\nstructure-property
+        relationship for BBB permeability is proposed in more mechanistic stud-\\n\\nies.128\u2013130
+        The substructure attributions indicates a reduction in hydrogen bond donors
+        and\\n\\nacceptors.  These descriptor explanations are quantitative and interpretable
+        by chemists.\\n\\nFinally, we can use a natural language model to summarize
+        the findings into a written\\n\\nexplanation, as shown in the printed text in
+        Figure 2.\\n\\n\\n\\n\\n\\nFigure 1: Counterfactuals of a molecule which cannot
+        permeate the blood-brain barrier.\\nSimilarity is the Tanimoto similarity of
+        ECFP4 fingerprints.131 Red indicates deletions and\\ngreen indicates substitutions
+        and addition of atoms. Republished from Ref.9 with permission\\nfrom the Royal
+        Society of Chemistry.\\n\\n\\n\\nSolubility prediction\\n\\n\\nSmall molecule
+        solubility prediction is a classic cheminformatics regression challenge and
+        is\\n\\nimportant for chemical process design, drug design and crystallization.133\u2013136
+        In our previous\\n\\nworks,9,10 we implemented and trained an RNN model in Keras
+        to predict solubilities (log\\n\\nmolarity) of small molecules.127 The AqS\\n\\n------------\\n\\nQuestion:
+        What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4684,13 +4334,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5791"
+          - "6049"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4700,7 +4350,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4716,24 +4366,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//dFRNb+M2EL37Vwx4lg3bSZzEN2+QRX0IULQ9LFAvDJoaSbOmSJYzzMYI
-          8t8LUllb7WYvAsTHefP45uN1AqCoVmtQptNi+mCnn05PTbtK98fusDg+X5s/rPv+6dE+2dVv952q
-          coQ/fEMjP6JmxvfBopB3A2wiasHMuri9ubpdrm6WdwXofY02h7VBptd+upwvr6eLxXQ5fw/sPBlk
-          tYa/JwAAr+WbJboaX9Qa5tWPkx6ZdYtqfb4EoKK3+URpZmLRTlR1AY13gq6oft05gJ3i1Pc6nnZq
-          DTv1ZbOtwEd4fAlWk9MHi7CJQg0Z0ha2TtBaatEZrCBig5FBPPQona8ZtKtB0HSO/knIIJ0W6PUR
-          QTqEELEmkw1i8A1stlCcYCAnGENEKekyR3I1xqy9zkcz2LrCUMS/SI7Ov/hiMAap4MtmC8SgQ7CE
-          dRZkOuzJaFvYem/RJKvjWEIFnEwHmoG9TQeyJKdymw06mbLEZCRFhIhWl4iOAs/gs4+ALzrXuoIa
-          2UQKUs6C1W64WXiMT/ldjTaStGXQESHxII9qdELNaSSN0+Gc8905co1N2eqx7hn8dTHY0hHh8eHz
-          7yXj0+bh4c+RJoYObRiUkYPOfwcOaHIxR4kb1EPO7G6kQxLMEkeu+DiYAsbmnmrIlGfOiu9YmoSh
-          xkjP5Fogx9R2MjI41yo3Za4bsmCENvoUcuE/SlNBTRGN2BM00fdQa9GQOHPXiAEs6ujy39A+s52q
-          hk6OaPFZO4N7Nj5i7uj7nXsbt3/EJrHO0+eStSNAO+dlqF4evK/vyNt51KxvQ/QH/l+oasgRd/uI
-          mr3LY8Xigyro2wTgaxnp9J8pVSH6Pshe/BFLusVyfjcQqssWGcHXV++oeNF2BFzdLKoPKPc1iibL
-          o72gjDYd1qPYxc3y/AidavIXbD4Zvf1nSR/RD+8n145Yfkl/AYzBIFjvLw3+0bWIedP+6trZ6yJY
-          McZnMrgXwpjrUWOjkx2WoOITC/b7hlybVw4Nm7AJe7O6vV3dNau7uZq8Tf4FAAD//wMA6D23WhIG
-          AAA=
+          H4sIAAAAAAAAA3RUTY8bNwy9+1cQOiWAx7Ad7y7i2yJN0EW37aVAAtSBwZE4M6w1kipqNnYX+98L
+          abz+aJPLHOaRfI9PJJ8nAIqNWoPSHSbdB1t9WPzyW/Otax8/Pu5/bz5o/PYZ/a/Y/fzPp9uopjnD
+          13+RTq9ZM+37YCmxdyOsI2GiXHVxd7NavVveze8K0HtDNqe1IVUrXy3ny1W1WFTL+TGx86xJ1Br+
+          nAAAPJdvlugM7dUa5tPXPz2JYEtqfQoCUNHb/EehCEtCl9T0DGrvErmi+nnjADZKhr7HeNioNWzU
+          x32wyA5rS3AfEzesGS08uETWcktOE7z5cv/wFiI1FAWSh55S540AOgOJdOf474EEBiFTYNwRpI7A
+          kGZh76oed+xaCNFrEiEB30CPumNHYAmjy2hxSaYQMCbWg8VoD2CIwjnkzU+Pb09x7BLFECkV7VnL
+          4AzFbIDJv2bw4IqM4sA+ZVLdUc+S4mEKX+4fgAUwBMuj7lNBqC3qXVX7/ZGsVNfeOdIJKDvmML97
+          cUNSHHQaIlUh+kAxHSCSHfGOg8zgk49Ae8zTMgXth8zToE4D2utqmcaQ6Mgh+VjVmB09uR3pZDGN
+          rwa9t1SsgiM3k0xBBt0BCtTWe1PVMUfWGCNThECxp0JX2MTboWbL6QAhkmGdkRn80ZHQtbYQ/RMb
+          AiwhxXN2wm2XLhh7b/IEnc05CzzZJNPCnNu5fsL6AMb3WSztcy8Clnf0+mYy26jpOMCRLD2h07QV
+          7SONg/z+BGeTttxjS5KhBq3Qxr1cLkWkZhDMO+kGay8AdM6nUX5ex69H5OW0gNa3Ifpa/pOqGnYs
+          3TYSind52ST5oAr6MgH4WhZ9uNpdFaLvQ9omv6NCt1gu348F1fm2XMCr4x1QySe0F8C729e8q5Jb
+          QwnZysW1UBp1R+Yid3GzPDWBg2F/xuaTi97/L+l75cf+2bUXVX5Y/gxoTSGR2Z5n8HthkfL9/VHY
+          yesiWAnFJ9a0TUwxv4ehBgc7nkYlB0nUbxt2bR5AHu9jE7bLxS3d3qz0yqjJy+RfAAAA//8DAHzg
+          6CUoBgAA
       headers:
         CF-RAY:
-          - 96665cb0dec7cf26-SJC
+          - 96a9b563c97017dc-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4741,11 +4391,9 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:29 GMT
+          - Tue, 05 Aug 2025 22:25:08 GMT
         Server:
           - cloudflare
-        Strict-Transport-Security:
-          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -4759,13 +4407,15 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "1543"
+          - "1686"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
+        strict-transport-security:
+          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "1547"
+          - "1714"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4773,13 +4423,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998614"
+          - "29998556"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - 05cb6b95-a48c-40ac-9c83-0346fdf49e02
+          - req_1b21c84b7897732fdd91c56d85c53d97
       status:
         code: 200
         message: OK
@@ -4787,11 +4437,14 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 12-14: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nnterfactual
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 12-14: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nnterfactual
         approach, contrastive approach employ a dual\\n\\noptimization method, which
         works by generating a similar and a dissimilar (counterfactuals)\\n\\nexample.
         Contrastive explanations can interpret the model by identifying contribution
@@ -4854,8 +4507,382 @@ interactions:
         although both are derived from the same optimization problem.100 Grabocka\\n\\net
         al. 111 have developed a method named Adversarial Training on EXplanations (ATEX)\\n\\nwhich
         improves model robustness via exposure to adversarial examples.  While there
-        are\\n\\nconceptual disparities, we note that\\n\\n----\\n\\nQuestion: What
-        is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        are\\n\\nconceptual disparities, we note that\\n\\n------------\\n\\nQuestion:
+        What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+      headers:
+        accept:
+          - application/json
+        accept-encoding:
+          - gzip, deflate
+        connection:
+          - keep-alive
+        content-length:
+          - "6053"
+        content-type:
+          - application/json
+        host:
+          - api.openai.com
+        user-agent:
+          - AsyncOpenAI/Python 1.99.0
+        x-stainless-arch:
+          - arm64
+        x-stainless-async:
+          - async:asyncio
+        x-stainless-lang:
+          - python
+        x-stainless-os:
+          - MacOS
+        x-stainless-package-version:
+          - 1.99.0
+        x-stainless-raw-response:
+          - "true"
+        x-stainless-read-timeout:
+          - "60.0"
+        x-stainless-retry-count:
+          - "0"
+        x-stainless-runtime:
+          - CPython
+        x-stainless-runtime-version:
+          - 3.13.5
+      method: POST
+      uri: https://api.openai.com/v1/chat/completions
+    response:
+      body:
+        string: !!binary |
+          H4sIAAAAAAAAAwAAAP//dFRNbxtHDL3rVxBz6UUSJFm2E98MN2mcNEZRNECBKhCoGe4u49mZ7ZAr
+          WzD834uZlSU5TS4L7DzykY9fTyMAw85cgbENqm07P7mZf7qrPkb+c64fvv3x6Sw8fFjgI89+/7X6
+          +MWMs0fcfCOrL15TG9vOk3IMA2wToVJmnV+eL5dni8vZZQHa6Mhnt7rTyTJOFrPFcjKfTxazvWMT
+          2ZKYK/hnBADwVL45xeDo0VzBbPzy0pII1mSuDkYAJkWfXwyKsCgGNeMjaGNQCiXrp1UAWBnp2xbT
+          bmWuYGX+vr4dQ0zw7rHzyAE3nuA6KVdsGT3cBiXvuaZgaQyJKkoCGqElbaITwOBAyTaB/+1JQBtU
+          aPGeQBuCLpFjmws0GDqyLOUvVnB9C6UuAhyUUpdIS/Bs2AdHKStx5UkjNH2LQaZwE/tsXaHVHj1Q
+          zjrgPkQiQLinHWDXpYi2AQ5QBHYpbtlxqAFLPpl2DBxyDEsTT1vy+ZfrRgU2O2BHQbnaZRfbYKgp
+          5wkcul6hItQ+vch9iL13gF4pFdVF1S9yon4K72MCesQ8LzkstNGT7T0msA21LJp2Y7CvtAlYDCB9
+          XZNoJs0t2SvNDTgwiKbe7vOJgLZh2hI4Ek7ksvKOkjLJFP46NsrzPcHnz9c370rBb95Pfru7288A
+          pVLKXshlxpoCJVT6Pr8xPLA2e5IN5UoV6ROsQxRlW5ix6zzblzZuojaQqE4keRCKhfV5bl/EgaLc
+          yzS3DdBLhDy/CUVlCIduS0kw5eHUhBw41GN4aNg2UEXbCwnEMGQCKW560UAikFCb0iAMpwPHnnU3
+          XZnxsBqJPG3zSKzFxkTDirw9wLkka26xJslQhV5oFZ5P1y1R1QvmbQ+99ycAhhB16F9e9K975Pmw
+          2j7WXYob+c7VVBxYmnUilBjyGovGzhT0eQTwtZyQ/tVVMF2KbadrjfdUws3nb94MhOZ4tU7gi8Ue
+          1ajoT4Cz8/3teU25dqTIXk7ukLFoG3KnMc8XBxHYO45HbDY60f7/lH5EP+jnUJ+w/JT+CFhLnZJb
+          H7fxR2aJ8mX/mdmh1iVhI5S2bGmtTCn3w1GFvR+OrpGdKLXrikOdZ4yHy1t168X8gi7Ol3bpzOh5
+          9B8AAAD//wMADCbZMYIGAAA=
+      headers:
+        CF-RAY:
+          - 96a9b5641cc41698-SJC
+        Connection:
+          - keep-alive
+        Content-Encoding:
+          - gzip
+        Content-Type:
+          - application/json
+        Date:
+          - Tue, 05 Aug 2025 22:25:09 GMT
+        Server:
+          - cloudflare
+        Transfer-Encoding:
+          - chunked
+        X-Content-Type-Options:
+          - nosniff
+        access-control-expose-headers:
+          - X-Request-ID
+        alt-svc:
+          - h3=":443"; ma=86400
+        cf-cache-status:
+          - DYNAMIC
+        openai-organization:
+          - future-house-xr4tdh
+        openai-processing-ms:
+          - "1755"
+        openai-project:
+          - proj_RpeV6PrPclPHBb5GlExPXSBj
+        openai-version:
+          - "2020-10-01"
+        strict-transport-security:
+          - max-age=31536000; includeSubDomains; preload
+        x-envoy-upstream-service-time:
+          - "1761"
+        x-ratelimit-limit-requests:
+          - "10000"
+        x-ratelimit-limit-tokens:
+          - "30000000"
+        x-ratelimit-remaining-requests:
+          - "9999"
+        x-ratelimit-remaining-tokens:
+          - "29998553"
+        x-ratelimit-reset-requests:
+          - 6ms
+        x-ratelimit-reset-tokens:
+          - 2ms
+        x-request-id:
+          - req_a7cd006c947676bb45d2fa82032982ae
+      status:
+        code: 200
+        message: OK
+  - request:
+      body:
+        "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
+        the relevant information that could help answer the question based on the excerpt.
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 22-25: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nut
+        to models informs the XAI method.\\n\\n\\nConclusion and outlook\\n\\n\\nWe
+        should seek to explain molecular property prediction models because users are
+        more\\n\\nlikely to trust explained predictions, and explanations can help assess
+        if the model is learning\\n\\nthe correct underlying chemical principles. We
+        also showed that black-box modeling first,\\n\\nfollowed by XAI, is a path to
+        structure-property relationships without needing to trade\\n\\nbetween accuracy
+        and interpretability. However, XAI in chemistry has some major open\\n\\nquestions,
+        that are also related to the black-box nature of the deep learning. Some are\\n\\n\\n\\n
+        \                                      22highlighted below:\\n\\n\\n   \u2022
+        Explanation representation: How is an explanation presented \u2013 text, a molecule,
+        attri-\\n\\n      butions, a concept, etc?\\n\\n\\n   \u2022 Molecular distance:
+        \ in XAI approaches such as counterfactual generation, the \u201Cdis-\\n\\n
+        \    tance\u201D between two molecules is minimized. Molecular distance is subjective.
+        Possibil-\\n\\n       ities are distance based on molecular properties, synthesis
+        routes, and direct structure\\n\\n     comparisons.\\n\\n\\n   \u2022 Regulations:
+        As black-box models move from research to industry, healthcare, and\\n\\n     environmental
+        settings, we expect XAI to become more important to explain decisions\\n\\n
+        \     to chemists or non-experts and possibly be legally required. Explanations
+        may need\\n\\n      to be tuned for be for doctors instead of chemists or to
+        satisfy a legal requirement.\\n\\n\\n   \u2022 Chemical space: Chemical space
+        is the set of molecules that are realizable; \u201Crealiz-\\n\\n     able\u201D
+        can be defined from purchasable to synthesizable to satisfied valences. What
+        is\\n\\n     most useful? Can an explanation consider nearby impossible molecules?
+        How can we\\n\\n     generate local chemical spaces centered around a specific
+        molecule for finding counter-\\n\\n      factuals or other instance explanations?
+        \ Similarly, can \u201Cactivity cliffs\u201D be connected\\n\\n      to explanations
+        and the local chemical space.149\\n\\n\\n   \u2022 Evaluating XAI : there is
+        a lack of a systematic framework (quantitative or qualitative)\\n\\n      to
+        evaluate correctness and applicability of an explanation. Can there be a universal\\n\\n
+        \    framework, or should explanations be chosen and evaluated based on the
+        audience and\\n\\n     domain? For example, work by Rasmussen et al. 58 attempts
+        to focus on comparing\\n\\n      feature attribution XAI methods via Crippen\u2019s
+        logP scores.\\n\\n\\n\\n\\n\\n                                       23Acknowledgements\\n\\n\\nResearch
+        reported in this work was supported by the National Institute of General Medical\\n\\nSciences
+        of the National Institutes of Health under award number R35GM137966. This work\\n\\nwas
+        supported by the NSF under awards 1751471 and 1764415. We thank the Center for\\n\\nIntegrated
+        Research Computing at the University of Rochester for providing computational\\n\\nresources.\\n\\n\\nReferences\\n\\n\\n
+        \ (1) Choudhary, K.; DeCost, B.; Chen, C.; Jain, A.; Tavazza, F.; Cohn, R.;
+        Park, C. W.;\\n\\n     Choudhary, A.; Agrawal, A.; Billinge, S. J.; Holm, E.;
+        Ong, S. P.; Wolverton, C.\\n\\n      Recent advances and applications of deep
+        learning methods in materials science. npj\\n\\n      Computational Materials
+        2022, 8.\\n\\n\\n  (2) Keith, J. A.; Vassilev-Galindo, V.; Cheng, B.; Chmiela,
+        S.; Gastegger, M.; M\xA8uller, K.-\\n\\n      R.; Tkatchenko, A. Combining Machine
+        Learning and Computational Chemistry for\\n\\n      Predictive Insights Into
+        Chemical Systems. Chemical Reviews 2021, 121, 9816\u20139872,\\n\\n     PMID:
+        34232033.\\n\\n\\n  (3) Goh, G. B.; Hodas, N. O.; Vishnu, A. Deep learning for
+        computational chemistry.\\n\\n      Journal of Computational Chemistry 2017,
+        38, 1291\u20131307.\\n\\n\\n  (4) Deringer, V. L.; Caro, M. A.; Cs\xB4anyi,
+        G. Machine Learning Interatomic Potentials as\\n\\n     Emerging Tools for Materials
+        Science. Advanced Materials 2019, 31, 1902765.\\n\\n\\n  (5) Faber, F. A.; Hutchison,
+        L.; Huang, B.; Gilmer, J.; Schoenholz, S. S.; Dahl, G. E.;\\n\\n      Vinyals,
+        O.; Kearnes, S.; Riley, P. F.; von Lilienfeld, O. A. Prediction Errors of Molec-\\n\\n
+        \     ular Machine Learning Models Lower than Hybrid DFT Error. Journal of Chemical\\n\\n
+        \    Theory and Computation 2017, 13, 5255\u20135264, PMID: 28926232.\\n\\n\\n\\n
+        \                                      24 (6) Duch, W.; Swaminathan, K.; Meller,
+        J. Artificial Intelligence Approaches for Rational\\n\\n    Drug Design and
+        Discovery. Current Pharmaceutical Design 2007, 13, 1497\u20131508.\\n\\n\\n
+        (7) Dara, S.; Dhamercherla, S.; Jadav, S. S.; Babu, C. M.; Ahsan, M. J.; darasuresh,
+        S. D.;\\n\\n     Dara, S. Machine Learning in Drug Discovery: A Review. Artificial
+        Intelligence Review\\n\\n     123, 55, 1947\u20131999.\\n\\n\\n (8) Gupta, R.;
+        Srivastava, D.; Sahu, M.; Tiwari, S.; Ambasta, R. K.; Kumar, P. Artifi-\\n\\n
+        \     cial intelligence to deep learning: machine intelligence approach for
+        drug discovery.\\n\\n     Molecular diversity 2021, 25, 1315\u20131360.\\n\\n\\n
+        (9) Wellawatte, G. P.; Seshadri, A.; White, A. D. Model agnostic generation
+        of counter-\\n\\n     factual explanations for molecules. Chemical Science 2022,
+        13, 3697\u20133705.\\n\\n\\n(10) Gandhi, H. A.; White, A. D. Explaining structure-ac\\n\\n------------\\n\\nQuestion:
+        What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+      headers:
+        accept:
+          - application/json
+        accept-encoding:
+          - gzip, deflate
+        connection:
+          - keep-alive
+        content-length:
+          - "6096"
+        content-type:
+          - application/json
+        host:
+          - api.openai.com
+        user-agent:
+          - AsyncOpenAI/Python 1.99.0
+        x-stainless-arch:
+          - arm64
+        x-stainless-async:
+          - async:asyncio
+        x-stainless-lang:
+          - python
+        x-stainless-os:
+          - MacOS
+        x-stainless-package-version:
+          - 1.99.0
+        x-stainless-raw-response:
+          - "true"
+        x-stainless-read-timeout:
+          - "60.0"
+        x-stainless-retry-count:
+          - "0"
+        x-stainless-runtime:
+          - CPython
+        x-stainless-runtime-version:
+          - 3.13.5
+      method: POST
+      uri: https://api.openai.com/v1/chat/completions
+    response:
+      body:
+        string: !!binary |
+          H4sIAAAAAAAAAwAAAP//dFTBjhs3DL37Kwidx4bteHcb37ZND8YCRQ9FEbQODFrizCirkQSKcmws
+          9t8Laby2k24uc5hHPpHvkXyZAChr1BqU7lH0EN30t8XTH+3H5afOPh27zddPB7+6+/vX+M/H0594
+          VE3JCPuvpOUta6bDEB2JDX6ENRMKFdbFw91q9WH5MH+owBAMuZLWRZmuwnQ5X66mi8V0OT8n9sFq
+          SmoN/04AAF7qt5ToDR3VGubN25+BUsKO1PoSBKA4uPJHYUo2CXpRzRXUwQv5WvXL1gNsVcrDgHza
+          qjVs1efHTQOB4fdjdGg97h3BI4ttrbboYOOFnLMdeU0N2AQIA0kfDOREBiQAjYkQmYzVRY0EAxqC
+          /QmG4EhnhwyRQySW000YVFnSDDYCaIdUyKwvIiYq7AzCOQmgN0A+ZSaQHuWcBsgEjpC99R3owExa
+          QPc0WI0OIluvbXSUZvD5cQMaPewJMEZnyQC2Qgx7h/p5ug/HkbMQSYDsdTgQQxLOWjLT9FI8k8Pa
+          YW9jgm9W+pAFyhxwGGwqBKh1ZtSnoqn1QhyZBPfWWTnN4IlOoHt0jnxHCayvxVmvXTYETJEpkZf6
+          CIR2FNePbzZgqLW13auuphpevEGDUQp4mwNtKDFtS0xeALOxxclUqmPqskMJfAJPZFJTMwNXPd90
+          TBE1VRYdcmmnRS0ZXWqqL4YO5EIsKemUhAYUq6FlHOhb4OfqKR3QZRT6rrAZ/NVTIsAUScvopmYr
+          9VFMVZYhHKpGEiAy6jNWHNTn7qwH601OwqcGekInvUam5jwzB8vBD0VPB4mkqJNmW9WMa8Dk6FC0
+          2yUdmMZ1WMwveBnwnR2wo1SwFl2irX+93S2mNicsq+2zczcAeh9GF+tWfzkjr5c9dqGLHPbph1RV
+          DE79rixB8GVnk4SoKvo6AfhS70X+7gSoMnxRdhKeqT63WC2XI6G6nqhb+MMZlSDoboC7+7vmHcqd
+          IUHr0s3RURp1T+aWdP7LpYkyZuGKzSc3vf+/pPfox/6t725Yfkp/BbSmKGR21yPzXhhTOeM/C7to
+          XQtWifhgNe3EEhc/DLWY3Xhh1Tjzu9b6riy6Hc9sG3fLxT3d3630yqjJ6+Q/AAAA//8DALLlX+pv
+          BgAA
+      headers:
+        CF-RAY:
+          - 96a9b562cf1fcfd5-SJC
+        Connection:
+          - keep-alive
+        Content-Encoding:
+          - gzip
+        Content-Type:
+          - application/json
+        Date:
+          - Tue, 05 Aug 2025 22:25:09 GMT
+        Server:
+          - cloudflare
+        Strict-Transport-Security:
+          - max-age=31536000; includeSubDomains; preload
+        Transfer-Encoding:
+          - chunked
+        X-Content-Type-Options:
+          - nosniff
+        access-control-expose-headers:
+          - X-Request-ID
+        alt-svc:
+          - h3=":443"; ma=86400
+        cf-cache-status:
+          - DYNAMIC
+        openai-organization:
+          - future-house-xr4tdh
+        openai-processing-ms:
+          - "2100"
+        openai-project:
+          - proj_RpeV6PrPclPHBb5GlExPXSBj
+        openai-version:
+          - "2020-10-01"
+        x-envoy-upstream-service-time:
+          - "2108"
+        x-ratelimit-limit-requests:
+          - "10000"
+        x-ratelimit-limit-tokens:
+          - "30000000"
+        x-ratelimit-remaining-requests:
+          - "9998"
+        x-ratelimit-remaining-tokens:
+          - "29998549"
+        x-ratelimit-reset-requests:
+          - 9ms
+        x-ratelimit-reset-tokens:
+          - 2ms
+        x-request-id:
+          - req_90fabbc3b45940308ee922283b0fb777
+      status:
+        code: 200
+        message: OK
+  - request:
+      body:
+        "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
+        the relevant information that could help answer the question based on the excerpt.
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 8-9: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nrepresented
+        with equation  2.\\n\\n                                          \u2206\u02C6f(\u20D7x)
+        \u2248\u2202\u02C6f(\u20D7x)                                    (2)\\n                               \u2206xi
+        \     \u2202xi\\n\\n\\n\\n                                       7                                                                     \u2206\u02C6f(\u20D7x)
+        \  where \u02C6f(x) is the black-box model and       are used as our attributions.
+        The left-                                                 \u2206xi\\n\\nhand
+        side of equation 2 says that we attribute each input feature xi by how much
+        one unit\\n\\nchange in it would affect the output of \u02C6f(x).  If \u02C6f(x)
+        is a linear surrogate model, then this\\n\\nmethod reconciles with LIME.35 In
+        DL models, \u2207xf(x), suffers from the shattered gradients\\n\\nproblem.62
+        This means directly computing the quantity leads to numeric problems. The\\n\\ndifferent
+        gradient based approaches are mostly distinguishable based on how the gradient
+        is\\n\\napproximated.\\n\\n   Gradient based explanations have been widely used
+        to interpret chemistry predictions.60,66\u201370\\n\\nMcCloskey et al. 60 used
+        graph convolutional networks (GCNs) to predict protein-ligand\\n\\nbinding and
+        explained the binding logic for these predictions using integrated gradients.\\n\\nPope
+        et al. 66 and Jim\xB4enez-Luna et al. 67 show application of gradCAM and integrated
+        gradi-\\n\\nents to explain molecular property predictions from trained graph
+        neural networks (GNNs).\\n\\nSanchez-Lengeling et al. 68 present comprehensive,
+        open-source XAI benchmarks to explain\\n\\nGNNs and other graph based models.
+        They compare the performance of class activation\\n\\nmaps (CAM),63 gradCAM,64
+        smoothGrad,,65 integrated gradients62 and attention mecha-\\n\\nnisms for explaining
+        outcomes of classification as well as regression tasks. They concluded\\n\\nthat
+        CAM and integrated gradients perform well for graph based models. Another attempt\\n\\nat
+        creating XAI benchmarks for graph models was made by Rao et al. 70. They compared\\n\\nthese
+        gradient based methods to find subgraph importance when predicting activity
+        cliffs\\n\\nand concluded that gradCAM and integrated gradients provided the
+        most interpretability\\n\\nfor GNNs.  The GNNExplainer69  is an approach for
+        generating explanations (local and\\n\\nglobal) for graph based models. This
+        method focuses on identifying which sub-graphs con-\\n\\ntribute most to the
+        prediction by maximizing mutual information between the prediction\\n\\nand
+        distribution of all possible sub-graphs. Ying et al. 69 show that GNNExplainer
+        can be\\n\\nused to obtain model-agnostic explanations. SubgraphX is a similar
+        method that explains\\n\\nGNN predictions by identifying important subgraphs.71\\n\\n
+        \  Another set of approaches like DeepLIFT72 and Layerwise Relevance backPropagation73\\n\\n\\n\\n
+        \                                      8(LRP) are based on backpropagation of
+        the prediction scores through each layer of the neu-\\n\\nral network. The specific
+        backpropagation logic across various activation functions differs\\n\\nin these
+        approaches, which means each layer must have its own implementation. Baldas-\\n\\nsarre
+        and Azizpour 74 showed application of LRP to explain aqueous solubility prediction
+        for\\n\\nmolecules.\\n\\n  SHAP is a model-agnostic feature attribution method
+        that is inspired from the game\\n\\ntheory concept of Shapley values.44,46 SHAP
+        has been popularly used in explaining molecular\\n\\nprediction models.75\u201378
+        It\u2019s an additive feature contribution approach, which assumes that\\n\\nan
+        explanation model is a linear combination of binary variables z.  If the Shapley
+        value\\nfor the ith feature is \u03D5i, then the explanation is \u02C6f(\u20D7x)
+        = Pi \u03D5i(\u20D7x)zi(\u20D7x). Shapley values for\\n\\nfeatures are computed
+        using Equation 3.79,80\\n\\n\\n\\n                           M\\n                                  1\\n
+        \                                   \u03D5i(\u20D7x) = X \u02C6f (\u20D7z+i)
+        \u2212\u02C6f (\u20D7z\u2212i)                            (3)\\n                 M\\n\\n
+        \  Here \u20D7z is a fabricated example created from the original \u20D7x and
+        a random perturbation \u20D7x\u2032.\\n\\n\u20D7z+i has the feature i from \u20D7x
+        and \u20D7z\u2212i has the ith feature from \u20D7x\u2032. Some care should
+        be taken\\n\\nin constructing \u20D7z when working with molecular descriptors
+        to ensure that an impossible \u20D7z is\\n\\nnot sampled (e.g., high count of
+        acid groups but no hydrogen bond donors). M is the sample\\n\\nsize of perturbations
+        around \u20D7x. Shapley value computation is expensive, hence M is chosen\\n\\naccordingly.
+        Equation 3 is an approximation and gives contributions with an expectation\\nterm
+        as \u03D50 + Pi=1 \u03D5i(\u20D7x) = \u02C6f(\u20D7x).\\n\\n   Visualization
+        based feature attribution has also been used for molecular data. In com-\\n\\nputer
+        science, saliency maps are a way to measure spatial feature contribution.81
+        Simply put,\\n\\nsaliency maps draw a connection between the model\u2019s neural
+        fingerprint components (trained\\n\\nweights) and input features. Weber et al.
+        82 used saliency maps to build an explainable GCN\\n\\narchitecture that gives
+        subgraph importance for small molecule activity prediction. On the\\n\\nother
+        hand, similarity maps compare model predictions for two or more molecules based
+        on\\n\\ntheir chemical fingerprints.83 Similarity maps provide atomic weights
+        or predicte\\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4864,13 +4891,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5777"
+          - "6108"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4880,7 +4907,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4896,25 +4923,25 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//dFTLbuNGELzrKxpzSQJIgiTbklc3x1gHDqC9JFgsEi2E1rBJdjSc4U43
-          ZWsN/3swpB70Pi48sLprqvr1MgAwnJklGFui2qp2o98PqzxfFX8+ffx4Q/PyH/Hb/PDlr4fVtLj9
-          aoYpI2z/I6unrLENVe1IOfgOtpFQKbFOFzdXi9n8ZrZogSpk5FJaUevoOoxmk9n1aDodzSbHxDKw
-          JTFL+HcAAPDSfpNEn9GzWcJkePpTkQgWZJbnIAATg0t/DIqwKHo1wwtog1fyreqXtQdYG2mqCuNh
-          bZawNu+fa4fscesI7qJyzpbRwaNXco4L8pbg1093j79BpJyigAaoSMuQCaDPQMmWnr80JKAlKlS4
-          I9CSoI6UsU3VEQg5VGhL9gSOMHr2BbRFEWCvFOtI2ipIjI3PKCYbWftLA5RNhV7GcB+aFJ2j1QYd
-          UJLusXsCIwHCjg6gIThgD5/uHodQx7DnLL2HrZZEOQT2id/SyNGeUrBwUarA9gCckVfODynFlugL
-          ShqBfd0o5ITaxJPVp9C4DNApxdZx6+gX6Tkfw0OIQM+YBiU9C1VwZBuHEWxJFYvGwxDsG18CFj1I
-          UxQkmkhTT44uU/HPDKKxsUc9AVJ9aU+QkXCkLDmvKSqTjOHvS5Mc7whWq7v7922x7x9Gf3z4cBwC
-          im0ZG6EsMRbkKaLSt/qG8MRaHkm2dG7mCAsfRNm2zFjXju2phdugJUQqIolw8G2EdWlgT+ZAUXYy
-          Tm0DdBIgDW5EUemew2xPUTCm6dSInIZoCE8l2xLyYBshgeA7JRDDthH1JAIRtWwbhL4/bOxYD+O1
-          GXY7EcnRPo3ERmyI1O3Gu7VZ+9f+MkXKG8G0y75xrgeg90G7JqU1/nxEXs+L60JRx7CVb1JNzp6l
-          3ERCCT4tqWioTYu+DgA+tweiebPzpo6hqnWjYUftc9Pp1fFCmMtN6sE3V0dUg6LrAbPbE/KGcpOR
-          IjvpXRlj0ZaU9XIns+uzCWwyDhdsMuh5/17Sj+g7/+yLHstP6S+AtVQrZZvLyv0oLFK62z8LO9e6
-          FWyE4p4tbZQppn5klGPjupNq5CBK1SZnX6RB4u6u5vXGzheL+W0+v52YwevgfwAAAP//AwC8OdGb
-          YAYAAA==
+          H4sIAAAAAAAAA3RU224bNxB911cM+GQDK0GSFSv1mxD34l6cIvGDgSoQRtzZ3am5JMshHcuG/70g
+          15aU1nkRVjxzO2cuTyMAxbW6AKU7jLr3Zvxh9tt1s7y8rX9K7teWz26aj4+PNx8v9fbTdKGq7OG2
+          f5OOr14T7XpvKLKzA6wDYaQcdbZ8t1iczZfTZQF6V5PJbq2P44Ubz6fzxXg2G8+nL46dY02iLuCv
+          EQDAU/nNJdqaHtQFTKvXl55EsCV1sTcCUMGZ/KJQhCWijao6gNrZSLZU/bS2AGslqe8x7NbqAtbq
+          xwdvkC1uDcEqRG5YMxq4spGM4ZasJji5XV2dQqCGgkB00FPsXC2AtoZIurP8TyKBJFQXGO8IYkfg
+          A9Wss0CDbU2apfxzDayuoOgiFXgMkXUyGMwOBlUfwFkSMHxHUBN5MITBsm3h5PL30xKtDeg7sJQC
+          GrAUv7pwJ3Dy8/W1nFbANlLwgWJhlu2TrSlkeeryFB10qUcrE7hdXQF6HxzqjgTYapNqyglqJhvH
+          W8zMXlmf0KSdDAnakBu+N5SqfH5Y/XFaDeTG2FonkfXeuzD6/MvqTzgZwjoLnzv0hnZwjyZRLj6X
+          e8+S0PAjZv2OZZakO0ABQcNk9a5YC/dsMHDcQY9eJnDTkdChU9xnwhhj4G2KNFT3TYOiA7Y+RWgI
+          YwokFXBNNnKzA+69C3mwQNK26J5Vgix1BS4ADUMEvTNU+gg+OE8h7o5TDEKzgA6pDFmTXa2kkPsa
+          A1rxGDKlCmJIEgchkuCWTWbGFoZOGdZFFqmAxFMOZgrcMJlXkXVHPUsMg0CH0gp1tu1kraphIQIZ
+          ukeraSPaBRoWYzbd43mwN9xjS5KxBo3Q2j4fb1mgJgnmJbfJmCMArXVxKDbv95cX5Hm/0ca1Prit
+          /MdVNWxZuk0gFGfz9kp0XhX0eQTwpVyO9M0xUD643sdNdHdU0s3OpvMhoDocqyP4fPmCRhfRHAGL
+          8x+qN0JuaorIRo7Oj9J5aeoj3/n7w7nCVLM7YNPREff/l/RW+IE/2/YoynfDHwCtyUeqN4fxe8ss
+          UD7o3zPba10KVkLhnjVtIlPI/aipwWSGW6tkJ5H6TcO2zWeHh4Pb+M18dk7n7xZ6UavR8+hfAAAA
+          //8DAMGJcy95BgAA
       headers:
         CF-RAY:
-          - 96665ca80a431694-SJC
+          - 96a9b566cb40239e-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4922,9 +4949,11 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:30 GMT
+          - Tue, 05 Aug 2025 22:25:09 GMT
         Server:
           - cloudflare
+        Strict-Transport-Security:
+          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -4938,15 +4967,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "3810"
+          - "2023"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "3822"
+          - "2032"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4954,13 +4981,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998620"
+          - "29998537"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - req_d3676c7e64e3e1a476515e59402810c4
+          - req_a86d938a7af14fc4818bcf59871a2cb9
       status:
         code: 200
         message: OK
@@ -4968,11 +4995,14 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 28-30: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\n
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 28-30: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\n
         M. T.; Singh, S.; Guestrin, C. \u201D Why should i trust you?\u201D Explaining
         the\\n\\n     predictions of any classifier. Proceedings of the 22nd ACM SIGKDD
         international\\n\\n\\n                                      27     conference
@@ -5039,7 +5069,7 @@ interactions:
         2018, 16, 31\u201357.\\n\\n\\n(56) Harren, T.; Matter, H.; Hessler, G.; Rarey,
         M.; Grebner, C. Interpretation of structure\u2013\\n\\n      activity relationships
         in real-world drug design data sets using explainable artificial\\n\\n      intelligence.
-        Journal of Chemical Information and Modeling 2022, 62,\\n\\n----\\n\\nQuestion:
+        Journal of Chemical Information and Modeling 2022, 62,\\n\\n------------\\n\\nQuestion:
         What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
@@ -5049,13 +5079,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5809"
+          - "6085"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -5065,7 +5095,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -5081,24 +5111,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFRNbxtHDL3rVxBzXhm2HNuIbw5QoEZb9JK2AapAoGe4u4xnZ7YkV5Fi
-          +L8XM2tZcpsCvewCfEPy8fHjaQHgOLhbcL5H88MYlx/2v7Ttz+2FrXa7H/G3c/nj7uPvXfzwDb/c
-          /Oqa4pEfvpC3g9eZz8MYyTinGfZCaFSiXtxcXd6srq9WNxUYcqBY3LrRlu/ycnW+ere8uFiuzl8c
-          +8ye1N3CnwsAgKf6LRRToJ27hfPmYBlIFTtyt6+PAJzkWCwOVVkNk7nmCPqcjFJl/bROAGun0zCg
-          7NfuFtbu0919A1ngh90YkRM+RII7MW7ZM0a4T0YxckfJUwNCLYmCZRjI+hwUMAUw8n3ivyZSmJRC
-          hfGRwHqCQJ6Vc1oO+Mipg1GyJ1VSyC3c3UOVRsEEk44olKyG5GQko5BVPpahnwZMegYfewLaeZLR
-          ZjaFmMIWhfOk8DXLo4L1aEC7MWY5sNhSzONAyRqgLcYJS9uamgvHMbKvhkLqU2F1qM5LVoWQB+Sk
-          oJPvARUGCuwxAg/YceoaGNBIGKOCep6lKpGDTB0EUu7SGfxEeyi9EH6YSi4FTj5OgWYrqvGWKmtM
-          lYw20BLaJPTGDzBh3CvrnKMKCELdFFH421xGm+VUQo5s+1k7z0bhRSaMmiGw+kkVfI8xUuqo8HrV
-          KHVVP59FyFsinTvuI6eqwIt2c4aDehi7LGz9UCgqfKUYy78E6lktS3WlbY6HggKoodEyt0vraYli
-          hcTbIRjQ95wIIqEkTt3Z2jXzPAtF2mLytFGfhcpcv1+n59MlEGonxbKDaYrxBMCUss1yl/X7/II8
-          vy5czN0o+UH/4epaTqz9Rgg1p7Jcanl0FX1eAHyuiz292VU3Sh5G21h+pJru4ur6cg7ojrfkBL48
-          oJYN4wlw/f66+U7ITSBDjnpyHZxH31M4+h5PCU6B8wmwOCn833y+F3sunlP3f8IfAe9pNAqbUcoe
-          va35+EyoHNv/evYqdCXslGTLnjbGJKUZgVqc4nwHne7VaNi0nLoyT1yPYWnm4nnxNwAAAP//AwBw
-          i6r7CgYAAA==
+          H4sIAAAAAAAAA3RUwW4bNxC96ysGPKWAZEiybMO6GUWCCgVySlIXVSBR5NvdibnklsOV5Rj+94K7
+          srROk4sW4pt5fPM4M88jIsVWLUmZSidTN27y++zPj9Xnw5fDh8tH+8jvP/81TX/s/N/Vhy8focY5
+          I+y+waTXrAsT6sYhcfA9bCJ0Qmad3VwtFpfzm+ltB9TBwuW0skmTRZjMp/PFZDabzKfHxCqwgagl
+          /TMiInrufrNEb3FQS5qOX09qiOgSankKIlIxuHyitAhL0j6p8Rk0wSf4TvV2u/0mwa/989oTrZW0
+          da3j01otaa0+VSAcDGKTqIlhzxZCmhxLolBQRIEIbyAU4XKdlAK9PzROs9c7B7qLiQs2rB2tfIJz
+          XOZ4end/t/ptTBWXleOySuxL2uvIoRWS1FrO93hLNVIVbHChzCdFiISePSfkAPYJsYnoGGptKvYg
+          Bx27iM5kuaBPFQTE3rjWgh5DfBAKnrINUUviPXper/PLyZiw167t/uQ67+9WxJ5qWDbaEde6ZF+O
+          s7rIptd1f7cad4qKqGv0V+RzC+HyJPeV15cdqTxJQi0XtEqknQSq4TsFlCpQxZJC7K602MOFJsMd
+          j6m0c/AlJOs7mdB5/qMLY9JCj3Auf6WByS9CCaby/G8LIccPIAvDkutNETgaN6YCOrURvVO8a4/2
+          ZAkRZet05O+9TblWrnOTwA4FseP01L3AsF1McA4mG++eCHVTaeHv6Krmugkx6dwloaBaP2Szzl5R
+          itpLo+OrFa23iLnDbVd9CtQKolys1bjv6AiHfebbiAkRfWffnuBWYDf5SSEZKrQTrP3L2m+32+HM
+          RBSt6DyyvnVuAGjvQ+obJ0/r1yPycppPF8omhp38kKoK9izVJkJL8HkWJYVGdejLiOhrtwfaN6Ot
+          mhjqJm1SeEB33ex6PusJ1Xn1DODL2yOaQtJuANxcHxfIW8qNRdLsZLBMlNGmgh3kXl1en4rQreVw
+          xqajQe3/l/Qz+r5+9uWA5Zf0Z8AYNAl208Q8lm/LPodF5PX8q7CT151gJYh7NtgkRszvYVHo1vWb
+          U/X9tynYl7m3uV+fRbOZz65xfbUwC6tGL6P/AAAA//8DABj7pcBHBgAA
       headers:
         CF-RAY:
-          - 96665caa3eabd03d-SJC
+          - 96a9b56f4ace17dc-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -5106,7 +5136,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:30 GMT
+          - Tue, 05 Aug 2025 22:25:11 GMT
         Server:
           - cloudflare
         Strict-Transport-Security:
@@ -5124,13 +5154,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "3713"
+          - "1925"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
         x-envoy-upstream-service-time:
-          - "3732"
+          - "1928"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -5138,13 +5168,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998618"
+          - "29998550"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - d87ab3b8-1b1e-994c-827d-bccc2ea04d64
+          - req_6e82b7bec90e44b88ef56ad7b88bac73
       status:
         code: 200
         message: OK
@@ -5152,78 +5182,80 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 8-9: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nrepresented
-        with equation  2.\\n\\n                                          \u2206\u02C6f(\u20D7x)
-        \u2248\u2202\u02C6f(\u20D7x)                                    (2)\\n                               \u2206xi
-        \     \u2202xi\\n\\n\\n\\n                                       7                                                                     \u2206\u02C6f(\u20D7x)
-        \  where \u02C6f(x) is the black-box model and       are used as our attributions.
-        The left-                                                 \u2206xi\\n\\nhand
-        side of equation 2 says that we attribute each input feature xi by how much
-        one unit\\n\\nchange in it would affect the output of \u02C6f(x).  If \u02C6f(x)
-        is a linear surrogate model, then this\\n\\nmethod reconciles with LIME.35 In
-        DL models, \u2207xf(x), suffers from the shattered gradients\\n\\nproblem.62
-        This means directly computing the quantity leads to numeric problems. The\\n\\ndifferent
-        gradient based approaches are mostly distinguishable based on how the gradient
-        is\\n\\napproximated.\\n\\n   Gradient based explanations have been widely used
-        to interpret chemistry predictions.60,66\u201370\\n\\nMcCloskey et al. 60 used
-        graph convolutional networks (GCNs) to predict protein-ligand\\n\\nbinding and
-        explained the binding logic for these predictions using integrated gradients.\\n\\nPope
-        et al. 66 and Jim\xB4enez-Luna et al. 67 show application of gradCAM and integrated
-        gradi-\\n\\nents to explain molecular property predictions from trained graph
-        neural networks (GNNs).\\n\\nSanchez-Lengeling et al. 68 present comprehensive,
-        open-source XAI benchmarks to explain\\n\\nGNNs and other graph based models.
-        They compare the performance of class activation\\n\\nmaps (CAM),63 gradCAM,64
-        smoothGrad,,65 integrated gradients62 and attention mecha-\\n\\nnisms for explaining
-        outcomes of classification as well as regression tasks. They concluded\\n\\nthat
-        CAM and integrated gradients perform well for graph based models. Another attempt\\n\\nat
-        creating XAI benchmarks for graph models was made by Rao et al. 70. They compared\\n\\nthese
-        gradient based methods to find subgraph importance when predicting activity
-        cliffs\\n\\nand concluded that gradCAM and integrated gradients provided the
-        most interpretability\\n\\nfor GNNs.  The GNNExplainer69  is an approach for
-        generating explanations (local and\\n\\nglobal) for graph based models. This
-        method focuses on identifying which sub-graphs con-\\n\\ntribute most to the
-        prediction by maximizing mutual information between the prediction\\n\\nand
-        distribution of all possible sub-graphs. Ying et al. 69 show that GNNExplainer
-        can be\\n\\nused to obtain model-agnostic explanations. SubgraphX is a similar
-        method that explains\\n\\nGNN predictions by identifying important subgraphs.71\\n\\n
-        \  Another set of approaches like DeepLIFT72 and Layerwise Relevance backPropagation73\\n\\n\\n\\n
-        \                                      8(LRP) are based on backpropagation of
-        the prediction scores through each layer of the neu-\\n\\nral network. The specific
-        backpropagation logic across various activation functions differs\\n\\nin these
-        approaches, which means each layer must have its own implementation. Baldas-\\n\\nsarre
-        and Azizpour 74 showed application of LRP to explain aqueous solubility prediction
-        for\\n\\nmolecules.\\n\\n  SHAP is a model-agnostic feature attribution method
-        that is inspired from the game\\n\\ntheory concept of Shapley values.44,46 SHAP
-        has been popularly used in explaining molecular\\n\\nprediction models.75\u201378
-        It\u2019s an additive feature contribution approach, which assumes that\\n\\nan
-        explanation model is a linear combination of binary variables z.  If the Shapley
-        value\\nfor the ith feature is \u03D5i, then the explanation is \u02C6f(\u20D7x)
-        = Pi \u03D5i(\u20D7x)zi(\u20D7x). Shapley values for\\n\\nfeatures are computed
-        using Equation 3.79,80\\n\\n\\n\\n                           M\\n                                  1\\n
-        \                                   \u03D5i(\u20D7x) = X \u02C6f (\u20D7z+i)
-        \u2212\u02C6f (\u20D7z\u2212i)                            (3)\\n                 M\\n\\n
-        \  Here \u20D7z is a fabricated example created from the original \u20D7x and
-        a random perturbation \u20D7x\u2032.\\n\\n\u20D7z+i has the feature i from \u20D7x
-        and \u20D7z\u2212i has the ith feature from \u20D7x\u2032. Some care should
-        be taken\\n\\nin constructing \u20D7z when working with molecular descriptors
-        to ensure that an impossible \u20D7z is\\n\\nnot sampled (e.g., high count of
-        acid groups but no hydrogen bond donors). M is the sample\\n\\nsize of perturbations
-        around \u20D7x. Shapley value computation is expensive, hence M is chosen\\n\\naccordingly.
-        Equation 3 is an approximation and gives contributions with an expectation\\nterm
-        as \u03D50 + Pi=1 \u03D5i(\u20D7x) = \u02C6f(\u20D7x).\\n\\n   Visualization
-        based feature attribution has also been used for molecular data. In com-\\n\\nputer
-        science, saliency maps are a way to measure spatial feature contribution.81
-        Simply put,\\n\\nsaliency maps draw a connection between the model\u2019s neural
-        fingerprint components (trained\\n\\nweights) and input features. Weber et al.
-        82 used saliency maps to build an explainable GCN\\n\\narchitecture that gives
-        subgraph importance for small molecule activity prediction. On the\\n\\nother
-        hand, similarity maps compare model predictions for two or more molecules based
-        on\\n\\ntheir chemical fingerprints.83 Similarity maps provide atomic weights
-        or predicte\\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 16-20: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nssion
+        challenge and is\\n\\nimportant for chemical process design, drug design and
+        crystallization.133\u2013136 In our previous\\n\\nworks,9,10 we implemented
+        and trained an RNN model in Keras to predict solubilities (log\\n\\nmolarity)
+        of small molecules.127 The AqSolDB curated database137 was used to train the\\n\\nRNN
+        model.\\n\\n   In this task, counterfactuals are based on equation 6. Figure
+        3 illustrates the generated\\n\\nlocal chemical space and the top four counterfactuals.
+        Based on the counterfactuals, we ob-\\n\\nserve that the modifications to the
+        ester group and other heteroatoms play an important role\\n\\nin solubility.
+        These findings align with known experimental and basic chemical intuition.134\\n\\nFigure
+        4 shows a quantitative measurement of how substructures are contributing to
+        the pre-\\n\\n\\n\\n                                       16Figure 2: Descriptor
+        explanations along with natural language explanation obtained for BBB\\npermeability
+        of Alprozolam molecule. The green and red bars show descriptors that influ-\\nence
+        predictions positively and negatively, respectively. Dotted yellow lines show
+        significance\\nthreshold (\u03B1 = 0.05) for the t-statistic. Molecular descriptors
+        show molecule-level proper-\\nties that are important for the prediction. ECFP
+        and MACCS descriptors indicate which\\nsubstructures influence model predictions.
+        MACCS explanations lead to text explanations\\nas shown. Republished from Ref.10
+        with permission from authors. SMARTS annotations for\\nMACCS descriptors were
+        created using SMARTSviewer (smartsview.zbh.uni-hamburg.de,\\nCopyright: ZBH,
+        Center for Bioinformatics Hamburg) developed by Schomburg et al. 132.\\n\\n\\n\\n\\n\\n
+        \                                      17diction. For example, we see that adding
+        acidic and basic groups as well as hydrogen bond\\n\\nacceptors, increases solubility.
+        Substructure importance from ECFP97 and MACCS138 de-\\n\\nscriptors indicate
+        that adding heteroatoms increases solubility, while adding rings structures\\n\\nmakes
+        the molecule less soluble. Although these are established hypotheses, it is
+        interesting\\n\\nto see they can be derived purely from the data via DL and
+        XAI.\\n\\n\\n\\n\\n\\nFigure 3: Generated chemical space for solubility prediction
+        using the RNN model. The\\nchemical space is a 2D projection of the pairwise
+        Tanimoto similarities of the local coun-\\nterfactuals. Each data point is colored
+        by solubility. Top 4 counterfactuals are shown here.\\nRepublished from Ref.9
+        with permission from the Royal Society of Chemistry.\\n\\n\\n\\nGeneralizing
+        XAI \u2013 interpreting scent-structure relationships\\n\\n\\nIn this example,
+        we show how non-local structure-property relationships can be learned with\\n\\nXAI
+        across multiple molecules. Molecular scent prediction is a multi-label classification
+        task\\n\\nbecause a molecule can be described by more than one scent. For example,
+        the molecule\\n\\njasmone can be described as having \u2018jasmine,\u2019 \u2018woody,\u2019
+        \u2018floral,\u2019 and \u2019herbal\u2019 scents.139 The\\n\\nscent-structure
+        relationship is not very well understood,140 although some relationships are\\n\\nknown.
+        \ For example, molecules with an ester functional group are often associated
+        with\\n\\n\\n                                       18Figure 4: Descriptor explanations
+        for solubility prediction model. The green and red bars\\nshow descriptors that
+        influence predictions positively and negatively, respectively. Dotted\\nyellow
+        lines show significance threshold (\u03B1 = 0.05) for the t-statistic. The MACCS
+        and\\nECFP descriptors indicate which substructures influence model predictions.
+        MACCS sub-\\nstructures may either be present in the molecule as is or may represent
+        a modification. ECFP\\nfingerprints are substructures in the molecule that affect
+        the prediction. MACCS descriptor\\nare used to obtain text explanations as shown.
+        Republished from Ref.10 with permission from\\nauthors. SMARTS annotations for
+        MACCS descriptors were created using SMARTSviewer\\n(smartsview.zbh.uni-hamburg.de,
+        Copyright: ZBH, Center for Bioinformatics Hamburg) de-\\nveloped by Schomburg
+        et al. 132.\\n\\n\\n\\n\\n\\n                                       19the \u2018fruity\u2019
+        scent. There are some exceptions though, like tert-amyl acetate which has a\\n\\n\u2018camphoraceous\u2019
+        rather than \u2018fruity\u2019 scent.140,141\\n\\n   In Seshadri et al. 31,
+        we trained a GNN model to predict the scent of molecules and utilized\\n\\ncounterfactuals9
+        and descriptor explanations10 to quantify scent-structure relationships. The\\n\\nMMACE
+        method was modified to account for the multi-label aspect of scent prediction.
+        This\\n\\nmodification defines molecules that differed from the instance molecule
+        by only the selected\\n\\nscent as counterfactuals. For instance, counterfactuals
+        of the jasmone molecule would be false\\n\\nfor the \u2018jasmine\u2019 scent
+        but would still be positive for \u2018woody,\u2019 \u2018floral\u2019 and \u2018herbal\u2019
+        scents.\\n\\n\\n\\n\\n\\nFigure 5:  Counterfactual for the 2,4 decadienal molecule.
+        \ The counterfactual indicates\\nstructural changes to ethyl benzoate that would
+        result in the model predicting the molecule\\nto not contain the \u2018fruity\u2019
+        scent. The Tanimoto96 similarity between the counterfactual and\\n2,4 decadienal
+        is also\\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -5232,13 +5264,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5832"
+          - "6067"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -5248,7 +5280,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -5264,24 +5296,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//dFTbbhs3EH3XVwz4lACSIcsXOX5T0yD1Q1IHKdqiVSCMuLO7Y3NJmjO0
-          rRr+94K71sVt8rLA8nCG55y5PI0ADFfmEoxtUW0X3eSnzae6Tic/33y4TtV8Qb9++ev0gm/uPl7f
-          vf/DjEtEWN+Q1W3UkQ1ddKQc/ADbRKhUsh7Pz07ms/Oz2bwHulCRK2FN1MlpmMyms9PJ8fFkNn0J
-          bANbEnMJf48AAJ76b6HoK3o0lzAdb086EsGGzOXuEoBJwZUTgyIsil7NeA/a4JV8z/pp6QGWRnLX
-          YdoszSUszZ+LK3jz4TE6ZI9rR7BIyjVbRgdXXsk5bshbeguJakoCGqAjbUMlgL4CJdt6vsskkIWq
-          ArNXSjGR9hdoyA3aEsREFdvimECoYXEFvTUyhohJ2WaHyW1g7dDeTtbh8QU+gt9aAnq0lKJCxWKz
-          CAncY+KQBYoGjDEFtC3JGNhblyv2DTQJKyavkzUWclvijm+pp9mkUrHdtUFS+Xu/+DSGh5ZtC5ho
-          J20r5lAI+xIRW/CUEzrwpA8h3Qq8+fj5s7ztUwZtKe3UXCmgk+KjH1L0wAQbH0TZvqb59ZfF9ZZK
-          Lqq/thgdbeAeXXG9TqGDBjsqDoe0KTRRNfE6K0FNqDkRcBdDUvSWjuB3loyO/8Hy+GEB+/cEHZO3
-          G+gwDn4Id+wwsW7PEg38W25ax01bLKxD2rpTjO+Co76cUKHioV99MYX2TcRdoRxTuOeqVEVKxmKr
-          BmjDA7CPWbdCynntcunIwTUIWWNWGQP5Fr0tj+8aENfsCu2+UVMWLcVaXIFsRKmTo6UZDyORyNF9
-          cWclNiQaRuPd0iz98+EsJaqzYBlln507AND7oL2f/RR/e0Ged3PrQhNTWMt/Qk3NnqVdJUIJvsyo
-          aIimR59HAN/6/ZBfjbyJKXRRVxpuqX/ueHZ6OiQ0+5V0AJ+9rA+jQdEdACfvtnGvUq4qUmQnB0vG
-          2DJa1UHs8dlsJwJzxWGPTUcH2v9P6XvpB/3sm4MsP0y/B6ylqFSt9u31vWuJytr+0bWd1z1hI5Tu
-          2dJKmVKpR0U1ZjdsVDP0zapm35QG42Gt1nFlz+fz84v6/GJqRs+jfwEAAP//AwB5tFcoXwYAAA==
+          H4sIAAAAAAAAAwAAAP//dFRNb+M2EL37Vwx06UUObNfJdnPL9gMwCgQ9FEWAemGPyZE0a4pkOcN8
+          NMh/X1BKbG2bvRiw3rzhmzcfzzOAim11DZXpUE0f3fzn5e+33R/2ln85fvqpiU1aPNj016fDLf4r
+          tqoLIxy+kNE31oUJfXSkHPwIm0SoVLIuP1yu1z+uPiw+DkAfLLlCa6PO12G+WqzW8+Vyvlq8ErvA
+          hqS6hr9nAADPw2+R6C09VtewqN++9CSCLVXXpyCAKgVXvlQowqLotarPoAleyQ+q9/v9Fwl+65+3
+          HmBbSe57TE/b6hq21d3NpoaQ4NfH6JA9HhzBTVJu2DA62Hgl57glb6iGRA0lAQ3Qk3bBCqC3oGQ6
+          z/9kEtAOFXo8EmhHEBNZNsWpMdCSYeHg5z0e2bcQUzAkQgKhgZsNDIYJsFdKMZEOYgoxe0uplGiH
+          Txqgyz16uYCNH14aqn3Ukqf8pUdDKWoNdzcbYAGM0THZQjQd9WzQDXn74Mhkh2kqtQbJpgMUkODy
+          gR3r0xAthrzORVM2mhNBIocDo+MoF/Dn2QbHx6Ipl0IaNJrRvTkgJnHUkICK4R5f3UkEWUaFbMkr
+          N0/QhQeQSKb0YiJV8uGkoZjVuFy6My3hAn4bXsAyqTWgtcVuNGzZlGYfUNhAm0KOJUOZ4NKFc731
+          oNZQUmQPTfZDXnRvnKIXRYLhMvrwwNqdpQ4+FUM6EgL2wm2nAui49WPo0YcHf25FTOwNR0dSg6U+
+          eNGEWiTf3Wx+EMDXJmgAS4nvCXpCz75tsvvWxyaFHiwqXmyrepz2RI7u0RvaiQmJxqn/eIKL6zvu
+          sSUpUINOaOtftn6/30/3KVGTBcs6++zcBEDvg47Pl03+/Iq8nHbXhTamcJD/UKuGPUu3K94HX/ZU
+          NMRqQF9mAJ+HG5G/WfsqptBH3Wk40vDccnV1NSaszmdpAl8uX1ENim4CrJcf6ndS7iwpspPJoakM
+          mo7shLu8XJ2KwGw5nLHFbFL7/yW9l36sn307yfLd9GfAGIpKdnee+vfCEpXT/b2wk9eD4Eoo3bOh
+          nTKl0g9LDWY3XtVKnkSp3zXs23KceDytTdytlld0dbk2a1vNXmZfAQAA//8DAJL/AlpjBgAA
       headers:
         CF-RAY:
-          - 96665ca9de0f232c-SJC
+          - 96a9b56fbce91698-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -5289,11 +5321,9 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:31 GMT
+          - Tue, 05 Aug 2025 22:25:13 GMT
         Server:
           - cloudflare
-        Strict-Transport-Security:
-          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -5307,13 +5337,15 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "4335"
+          - "4245"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
+        strict-transport-security:
+          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "4342"
+          - "4252"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -5321,13 +5353,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998604"
+          - "29998546"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - 74af338e-d277-4e5a-a98a-f4300db168d0
+          - req_76a1b6e59e3701ede69ce05a0ef4cdea
       status:
         code: 200
         message: OK
@@ -5336,69 +5368,64 @@ interactions:
         '{"model": "deepseek-reasoner", "messages": [{"role": "system", "content":
         "Answer in a direct and concise tone. Your audience is an expert, so be highly
         specific. If there are ambiguous terms or acronyms, first define them."}, {"role":
-        "user", "content": "Answer the question below with the context.\n\nContext:\n\npqac-1026f14b:
+        "user", "content": "Answer the question below with the context.\n\nContext:\n\npqac-0779e397:
         Explainable Artificial Intelligence (XAI) is a field focused on providing interpretations
-        of deep learning (DL) model predictions, addressing the ''black-box'' nature
-        of these models. XAI aims to make DL models more understandable and trustworthy
-        by offering explanations for their predictions. Key terms associated with XAI
-        include interpretability, justifications, and explainability. Interpretability
-        refers to the degree of human understandability intrinsic to a model, while
-        justifications are quantitative metrics that defend the trustworthiness of predictions.
-        Explanations actively clarify the internal decision-making process, providing
-        context and causes for specific predictions. XAI is particularly important in
-        fields like chemistry, where understanding predictions can guide hypotheses
-        and ensure models avoid spurious correlations.\nFrom Wellawatte et al, XAI Review,
-        2023\n\npqac-4fc44825: XAI, or Explainable Artificial Intelligence, refers to
-        methods and techniques used to make the predictions and decisions of AI models
-        understandable to humans. In the context of molecular property prediction, XAI
-        aims to explain how models learn chemical principles, enhancing trust and ensuring
-        the model''s decisions align with scientific understanding. Key challenges in
-        XAI for chemistry include representation of explanations (e.g., text, molecular
-        structures), defining molecular distances for counterfactuals, adapting explanations
-        for different audiences (e.g., chemists, doctors), and evaluating the correctness
-        and applicability of explanations. XAI is increasingly important as AI models
-        transition to practical applications in industry, healthcare, and environmental
-        settings.\nFrom Wellawatte et al, XAI Review, 2023\n\npqac-b0844ec6: XAI, or
-        Explainable Artificial Intelligence, refers to methods and techniques that provide
-        explanations for the decision-making processes of AI models, particularly deep
-        learning (DL) models. While DL models are often highly accurate, they are less
-        interpretable. XAI aims to bridge this gap by offering insights into the underlying
-        mechanisms of predictions. It can be categorized into intrinsic (self-explanatory)
-        and extrinsic (post-hoc) methods. XAI explanations can be evaluated based on
-        attributes like actionability, completeness, correctness, domain applicability,
-        fidelity, robustness, and succinctness. These explanations are particularly
-        valuable in fields like chemical property prediction, where they can align with
-        known physical mechanisms.\nFrom Wellawatte et al, XAI Review, 2023\n\npqac-c3a673e0:
-        Explainable Artificial Intelligence (XAI) refers to methods and techniques that
-        make the predictions of machine learning models interpretable and understandable
-        to humans. Counterfactual explanations are a key tool in XAI, providing actionable,
-        instance-level insights by identifying changes in input features that would
-        alter the model''s prediction. For example, in molecular chemistry, counterfactuals
-        can suggest modifications to molecular structures to achieve desired properties.
-        Techniques like MMACE and CF-GNNExplainer are used to generate counterfactuals,
-        with MMACE being model-agnostic and applicable to both regression and classification
-        tasks. XAI also contrasts with adversarial training, which focuses on model
-        robustness rather than interpretability.\nFrom Wellawatte et al, XAI Review,
-        2023\n\npqac-e9abf8c6: XAI, or Explainable Artificial Intelligence, refers to
-        methods and techniques used to interpret and understand the decisions made by
-        machine learning (ML) and deep learning (DL) models. In the context of chemistry,
-        XAI is applied to uncover structure-property relationships, such as using counterfactual
-        explanations and descriptor-based methods to interpret black-box models. For
-        example, counterfactual explanations can suggest actionable modifications to
-        molecules to improve properties like blood-brain barrier permeability. Descriptor
-        explanations provide quantitative insights into molecular features, aiding chemists
-        in understanding model predictions. XAI bridges the gap between complex models
-        and human interpretability, making it crucial for domains like drug discovery.\nFrom
-        Wellawatte et al, XAI Review, 2023\n\nValid Keys: pqac-1026f14b, pqac-4fc44825,
-        pqac-b0844ec6, pqac-c3a673e0, pqac-e9abf8c6\n\n----\n\nQuestion: What is XAI?\n\nWrite
-        an answer based on the context. If the context provides insufficient information
-        reply \"I cannot answer.\" For each part of your answer, indicate which sources
-        most support it via citation keys at the end of sentences, like (pqac-0f650d59).
-        Only cite from the context above and only use the citation keys from the context.
-        ## Valid citation examples, only use comma/space delimited parentheticals: \n-
-        (pqac-d79ef6fa, pqac-0f650d59) \n- (pqac-d79ef6fa) \n## Invalid citation examples:
-        \n- (pqac-d79ef6fa and pqac-0f650d59) \n- (pqac-d79ef6fa;pqac-0f650d59) \n-
-        (pqac-d79ef6fa-pqac-0f650d59) \n- pqac-d79ef6fa and pqac-0f650d59 \n- Example''s
+        and explanations for deep learning (DL) model predictions, addressing the ''black-box''
+        nature of these models. XAI aims to enhance trust and usability by offering
+        insights into why a model makes specific predictions. Key concepts in XAI include
+        interpretability (the degree of human understandability of a model), justifications
+        (quantitative metrics that validate model trustworthiness), and explanations
+        (descriptions of why a prediction was made). XAI is particularly relevant in
+        chemistry, where understanding DL predictions can guide hypotheses and ensure
+        models are not learning spurious correlations.\nFrom Wellawatte et al, XAI Review,
+        2023\n\npqac-6997b9c2: XAI, or Explainable Artificial Intelligence, refers to
+        methods and techniques used to make the predictions of black-box models interpretable
+        and understandable to humans. In the context of molecular property prediction,
+        XAI methods like molecular counterfactual explanations and descriptor explanations
+        are used to explain how structural features of molecules influence their properties,
+        such as scent. These explanations can be represented as chemical structures
+        or natural language, making them accessible to domain experts. XAI enhances
+        trust in models, aids in assessing model learning, and can inform data selection,
+        model building, and feature modification to affect outcomes.\nFrom Wellawatte
+        et al, XAI Review, 2023\n\npqac-d5ac699e: XAI, or Explainable Artificial Intelligence,
+        is a method used to explain predictions made by molecular property prediction
+        models. It aims to increase user trust and ensure that models are learning correct
+        chemical principles. XAI can be applied after black-box modeling to uncover
+        structure-property relationships without compromising accuracy or interpretability.
+        Key challenges in XAI include representation of explanations, defining molecular
+        distance, adapting explanations for different audiences or regulatory needs,
+        exploring chemical space for counterfactuals, and developing systematic frameworks
+        to evaluate explanations. These aspects are critical as XAI moves into practical
+        applications in industry, healthcare, and environmental settings.\nFrom Wellawatte
+        et al, XAI Review, 2023\n\npqac-5f671746: Explainable Artificial Intelligence
+        (XAI) refers to methods and techniques in AI that make the decision-making processes
+        of AI systems transparent, interpretable, and understandable to humans. It aims
+        to provide insights into how AI models make predictions or decisions, ensuring
+        accountability, trust, and fairness. XAI is discussed in various contexts, including
+        its applications, challenges, and opportunities, as seen in works like , Gunning
+        and Aha (2019), and . These studies explore taxonomies, frameworks, and the
+        importance of responsible AI, emphasizing the need for explainability in diverse
+        fields such as chemistry, energy systems, and policy-making.\nFrom Wellawatte
+        et al, XAI Review, 2023\n\npqac-0277ab97: Explainable Artificial Intelligence
+        (XAI) refers to methods and techniques that make the decision-making processes
+        of AI models, particularly deep learning (DL) models, interpretable and understandable.
+        XAI involves providing additional information to clarify the context and causes
+        behind predictions. It is often implemented as a two-step process: first, developing
+        an accurate but non-interpretable model, and then adding explanations to its
+        predictions. XAI methods can be intrinsic (built into the model) or extrinsic
+        (applied post-training). Evaluating XAI involves attributes like actionability,
+        completeness, correctness, domain applicability, fidelity, robustness, and succinctness.
+        These methods aim to bridge the gap between model accuracy and interpretability.\nFrom
+        Wellawatte et al, XAI Review, 2023\n\nValid Keys: pqac-0779e397, pqac-6997b9c2,
+        pqac-d5ac699e, pqac-5f671746, pqac-0277ab97\n\n------------\n\nQuestion: What
+        is XAI?\n\nWrite an answer based on the context. If the context provides insufficient
+        information reply \"I cannot answer.\" For each part of your answer, indicate
+        which sources most support it via citation keys at the end of sentences, like
+        (pqac-0f650d59). Only cite from the context above and only use the citation
+        keys from the context. ## Valid citation examples, only use comma/space delimited
+        parentheticals: \n- (pqac-d79ef6fa, pqac-0f650d59) \n- (pqac-d79ef6fa) \n##
+        Invalid citation examples: \n- (pqac-d79ef6fa and pqac-0f650d59) \n- (pqac-d79ef6fa;pqac-0f650d59)
+        \n- (pqac-d79ef6fa-pqac-0f650d59) \n- pqac-d79ef6fa and pqac-0f650d59 \n- Example''s
         work (pqac-d79ef6fa) \n- (pages pqac-d79ef6fa) \nDo not concatenate citation
         keys, just use them as is. Write in the style of a scientific article, with
         concise sentences and coherent paragraphs. This answer will be used directly,
@@ -5412,7 +5439,7 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5686"
+          - "5327"
         content-type:
           - application/json
         host:
@@ -5424,39 +5451,39 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//hFdLb+PIEf4rhb7YAiitbHk0tm4D7AQxMrcskAXiQCg2i2Ktm92crqZs
-          7sD/fVFNUpKdCXKU2I+q71XkD8OV2ZntZoMPiOVyi+Xn5V15t11iZXFp8f5+vcHtA1o0hQnlH2ST
-          2RnbYFrZ0HaOEgdvCmMjYaLK7G4+f9p8vt1+2twUpg0VObMzFVEnRM/LSCjBU9QNTWBLYnb//mHY
-          V/RqduvCtCSCBzK7HyYGR2ZnUIQloU+6J/hEXgv4+to5ZI+lI/gSE9dsGR08+kTO8YG8Jbj+/cvj
-          AshrnShCAi2lJlQC6CtIZBvP33sSqEj44KmCFKCL4cgVQdO36Je9ryjq7VW+iX2i2EVKqF0LhBq6
-          SBXb8aceW5Flyb9arAjKAfBcHl+Wl8GRAjpdYHuH0Q2gSIEjjJ79Aa5//bY4rXtp2DaAkSDUiTyE
-          Dr/3BE+mdGifoQyvJE8GrrvvaJc369ttfXNXFpB/luv7uzuy28UKHpNAF7nFOAByCyzaNvkGtagU
-          e0kjPhG9dBjJ20HbsA4j14OWlZoJCo/u1PCyxWd92MVgSdEuINQ1Rf0v8/Y6nmuxVyrqEEE6sgrN
-          OxBTE0N/aICUYT8B/dOm7mp7d3d/+2mxgt+/PAJWVcwXgw2RwDboHPkDCUivwMkFfSU7TgNcT51E
-          9sJ25BwuOR+XhRpwpGExQtMQHMhTzNXp4z96ySTP5X7v0SdWmRxJVRfZqsxq8tWI0dzvCPhLiKlh
-          TyKLD60uVvDbWahKfhkDVm4Ai4kOIfKfVGkL4aKPayFXL2cAQxzGsun1tKALkpZNsIuTJa4/6uQf
-          NFxyAGnoSCG0rq8IbOgVzBpt6vGsztk97BVAS0tHR3IFYG528pDwoUmimuKKfJpE1bLnNnuk65Oy
-          l6lLDSZAlygCXsJ2TavDqoA2OMrWUXouCEgB0DZMR8rmjlSBbahli05r7CgmphPYdoPbzxtaT7qi
-          Byzr+wzC10sRKvp0RNdr0EHwgClFLvtEAo6fFZQYySblsYCaK1L1FBBD2cv07wwEj49yYoQW2QN2
-          nWN7UuZHOlThLGAjJ+3CDeOGTD7UTK6ai9A+JcUhW+wM0NT2cIGiskaRgBP03oYjRQFJsbepj7Q8
-          bYjkRgQa7saQQ8cHTTi1BJTU4JFDhBdODYjlzGl2NXvLnaNZXbNh/wtnU5hxMrA/7M8h//e2LbLZ
-          eqEIL+iTRriSW1HNnjIoJQpVIMGRG5SUSYLVKXTEc9dREqhjaOFf5By+YEoEpMq6Eoh0ZHpZwTdK
-          ejx6dMOfpBcLgeU57C1GqnvnhtWTf/LfQshxhylXOF1WADoHtbr+dK0NXicY+ZTz/VQ4nufR6N4x
-          m+DL4wTsaZSs4Ld8herPUUuKA3YdYdSdJe3gZjHH35zPeSgsy/CqgJSOWtXJr9+mafIxZQq4XUzI
-          6QE/HX4/CeRTAo9K3izOqZQtPY9bOYfTL5cp9EHjimvudJbw8jQhJnecjdjF0LInnwBtDCIgoY9W
-          pw7lTdki71PqfQfv3fE+Pv5fLGg2fjT/7PfRIJc2H4X3rlmQJvSugpLmRK1AOA/ghoarc9Bo1s3z
-          Q4nn4EfLX4huBO4RfEg8nnCh2l4IxGniqvpYJ7KC1jURVSsCS5DQEggO8GQuJqSjJ6Op7ghCanI0
-          5CXvVfFkVvDPXLgKX4k5ouMKInWRhPxURgGPV86BDF5dxaO72pwYoU9Q4zHkF4Xgz7WDBhKgz7dP
-          FphnOryw0+SBnFrZh87Bs9Ki7CfJR7OH2/UaXkKsZGplDDV3pk/3EtpGwbTUpezFNMspW/1v+qoy
-          x+LcScKYxsTLbtOhPJqbx2RNDfmxuskGwYXDeWpnoX6crqOLavZZvONc+B8WYH92yQq+agOK9vhi
-          qa9w5KuxvC5qjFxKosVkmzknJg1dyUkSM9TnMWL7OW/yxXe/fDqlWwGieOmUpXgkAWq7BoUFyj5N
-          KvdXSZthj4ku3p3cBV4r81YYFw6aVWJ2vneuMPpQmv04GczOSAqdeftPYfr5E0FDoEv7FJ7Ji9nd
-          3Nyv9SNh/iw5PdjebguTQkJ3Xnu/3hbvD9hXlJCd6MkWbUPVafX67SfnXq4/j695y+bm9u10QT5u
-          3/C51vWHZy2LvGvkrTAySKJ2X7M/qCs5D8W6228eNqXFh+0n2ncxVOvt7WZfd/f752M+y7z9BQAA
-          //8DAEbzdXLWDQAA
+          H4sIAAAAAAAAAwAAAP//fFfBbuM4Ev2Vgi6JAdmTxIkd+xb09gLBAos9zGGAzcKgyLJVY4pUs8gk
+          3kb+fVCkLNvp7N4siSzWq3r1Hv2zIlOtK3y4a/Ri9TC9fbi/md7f6uX0cfFopvp+3tyZ7d3jvJlX
+          deWbP1HHal3pVsWZ9l1vMZJ3VV3pgCqiqda3y4f7+/nd8va+rjpv0FbryiD2jLifBlTsHQbZ0HrS
+          yNX63z8rcgbfq/VNXXXIrHZYrX9WwVus1pViJo7KRdnjXUQnCXx/760ipxqL8BQibUmTsvDsIlpL
+          O3Qa4fqPp+cJBNxiYIgeOoytNwzKGYioW0c/EjIYZNo5NHmJ2iPEFsGgJibvpp3ak9tBH7xGZmTw
+          W1CnA+n8QD5wxI5r6GWFTlYFewABDxZVcBIol4TrvDH0AWOGICklZzAIUpNfRQ9t6pRjuO5/KD29
+          WS5XOF8ta8iPD9vF8nZ5v5jM4DkC96glJWXtAZQxoeQqSF6qxiq9nzb+/aUCp2IKKCBK996HhKA5
+          CMZXMpLkKTnpLtfwZ+IM+fgs+aK0wJU3sPVBTqMAfUBDuryNHtC1SmoTQ+JYg9LaJxdVQ5bioQRK
+          PDz+f6TD42K1WjYrfTeZwR9Pz6CoG85hATagyeUG7UNAHUtl7aH0kZym3iJDULFFyVo54D4F8onL
+          FjuAkuRKUZCBHNOujfIjenhrD2PRLyB7rVP4Goh5UHqxWuFk9uJe3LOUv0NXagx+G9EBuVdvX5FB
+          QXzzU47YH6m3hi0FjmDwFa3vBYxyUs8UVByA19IDB6rvC1pyMUjeGq6bRDZOyU3AB8D38X3vOU5b
+          ryfjeEQPO3SYo170eAB1t1yqZrWczOAfeMhnHXkB5LRNRpKxmOkvyfcY4uGsRkAOdIsdcQyHGt5a
+          DJhbeUzA0v48RCYMhq3SMSl7mRLapMlIpq1/A44h6ZiCsrDFTHRJaWtTns4hFUIGTroFxcAaXRxw
+          HXn1S7Pg9xb5Uym0ctAgBOwDMrqIRsJlVFrZMRGRi1BmTlmwyu2S2mEeFuM7RU6iYoj8KYfCkG+B
+          Yg6nW2Utuh0yoJO5VVxmezz/yKGLLGswuKVBdY7VZOrIqlCGz6g+yudfRtnQKwZGUMmQFC/jCLhL
+          VkUfDuAQDdd5nw8S4QS9V7ogvOzbUTRYBI+4lU1FL1UkDfiqbCootkF1+ObDnqFRjAa8AxVjoCZF
+          HNixJYNFQIYZd8hcQ/BN4uG3nKYy4S7V5djYoc8nOld1VeyJ3G5zcponG2Tua7AYrxii0ntR55YY
+          fiQMB1CNTzETOLGgksYMqmEgx3mPmUSQGAO8KRdlvvsgFoNDjzLyN4otaIrHRgTfnURmjweegfDi
+          76IDNTxfWQssTFTWHg8CdtT3GHloQTg/gGdwoUtF4M41Fhof27KjzKTic10YfQGenkeL5PpL+zmW
+          4tJ/+uAbi51owIUvzi7FHQKS2/qgs4lRtifRSsrdOTfP0loBW9yylEiUqVxWSNQUOww7XA9mlDMT
+          O/papschrIupZHoPTjLa+CWXJke6nZvbsObMqIUDo6zTpf5fj1r920mdh9JPChUu+AqM2DHokPI9
+          JPtvvtPIEBZX0IeBLz6cy/Qn/b3AfMX/U3j5xJYj7Cs+Kd10VPrRP1vqGVRAkMuIi6TsDL6dpCwP
+          8pnyfFazTyUu8nGSiQ5jIP2LKwG3Plkj6jy4kYEmRdhjH2VIZOZKWcr8HPPPXFWO3zCs4cXdzuDb
+          5fCU6ewDdSoc8jTCi7ubwb9S6D3jbw063FKUt/MZ/H7qRd8Hr3SL8uV+Bn/Lwj8d5/rCQF/cw0WN
+          JM9/YrmbqldPRnKbwrN7VZbMqBXS/k7ls6fw/V0jmsLaQRO07zF/e+oa2iW56EQMHWdMol5nGpHP
+          FKaOHAHldhaBWKa3IyeeSQ66ZKPcpHItamBfFAm7vlVM/5WKqghvLeUl2B8F4aykTfDKQH4S6ZjB
+          91N/zyT/C/ZfnxzgJPqAUc8m0Mlon/dfZbc8XAWElC/9Us6c3R4Ps+qjrqzfiTJxtXbJ2rqSDLnd
+          FD+o1hVH31cf/6mrdPx3IqXo4yb6PTqu1re3d0v5f3L8RzR+WCwf6yr6qOxp7ePNQ30ZYGMwKrIs
+          kbWQxYyrbz6+iHu+/mRaxy3z+8eP8YAcbtPSKdebT986Yr4A8lFXxZo3W3I7UVrKVrjtN/PVvNFq
+          tXjATR+8uVnczTfb/nGzf82xqo+/AAAA//8DAHa+XBBRDgAA
       headers:
         CF-RAY:
-          - 96665cc629a9cf12-SJC
+          - 96a9b58bdfb9fa8a-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -5464,12 +5491,12 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:31 GMT
+          - Tue, 05 Aug 2025 22:25:14 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=hRoEQwJ.K75V4H.Fi1GOmNEGzeiEOyUxEj27ubsWQ8k-1753726531-1.0.1.1-40ymIRCSTikw_tlLCv7_W_8mSX1TsQoXU1HPwm_XyO3Jhh1tFPCUvf8fctPfd337_Ue.D5X2ELfYiVrLpKJZRoceJpaSHGMBfLlNq9LfeN8;
-            path=/; expires=Mon, 28-Jul-25 18:45:31 GMT; domain=.deepseek.com; HttpOnly;
+          - __cf_bm=BsYnd2DIn80ZJCXQ2ncb8sCXxS4XiyBCK4J0F3f6lXs-1754432714-1.0.1.1-Ag5Gi1dBQpJKgjIA.Wxv02yIXg57R7MyhXkJVwp1L8PYi1mRYh0h.yZpbandLm5jzXcBSp0mm.XKr3vUu20yhtXnM3L9KnezvEbTvOR7zSM;
+            path=/; expires=Tue, 05-Aug-25 22:55:14 GMT; domain=.deepseek.com; HttpOnly;
             Secure; SameSite=None
         Strict-Transport-Security:
           - max-age=31536000; includeSubDomains; preload
@@ -5484,7 +5511,7 @@ interactions:
         vary:
           - origin, access-control-request-method, access-control-request-headers
         x-ds-trace-id:
-          - 98250a3e5dfa3128a34efc0c0bc6af0f
+          - 89175579f5f187dbb4d4e79873f5b8b9
       status:
         code: 200
         message: OK
diff --git a/tests/cassettes/test_get_reasoning[openrouter-deepseek].yaml b/tests/cassettes/test_get_reasoning[openrouter-deepseek].yaml
index e0116a0e7..5aab40f86 100644
--- a/tests/cassettes/test_get_reasoning[openrouter-deepseek].yaml
+++ b/tests/cassettes/test_get_reasoning[openrouter-deepseek].yaml
@@ -19,7 +19,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -29,7 +29,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -45,18 +45,18 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA4ySTW/bMAyG7/4VAs/OYDtpUvjWYR2wDb0M+yhQB7Yi04k6WdIkulkR5L8PspPY
-          3TpgFx348KX4kjxEjIGsIWcgdpxEa9Xs7fNd87Wd7zfpl48/27n4dvO+ad592t/J+9tbiIPCbB5R
-          0Fn1RpjWKiRp9ICFQ04Yqqarq/kqW16lyx60pkYVZFtLs4WZZUm2mKXpLEtOwp2RAj3k7CFijLFD
-          /4YWdY2/IGdJfI606D3fIuSXJMbAGRUiwL2XnrgmiEcojCbUfddVVT16owt9KDRjBZAkhQXkrID7
-          mw/sMz5J3BcQD5R3tDPOB/5QwHdUiu85ETIkxlUB61NebWTI0Z1ShT4Wuqqq6f8Om85zdcqYAK61
-          IR7G1ztfn8jx4lWZrXVm4/+QQiO19LvSIfdGB1+ejIWeHiPG1v1MuxdjAutMa6kk8wP771bZUA7G
-          JY5wfoZkiKsxniaL+JVyZY3EpfKTpYDgYof1KB03yLtamgmIJqb/7ua12oNxqbf/U34EQqAlrEvr
-          sJbipeMxzWG48X+lXYbcNwwe3ZMUWJJEFxZRY8M7NZwf+GdP2JaN1Ft01snhBhtbiuVqtbxultcJ
-          RMfoNwAAAP//AwBcX5ZyjAMAAA==
+          H4sIAAAAAAAAAwAAAP//jFJLb9swDL77Vwg8x4OdOM3gW9vLigE7dMAeqANbkWlbrSwJEt1uDfLf
+          B9lpnG4d0IsO/B7iR3IfMQayhpyB6DiJ3qr4Ov385Wdysabn9GN3semus+eHT19v2rX9dnsFi6Aw
+          u3sU9KL6IExvFZI0eoKFQ04YXNPNOstWy02SjEBvalRB1lqKMxMvk2UWp2m8TI7CzkiBHnJ2FzHG
+          2H58Q4u6xl+Qs9FmrPToPW8R8hOJMXBGhQpw76UnrgkWMyiMJtRj11VV3XujC70vNGMFkCSFBeSs
+          gB+XN+wWHyU+FbCYUD5QZ5wP+F0B31Ep/sSJkCExrgrYHnm1kYGjB6UKfSh0VVXn/ztsBs/VkXEG
+          cK0N8TC+Mfn2iBxOWZVprTM7/5cUGqml70qH3BsdcnkyFkb0EDG2HWc6vBoTWGd6SyWZBxy/2ywn
+          O5iXOIOrF5AMcTXX0yRbvGFX1khcKn+2FBBcdFjP0nmDfKilOQOis9D/dvOW9xRc6vY99jMgBFrC
+          urQOayleJ55pDsON/492GvLYMHh0j1JgSRJdWESNDR/UdH7gf3vCvmykbtFZJ6cbbGy53jWrlUhR
+          ZBAdoj8AAAD//wMA7qc2kowDAAA=
       headers:
         CF-RAY:
-          - 96665c683c127afd-SJC
+          - 96a9b5391dd1ed41-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -64,14 +64,14 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:17 GMT
+          - Tue, 05 Aug 2025 22:25:00 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=_aOWzp2Ts0hjUIVqpz2FbZxEqY9wvFkIIkIunDmtqfw-1753726517-1.0.1.1-QWz73omTXi4ThxlZixtBZGyB0DiBq2Pl.LxiEAzHi0y6IHR3NqNQGCPZDaU2UOrEDshfAL0wzKMQsRmCXxOrjp8JKZjSv5UeCsqYhMPhK0M;
-            path=/; expires=Mon, 28-Jul-25 18:45:17 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=Pj26ca0RbMtlnDj_ZjyjLdBu6eWs4zuHi.6Icqya0Bo-1754432700-1.0.1.1-0_nafmcN_I_ocpYRQDCsmiGLKZ8PZmwwBd4jWbh4hRvzjNwVICA2lvK5VOxXxQk.OmXqR1zsq4D6b1D1BLWJe_aFJIgkzUsBTGwEtwVNz28;
+            path=/; expires=Tue, 05-Aug-25 22:55:00 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=ezQLSnCmmw4r2mKAy0QcLkziFEA6Pd03_g7WQgRzqEI-1753726517455-0.0.1.1-604800000;
+          - _cfuvid=sA4acE2QX16cmYOv3uZs.wRf1iqmmAAu4STD3hJ8YNo-1754432700805-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Transfer-Encoding:
           - chunked
@@ -86,7 +86,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "834"
+          - "369"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -94,7 +94,7 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "854"
+          - "388"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -102,13 +102,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29999934"
+          - "29999935"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 0s
         x-request-id:
-          - req_9f62b44ff5c585f517e08e7042732794
+          - req_4f3a5b87147727c68059bd080ce18438
       status:
         code: 200
         message: OK
@@ -130,14 +130,14 @@ interactions:
           null}, "publicationTypes": ["JournalArticle", "Conference", "Review"], "publicationDate":
           "2024-11-20", "journal": {"name": "2024 8th International Conference on System
           Reliability and Safety (ICSRS)", "pages": "511-517"}, "citationStyles": {"bibtex":
-          "@Article{Nguyen2024EnhancingPW,\n author = {Duc An Nguyen and Khanh T. P.
-          Nguyen and Kamal Medjaher},\n booktitle = {International Conference on System
-          Reliability and Safety},\n journal = {2024 8th International Conference on
-          System Reliability and Safety (ICSRS)},\n pages = {511-517},\n title = {Enhancing
-          PHM with XAI: Review and Dataset for System Health Indicator Construction},\n
-          year = {2024}\n}\n"}, "authors": [{"authorId": "2167438752", "name": "Duc
-          An Nguyen"}, {"authorId": "2184162840", "name": "Khanh T. P. Nguyen"}, {"authorId":
-          "2267984723", "name": "Kamal Medjaher"}], "matchScore": 57.629684}]}
+          "@Article{Nguyen2024EnhancingPW,\n author = {D. Nguyen and Khanh T. P. Nguyen
+          and Kamal Medjaher},\n booktitle = {International Conference on System Reliability
+          and Safety},\n journal = {2024 8th International Conference on System Reliability
+          and Safety (ICSRS)},\n pages = {511-517},\n title = {Enhancing PHM with XAI:
+          Review and Dataset for System Health Indicator Construction},\n year = {2024}\n}\n"},
+          "authors": [{"authorId": "2167438752", "name": "D. Nguyen"}, {"authorId":
+          "2184162840", "name": "Khanh T. P. Nguyen"}, {"authorId": "2267984723", "name":
+          "Kamal Medjaher"}], "matchScore": 59.36049}]}
 
           '
       headers:
@@ -146,31 +146,31 @@ interactions:
         Connection:
           - keep-alive
         Content-Length:
-          - "1448"
+          - "1439"
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:18 GMT
+          - Tue, 05 Aug 2025 22:25:01 GMT
         Via:
-          - 1.1 0ac5a786563da4ae2e2a28a1fe210e04.cloudfront.net (CloudFront)
+          - 1.1 fb1699c4cb8ff04b39762e99ca06e3d2.cloudfront.net (CloudFront)
         X-Amz-Cf-Id:
-          - Idkt5gjlDAVpamIvc3bI3Q3HD_2OB-BAjGCcOXD4XZfE0YAF3sdYzQ==
+          - 4yZAUH24sDK8Gw7AliUu1iU7u1v6OAdsPmyT81tZP3T5Suolougpiw==
         X-Amz-Cf-Pop:
           - SFO53-P7
         X-Cache:
           - Miss from cloudfront
         x-amz-apigw-id:
-          - ObqocFO-vHcEHBQ=
+          - O2mtnGYfvHcEqEg=
         x-amzn-Remapped-Connection:
           - keep-alive
         x-amzn-Remapped-Content-Length:
-          - "1448"
+          - "1439"
         x-amzn-Remapped-Date:
-          - Mon, 28 Jul 2025 18:15:18 GMT
+          - Tue, 05 Aug 2025 22:25:01 GMT
         x-amzn-Remapped-Server:
           - gunicorn
         x-amzn-RequestId:
-          - e9f78903-4a69-46ef-94ce-7e7955c6434b
+          - 17ffc548-9711-46d2-ab7e-460bdac6cb4b
       status:
         code: 200
         message: OK
@@ -182,11 +182,11 @@ interactions:
     response:
       body:
         string:
-          '{"status":"ok","message-type":"work-list","message-version":"1.0.0","message":{"facets":{},"total-results":22730,"items":[{"indexed":{"date-parts":[[2024,8,7]],"date-time":"2024-08-07T06:47:19Z","timestamp":1723013239657},"reference-count":0,"publisher":"Transstellar
-          Journal Publications and Research Consultancy Private Limited","issue":"3","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IJMPERD"],"published-print":{"date-parts":[[2020]]},"DOI":"10.24247\/ijmperdjun2020913","type":"journal-article","created":{"date-parts":[[2020,8,27]],"date-time":"2020-08-27T00:33:19Z","timestamp":1598488399000},"page":"9575-9582","source":"Crossref","is-referenced-by-count":0,"title":["A
-          Review Paper on Food Security"],"prefix":"10.24247","volume":"10","author":[{"given":"Ansumansamal
-          et al.,","family":"Ansumansamal et al.,","sequence":"first","affiliation":[]},{"name":"TJPRC","sequence":"additional","affiliation":[]}],"member":"10346","container-title":["International
-          Journal of Mechanical and Production Engineering Research and Development"],"language":"en","deposited":{"date-parts":[[2020,8,27]],"date-time":"2020-08-27T00:33:19Z","timestamp":1598488399000},"score":24.880903,"resource":{"primary":{"URL":"http:\/\/tjprc.org\/publishpapers\/2-67-1598502792-913IJMPERDJUN2020913.pdf"}},"issued":{"date-parts":[[2020]]},"references-count":0,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2020]]}},"alternative-id":["Arch-13770"],"URL":"https:\/\/doi.org\/10.24247\/ijmperdjun2020913","ISSN":["2249-6890"],"issn-type":[{"type":"print","value":"2249-6890"}],"published":{"date-parts":[[2020]]}}],"items-per-page":1,"query":{"start-index":0,"search-terms":null}}}'
+          '{"status":"ok","message-type":"work-list","message-version":"1.0.0","message":{"facets":{},"total-results":22876,"items":[{"indexed":{"date-parts":[[2024,8,7]],"date-time":"2024-08-07T06:45:10Z","timestamp":1723013110666},"reference-count":0,"publisher":"Transstellar
+          Journal Publications and Research Consultancy Private Limited","issue":"6","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IJASR"],"published-print":{"date-parts":[[2017]]},"DOI":"10.24247\/ijasrdec201724","type":"journal-article","created":{"date-parts":[[2017,11,9]],"date-time":"2017-11-09T01:18:32Z","timestamp":1510190312000},"page":"173-180","source":"Crossref","is-referenced-by-count":0,"title":["Bovine
+          Mastitis, Global Quandary \\\"A Review\\\""],"prefix":"10.24247","volume":"7","author":[{"given":"Kamaldeep
+          et al.,","family":"Kamaldeep et al.,","sequence":"first","affiliation":[]},{"name":"TJPRC","sequence":"additional","affiliation":[]}],"member":"10346","container-title":["International
+          Journal of Agricultural Science and Research"],"language":"en","deposited":{"date-parts":[[2017,11,9]],"date-time":"2017-11-09T01:24:49Z","timestamp":1510190689000},"score":24.770569,"resource":{"primary":{"URL":"http:\/\/tjprc.org\/publishpapers\/2-50-1510208310-24.IJASRDEC201724.pdf"}},"issued":{"date-parts":[[2017]]},"references-count":0,"journal-issue":{"issue":"6","published-print":{"date-parts":[[2017]]}},"alternative-id":["Arch-9275"],"URL":"https:\/\/doi.org\/10.24247\/ijasrdec201724","ISSN":["2250-0057"],"issn-type":[{"type":"print","value":"2250-0057"}],"published":{"date-parts":[[2017]]}}],"items-per-page":1,"query":{"start-index":0,"search-terms":null}}}'
       headers:
         Access-Control-Allow-Headers:
           - X-Requested-With, Accept, Accept-Encoding, Accept-Charset, Accept-Language,
@@ -204,7 +204,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:20 GMT
+          - Tue, 05 Aug 2025 22:25:02 GMT
         Server:
           - Jetty(9.4.40.v20210413)
         Vary:
@@ -213,9 +213,9 @@ interactions:
           - interest-cohort=()
         x-api-pool:
           - plus
-        x-ratelimit-interval:
+        x-rate-limit-interval:
           - 1s
-        x-ratelimit-limit:
+        x-rate-limit-limit:
           - "150"
       status:
         code: 200
@@ -1283,7 +1283,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -1293,7 +1293,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -1309,1691 +1309,1692 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA6R7Sa+rTLfe/P6KV9/UkUxfVd+M3vSFwcY4I7AxBowxXQF1lf8e2TtKFOmOksmR
-          9jbHQNVaT7dq/+d//PPPv7q8Lm7Tv/79z79e1Tj96799f3fPpuxf//7nv//HP//8889//v79v64s
-          2ry436t3+bv892H1vhfrv/79D/O/f/N/Lvr3P/8yDw81WFjmBTZLkznwkaInNsbcirm6RhWkAnMk
-          j+fH19ZbDAqYBpOIg88x1FYJ2RVC9FSQgiVbvu1lNYKvilWIwtGztt1fTgNt1UfYdD1X44eHkkLY
-          Tzk+DEyrdQb1RsgONJnZPVEGLnFvOnzv6XleTFBTAsdnAMpGYrBndyd3fWXBDlrbzSEZLT2wCW0+
-          Qs27SwFc5k8+DfZYwWpIbhinu8uwxO4cQDc/AWKps6ituxWb0OhCjM+Xjg7U3ZUSuo1SPDce4rRN
-          KjcJzeChY3O8V+5qvL0QQCuMyONdDPm4kbiA3Wd7BWJWzcMSSEcVfddvbgNVzVkaLhzw3usJJ/5u
-          7y7vhdkQfZ1nYtjWni5C0HOwPkY2UW+6pzGHaZAgY2k6tj5oF9MRhjtEl8klVzs06VReLRkxaLcQ
-          +6jFwxg2lQrBp39gPBwFjd7AViGeEc7kZn48MCfuzZQsqjnYSitzYCH3lpA35jxR2iFx6avqCwC5
-          QcBm32k1/x4fMnyU5zM+YGWtaWTZIUwFx8a+9Cxcji2ogyJtOeFj3GCwNr7fwLwej1ilbuRuviE3
-          sD1NztwG+wvdlgfO4GnUDRJya56z0zMPYfOSc/Jg3nXOfw5Zj9ZBJ9hxrztAr8wRovLYFSR3itrt
-          92oSwibzdzMyGa5eHnzVI1UdBXxTudJl68d7g0w2ZthwbKHejmKgA/TeK8Q009fAOEY0I3bwEnL6
-          7uf0jiILHtW7S7JnAvOtHGsLPdIiJ97O1/NNPe9DmDwUh8Sl5ObEeV5PkLvKEJtFfMzp0a90xE9N
-          TrJ7jYe5HRJTeurGm5gvFGjkcBU2xHpNjS9r/Ix5dMtluD98FmLZl7u2+PAlI+SnK/FzNOfLK9lU
-          1Dg3F1+xYbiMOGwZ8sJHTrApOC63OscARXNSz1v0CsHGKHUrXjM5IvLs2NoE+G4EzafbkaubW/lC
-          3puEmLbniQ44nq5R6ThgKEeHhNZSxb9+RJd355P7eS+7k3SbLGDY3Q1bonUAiy0cQ0QY/YDvGWTr
-          pRH7DMbk+iDK/p3VXHXqLMBggLAjuOqwxNcyRVNtYuwyj1u8vpy1/bu/KYlRzZrNEiJVOwpE3t9C
-          wFi5PSNZz5JA+v5/Zt8dT0hYxRk7jVYPnHe4tVCbm5DIqyzHzBCnEIb7lcGnU6GC9fAcbtDJmD1O
-          jVzOWV+WOPScdW4WsWFojHFGDgzfEsDOFw+W8iXfUC9t95kb76rL8CHHIWeqzthOmzLn04uTgXfT
-          mdj76GO95GezgjfHQ1jpLnrNJbZsonZ/vOJ76DzpRk/GAgOxvc6bQO851w6JDqxbdce6hxKNxyqc
-          4cV48tj1YmXYEmHl4B2lmJwZ4+UuJXPn4FDODtG6ptOo1Bw49L7yH4IPSuY2jcUz8HCAu7lOMj7e
-          jte+BC8rXLHekBVssyr06DSaBj4bZ+oubPviYM/N0ryOzpqzD7BkiBfIEHDD5sXcOa8i5DN4wnoe
-          JvUo7G4tuJDaxDFJAFhe3LmFgiQz+Cakpss6YhQgM7NS4ux32OX6MAiBHaQtSTm7rCfltEvgLTrw
-          gXAUq4F6gy/APZJv+NQ3obsqL72CgyBrOBNctd7mrU/BFz+JkRQOmGwX3iTFMVjiU+K42zOTIVKP
-          4kAOIRdqmx6oBdIwvONEq+Z8y/s6Q+oUH4icayrljvYtAcXbjoi+n95gCS5lj0CUtOTHTxsUnj1y
-          PreNxFPq11v1Vk/A6bUNO0ZX01kK0hnC26EL+qLp4mWfXSw4Pd5DINlROVCZMCpqrW0izjTM9eKw
-          egE9H+v4cJpKlxWO/AYtqjjkoqc+XWK39aDiHFjsAk/LaXu3OyjkVwUXVIopHznnCDgZt5/3iVcP
-          mymRDJqmoGLHvasauzvWFarLAv7eD5AQ73TxKRYxsY5iVf/wHs5+V+BrpHv54qq+AOId85ifr/6e
-          L6EidpAmFsC6EYv5qHPHSlqewkRi6g9gO16rCuGbUs3iS62GbQChjAp9tompNLLGtTkWoNKRLqD7
-          YXbpSw1M+OOPwxtx2oKWwALhsCsI3odv98vvJYRWFJFAFyqX2fafHn6qXfftJznnVsmBsDluRrCX
-          bs96Ed1ERufn+sL+aXfI2UulSIgs7yZggbAO4z57WMAyvAYXj48OFsC28NdfX7xY6KqqJwsxlXwi
-          lyUX6p8eQVOt4wBpygcs0W5c4Bdf8OPGvmsGt5yM/viwDx/5uibHHmor8yaFLlTafAoZAV3I08S3
-          8npxOcX8SLDKwY1kAebcJb52KfrqCXxuVndg0E5OwPR4DSRfJyvm2/2mottCemKRjzdwpoM68GKr
-          C8H2rOSceuajv/q7Ohmul32qOeiHZ3kCX/G3HhMUKnv81VNd3JX6qEN7CRccWuKWb5ZmcT+9hE8C
-          v4AlbCoZAWxijCn060XMR08Sb3cUVDWf1qtX1wG6RuMVm7z5ApvpoB6wcncg3o1k+R9/z6ddSu7b
-          vMZbrV93aN32n7/6Wrn0mCIt9hgivx603prdqqOQRilxruOjXpWXV8F397jNPM0eNZXvVwbuL/oy
-          M4sH4v6z9D388kEAg9u1Hr71juCjcXEO7atLkS1wsMPBGftW+dI69OChqF+WmGjf/uTeKRDgfIIp
-          Piy5MFDPrU7o1TsSsV9qVdMn2kJk31/XoN0iOZ4F7inAiTu9yMFzs2F7HEkD7eYUY/Vy5fLl2n3K
-          P/zV+oqLZ3TY6+B0vnjz60jZeIlHtYXJOj/nfaqoGr8T+wZSE1rYhmjOV2juHGDVzIUkXIBc+kRS
-          CPN6Pn7raR+PZlmVKEgSjfhSGriMET8tNMmZSDw6LflyfCgetMhGsS4nGiBn/GGga5kvbI29MdD2
-          rnTgIPIHoj1B424cdEZ4eIQ8Po7OGq+SV+xg9Q7FGZ6RmfP1YZYglj5nbJqpMbDshFsQadtp3jh2
-          BsshsBI4EXUgwbYM2pqcrQochT0/89vd0ca8tkZx6N4asd57AuafPpvkvUiCk5XmizhIGUxQh4iW
-          R7LGBoWTwfNVGmak8E932w56Bje+tgK+2RF39O9KIfY3qych3b9j6hxdB2aulGNDVLN6uSxJIpUw
-          mIjdsThe3Oik/+mHJGSSnLEvUBIicZbntZwOlDrLPMIfH3jvdOcOgamUaO68eQZ04IbJmhsGpmYo
-          k7PrrPlndxxKeLyu+ryI762m8SvkfnxDbP2m5NSmoIL1MbSDbUiNfBUWy4Hnog7wobJ4QJpt6SGP
-          nx1OHqfeXQwtgz9/QxQqf/mgOyZo3x5rHN9MgxI22mYAoMCS0+qfvvx9ZuDv+X/r/wkFw4KeEnbY
-          ySYh3si+Xn74RwLm9Rn6x95IYBavZ6LdtTymez1U4fgRPHxXeMXln0Oqw13wmIKtnq16tsswhQjg
-          C3HOYgTocwhNGPT2lbh1dqj5yscj/OrlAIZWp23Hg5WA3/O6RbWj4+ztT8BT+I7YgQ/zVdvjCPD9
-          QSB4MpOBZto7hY3Arhi3J3XY0FkM4Wdt8j89sEGJ7eAX3wIRB6y2BFemg6GSYpKsx3dNPdEy4fKU
-          JuI/b3bN5SOfwdluwcyfe3ag8lSqiKjbLljAR3fZWwxucBGDFtt5/c5Xu0g6cHiwfIDiSADDp+U8
-          8dc/5uEM3O2x9xP4Zq5isD3FPu9HwQ2gmPZvclM52WWVl1eCiyyeZrRzJUpeoRyhKd/OAXXQBLa0
-          Vhlp2vkFTr/4zH79KITz8qen8v67P4gzzy356gGwig8RQmPfWthvQ6KtlXZVoXETNpL32lCvgO9m
-          dBdPI7HU10ej2eo7cLvtBGLBrakXFUgJzEk/BcIH7fKxQztHyjdRx9Zyj3K+4K4j1C7hhWjXazEs
-          TzU0JWEeLr/+yNcL05hgPV04crA9Gi87V1ng4qbydz9yOo1NqcNpHVX85b9hDdBrB6/RfCXmEgZ0
-          ZadD+6cvk/FeaWTdhSP81iu20oEblo3kBXjwQ4DNr/+cff2wQF6TPeIIblWv76o7wXsyJESNbiWl
-          hOUTeGdjexa5u5RvT1va4D66+dhZepa+zTHaAaMJxmCnKJG2XdX5Bh/+Z4ctzi6H1bHLAlCHFnM5
-          nFY63vA0A6ks7uRQy/KwLnYYoccYr8Qt2solRatVCBnxQJR90Q+LfWhHKN4eiBiuc8yn7FPdIIAS
-          i/0mKsGaXG4VGrOH9+Pfmve7VYDXZ+Jh3z/1wxgFQJWYtuPxaQs8Sg2sZegq+kYgCAfiroudhtBM
-          dPbrH8p4ndZPASVD8okruc6wklHuJLkXVWxlJQOWGDx28McH/g8Pj9eqhPopXohp+Y3bXZbkBIVH
-          0WCXLc/aKlT+Dq4PEgWwvj3z5bIUJyBfupEcTYOh45fvgRNIFXbOiwq2+ztpJT6AHb7qt2e+odQI
-          0fV58mbEVxylCmQz1L0clRhK1ca/94Hd8/Mg+BN/wJb3Qwq/fvSrH6J8+eUB4jvKsG44n/zbjzL6
-          9Z/8HFu6vLh7C83PxSXOR3iCdWWRB+eUzsR8AxhvslgWqOXeMT58ttPAne+7CGgvwSEqxwZ09bSA
-          g9zdfGCV4eWacTehgOuWfX71pS1c8ugh4lMFyyzzopt3uDXwy7/E5u5SvDzEXQ9I4GT4h4fz9u4W
-          eK9k+sVHEC+HqRYQjYwUY+m11kteyzPcr6ZMPHzfYqr4bgfSkgnxYwMM3c53LgQbaAYcleRZU8Ex
-          vnkOi4nqQg9wqMtvcOUKj1jMSc35EHMmMEKY4Iff77Q/f7gTCvRbT7q61ruD3F1/YDeBRr5aOyDD
-          Va5KLJfMzl0eItfBb/1hX7yw+Wrw2Q3w2cmYBwWiens8ngs02WM3N5e7DdhfvrTuzgaxydDm84BG
-          WbpPexygr/+e9bLp//rH4PVRW95zt4H9ZzGDiOfe9fxZ+g5ijyH48OXfDeycG0gdI8GWRo50qYqZ
-          g2NREmzIT7/+8sEJ/PKHb//EyxppAiouu8dM4mSMl/4FLHjq5IpczfpRf/16g8jyavBPP0yVdlT/
-          /KK/zceYPdT7GwhQToLdMdHzRRLo9uf33Hl+/p6nAIcVBti9NctAijObQG1uQ3yjjZwvRZbrMCfd
-          NM/ZjsRff6HDYqvIj69A9zJrBj6uWo4PTNQNG961Atox12zmtjyMF8wsHhQJLHEWAz3frovuwV3Z
-          6MQhX79YwvUEvv0RbL96XKvIlODlYRDFb5WBp91zBE2Gd3M8Xk5gUT25hbu7NGGfbaya0xEMoN0k
-          cdB+8XN7p1RCgdhc/bUrmXpTzI+AYpI/sObuzGFkOUaFjLZ/4gMXJmDTA6f4+b+ZJ0UF2I+3JjAe
-          Ghs/jicJ0LKsZthzoxQwXfukW/hZRqQ+k4HI1qLGa7SNJ8DkxyzgP5+m3jpOu8FwgAXRfGpo62fH
-          m+Cr77C1l/Khu+zCEn39GPY1xQYbfrgJ3PcXk2jRktUbmJgQnew9wZp8clzuGpY9chLvRtQvn06l
-          tG+g4AUB1nhZqDdjyFS4O8IjvnH1oWa80NKhqs7C3LyWvl7hxTvB4gIfM6s+ACDv8SLDsb1+yJU/
-          ay7VWLn5y8u++Fsv/Ys6aNgLEF9i9qJNoeBb8NPzz1n67hf7ZIsNRlV6IBncd9pXf0SgsOpHsO0e
-          w/B6V2WCrLBRcDHpkE7iQ9zBb55K5HPQDPTaaAsKIr4kTgz0mGrNlqIyiZV5bwq9NuuICZBtxmXA
-          ccHdnTnozODr34me3I7uOlF0g+bzFmHzdtyDb14mwMtG71h2BTFff/5+d+Ud4oa4yOmJmT2gsQP6
-          5XfuQm9i+Od3QLrWdD3U+wLe3cokykOs3O1V3EZJJ4NHTC5A2iyF4/jT01i2xChevnkk1PDuHix+
-          86Zrbkq6mN/2lwDd1pM29tbQgpHVblg9551Lj/3aweOkZUTdFDvmb7uggsPkt8SlQzLMzTE+QQ0F
-          GvaDgub0QT3hV9/EgCMfj1+8hToxPaKKg+ny9QXpQOdSE7s46mp2OIQVvDRuTzRvG2MyBMCT9B2/
-          x7olywNnSu8Mfv0lud9WRpuP176CaWCJOAt1pp59HS9wZ7YMMe9ROrBfPwrOB/9MlCcfDjRyziE8
-          Ftc3sZfHqG2P47sFP72SHpUrHetbxoFhwi3G0u05rPaYMvB3f7e6BuCXd4g5UJ3gkzZyzK2SCqFy
-          EiKi70aiffN2Gf78hhExgbs0V1OGej0v8xJfW214hVYk/fJJ+5sPstYOqOBXH+yaIrDWZdlA95EW
-          xB3Z/TD98ukvHk5SDJp8Pb/7GZyFspjhJEQaT0y/ge5OPmL7W//cNew6qJpLTjKNvuL111/bDQrE
-          2iHTpek1SeH9A0biwL2lLR3iHPjB45FovJzWq3yAEH71e7De1C2m4Ak46N06HYdf/lydFHGSZQTN
-          3PKmAfjIUqKfHyfeGbX5NIhJBHfH3TGA359pFqMW7vuzidXdwx02V8ot2Ne3Bj++/Lb4Y8EJv3zI
-          uDzqgYSlHkG7EWOceccDpdO1bOHoRtcA1NlhWJnvJCU8dS/s48eoEcXXerTGC0dOn5rPh+F48H7z
-          D2xEzOwSuHdT6avfcfCOlXpGqR/Cxc1kbGvrMNDKATM8v3iLfPP6YYh5oALOvLREFSjKb1auzOiL
-          x9gTz2D45lkl/OZDBNdvEBPj+MigsywjDkBe1t/8qwe9cV2xMd2eLsFyN8IvH5HLGyUuZV4n4U9P
-          4+bRxAus9Q32fsUT9eu31u98Br56SyL63WC1UQSHDSaoR1jdl02+WA4t0fvsm795Qj27zicFX3ye
-          S5d88vWz25vgDFuFmDxmwbqfqQ5lGdBg9op8oCJfMcD1LAvn6arRbXqzKlpQaOODr9X5eBK1HTz5
-          t4hoY6ADdtS1Ec30vMMqz72H7XWzZhhz0xXb5/njbpfEvoF0f1OxiXm5Zr/1Aa9MU+HHQawA3V2D
-          AGoPN8R+HKWUztzigIcRGCQ4y4rLoqc4g9/84uv33YVfcw/Y9qLhUA3n+Jf/Sz5M6IxGMdbW1ggX
-          9M0fyEF+Pd0lzUsLdhNjkQyHAd2Yo7jARL6m5PDlv1/eB7ajfAtooZ/c5XEQBOltqjFWR8/95hHy
-          DL7vM/OVdQFDcE6EX7/NTwPqgKXqrv+9H5ELjgwTNHcWuIrYCNZ6nOkaps8GFN5bxea5y+mCdvIJ
-          /tbr62frbXp9ZMAXdo79H14NIJXBr990DJG7iOCw/Ph73g9HwV0NPrqhS4r9WVK3sV5k7R389Bu+
-          m0NbLyLAG8wGI8COcNqG8aytGbznRxE783jIZ+x9HKgD9TOz33pes2FXwDj0JoLLjokXvdVO0IbY
-          +9ZbVw/PadTh+xp9Ar507GHrOPcGT51aESfUT/VYFS0H/ddLIsb1pLpMF3gprPefGdu0iF0i9J8b
-          /PLRj/8p/4ifCez3uRxMOtfFI3tJC2ibxzIQpw9D6e4tZ0gtzBPBXnnO++rtJPCT9/OfPl8ythKQ
-          EFcW/vqzmmnjl4cW0WtnFo6X+Ju/p2hEToXlppDoYLWbBcvzrGAlWPp8abXHX/8G16yaazKsQ/Pj
-          J5zZ+VJv+bhPf/6YKJkf05WZ5RR99S/R7V0AusxYLcj3hhC8n24Vz0NGu7/8Th2ebLzhh3aSvnk3
-          Nr7zE+oNhgQd/KiIUcFJ47br3EK723Hz3n/O2npfXRmG6WwRx7nZMf3iNXiQ5YSLNurA/JunaH01
-          EXX/OeX0jJ8MemldgC/f+4/nJG5h/y4PODAnOZ7f5nWG0UXuiQcukzsdz6cFJXPh/OX920cWBJjt
-          zg02BaDHq4WVCKrLsyCqvw4xmUTdgTIXPrBVXHTAZsPuBvCjMrD6zac4GTDF737Bsz9HtLsj7YaO
-          R8HF5+t9c7fmGCdIGT2XPDr76a7H2FrgQTYNrKb+TaNhqYcwurz6AF0B/Obn8w4qdsniXz+Oyizs
-          JDObUqyxqpOzzT1eEH6UBnGsjKdkeN4zkAL0IG5snHNq3rwF/ut3KuB//Lf/hxMF7H99oqC3RhXL
-          laBRlhIvgmduX2DrLuoav0mkAA/vfSeO0EOwPHbbAsMt281NIFZ0BdpnRvycnkjWWXLOG3e5Rc/h
-          fiLOHfg1/4YbA7H4pnOVh4rGEP7KgeGlBkSjhp6v79MJokkBX8bbLoC7jnMIDW6OgtjWaUw26X2D
-          LvMJcR6t87AdiqMqcpfnSHQjH8AmnaUQUv/EEm3ZVndNGMCBziogwZ/85s5Hq+lgJ+giiQ5mAKid
-          lxmaV/eND5KkABr3HQeFQj6STBYHbRTeTgZ9ObXx+cS/c+JISwSOcebMrD1qw0x9d4MdDT5EXuRD
-          To/KPoKPqfmQwr/rLtOdpxTY273DZ+YcuexyaSu4UwNCgsdJpXyZmxmcy+ZEInYfxnyOkg7O+H3A
-          B1DttW0+WRJkR28k10j+DOtOeC8onKqJOPtxGIgSiRKcTqGFb21SUBpJhomqj0OIKvctIMd2NGEv
-          hU+CXfecM065Sije+BORHxxDFyhqHgz5Y4C9A9Ji3iZKCgPSqYE0aAsYAjHUURbwBLtMNYJ16bYE
-          ecNhj9UU+XRbs4hBahcUwZaiCdD2c2Tg4Hs44O5HStvoFFco0tQD8XK1zflHunro8/mkJPo+H3Oo
-          MhMdO34K3hOg+ep56AbbKg0D8fhUNEZ+ZTe4L0oJH+5RWtOjwkfIu90DHD0Tdlj0m+VAsz3hgK58
-          GVNFaXV4DfoH0U9cNTBdHAZoUbaFHB6vAfzqC7ZVFpLUDz859RU1hDkf8sSRikM9cYpvwWOcOiT7
-          /ryEpwsD3SAosXohn2H1CiVE6yWySfDJLvm8RuUoxoN2Jn4svPONObQcYsPDig8f70K54fzUkf+E
-          HckFT9RmrEYjKk6eR5Ruj9yxMMcdTI6Tiy/H9Jrzv3osdvWeOC29A/Zu3Cr4gS+T3NMjybdTwLVQ
-          8vkCOyHca1swrx3kaNeRiz1PLrlcIAefmaCQczubGlONwQYVdTK+/dXEXd2ceuTIT5Ocx+FFt+Hj
-          6PCNzYLIhd3lH/bJeiivsje2I/lT0+nGb/B2t2Jy4YWIsq5ThtA3oitWZ2YcaPzEFuSkR4aNE3+I
-          efe5JejtrE9i3EOsMZEOGji/ioaYyqkceGWfcmiUjJn42HYBb+1eHcyXm4HDt5IO3EEcLfC+Txhj
-          xVA0Xq4KCwx+gPFVNfv82+8yDG7jCR+jvZ0z5Qty6GNaF3zi3nLNGn7XQxu9Llj3mXjgwSHaoLMd
-          XljbhZbLma3mIH97gnl7fzrACT24wcPzg4ldfZp4GSbXgbdWYLBj6TOl4b5SEahPAjYnaa6XYdIs
-          GH90itXxec35lPdGSdQfSgBjvqoZs9cDJERCNUPBR2C7+EUGf/td+K0AaFiRREqOxCXuKJT18vR2
-          EpzgcJ1zzde0dX3KHkwOsJ8xSfN6O5VqJkbyZyQm5ygD5dNLgFQ5LfHlatd03Swn+l2P5fg+uVzs
-          fTJoq+yAtZWXY2bfqCoqds/9LJxy36X0ymcwV/IchynQY84e7yZk906GAzM/u4tBng66G89X0Nzn
-          CPBeJnvIz5wPOSiFpJFF9gq4r68uMVMN04k/zRYUbEeZIcfpYCqOWgZK9i3gwEtPtPNdYYRZwBLi
-          hgWI3z3IejgrdT8v+SS7zNvelfD1YBp8uJwFuskXfkbRQbgEb1A93OXC6RF6TYJK7mx8pIwKBQ8e
-          HweP+PAsg6180g0N2FmJsuJFo3PJVOj3fI/niun8zJQOng6kwX7Qa/l6S7oTdCZ8CngGVfG22liG
-          72GbZq6MK7Dyp9aB1azuiVuLR7CymWsC0+5kHOmn2GUCM95gvessElNDj5mPlHlwH97v2Liprrb5
-          tbqgT7Na+DI8XI19makATbuX5/2lubtrrjw2SCKLYs+5uvWm8ocWAbcUyQWx0cCKoyZA+RaeycHN
-          r8M0VWWKJGsfYCuZ5ZzlXiKES/4MiX98nVxiEzuVfvxkxv0bLGqqpSgU/YU84rzLt+jm92B67B1s
-          Xk4ndzNYkkKETgU5S9E6TI/bFiJ/kU18ensfsPTXygLLg/HxkVkOw/L5BCkw3aYixqtUXVbXA/jH
-          N/JwHAZ2v7A9EkzaEocLmHry0dUBXZ712H4ZY03nzTqhTrf32HwHcc4e5GOLpJD6QZux7rDZR2UH
-          9/vS/K2Xy+WdIEDebJMZeAzWeAuLJdLquiC6y1O6Ors1gX2OeayHIQBrMikjanO7wdkteGr8p3hA
-          SOeTguP2VeQLn8cV+O4/sfZ6nm86jTx0iOXiV/85x1+yApogeJBrR1I6Tg9GR+ndlHFsdq7LbKje
-          QL47fIj15rSalw+uCnVpfyVuErXuyjw6GZpFGZLrh7Hc9ZA3M5r3vk7OrbWPh4cRfk/o4JjYZXJ0
-          F62WIsiGxjrDaKPxq9TgCSaanuGoyhuwiJsYwatxj/B5r/juH/+uuHCI2wKZbqr8OcHmZQrYOxot
-          papIpd/+zkKWjsPW9X0EovTRz2CMrmDbB8iCxsE8k+JQ6jGnyp8EWVuYkvjHl/W19RC/njRy3bWX
-          epW1K4R+RzMSXOo0p9m7LJAGLrdgjrO6XpbWlNH+fJ7w4dx5MfUaXIEdes5YPi5lPQrquUVFQx9Y
-          dZbc/e0vYBl5IIaOOnf5rie0Mz0hJnnnGn+WLzokdIqwRhsPLFp4DBFvNglJ+2igKy9GEJmezWJP
-          N4WcdvxZggebD4mvLo27nn2RgcNumYOX8Hq5S3ZNIbSfwhnrY//UqFqUC7TOrR1cLa3XVvdFM3gP
-          5g3rzMmPaeSmEL6O9DDv95EENoYxNuR7hkvugQ4HSgc3AJiwEZYxPA/kEqojWh6cTzDBJuB3/jhC
-          79zdiFsNPthe+/oGXs3MBtR7OWDh408KH04PgvMq2nRbaLhA2is60RuRq8fxPXvwu79YPjpXsG55
-          Hf3pBWsoMWCz7D7D9eoNOGDpwyXtqkForIxKjGtbgbVv5BaJUfUIBD77ntJQ+R0EbiUGK/DIMFeE
-          ZyBfR69ZNJRc2wJHZ2BqhCqJzW5wiYXXCk6m1wdzWR00vn+vPXzMNgzKbpfmyxl8Asheo9080rOf
-          U569CtIBXAmJk6qP6THWTGRvj25m7+RSbyLXn0A+VUnQXhcY//hKitJ7T8LLQcw38MlDeB6ifAbG
-          Z8g/71e5QEl1bvNq8Kb2KbuPBX96WUt320A+xQXCSCy4GR6Qlm8qjxvw1ecz3MJCW9iubwHLqAPB
-          JAV1Jz7CCOVZ22FjFT9gNAPRhJBCBSdLkgBi3K0Wsnsrw8eZHgC/v+YtlHIBElt9bZRiChyYWTLC
-          j9Uz6y2JaAL9bs1IJOTpQAuwu8GYd47E2NTd8DledAtG4o3D/mwFGte6UYooZ8l/9bswy20UdxDm
-          +PB4uXSK56cAg+ws4p9+3xrDukFDJ9IMLvutbjK8D4GZ+TGRG7b+1p8WQBE3y58+5ZOP1UH7YHXY
-          3bCYj/21t8AWvW3spWFfz0gWTRjGm0/cs5+A8c3dC/Gj8BlRC8nPKb3uU6g9W4jNAQVgq07jCItc
-          3WGvc2zALI4kSGN0XIgy0wP96gELGcyFJ17nfCh9tiiCzGlbifHTl82tr4AjnsKvvnYBm0zKDDrB
-          FImiNduwwDefQvjmDay4jDEs4FUFME6N9x++jOO7DeAi4Mdv/4YlHr6JaHi/B0xazoCGGRtAwVxb
-          cto/FkDvrRVCTVbvxNqGeNgAhBGMyHwmX7060Av8dLCF1oIvL01z+c8p7NGTb8OAAbtrPJ745oZu
-          DIzI2XFHbVWMtwXaal9jT6BNvcXgXkHjVgvk4ONm2Khx7VHKX5m//vn6kwIquRQRWV83bdKavQrb
-          2YiwEhcr2L76EkD9pOPj/qYO7PkZMfD5WI5EkT+Vu0i70oK6cUBzdTBn2lmzBWEX+fdAOO4bukr7
-          RoAfjUIs//pPpUklff3E3+ef8zPioLS/onk/zDje2naqoCe6d2JWMomni1wtsFikHMv702NYcz7M
-          4PWtrcF2jJWB/Qzb+MPfYPfVe/PnZZzAl/9JsviqxiVzUMAD2hwsjw8M6BgJBUSvqsTmXj3SRbrS
-          BGrjPGCdZ5t6iioqo6ByhUBs43CgyaPdwaYMOqIe4+ewfHwQgauYekS53eVhBdJhBz/5WyLGc+/U
-          63V4VsjPrA+Ju6udrxu3S+GF7AXi3Lgi/+LzDtCS5fFh62dAARVnpKF4N/diqNVspwszHOHNxg/0
-          mnP6xWcQa/4FPwQuAvOgnyOE9kc52LZCABt7lEskvYY7tu+SBeZXEybocrxYBPdKBVat2cvgYzqX
-          QHpfm4EWvnMC04cMxNm9b/n2OTaJ9PXDwfQScnclipBI0HU4bO+HltKXwJmwS8psZspFARQpiAEV
-          1O8kE1wmX7bxsQOIO1DsZnlGF2VVEiiwXPinpxY9am/o3GbuH5/y6Y7jkLedhgC0oKR0Ohv9Tx9i
-          5WA47uK7ywj7a3zG2gs9tfk1X3q4rb6ELYd5ussXT+DXrwf7g1oOCyr9CO6RPRBtX3gaCUdbhty8
-          XEhKl4r++A6KTzfEctYt4OtPBPDTl8ojXbV+sjL15zeCZVhO2nIi0glIdAZYfTtlvHx8GiHDBwfs
-          Ct5VW4rhuoHgwujEqQVfY7bxsoOTq0vYK8atnh+SOELOX91fP9eLci9PaPKVW/D+eDxd3vauApLP
-          FsSvQR4vjZuMsEicGRsD7LX1VmQ3GPWcgb2UUI3S3RjA5/A4EbPc3Jx9sYYM2D50iFdLybDuVaUC
-          cRPUuLBwqW3PzO6+uds9WLVmq9evvpbq04MEjPnBdD3k4wjfIuPiM7krA7/I3g3qwdHB/mNrhz8/
-          eurGBcuXgxhvm3ISIGueQ4Lj3Mq5kRF2kJY8j3/1MxTi56v/DuuMsqYaNid+OIC+LAYf7eeWz69R
-          LdCu012i3QOjXq807uCVPWBs54U8LCiNHegsrvbTf9/6Yjk0pGIUSC290wldHxLYP4sPNo4Jpusg
-          cgti+8jBwetY1Qs/MyVkpM4nZ+H10rYx1C3kNUc/mGFi1Ivtp9Lf/uPI12NmXRkJaBF0Z7q7QjAp
-          fb4TEUoKrDDLe1iYG50hCvswKJLzKV9nu7+BqH1aOLcMWn/rHUJy0bSZkccl/vJ1A6ZTZAVcqmFA
-          WfEQgW/ehB/3JwRL14oemhRxInp4eLsfNtN0+JokFcsG94ynRvpAMFWS+OMbSm7ZNQUkcih2HvZa
-          E5soGbymuTxfVdPJ+aVZI2ih9PB9nyZfkL45kHHVEFvl2YuZZxK2MHWsKNjJBTt04nMq4P1TPIjV
-          rkw8JdZJQEmiWtg1HBWwjyWPfvnLF09SsNWdP8PtmoN5/1wx4MdoKdBpexyxentL7irFBoQXozOw
-          dWNO9I3TToaXQ/sOxsp33F+eBO/pPf6ecGvrJRjkBh10JGO95xKXKtEqAPmSGV+86ugqppWAaJ7F
-          waj0i7uCxeqhG3gliSeroeQsNAKShJeLU3Gqho12hQRpGDXz7pPx8ezcFgYxWCpJMNsRpQy3mRBc
-          fIgtaDuUu8GjADmhFHDAnrh6RlMow6++C6rv+0xhsQthsPgJOY26lJNVyEL48F73YI9e89dPH9Qf
-          3uL048N6+9iSB5m5Ps7SoIWUaW5VhQQur4kShBMdyxfDgOtbWWeUHnG8iUdUwdJKEqzL/oHOW2vP
-          ML3rMrmstx7Q6ex3f+vfffmfakbFoC2RTBLcy9RdHCaTIDRZEattIw7kp1eT9ImCfQfjYW2nTACB
-          vbRYp691oK3+UqHep+bveYf5tR8KiPT6RPQwzAElxieFL3ppyY+PqaP2N8kudi3WdldI16WTElje
-          g5boQQzd7b2dMiA/rZbgvaDUrF8WN3CpktNPj9Kl4dke7p+3D9G5d1nTSioiSDlHnjV8aYfpkrMS
-          dFv1geXS7Nx1J5AF4MYKSRxEfLw0JyGDHCYTMUNhdPtLUQUw+6h2IC77AaxRBWT4zSvIb/1HLTz+
-          L7zp96M7MPC+6H94pjl2VbPX4VMC61AORC9gQlfs3iPQOs0Rn1bBcPnt0gdSplkTOVX3OZ4Lp1D/
-          8q80lNh6bE5CCl3E3En2+/6uXT3IXeoxuMhh4pJrYS7wy2dzq9zP8TpdgxGqNXgSm66H4d0nEwfa
-          +RBh5XYvB/pOTRm+I9XH+olT6+1uyRyUW5Wbma9f2jL7qQLBNeuAXBqkrSvz9CDw7zk5hOICJt6D
-          Mnw2XfXtR0LH3PIdcMw8iDF7BvHKskkAv/2DPVvS3IWfYQkUVKxEO57MuFMfqAQLPeOZ26tHQCel
-          6X/PG0Bb0jTGUioZ5so1x988AtACcDfArHeZFOOtjWlnfRb41ZN/fnLIP2wH2Ojm4QAsu/rnz6Fy
-          UFXsGe6hpgN6M9A8XzRsrnmrjdKudNBtmlqsJk+j5s1vBGGf6TW4jBPOGfFzaeFiDjIx9srkzs3h
-          3MGeTY/4ju0BTBdhaRG/Jhp2YrzlkyXBBpxipsRRb6zaQvYuBOXda4muJ7dhtalaQO94fuBwDdyY
-          SgmvwualCwF4OEncTXq3g9/8I4Cj+MqnghUC+Msvlbtxp9T9NDdAL4cEe+98dgdyVTPwy2MMOfAG
-          +stj5nv/wofkwAA6HJQWwacck2smXMH61bOg0coruV8qOowONFSUZEKKb1+/R833QQAHfS8T6x6j
-          mM4Q97Ba4IxlQbjU29NCHEhksg/W42rk9GYbDVxgefvmc02+OU0P4WcRdPLlz5pFx9cG10IH5Bo8
-          Io0Epzr88+/YCFg6mXXBSWNeIGyU7h4Myj5k0POecwFXxup3vuLr4Lm7qjNdeTlnU5WHEPnjkxyu
-          dg3GL7795aGccpLrtQMFBNdjE+AHl/X13LavCq17/YAjJtLo9hn5Akp0BLPAHnLAze+PAA+vnYd/
-          87VVrZ4c7GuN++W5OZs7YgG/eWwgfViPbiq5F3DGrwPODacCmzF0Jvwo0eEvj98k9RjBy05JZoY5
-          TTlVZl8GD/7ikAO49vnCnO0Gaui4wz+84tk7dGDBE/+b3zJ0ER9pCL/zmZmXpCfYHp9J4lMwrqTg
-          3zZYoOh6QJbtjFh95H7/go/hpK3s3thut6s7AQnvwPARCuxs092l4eEWgBy6h4AOUjmM9lmTpR9e
-          qt95Di33IIDRsReCt/Ce6/k8rBECxe0doFh4x1//AGFXRiDY7qMdE0etCpgE6+37ffKwnUonA+pk
-          zsE3X4hXVzEbmC5vlRyGHR3Wrx6E2zzqWBEftbt4M4KwbPeHeUlBky+Cem6k82KN+BSGgFKxczhw
-          DhoPe7lq5vzPX15ZA+MDg6p8dXZiAjPxdZsFm6zg2z8ylA7q+Jsn5JVgmh1c3p1HirXv4++8TgBY
-          fNGZGYCRj6HodTBa4s8sSnCIp7DgIkgs9CJBMFgu/5tnRQGxZxCtQT3yeVyi3umbb39HMbfTggjs
-          uUQMhHQo3bkzUAFZXN2xK+fvmH40nYNHRSGzKs0DmNeom2HdMg6OT5jX1vpOQvC9f0A5mdXWqY/K
-          X348X05ZFy+SGZ1QNyA0M1++X3mPkeFctqcASDsGbJm7CwAbFd58DMdEI9Ntv/x9/6OsXbDcg4cO
-          q0cWkK+fpASdCwYSKDzI2bNcMFmPZyH9/PQh9hrtg91zJHweEo/9F2ljKigWB7V+sbGbK8qwrMln
-          g0ouRMT+MJ3GT+JWQMMXD4HULfd6FdaOgaWxB8SU4uTrR0AIPXnxcJZOn2GKdoHzlz+eH3WVL2ZS
-          e6iwDiEO2h3ROll+9eKxY6dZPN1W2us0C2A9zBlRRzUEr2+/oySRLXwR3kE9f+cnIjUrDTtTosfM
-          d34EbJUfiDXOjDbkQV/98knsf/PEdTd3BfBQkWHz/D4M/JJfWngtvT0OHqeKLr40Mn/zkOsF2S4f
-          joosZImZE/9SxcO2KTcBLmH+xHYIpHgLA6WF2SRcZ/6bd2zVa6ngW4OEnNR9PazoepFAi29PrD6T
-          88AX4jNEaUe1YLKmGozUjmb4zetm6L5xvfiP6/j/c6KA+69PFOxm+03kfn3HKz3rHqy7lgTiWRLq
-          1aq9Ai6pOxG1vb3dsQyaHUxep4BYWXXWNjLhHsrd6TDvx/sGtgGTAp6luCG4qYKa7UC6AXA75wSr
-          INT4I+OZMKblnlj+e9KWQ7E1CHFrg21zbwEusz8c3EWOFIBrvg6jfa97UF09Hz+O3B5sn33VA/m0
-          D4nPbdIw3nelAxVrWGaBxFw+3RmXg2yYcQHKlN6dElvPYB48M3LE6ROMTLhzoMi+KMYoJmBtlouK
-          /AGciHvJbcpUo53AmJQk4C68DSbFBAFosfYiwSLLNVcctJ10uwMYfHSdB8vLfrboUesaCSrsDMxw
-          0TsY816FI+iP+QZrb4Z3OJhEj4+VtigxUmF8cwVizhOiC6GRDspqPeKbK3iUMrezBxsNGiTjSJMv
-          mpKlSNhhQuyufecfuV1DxOXFDeMDoN/r7wH0bmxB7FA6gbWIqIq8ri6J4lYXwAb784zm5+E6i/T5
-          rte+lwWYRcwNO0Sc6eLlNoSRtR7IYZa5/Pc5mtg0I3qjagPbD3GPrtrS4ohGujaBYBVQZ99W4qtn
-          fthW+5oANZ9YLF8NWC8RnwdwTPYw2OgFu8yk9zJkrfRCkt316TLWDCrE78onvrrSc6DRtp4Q62Xt
-          94TBEL8S0U0hGEYN2/CiUw4+1x3ai36A1eKpAn6dPiEcg/2daKCX6vE53XQgL6eWJIbyyrm4qBJ0
-          FA80UB/tcZjcRyij3PEe5ChKVjx7W6iDg7ZTiI+m3N2iXZdCYA0CceujpvGVxViwsWIZa0tPcqrB
-          5YZc21JIItY0J6JV9zCmQTzv7PqUM1BMTaSpS4ijx9OseZy8R9i890eiZPGx3ty4hCjiGZko0V3I
-          aVoWCdQ/TIOTrdsAS1+Gg3TV2khy4T+AW8UwQepg3smR37sDfb2JCXsQyTh5bWfK2nInAcfURZKP
-          hzKfVdc34bwsE0kMM4wXle4kcFYjGwergeNlNJgNymF/JoFdM/EyndkRWdtoEfdUCnT1HzGDysdh
-          wraqL9r6QluBTm1zIGmTJhon1ZYAn8l9/auHjV4THV7sSxas5nuvbWgpHaTc5De543Bfb30QCsi1
-          HYXgNXUHjjhlhHYv5h6AS8kA+lsf4xXwWEniKmeOH34B1kU25ykyDcr+T9KuZFtZGAk/kAuZJMmS
-          GWQKAiLuABEFUcYAefo+3L+XveulxztgqPqmClEI5xjaa4HxHd4aOhhO/wYo2zjsZ3Gbz5ZxnlEP
-          Agmn3tdTGUNySljGiod9IDmAr4UDhHv94sdDZodl0JcDOs/TRpR7azvMtVka+OqYH/bARKPF6xYb
-          9fTr+PR58ujmNLCHmfpkZq7k2WjOfwFEb/3d+y+qhWCtg6sNl1t19RGzRSpzfgy9aMZRTxyrfw4r
-          WPw35G9OTqSLpOQrjdYC1L9QxcawoXyMWycGq9hkxD9IfD7u/Q9j9QXng1H/AHWTyoB3szzP7LmT
-          AeviUIQ4G3Lsz7YPpuSkZmg9gOvMbtdkYF6FlAIte0IfbZMUcb8FluBhHdoZ3ubWIahZDbS/xpfW
-          Z+lSPNcd75rfjFr+60yqcZTAq+N+PpIC3eEg9X5gU5oTyeMqpcuOvyjbfHlGaALqWsmDJYZWmc+x
-          KPzUZWQ6BYLmJmLFYwFYljo+wAmRhpgufjvEv95tWKZXi9ixV+wTmpiBlbLEpJRqOV9Pr/iNGIGo
-          /tHkfrTzZ1uBpHl6Pv/U25pfp1cAH0uXkPOOX3v/Gsit/ISUQHIoR++lAR3IvQj+44cACwkc6qjA
-          5mG18u15cwVoXzeP6Ap1ow7PHgNqt5XxeceflXxrBd3uykx8o/7Rgc+NEM5ybs11ePgAfps+Bip0
-          UBJv3GqwjulJgjUMQpwLilYzo2lWiMuLAqfy+KPrtVMK6Fyf+XzyxcmZHm/BFrmcWljnHGkYC7wY
-          6GvcO3KL7TZfAx8voOUMiXgV7GuqumuBzgcjI9Y3Mp1V08cAzl4qYN8p5WEu7m0jho318Hn/fhwo
-          TsgI8dz15C4fP87yHuUE8diGWGFCS6Vv85Wi51cLyTMfL9HWRHGA0Fze/V8b1fn2vIo2PJfZGeej
-          KeWc4ugaqn+BSrwfPAEar30KDUk4+S2sS8D+3a/jCft+lUjtzt9iA6t3+cH4/W4prW/vGFk4jnDa
-          RTPgQSspaJxvyowcR8/5Ywwg3PEX35lC+uOXH8p8xsXm3bZVpkJ3F6XHJSG5/RkpeToPDm5+2xF8
-          WWOwbafCguA4hzhelaTmn86VQdv7y2FtNB+Ath+5QVZ+lUm2npJhnW+99tevOJp7c2AvoAshyvU3
-          ybpDGO31EqPWUZM/PMhpOUQCeiWOTnyv7CKmLHsNFlPYEKd31YHXDqYEnYVbifPM7w5baGcFLaSi
-          JFjXDdBRzkR4fcGMXN1ZyGn0hi6Y0NT43GNr6037dQXc8Xd/pguB5fsLZvRVtBlHS08ieida9fea
-          OGrvRtzxGQQoL90Zq+zHchZagRBWXMnseHMc1knYRGTETYHtlRmGkUfWCGvdE8izedT1cvhZPirY
-          SCX+166jQcKRheaLVpDItY50deRjCQ4f7uEfOOnnkALWG/JFZBNXRVZNJJxbf/cDx/2+g+bdjBbc
-          9Zov1A/kkE43EwRi6YKDu14MtDGyEvp5eCLOdBEdIn7DDTpXixBH72OVtqxSIb7nDsRIeQ+wE19Z
-          CA73C9E/z1NNT2Yg/PHT33rk67Wz/60XuXtil2/MTwvhYH6zeRlX4iyX2yUB/THExLEredcLJwiD
-          V2oSk+EcsBnfsIJqts0+83YvdKuaKgbFAi1yrZ8r2C5HYKOft47kmX3fA3eR9QP0joGz6wPgzCAu
-          Kxic0g+5rVyVb/EUVPBvPfWgeecbl8AUwGh6+nyM13zxBlsE38ui4JgRpIF+2jgBz4jkMyrhG2y2
-          +pNg437BX707m8lZM8Lb1cdyEZ1yuvroAJ9OFvjw0b0Bcb90gcjfrPn0Xthhr+8Wlvw8+UzmKgN9
-          jG0PkytqyXl6Veoi6T4D1Ug/EdcTmHp7mJL9bz3O31aN1sRcezh7mUDMwmbrLctQJbKLaWJZe7wG
-          Oqx3EVTGVhPXQgBQNrcL+OFlkwQDl6iL595aWP8aQuSdD7el7y0ABPnxV481HV4phEkXP0hxvLrO
-          CiqvgSBWLjMbzk86vgorFY0x8InUc1tOi/HNIPu6eHN50sthQt77AA4r6xD17B2jAcWLhdhAs+dF
-          GGk0PV+dD3v6ccjZ1p5Rb5pxidSWFv7JPwrOip+3XgxWsmCMudQhf3qrpGOJiws9D2wPSQLeCN/9
-          9SgSSl8/0YcPC7b/9NXi5/oChNLfiJcbKZ3T7DSe6LkPiFz6bU16pWPg5Wp8sd5nW0SXEmjwO5Uq
-          8eBvzKdTJS+od7E5b/fWVv/0NcDz0Pt8J23DgtNf8O/+GLv+WwU5/IHH7/i33jnY5E7I4JKzAVE+
-          hRqxORsu8H15pcSJPkfQL+KgwT88sIzRrafdX0E5lVKSD66Z81wn9HDX6zu/vnLyq8UfvHiahRVw
-          O0TTZpQSDJezg7X8csmZPLw28HcuV2IOQK15Swwb+BBih1wq348of95adFqgS+7v5Voz6B5KsHke
-          tpmZkQqY56E+QDO45ETHXZdvbHuqxHnSz/Oms2O9Eus4in4UpcTMcEq3LGPf8GBYLgnAV8u5/Jce
-          QFWuMjbIz9wn6pwILZxE5Bwcn2Cj54cF34JQkkdo6GAp/ZeNdn07Hzb74KzZNRBQYF9HfNXpDXA1
-          MxlQ+Wo+kb33sV7Lxh+h5MKVxA93Vpc4UTKoKY8Sy/dUoNtYPnpoMMFEItMa69lT7gYqU6nAlvPV
-          o2VpvqMYGNJKQpXVAD0kxgx3ft796g8sbi4fYMtpEs5S6zLMIWZsCIZZnVk+IDnRh18Lb/rNwU6Z
-          9PnW1FcXWvyVEGX3U3/1A3Z/99f/0RKlX1H8lnqOlbg18u1mCxs8H+4GcavOj+hVHlJ4E+ajL1Yq
-          S8nmDS7gnusN+0egDcwSyiHUJobivX5yxi3iEdwetw5b4tkBf3oMzBejIIq3BJQRH34J9XPF+ieu
-          +Knrgy1bkJf+TMxGP1EaaroG8JwcsLXrm6HhGxd5kcj5R+jLDjtYiwbtVuf81X50zorKQw/11fsQ
-          +R5NYPrrN/pJQuLdZSMnZWb5/+rzzr0ideLXew/oo38S7K9dTr+qsMDoVl1wUvEPZ+1B0SLnx4V/
-          +jlaPhkKocpZGr5U5zofdPCqwAfHN6xfm9aZjFrqUb0Eb3JeXs7wz79drtoXh/ai//lbER1KuydW
-          lRp7ntD3cDNOjA8vtBvow8kUmLrZm9io1wBXWKGIsoQbsPuHH1OGW2hUHxs7bXKv1/Zp2vAkt1//
-          VAy1OgreeQS6IGbE9O/Pen1sqwHn+73d+QxHy6U7bqDTz8vMnPNg4PjPu0EXWUvxn3/j9v5An69i
-          73zu1FM4NDYUWOGDNeWTq0uBBQOKQoL3/l2jLSr7GHDUxkRSbTlfxYMggB1fseod3s6cFFYPb508
-          zHueQddg+MXgWn7fxAuND52Yesv+Pi/2tV8+rK0HCsCXXIB1STiqY6rWLbQz3fJRCzeHTK94g+lo
-          UOxqwqvuQshzqHrTC975gi71ozz8w1N9yBJ192MH6GpVjF3hWtHtxXQhOH7qgXhv+R6xEfhoKOJr
-          f2YmVKl/+A0zkef9+ku+DvX1wAc7H2JPDizK5OGjgbUrTFjeNKKuz40RUcroIcEauAH+oj85cJ7J
-          NgvFG4L5cTsbf3kU1v1Or7mzob8BPBrMfOqGRJ0YChLA3865v85n0amK5yoh7PcTtqJ3N9Ays1zY
-          VX6GbZZUA9nzCigYteRPFbRrvhxyATLCpM4olt8D5c9iC6VfYvqn+/Ielgpd/D+89peb1IHVep1G
-          iLnbAzuXXMnXPZ9CP4FL/NO3raMpN9oKPmk+/uF1Th7j3MMgUWesIfobNpbbSnQZtzs+N8MWzWRg
-          RLHh9wmOWT6dZe93ZNt3Bp/vi1IzD9kboSj0gOjVk6XLnbYpghF5Eq28QEpz93cAf/rRKRM7Ymfz
-          ZUFWIm9/nbQhp2Z/+UG5knjf2/U7s0Sxgu6uufifWyfU//IFNSsVH8aUVcfzzMfw1z3neWWOpJ6O
-          6LPAsY8pcQ0lz3/PIzfzrvTWMdYAD9YITBq8P8sv8f3Sc5aLGaSw8JoIm9LdolQidxcacVtgf7S5
-          //JnfGg+BCdrOmx9eefghdkEbPeHxGH3/AuaINZw1qoopxkSZ9inmMVGwrfOwh1bF7wFscS6NndO
-          p10nH06hWvucYTDOxG9NiijbAexs5hf8y7P+/Krk2MuwIdX/iebxOBAJvV8Ov81HDvYU8L5HN+JM
-          f3+vZsLVPyT3S77MLN9Arl6/2Pvwas54wGEAaawX+dcPe34Cd37Btu99Bip0sBQVkqfYLPU3nS3+
-          GcC7uu2BrF4O5BedZtg84UZy7yPkxE1+GryeoICDN/CdhT6rN3TPZ0CULayitfVo8S9ful9FYfh1
-          wrWEwktI8F9e88/Pbkrg4SvvNPkYmYUrnu25/etXQG7H8w/ipHPJ5eWHTl/TMAHq4eET+ZJ29QLt
-          mkO738carxzU5ZfmDVwjL8ZKsBQRZyV2/3d9JDuHI5j9LRLEEDAY20+rAAtOq/Cf/jXVewOmd5WF
-          sDhF4b7D7wM2u8UpzD/iC+NgXR36FG3tn3/CFjeCFaHXAeYPQSH2xTKc9W6857/8dz41Q5hv/PX1
-          g8otFYl8YF/OPOWLjcRzlOA9v6x/HnA4qCnPktjYO+d0efwScLhwPjFju43m4ZDPEOXm+6//anoL
-          rqkYETYn0ofzB8aexhb2T7GfH5Fxrte9fuDuJ7AkxSBa8J0q4MgIEdHX4EsJB7QEHQzbnTny+0bD
-          y/ZF2EnVGZezrtJ5PaUJTEeNktQfc7pa1MtgKqUbMfafp8fo5UM+d3p8nlsf0OQ1BTANNIOUne04
-          M/Nzg9P5oGXk+s5q54+/4at34r3+6oiNxDABf/7OiT5PsPLKlImDePSxkrUT2I4rN8Pjqy6Ik+SG
-          yr10VvnL9//ySkqs/Bv+8xtydVbzLTxUKUQcbfyNDQcw3B8CA/c8aj5MeqXS4V2P8HQNbWLs+cHm
-          hZnEdxLDYmPX2zRWDrufH0Ofgg8exu5mcXBRmc1fWoAiyo6lDTfjY2Ezxmu0vl6TBPZ8lkjHLx1m
-          0yK+GC4PHqt/+o3nuABwe3arPk8TnTm9moGV32TsTdu33usFwlit4c4XPO3++Pv5zJ5ER8ZLnW31
-          pwCbpTaxvummrlXIccC52NUMeHMYaNOmb7g4TESSY0noWmiyhAAj+0Tf+2f3i/uO+wpjiYySymbc
-          9IYEWkdfeKb7GUWmZCGk6hp2dPJUt3OzcOh2Ub2dH86UP1lD/09fmE/7MnCb/RXBn5/3jMAdlj8+
-          3v0Jtl+HWt3zOwU+nTSYV+74BdtLRxKMSyHGtygWoqFAXPU3TyGeGSxRd/TaBbKNdsdpG9XRelqv
-          BWT0K54rB4NhdmP+DeZxe8xHIXLVRa/Obzi89zOSd7+1GIelh+yo8ThZ68YZKy904dPM5JlNOgas
-          F/AKEHHXDDthtaqEze0S8iUTkMfO3+s+nwGoL/czToy3s03DyYBsnOu7fm3qEQDUwF1//NMnG//p
-          W/gofmecYKjXe34pQdnqlpnNG69eRhYsQIdiNwt7nrVaVE/RyN1e/tzEj2jN2rGFtEMncu2v3+i3
-          610gWYc31t1ZiDpRcmagrdVt5okkqr8tejLQbk1u19dXuv3ph9X0Wx+i7ABW8bAIkJ4h3OclSTTq
-          i13At84m/rocq+gffzUmjP/0ZjSczEBELvRU4ms/UE/tE9v/rkeJBi/nIwYLYJEcPNsc0XLuJ6ES
-          3mvLwckdFsPy60gIVEfPsHEoXXW4SUUJ+ejsYlWyXjmNN2T/6UEfvrpNXQ9Tw4Cdf+Y9n3Y23apT
-          +H0tRxI/zNoZRxNXYNxuob+GqRcxGOBevOdFiv/wiS77Ds+/+RllPk2+0IVxxT0vJulRa4ZN7pYU
-          FNdm+zfPmE5yUkJpHJQdny/O8pC9WZyZBWJPnHC9HV5oEU2QaFhfgkadhHBOxLgVqc9fMetsjzTn
-          gCjEGJsv+57TzAMzONhWPh8N7qNuApMucM+jZ26eHnTM38kbPm9TPx8WSRoYokYadH+vCrtiRqNd
-          f8J/+l2++2+w2EUgAFLZIbZxyOSbJvQpXMU2I+qLsevtbsMWCpNg4OufnjEOwk9UI/OEzdPaDKui
-          /wqoSDKDlVur5PzwSg9wz4/8Rfha+UbsKkD6y//5p93PLDN7bKBbvDeiLkSMfl4nWBCa4zoDXqmG
-          aTFHCJUoG4jkYFAv/ulkQI+AmshuqgPWDjoG7P1DXFdY6Gq91hH95SeYjap6/cG6gsrRGMnfPGEe
-          PfkA0ziXsYT1T00lqTnAsVdM7LSXuR436ZQBKf41OEn4Vp1CeGRgV1QctpuDX3PTsGpo8q9HYtjy
-          2aFy1TKwUk2JmFGc5uvE/yzYgB+D9Ut3cbbBL2bI25yM5Ut6rv/yL1iJlGCV7akz3IJHCi9M4RLz
-          r592vYU29jARg/2V+Xbcd6ygubgTyx8BWN3jRQBnUXP8k/AS6aym1Ru9EMdjrTb7YZ3NzgL5qdH+
-          5kk1FzzXEP3NT/d5tLqoyseCblNfiJ/FRr65MV/BDFsCvv3lDX/z6Gour0R9na71CO2aQVclOOOk
-          2jZn3PMMceFfIfYiJx7m6gMMmBGHnac9Dx4lCvb8mJf/8hxnYlhDg6FLk3kKnLT+m//C5Ny7xDTL
-          ozo/fvHyN9/yj2xUDYuqTBbY8765bpk5X12ciTD82aN/tMEz2sKXycA1E8747IuewyzNd4b7fIfs
-          9ZyvV7G04LKftSalkkfZh+YykAjWQjS5cvc8vwihpcY1ieCH1uRgeNu/633nQzHQk/DugSF8fsQc
-          gVmTo1qFqCSjT4I3V6vbKQ5H8H/sKOD/944C5f49+8dVF9V1ddsCnsxt9Rk26+kGGd+A31QdiNT2
-          izOS6GUga3hGM8uxCl2Av72RDVFAlLojgPZZY8Cm+iWk8H6tuhzcWoFbnKw+RO255n+btMGfQ/AM
-          JN2PNk7WJMQmyTTT0QoAo2lBCFOznbF5nUJAxTXUoKJFHLbexxHQ6Jf2UD29AmIHXZvT6z1joNwe
-          C+KdXY/Sk2PaMH3TDFu3vs1/bSJZ0JlfBnm8iila7xfGRswXFj5/PHcOvT7eIRrHT07kHFkRr0ij
-          CCfdiLE1f1Zn/Xh6C7SW74nv6lBdjYWE4v77RDsN1rCwWDHg2ogdke2KAjqvq4t88nCwk/LnnOmc
-          +waVYplJnBhKzUQotGE7HyZi5tEnolPsWYAjpw7HJ9jm6wlYMzwqWCTGKB3AvJzFAm3bpJCHyadg
-          uS9xgGSUtPO4ErNmj7PCITBVNgnXm5Kz+uvXw8t4ZogFo3rgrtorQTeJfPwVs796sLRxg7EMfCxT
-          2KvLXQ411I3Ew3ro6oC3oM7ATfFULCn1PCziamTIjg82dreL6vBzRQuUPxiDmIsuDWz8TUUAprc9
-          81PO0RXkug2zKTZJ+OaSgfu6lwpFnWgSaTg2gC/6s4Ks9DJj6wFYsLXzJ0VC4gNi5w6NtqD6MDB1
-          GNs/skIUsZdPcECaO+6nVg5cvn6guYD1Z/hEfqHMmbUCvcV719czo08fhwVfUMLx+nNn8OiVgfO/
-          7gKDhNsnmL6ojuJqpHDD8oOYB8GJxutrlmCZty2RMjqra3sdy5P9IwyWRcI666M7B7Clrk0u7k9R
-          t1KxQyAFjU4ux7efs4/1ZyNf9TtsB6Km8hXADXTGPCX2I3rVtPjspwyq6Ygt6XvMu+oKFfg8bmd8
-          y1kJCHx/VKC2FhdS5vAB+NdDtVHhhT65pjN1licECswSiHBxsHSVqUIrhCfePZG7Nb5UeuBfJZrL
-          ISCSqvcqvckHS6ziz4BV/hgO1I0iBTbl+0v8zorp6mebhV7nm+FXEfnsE4ZrjIQurXCa8JO6HdZD
-          C80l+JE74AeVE4tfA61q+2LnKNKBDuQUQLhSBjvS90w5OUxmxLTCh6RLIddMUx5TKB4MhWBJPQ5b
-          aXc25J3GxBbb3QG9BvtEsZ0Ef9Wdp8NBCRagljeDWHNMo2UQtALygvDC4RgN+ZYR9/3v/yliI9aT
-          NH4zxG1yis/m70y5DGwcah7lgo3bcnRoHlMGxfcY4+CloYGmIRDh9E7nmX8p95ybgGsBWEpnf2FS
-          pWZ4yd2feWNTbE/5K9qYYrHhE7yn+dSfU7pKI0nhG9sp8V6T4yz40drQQ6crVoVgqRdNkjhIxud+
-          Ugtb5NsrWN5Iqi8CMaeRGRagaC5Mrf6D1dNHz/vUFkbIMWfDp6bPgrm3FQlxmXQhez9QwpPMhvzz
-          tvoHXab5H94Be98BlJYurWdVDxQUWuKZ+LPrAG5SawMaQfzyedv0a6askwz4RurP/IlhItr4XgLL
-          7E1w/BgNunw//gY/h2LG55G+6+4Pb5ZZfhLDylPAsiHiwPmdPIhzkwQ6/jjtjVhZzrBaSKszVe8+
-          Bofr8YbPcpVTWiZ0g5QvH1hC8cPh02fvQ001JCLlmgOosz4L0BTZyduq4hxtWFv2U/fliPi1pTrU
-          wJGN/q5PLznZocM33L/1gPlgJY67mlHfxSjG9KHteKbWv8iW3sgiI9nr01FZWYY9PGUwIbe/9f5x
-          WgWvgv/EWt4Iw1hmxwru/IL9y+DXyzAktviDno81Nb476095+vAPn5z3vg+Aie8hjKI32u/XJ2e6
-          aOkRzLiGYK9i1Fn7FCk8FZKHr/2kqmyBmx/SBt3Ad8E6q8v1btqQHL43oj6920CbMyuA+7di8eUo
-          bXSNS1GBF0WWyOON2/png24DOmFeJLxtHtjkDDVwDQbP78qaGfpeF1vx3s+ZP7OvMKIHsAjw6bsR
-          NrKhHdbz/Fzg+XQZycVe12GFzGtDByyZODqYEqW+Nf5gzLYbMe10jLZvUEHkxVw6f+TLCDbOExP4
-          9KwC3yPfdxg8DwI0Q+hPJ++GwCq8Gg19mtORqC/+lXO6+GpRnXy+RJdKWg9GcS7Fl7D6OBraq7M6
-          n6pCr6Ia8D0OPzljjPSNNAmGfrzfj/XRnBgYXGoZu0R/gOVtXgTkSQ3E6nqUVdY+XivUGDIh5z5L
-          6i9+rByCxJAIDhUyrHc4C9AIkheOeSZV+c9iMcg9nUW846GzJYXBQNt7WSSsii6naiZX6P2RV3zT
-          gkO+RAwNBWG5LiQZpRIs0UkZ0dD6Gra1TcsZE50CeOW0fQI1tpQ7cdcGTroWk9uriQEnZ2wDTaQr
-          RL/dD85W2i8b1ZV1I3fB6pxNGUACv68mJ4nnbw7VlZGD+3r7gj4k0Rp/AwFpW2Xhi7RKDi8TYsAv
-          KUYcnJN7Td+zJSBFu3BETgxBXRKOumg7PD7E1LYTHa5u84ajj+2ZYTMbbCi7GKjrv0+sPzZbXe7I
-          bmA2JSa52Jhz2v6k+WjHT2K31YnOGY0geoolt3/LhVQzphG40K3DN/F3fuxL4zKj4ntX54P46SL6
-          /vQSWtjihfNvH//pn0B0+ks/v9RicL6+dvVRvz8jp5ZD4SxnaotwPl8RsZ8Hd9h8qSthi6sPManc
-          Rev2GF34Yr/dfBovpsPa6iTBkA8y7L8116FJHzSoqoIzvlv2Z1i2aydBK8uLuZVPzsCjnlrQYQ+W
-          f9rfH2fRTeGO/394W2/RlS/gH5/ZmXit18OrqGAuat18eFA/Wtjo8obZFRjEoVsE1sPDM/7+H4kv
-          zddhb9MvgO0kckQvuZdD0zxcEImLmjwOT8NhmzVtQPNeNBJQ1RvoBgMJxhJ7wyaVz/nygUMBlHzk
-          cfBsvzmpgNnCarYzrBUGk9OkPVswvrkvfGVXB4xhzb7R3m/48q3dgRF/5A05gxvx+eTup777PYSW
-          9taI0Uqew+Lt1sM6+X7nsihPdf9Y3RJWCbnPpzj8RFtoLAyaxRMlyvqTnPF8b0PgNvEdSweTy8ki
-          8QKoZ+VHLOn7jLZLg0Iwn2/ov59XkUYBkIUYRGI6KefsYjqA+7Bc9vUxVYpKl4Oj1lwJBuUHrOzN
-          C8CII5nIunNUt7jd8xN0upLik9gO54iZBA2T/c0lM98dmvF4BDu+kb3fHVa4ey6cyy4gzoVKw6JH
-          iQj/6h2Tr0oXDZxieCsyZT5Vg0KXjyQeINTry7xZbgc+1SD76KL7H5+/ThvYksLn4FcHCMtzwg/j
-          uR8V2P0eK77Kb3tg7uflALNUYuY+6NqIZse5hdL+jF+43t7Rrm9C8Ocn9s/j0Nh6GDAQPwf/VJ3N
-          nChe7sLydODx2TTmaPvjc/uoHcgzMox6ud6xBSXIfYmc/KJ6bfv1Bz5Yaok/P84D4y4ghYdDieeT
-          cG/BMmgiI5o/vp3RKcfO0l6WGR7puvj1k+w7Fu9zAJNBVInTBQbY/vRSx7rGzN2Wp7p48urDqocs
-          CVqLGdYX/RmQr78+0ZJblS+mIi3oxJ0VLD15XR13vXOSB1/Asl1FYPnTv3otmdguKivivMuxBaUR
-          RTs+VMNKGtVGSXiLyblkwogpRKBB45SXPrpyZ4eFyPqBt6f0c6DU/rDhsxSjN3p3/v756GZ6/AyC
-          eG4JLpFVr9X7nYBzsJ/ZRgIFMAkHfPj07AIryczR7aXZ+w7p041Ek/sFi3C9FEjzlpHkzWMa6OuV
-          tdB42GDXD3ti/lwXtIYLwMGuP1jruyjopL6uM9z95zanS4D4c3En8e+5UkqLooDsLW38w3z8qEtx
-          tN/Q4C8LubwXGs1N+Hci+hkT3IxwGBNg9//wq6iXW/SzNWmGUPIO/piFr3ypY8kFLRPIODiIZCC4
-          DWZ4NefKX3b9Ob0dr4fHfOt2vyI6mz8U7uk+bBfiNC4G40/ZBGixioolT7iqKxRWFxJPfu0TwTxf
-          8IttQf7gDKIF6/6tS/qqwU4qBizTbK7JEZsSbF6jhxXrYQ7c6yiOwnZ7JxjLJjf8XR9M9YeENU0c
-          KJGqqhCd1IHEPDFxRI+3SYRG/TawnPpGRC7o4J5yYFfEa1dFZYyc3cCfnri4tyYfm28SoA9K+b0f
-          vvlysKIGaj5TY1zmk7OsUsTBV/EeZnaZGDDaVmqBqYokogqbq85agH349rUXufYHCWyNfW/AKeV+
-          xF3LDlD1KmoAecUF5+KnBOP1jm3UKc19v36Hct0yJWB4Mifs2YFG10eRHcQC9E+srIZWc2atiUg8
-          cS2RlDAA/IHvCljmTYszHvnR1I1ri/71IwneYD1y3QwNk//5p7u/P7GwfBLAC+LLo/FgD9yhu83w
-          W6Yswbff29lua6FAL4EGzmkgRGvd9OEfns7oKMvDlvVIg2u4AaLyD17d8W4Wnziwcc6aFl1sK7DE
-          77sPfeY4iM4irWGF0pPdYNdqsMoEIPSRiZcfVsLH3RlNI3VBNobBfNr1+nhn6hlcuC3He33W/+6v
-          Pi/9P/3/b/1yhib+Bm6UrqY8GdBntuckXCjnEAFKJXS2pSahYE85cdZbIfqKvWETT7LD//HD9aOI
-          HjqNIyUXRZ8hctMM27+fpVJwWiAsLurJX6VVUrertWwoU8gLY96z1CXYvxVL9sRyrq3appz81BXY
-          HTiw++NG3ejRlcRjwVVE/tZuzcr5JUAdTumuz+f6z5+AXf/NBzuxHP6vvjfJvOKz0Gs1296BAEUk
-          OTg6qR0lUnFR/vkBOzNhvoRLzcEvKUdi+GHnbNDrN3B+2jw2FPSuN1KTAuz5FTHaiw3IuTxnUOJf
-          MTGiRFK3joELjOKWwZLCjDm9nZ4uPM4Hf0bbrY1I/uot+DzMIVbNfYe2fLAYCKJj7g9S4jrbZR1a
-          6JAG7XlR4VAasS3Up5H6h/31pgjdD7ptafpFPM60Ox73M3SuIffPr/coOe/vFyYud77Y8Sv8y9Mw
-          Ng4tWIKNJmj/+zOTN+lAV8Fq4Is+AHbSdhi67UV/KLhuj5kYU6du9/sxhu+v8yTnnT/Zo5pqyLsK
-          mHgWCFRqVIGIrPFlz4P07Sg9X3INCn1xmwecxQMdhNQSe5KI2JB/HmDZmrbIXiwOy8e3H/HurS9B
-          /R8AAAD//6RdydaqMBJ+IBYyScISmWQOAiLuABFBEQQSIE/fh//2sne9vOf+RzGpqm+okBpFCVXh
-          V2jprJ0zGHBsuuv5OJ4aZoPQ5aCFLMOWANnjR/4WKER7fXG5nNSerInPO/GxaY1EqfsSfJzV+Zc/
-          64yPNdzjK2Akzac/ceYneFIUn3jMvaEE16CC1ZNXkfbqtH2KopLBXZ8gpzencXHAa5Nvd++B4vOT
-          jmR2cXg8LY2H7Dr/uPhrOSYE0fH9p6fa2eb5XvahUCFdz5YCB199AzvfDrhdL1EshSbstE3d460s
-          +KE3K5i95hYp22tzh6eCefjsrxam9QXvf5/psOs/LTopa+2ud8NgATM4kCh3ImiL+S5/QGg/AbHa
-          l1Vsb7kNYLuFBvrTV1R6j8u//TOiRzvihCTmH15iiV3OYOMhrGHavRf0rCYHLO9KyP/8FaKcv592
-          KSWqy/eLOxKUxQf3t9cv0CArQ9lFgZQueVlDWdpGpF6A3JK4z3pwCpgrMjtqFmx5fpXyHg/kAnVH
-          +w3PbwcO3LEifrApMft6BRLk9YlBj67W44Xcllp+CeQWLB8hpVzS2hUsQf/800facKuPEMBCKgJJ
-          vWcFvioXSVaqKg2OQHDdTYVHE07yGBHvHjgF2f1RWAgQ7P5PPhIshbo8gtMnkMLBLDjHcyV49zMF
-          nVDsuxjCNwPun+IcsP1Cxt2vC6H9rlycM4EMfmE983BucozcN3aLzTtbOeRNdiJFdp3BlnefHygE
-          BhBrcTW63vZ3kB0FY2RbRRtvmpEpcDjVDblpn1PBWwQEf59HHO2B4o3N+gnyMwUBL95hTOWmjMCv
-          tK/Em/KIUnT/5bA4qhve3PsS473eSGRjWXSDvlJwZPqpYOevxLamk8Z/kyWX//SGluQTnVZ9jCBP
-          zwj583vW5mVZS8g6P3XXi5lLx/HFA+18yMmffmCZR7/ARYfNv+fn8yb1YJo8VHyclKGYnsadgdLn
-          46CQzdSR3/0rSLmUEFV05pgqseRAWexllCAXaDNkhw0ecyYNnrHZjQuJB/3P70U6mvnxMyrnCfax
-          c8Es+13bZRS9EkqASYj2axh35iKZFZ1+ZpGDkktM4aWq4c7PkJMPAlj9qIjgQ/0Q/BD7K2V3/Ad7
-          vUOu8y3ijVtMD6pp8SZ6MDzHtSZZD7Kj9SbGdwPtvOtTEH2cCaneQXPHLKISuJ3WEnnW/o53oagJ
-          iAfxTO7P79huctXXkmkQiTi0iCjV52cn0JEe9vqiUnbpBAvsfIzohSq7Wy3EPGSb0Sa6lhy1RaAP
-          Ee7++P7OeVXg31H3IMDvE1bKe01p9DxuEIfLATnl2rl0rz+wsQUjOPgojfFfPJ6miPzDTw7Kyk+2
-          5s+JmNvdGYUl9U3pkE8bMT/VZVzFD2ig/EIwAAiG7rJdLp38rZs3pjV1CqrqMPnTh+h8La8atbs+
-          BPP9yO18k1CiKm8Rok9YIeQxYovdj8eDqFUZTH6nS0w+bd7ANLome/yW+52E4Y7v84tYmMdgen+r
-          8G/9kR++kniJj84ERxEIATOAwP2Hl3v/JlgKfaTb0AcVpNpyJMX3zsezk/wWOD+MPPihZC2m7XLv
-          YAaiCBnvRCjm33pSAdOfjZ0vTtpq8/zvz89AJ3j9UswMTwx2vz44+sdj8XuFYg1tW3gT8xwohbD7
-          31DMkoGcr64QL6JeYAhz9k3KE+rbRbRGC/z1E1A1/SgpEspDwnxuSNn107YYrSL/6RHjHuJ2xz8H
-          eu/0HvCqYxdboLzKP3wgZ/tDXZLiXoExM/dEpxFstxvXVn98DEUpTildG5rLIJaLgB2iUcO/qJbk
-          ixmWxPCu3LjroQT+8tMXizDWxl0PB5CLXQtZ3lVxd346/fFh3LnHh7ZEvcfAsfN0lO14+qf/ZJCe
-          bcxxNdbWMGt66azwNQkueQioKhrwH9/+87/mOHInwBMwBBIbeDF/apMKgtSwkTaBn7YmY53Iipf0
-          f+vv0kDALIh9dyPqm3kDTNiPBwW7upNzvg6AFl7JgPfzXSGXO1tg15+6FInpJeAWJmpX8ZH8oKsl
-          RSCd8ewuxNwWqB/S5h9ernlzlEDv83LA7n4seQdGCqXvJJKAvOp2udsihNutTvG/9b9wVIHrZdZx
-          a6dvl34WhYeoS664yfiPtvdLFKhdaodURrS6y/XDVWDfPxTs8bl1ghfCqPNa4mZ6qNGL6mPA/Dz6
-          51fGy2Q7ESzs7of+6unUnO8SYOY8C6TnHIA1EokK1fT+xg3oL237V79ASPlg7cVXuz5FJMLE0Cm5
-          pLcraJthLf/hny75b5f+9c/2/hq5vNNOW5rN+8HqyaokMdxhXBuT8YD6Ulxy2v+fDPAA//qnCMVm
-          15KoaFLgHV0J/5jgAdbSvXpS5vIOOh/btKV//YPFss7IajdT+/MjgS+R5V99na64TOB5TZ/IIoOl
-          ccO0vuUYfiViBm0HtrqrdGgXcvSPH223NVHgxxVO6Lzj06SURgqEI5MhxNurRp9LlkEwlTGKp3vQ
-          run5kcBDrZrEs97E7UNBXWQSnDNiuE3p0v1CXnnHW+Ts/GVZSarABlYb0p4VFy+llqvQfFi7Xjj/
-          wEqKdwh+3GT8W1/hT68SX3sR+5nOlBaPfIKulhZEWR9Aw2y5WNDapCvymaepbfvNhzJ9GTEJbpeX
-          Nriviy7rsseikm9qsMZp4cDi9VmRzhqpuyh1uUGVMDEeh9d7nBkPJXD3YxFqlxegVaEFkFONC3KH
-          sANTej4qcNdPu99Dxy5Shwx26YPDfHByi41KsAPvZtOJ8eYjsJxVa4N57jRIrcshJtnNTWAJpxR3
-          mzWBBT6XSEaNfiSpeIfFP73Q+MoPGdFDa7mgvSvQqMU7cbV2G3tzVRk5we5CjIXZ2u0ofVSQqAcj
-          4Dzh4q7Wd1Gg/bQE8se/pzpSItk+xhOWmaWl62+1lf/njgLxf58oUFqhIabgmi0N408HW5PrgoU9
-          HjWsuE4Al5Q5EWvt4uInihdeBlo+IcPODmA6qzwvPxdkErM1FZetb14IP747kYCbMV1MzmCh8zSU
-          gJPlW8uF3zqXxDa44CkND/Qd4TWUb1TXibLJOmDNdl2g07Qhlj7XIF7XgSthf7zVyA7tp7tqL7mD
-          Z0s5k6SDeYGjs8jCywtcA/4A3u7sJ3wKbuFBQSqcpXGhq5mCctk7Uutz1qie3lTZTo0KWet8HAes
-          JgGI/TIhyq+RtPUr4ggy8D0QH/QFoIkl8JCeO45oUxG09PR4/WT9YLXIBHodzwaYJShNTBesp6QF
-          64W7J6A77AytJY3LR+eFhx6MP0RRnbpdTVHdZJ3BTvB17kO8Nlyrw4T1WGQ1Fydee33GEArulbjt
-          6xzT7+PjyUiaShLbllFwYy6J0DLRI3gN0yMW3gJXweqN9qkRhBnnhh4beGZKL1iH6VEI55/0k3N+
-          wvizPOBI54phpehW9uiRBF7La+xDgTCMBHyQ75q71B1ewPOwTcgL/c8oTIwVyL+T+EMhuT/A1h3M
-          RH7W4xMF/PZ0V2ccK3gkPxXTLO21jeGtBOwHUshdAWXBt4kN4XrQRmI+m7vLWRfgyCeD6QIAslPB
-          +sqcwLc2rwRdHuo4289Jh6nSd8EhaiIq8OrDhKykxCiTLL3g448XQn6SHAxMarfbvdkWWIVrRQov
-          1gFHgriToR8ryHhTBXD34LWBgawfcr1OekF/XAtBnbeU+CdloOuB1KH8uN0aosgD32LRniE8trKO
-          CspKxVrPJguVrbkRW2FTl0v80IHJI70S9CkasLj8rZJPBuxQPDUspcWhVyBd1ZXo9PbR1vBbZ7Ih
-          WiaJeZWA1Y+9DfpslqIb8OaR+5q3CCZ0Kcgj1/KC/dabI3fZNJFSrESwHhSYSa7+S5AjVHVLO1VL
-          4fhKI5KKqtNupAYKoO7zsU9LYrX1dGU8+HTQHTmjdqWU37ZGxv5PJzZju8V7e0qRdK1kCa/X+6+Y
-          lvv5LQu/ACPPmZ2Y3fNFVkX2QQy1erWcXtaiLGhJGIh3/tAuRXLS5e+1E4i1z8Jch5PayRP3SsjT
-          308xDF+qyOdVKrFw/3qucP++QmD03TMY6Ni5C7fPsS7hu0DnreULyrWZCo3heSHOnu8C+RY6BB/G
-          RqF6MfdDDSWEQJ7vKNcU4pJabSPZMtIRnQ5fTeNf7gfLT3tVCNJMbVxf382SH7HyQuhTqIB9e2kj
-          M4OVI8WSs1GYp+8G//IPvC8fbQPVqZM/T2Lj4+HMxZMvOw68y8QNpmr8tdSr4kR+dOqN+At0CgHf
-          lk4mKB7wYTWYluqGAwHnlw3ywsQF7HG8Yxi+9xMIzXQZF6aSMnA/k2dAH/5I6WFIc9i7jkOMWtOK
-          TT6OJbzxzPQvnicihgwUVvmA7FgPRpa/6o28PaIzSrIbcokQBirkWV8lnskuo5Ddl1QO8KyhYPJj
-          yuELjyFKohsyxFM7zgfSR/DwdEx0ud6dYmu/5w6eC4kg43MQilUKGVba6ydSLk973H/fW1YDaJEi
-          FxcXSx/Ew2c9PANo8N24fj99Dx9duSCNPPSY5sktgvcIGsj4iBeX4x6pA9PszgeHJ3sqqJeLASTo
-          MhAXok+xKlIS/svvNxcfiyXD91RO6u2DdG2fuUb1YwBh8hWJuTUoFp7UwvAYysGeT4ZGlGcRQDR+
-          ApLYqqnx90ZaQEeflATGQaLkMFS5FIeViwyZuBo9F5kF5f30vvQ5vyixpKslv8X9nTNw74ufLKyS
-          3CtXhjjnxWsF7eg2xy88YqQX9eiy2uNs/quvriohuiWr1sODfpOQbT3beFFRasm4v7vInoSSbp4V
-          bvArvmdilBnvbu0VNLDvrj/0DIajtvjKJ5Wb8nAhquAfCuqGv/3y7vpCLLMv6TdhlU5qNobBg8a7
-          dB3zTQTX+UeIaSXpuP7V//HjSUi952uxMLgUYTdZgNyXISy2234m+tX83sisUnVktQfSYdvDH15Z
-          Dcab/70ustndSnxIfcXlC1/wQCEAKeD258X1VP/gXN43op3eVcHnShDCRc5NDMUqA3wpHRv4agKX
-          BA7PjNvtnIlQ1F8lyunkayxTbZl8Lc4RUtfLWKyKZUQyYfM3cTovcpdzSBYgtt4F3fJPqvUzYSc5
-          +GUq0vV7QDlVFd+yXn5NfPSRQieUSg0UlaDA7B4ffAoupXzqmRJlXNZR9nzyQ6jQu43ue30h/pHz
-          5EhSDWJuxeAu4VWBcuVfAxT/OmVkWZBY0Ixv3T/+MM/bsf+rzygdthVQWs4JOByPPnGedR3zPNZV
-          IM35gaAOP0f2mgzqH94RZcd36k96JB9pPZJsr4/46wkp/BSZRarXq9GWy0tJ5SQyFrTzC5d2lw8L
-          Xk8vQaVXRK7g1p50JNwlRanMMONy+B0XWTJ1i2QTIe30t38g5QEWBg3E7+ZZN+B3TY/IXoYwppYt
-          YGjMtyMGV6lzifviS5lrpZ7Ee735ngrRlDc6+eTJCpG2vNjbJD8zKSDeXt+W26KK8uT2B+JufFhs
-          17mvQRh89X/5uJS1Vco7PmNYfyuwlG37BqrIPwJJPVpgerE3LHf2pwr2PvBIh86uYHaAAVKumVas
-          xbFz4If/9MS3h5PG8cOjhtfzqCEf9ACs5flpwd+pvBBTuFNtq3HuwSPVI3TFPW5nmwRvOAYbuzva
-          OmCBSXvoModTwIqCUqyfh55BsfvVxLYfc7yW7L2EoT9FxFgeZbs1Yq3C5ZeHRON0g3JZ5EkAX+9X
-          4vHNu+2BcDdlM752//CXlx3hDbXnBDCN+jreErZb5Odj7IlS+BeX+9yHEPzVM+dzxcV898RAOodo
-          RQqJbuPPXi4Q4quAkZUcCjCUXV3LjGZhpCo3Bcx/z3tNlwu6Gc2nnSmrRbLScrvCNk8uf7zOKSQs
-          Lwdr2HQjvZ7WDqRM0RCDXj6AbnIdyf/4B7uMdNOXVZTiZPWQwVoDpbmJE/jHd80tacAmRfvbg5/X
-          AYPUisb1ZBaThNXiSTyUvtptr3/Qeb7uyJZuAljVKDZh4xc/fGCFyF2p8cyAuwk/4rJjFW/H6yeF
-          z5ch/sNvTM1mg/Zle2BmREO7aP7UAam6rORZ6BpYPkYYAn13wlVpNWMhjOc3/BT8C9fUvRSs0rkh
-          1FrbxaB97XMKyaaAB+ouePHn1F0CEL5l177jv7+P16mTItiU8iWQRge6k/K+Yzg6ooe/ydC4Gxyl
-          Bj60Y4xOlXoap/w2BUC61zd0E+IYsFPLJWBagxgflCiP54cxd3Dn/6hs3nq8vaBrQVo9Y6IdwTB+
-          RC6RAOcrNjm71X1cmmdfw2+AM6RFpGsxOgUW5J10JvZjkLTl7/f06OmSQNFMuux8QvL6TCaGOgUu
-          vyzXGmguHxMtQy+w2csdwh2PyM7vxgVPlzfkzcog1iuuY65CIIPseKuwnGtSTJoyfsM7Y78xEYfP
-          SDsEI5jdG5uYzyICaxYzFbhrWUWyV6wUPNd4/R+/RDekKi09Zk0EuITmO/5ZBZdUkwdaAGvk7vyD
-          /3qHVPxl1CL6Z2np/jydzL9877/5Y5Ogk3Y+SHTyWcf1HsapdKlAhQK/fQHMnMsS7vUBBdnp624O
-          7jNAly4jFr/P9FvMEwN2PCLJSzELlteWSs7YeUB2tqXxQkm5wcyNZ2KcS3ecF0esQLZkPSk+KVd8
-          1HBo4M4viV7cdbr9tq8C5eBE0b18aKOwrL0ErT594WnPb1qrYwjXJPyQe6e/6R6vC5xWL0bxvn8r
-          5zkmlB/fnGh/+V4v7ibrFDA40Yp3u8ngqkIHewHJmEOvLYdBCQB5ZRiDXAzdXm4eNYDtbAYsF9+L
-          VRGbCYaipuDJBDygqb8w0IyAj49BsrpzFvMVWL9hgbvhmLnU5qwKKr7AofNadRot/bsCiFMjoh1Z
-          reCX56kE7+P8wHN4jeiqz5CB8OQOWEiGRpuaZ9/Am18vSDPSC9gQkVRwxuIHafdkczevSLM/vYB/
-          UbDSv3gBv2ty3PPj1c7FETvw1epy8P2ev+7afaNAyu6wwdAd2WL6WCAHQfOdEGIZMSZ40HsZAYcL
-          5Nf3VfzpQaj4RRkcu2gt1lQsa2A+1TdR3ENQ4FpmsSSk23P/Ny4mAlhTFtLEJHe9s1vikW8Opnzx
-          SZatKF5aSCF8Nf17v1W3d4fz7n2sxndBrtD5lPWDNITgFztIXR1D21z/U8H1GxWBPAZqK1jDyZSb
-          8WcSPVBCly9gq0PKXTdUXvUh3oLGNuHOL4Jj+WjH5TEfMxCfEhmhqJhGoiBxOu71m/iZoAHuJR02
-          8OG/fbAOn5tL40FaoP7QnoHAB1rLj3OUwqP9SzFVuqSYp24L4WCVmOQFRzUqdbEHo/EWovPbWItx
-          PI05QJlzQLs/AuiT3UIZ8mGL1M0i2sIsQQUtY9XJef987ufbDAC3OCHa+HZGQWBWEZqcckEF+860
-          rblXPeBCaUMO/Zh0+een2L1FzEitNHLMCx7uepgoaTrHW8GUPeTaEpH0DbA7jpvCy4N1PCNv3296
-          eU4QJNF5CaSBMdpZ+Eg5HPS3iZ715o59Yr16qJT6Be36pt0YXknlMUgL5A6nZpz2fIYHzwqxdB7M
-          mHsW0ga0lVWQQu4y3flBCc+FSIgWEbPlpuNBhGZUsYF0Hrr4T49Br+dSYn2tQNuUsgrhQbp1KAhe
-          LuUBefQw+okCSn7fb0t5ZWxgZN1yYnr+UtA2W0oZyOROTN/NYxpeWAXO+q9B557DdJmrvodPx78T
-          dUtBS3TtkkPzeIsC4QBHd0z9BcrwyT6Q+oJq8Y+v/O7ohYzOfIwbLDNVYq7YDebmM8aLVjxquMcv
-          KpWDCthX6/6AUpoXEsTPK+jrrtsg5CeINF9sAEk4yYIkmo9I1wq95dsv6sCFPbpIecpBuwrfEcKp
-          aStkv/yTu8oshHDPN3Qy3XO73r0lkB9dtQTfK5eNlO2nCZ6mfMTsEgzacA9ei4zfukM0oNzpOt7t
-          5L96qo54uuu1UL69C0T84fLQppiJNxkULsRirH3bFRUvRd75APnzdyjj3DNJvod3lOeLCrb2Npsw
-          Dkv3Hx/b3u1UAQ3zNno8OJ+uxc/tQeB9D5gm76Td4E9hoP1eYkyRfi92PqbAP7wpv4Xl8itjlvLu
-          NxKE3bogUXliYfi2MhStFzdeZvpkofdl3kjJLYXSxlt1eVMfgJydu13M7rHMIQ70B3HcpnZ5rtF/
-          Ujd9ErTHv7sdb5UEpTuzINt+x23/dD1R2vkMBif4oTTKw0jOE3vXI1MZb/s7efIojjbyM+bnbtPx
-          jsH+e4mz+w8011wP+FU+7HOYm5GLUsRI9fn83vGL3+fa73p1n+rg5EtD1+KIrT++hA+vhRvXLPJE
-          qVzAAyns8a5NivBMAfpGPvIrKlA6nsYMKBuzIBMPXLs4t7yE+/4Q7cO37rTYHxM+4koiKK5LsDzP
-          r0X+W8/0yq7j7jflMOfDKylA9opX4DaWfPQPNwyb9z7l8cYxsCoVDUXtKWrZ/nCwYK/cGKQ81IM2
-          P4MTD5nlOAY0QyewDa1ewz99mtZRCv75w4emSzH3p6+Vl77JMfAvKD69q5g+7mEmZ+5lRjrvyNo2
-          ATeCL+V8JujJnmIilkZ2/H2lFKG4hmAsDrUqFyXOETriK53//DC5/l6Itbw0l0tKqYf7fhLXnNp4
-          vU4PCyIp1ILt2stgtR9bI5fzww+42hnG3873wW9CN+Lmqw24T5GzEA8XOei+cguW2wqrv9+HlCK5
-          FVT0GglS2bHI2Yk0l3b7LfQ/kHzIiWdc2u/6Hs7WoUW+vZzjBWQIgz3fCHqC7zgfNUOCxqCK5LSd
-          e3d8TJUK/n5PcHw28YqKQf3n1/JWwrf08ZsqGPAngJfL025pEc0TiIVkQLdBK4pN7a+9JMjem4SW
-          LLZr4mcWWH77O+i7fifIYH9wZFMLGWzKx5gLl7e0MJFIUHlcKUl9kYFa6F7JX71ZnB7XUP07ccwW
-          LeWHN6MD6gKe6JbbxPN1ejhgKUiB/q33hbunACcM3PV9NWK25QP4/mTlP361VkoowUWDLjlvZejS
-          7L4kUE1jBZ0H6emuWcyX0pApV6TGHy1eP0XESmqUZ8GE3tglln3AcO1dH3kGqYsui7NQuMxdjIJy
-          1UfaF6AH/ie2iDXxHP3HJ5DCW0hPm5NGHfVYglVhU5I+iw1ss8WGMOA1gBl+O7iT5KQbvLrtiBRx
-          nNrtnBUeHFNjQp58LduyuEo98B8wJKE4TiO5YVLDXZ/g0uDNceZHpEu3k7ARRTcTyo1TAEF/vNZY
-          0oQNbOcsDqB/dBfif9trvEQnBkLLHidy4gVjXFriKrDNLxzZ67fGLQ/GhJezExD7BrZi6tIp+auH
-          KEKUUFIrK5an5zlAXjuw7erAQwPtIPUDAanKyM26v0H/c7GC6/lqFFuyuj3IHCHZ9aE54mMM9qma
-          l5roHZRiUvWSCOdR/6Iq6CON3r9D+E8fGOmatHT3l2CAiRYcNX4/kPa2WEmVv0IgyYaqbfeXHUDQ
-          ZQ2WUsWIlzHV839607u6ZOxRKtXSCj4lOvX3euyNz5jAhjQYebs/sIHKfkPwgTb+6C+o4V1PwKAg
-          PnGOeULprRgkaccHYr39aqTzww/FnZ/+q5+z/ZBqEOnVtsevSlkGlxLkPwaLTDed9hO/WXb8it2M
-          NNv6FJM09NZf/4Wo5wsDcPIaLFgfvR45GJJ298MDiL6hT56fZtCWxIk9qMWfNzJ8bdbmn3+C0Lh2
-          Q9Dwhwgs23ML5b0fhklcl3T583P5eqP4expGQO+Cu/z9PQleJBh3/TkBYViuyEQHrV1y6zTBnqtE
-          5AzMp935dy6PQVIQQ/uJGnk9GfyHx/jwi8yRM44wgfhtOsgp4LHY+WIAFyYU0YObvuC1+wuAU5Ec
-          kCnI91v5vQ2+vgWDnM7b3OlqLua/eqVcnkOLC18I4KV6GMHhLJ8LoakMC6LDkyeBtE+l4zWxhL7T
-          nJCtz4Su++f/+WEYTqCNt/vr5MkC3FwslVQuFtsealhtYkOisW3HlUWfAHin5kl09tyAxdpv8gIR
-          jMlJvSfuaKBsg8YFnwJhx+fFXJsJpl0f/fW3Clp97eQPP5A3q8O42g+pgX7NnMkJfvwYd3K5gEVj
-          XEwPwqNdUu72/qd3Fp8tiz8/Ct5qPUBertL433qtycQEHH5ZxS86LyyUpwgRpUiEYumXsgd7PwxV
-          FtULlq5BAqX8bpNTWuXtii8Mhrkyq6iyVxzvflIOx88YI8vsIcAFk/zgXQ43dBYvjEvFB8MAeLIH
-          dEJbD1ap+i5wxx8s7PyS+EfZg8GZ/ZA8Ci6Uyj5b/9XHoG6J6k7XNHrLg1Vh4t7Msl29s5HLYvz7
-          kdNQhYCK3k8Ee75g9odn+qeH5Om83f78+Za1wNLAMrVStPvL2vz6ShYMbpm3+7VMjKXXhGGzQWb3
-          J1WNWOLYgb2fioLvGxYYYaWEer1fCZcdVm1RhFsiS4uAyJ9ft+NTCXlnNTFQJQKWQfqVYCliNRDX
-          p+9yO58EqS5/kWPddXfj6liHozjYyBAP80j6LKxlyeEa4sFwiie3vbCyf7QX8gjbZ0EnVCZg93sQ
-          MuzKxeSS1fD+I1eknq4OXY+HZpL+8ssK4p+7fWo7ARvrdFhOzt3eD1CwzMBuIBqz+TFN0a8G7RI3
-          eLSIDMhhSDNofo4a2vFc20LfaGBVliXJ1vk+zn/6fP1OLLrd6w/47v1NcF7FkpzzCRZzdAq3f/6M
-          vTQ/rX/FL0sWcEuCwy0K45Uatwwy+3j4Ze+Hrp4R83/4SiyoZ+2cfq3yb38wjYHsblcGdmDZ3Fcg
-          HXMW7P2gHB6kaxcIu/+91Yu2/OkxFMK3Ga/dj0tlrZm2XR8P49LjUYcDlihyvXzUFuahsHDvf2EJ
-          c3YxLbdKhYK2DUTLUsudvNXZoIMOW1AOYeoK38fswVhIh70/X7vL25kYMFCfI+qGuGJ6urr4/5wo
-          OP7vEwVPu1ZJxgYXbZbxM4F9pfDBfnecS/GZNDAajwOyHOEbb8f6lsKn3ahIf5JvvA2JzsvX6eDs
-          t0ydAEv4ggdW7FokmLWt3b7jnMP0q1yJloTnmBeujgpn5XJBipjrGq1ACaH+zo5E169fdzNf3RuY
-          L29GEZ4NbZbYLoFWYHjoirlxXA/P1ZHPbywSNLNlu/7GC5Tp622ge5R+xwWspAbJz3uh3BAegFpV
-          lUPZNI/EW1QQT2/z2sjrXRKR8VWJhqXL0ZGmzs6JevB5usw+W0GvefvBGrSfcY1le5FW/VwQZWxO
-          2mJt/CZnePPxhm3Q9mZY/GAKzZBo7RiNfJDILHz19wE93J51l/yHedDdNhVzRjS0K3tWfnJyCMtA
-          jpyVLnEjO0Cijx5ZuSnF1OKmAF7FziAIzkrMstfOkrG4jCTQ1/0W/nuoy5uc3gLOiOxWEIcol+jU
-          2Hij3QZIzSip/NL5jRhG1bicDvVO/tsPnc07ugjdG8LGFI5Imaii8dnyLuV3PM74Nt0OAP/6pJa5
-          8lAjH7yKglveYiA/i+9+S1X/askXCr3sS/mKE1VEo3ANVBW+8h8i+l06aGss+DUcevNE8i0JRqEV
-          1gT2FgQky8284A+s3cgaj12k69ezu1ybXwZil5fJefMsd2SjWwIFfJ6JdwNGzC7nuYLCnSbIUYtz
-          wZLjHUKrRxo5nxVeo7dltmB3gR9y6z66NifuNZLB4N3RKXDZcR0Prx7s+0EqVTu6kyjwDVzpxSeh
-          nzstFYc8g/7zlZFH0DSAHIWZhca2PZCHFgFsvtlkMheUb1L0wHMFhnU82EXTmaj7eq/b39xO9mUj
-          +yHvt7LIDw/yh/FB7Mcnb9f5aU7Q0FOfuJf5ra3XEvOQX0UdlY7wc9nkEvLwXpA7OYmH68h5hedA
-          MdZlco92hYzpU4dzIbnIP6YaXRyucWRmaDIS4WvibsqqYjjEihlsKEHtBoU8A2JsykjpLmqx/oKz
-          JN9reiX22r7H1TXZCt4tqyHKKpox0X+vXr6gbUTu9+drwvkbbvLr7roEXa5Xyv09v1T/YmS0wXnc
-          3uxWy6Z5S4hlB7d2u/TSBubRGYhndGNMGUFJZP2dH5G1rfeRvztidfQ7M0MG3ueoC1dVkZOraCE9
-          N28jf1LLFFp4UUngqIHGgvVbg604HlCEX94omKEK5fqUfoOFl6dxJufOkatN1tAez2BUjNiRNQ6d
-          A3k8DeN2lNoQ0qm2SRB6OVjZgW/kLsJnpNtBAtjKEyJwt5yG2O/nEG/FXHTw5Ewf5Lt9om388K6g
-          BVaJWJe74bLv+lXK7Qc9iG53BmDN76uEDNPFAc3ksuULfShh3E0pCVS0jjRt7hjSV2fg12udW3pw
-          MxMOl5caLFU0x/T8zRbI+bOKt+PyHJeb5U5QKMKMWNt6bMk9DSpY5btjv9fD9RcgCX7srEKJfC41
-          LAiCtN9iWpIgPbWuMB3CSN7rRcDSqxILi5NL8P5NE6QItuGu+xEJeG0bB11E+nB/3pTm8MdNGqr+
-          6gk9qyqUTbZA5uNSjGwsn3Q5m15fcon7kU6uJ6rQv6gPdF7Zm0tO8V5Abv2AtLRg6GjGtgmxtpb4
-          aJgx2MZYMuGy1G/M7s+7HVYtl/mJtYjpTlpBUV4rMtNOAKkNDl3WqcpU7ia9COpTcR1p/Pm9YXhJ
-          S6KdPZey/DXOZf3BHYOveu7jbTx2JoRbrZIk/500rvbmAFDFnskt0S/j4hVaKR819YLQo27j9TPu
-          7yzL8ErSTzm4kyVSRVbqN0FKzbruxnc4hbmChyCtbw1dxqYSwYvdTKJG6gfw+XKZ4P1VC8TNt1+x
-          WL66AdBXAjIXIWnZ/Nex8rkCOUKgPMWCuxUplIruFmxD6AOaL1EOL/G1RU9D+hUTbE+pvIpXi9ii
-          oRbctRlNMH0KkfguO7WrIIMSTp0Eg8PlytEtadoNzsPwIefuo7usqkZQVnKNYKZ7JUB42yqUt6g8
-          kILuc5uLUvnJ3nTyUVrfVCB8LdGDr74YkPoYbLpYHCvKv6LziM5AvWChJeeAH286cjv/SmlOsxqG
-          OS3JecvqUajEioV01AwM8u4ChOL0VuUdX4n7Vw+79ZdAR9I3lGYeF9NfuogynCsTPRLRcjmZLTro
-          IhGRiB2PxWL6jxoebshEqBhzuiqsZcl1+DWQIgZuy13LfT0vzilg9/zGavjdgFi+ZGSU089da5vo
-          ECBeR8Wjrlv2ut+J0FdLhm6bQtr5EbGe/GODA17VLHPX+RlgEErTgpzanV2sWOFb4G/qSJLp+yuI
-          YhSWbOm9F3DvOHD5PEY/sJzSjdj+5RULj68fwkPkukir8hhwaHIYqBQNDYTx9ytoOK6d/DxfSlIJ
-          n0VblANp4Do/CTF/LxTv8VjJ2EkLzLeHy0jvpTNB455ZAbUuUbFk92sk8TdlRDc1rumcXO61vMc/
-          lgzhQan6gxnQn6lMLFXc4nGMpU0iEt3QqT9HYP31BQ+KGx0DHtwjd4svkSe/DSkm6dPU6PaMGkfe
-          uhO/X6T6ots5tjLZ2JYHibSr1OLgRqFcZZyHdDv347/vg0RaN3Savbu7MG6/yIwlNFjE+jt+V0jO
-          IBZPNvEuPikm8/uq5P5xSQLBMspxWSsQAv7afrFUSYFGX9uZB290jUneHth4cTX5B6ITpci1U8bd
-          rOo6yfv3kWAo2pY6r9MiN/tc1Ih2Ed2O7aOEcXEugmV0Qnd7PYRKum6aiZnQk8Cqq4ooF1nrE49c
-          P5Te5q8Ct/OnQ+hhbfEqXu0EfgftS84RjVrhtSEe9K25EfvCpONffOxz5CZylcKb+4+/OeIXo0DV
-          7trGT81bFowSkOIv/sNxfcNU+nroPPIq4BPrxgMjqAn647tYVr8LzA/OH79E7nKjAEIlP5FgO3zO
-          dAnMQQGn51Yjr2MOYBr72JFP9kHbFRoY1/s8/+BH7ELkpVXgkm79pbC0PS24zecTEH6HUf37fnI6
-          3rp/eC8VOPkhV8sGuli+s4AJTvdg8btW64+R2sAdT4lqC1m8sWc7gVtgfYj+bTTK/z0vzHIXqV+a
-          FrSp9w7N7RPiQ/A9jH/1Hvx+Z4sYL9miVGxTB14SwyCOtU7jDGtXEjLNjfF6fjAxTZqpBKWJAPHO
-          WjNuR3k/s31bVFIehoP2F09wCE8MQj1Ji00O6vAfvp2v8ykmVNckUGjPCMsnM22pdvB0yKtTRPzR
-          b7SFVnUDI/V1Rci9vVrSCscUDt/0SYxw5tzphZoe+GEEgu85+7rzZzxU4JYQDel51Y+bgvU3VLPk
-          S9AjcQANx2MHfRe+0L/8yJcLhj/WOxD1cEfxGrPnHNhNFBBDYutxy2jTwcmjBgo08VV8g83qpPop
-          NMQzAileBsi9ocT6Dkl2vj3VzEWBytnLifkJw3E9GfcNfqwfxOLl/nGxFAAFemM8kgA9jWJNj4sE
-          t28j7Xzg3nLj4fX7V39X3a9H6nKXErpfjdnzUS+WBzNuEn9CAWZQ7IGFcesNHtfyRS5lvc+hf+wD
-          k5dTj9xQT4o543sVMs/8hXzTLYGQGqgHgezmxIS83bJaPDXgjz9dEPuOF5TACmZT+w1eAk7bbTla
-          b1joAYsBHVSX2/URjNnHgpzyZozLOF5C+HCRg+xT9QYLLZoMcsgzifMpR7qIisjA8jwdiG0SbRQ0
-          aDDw3CIWJ8E4aZPpXxv5D3886S2CZXaYBD4WLycX9b20S2dbEhzfTkzOyG1HYjBrIhsQHlBsgNld
-          h8RzYLDfsntSs0xbXPsoQvajh6SUoA2oHhcsXN8KQbfr/IqX5bXm8p8+KS8+iZeLcOtkmNofpHrr
-          2d2Wo9LBX87/iKt5a7xoIc7hFZcBOSVHhy5a+vtBerhVAX3DtZ1gcOwh9+WPgZQoh3EB0iDBstSy
-          AEwt0AaPy3R41dMMH3e9SgIzS+HkN1+k3N6CtvDoaYG/9Wa/TUsn7d5v0Mh+HWbOdRgvjNX8YEAy
-          hTy9ELWY77oU3IaNQYh31ZG1RKDCjLI2/qpxDbbL5c0A/eMUaI93ygtXVQWtmlvEqfC1pdzURnAu
-          RJeUbRpSOoxFCvR2uJFzztTu9tch3vkgUlvzpf3lj2z51x9xLNnQaALFTNJefYhO6jJog6CsP/nz
-          wAKxLr43bnV35P/VHzcMJ/o7ttdKVrVwIjfrwmkr4w4lfDYOh9nuYmjkkyc1/NPzl+R1By+7eqgy
-          dpICGZnYtMLV1UT4p1ecp9mNk1Ief2BGLQ4O8Eriv/ogtx/mQLRzts9JVUQG3LtuQ8pwXOOVJ70K
-          JeOUEfPOGS7ftj38y4dA5EarWCS5V+F1jd/EyER1ZA8qzuGn9ClCXtm7a/e4hHApAwfjGY3aWix3
-          Faod1xI94TK6XZeDB125k8kpTVVt+T7CDob64YBcnGnFlEgOAzV+cv/5FbQSU/Yf3p2roHKJbVMG
-          Wugd7vxIohTJD1Eaq5QjVtln7vZdh59UPpoSecbW0rlGYwpcJKHgMLNwXJeKQrmVLlkArq+x3aqv
-          3kF+dRrifnmjXVsvr0CAyolkzDt1aRnJb7DvJ9ISHQP6hYcft5Seg855R7Rpv1IB7t8fUPlwoqt+
-          UDG43vMYaVU3FfPy2sz9cvkOGZV9inkoRLn8QqckSGEltmvMohxqr1+I3Ofz0G4lMztg51t4kMot
-          XiWpect5+2KIZoBZw7fmosr8KukkEKtK+w03BwIvlSP08I5UW6Q20eWKrc5Is4Y2XtX7q5JPVVwS
-          a+kadzXS8C0XOkxJKb0zuvjbPZL621MIlrwyKOtPDwW6UCiIcZiDuB+MRYWm0go45NdEwzbX1jJ7
-          zAbcIEvQhjp6lbJ3AjeEvL4e6SsKdXDVOj74CfcEzONh6KHwyRAyOSiB5S9f//iRURSjRgZ8KQGB
-          ER+wZ5kDWDu4GWi99y/olC0BLS/xGISgjFBVmEM8R/1mAtE6mAhl6D0ub/NRg3v33pCjIkxp+5hz
-          KEilgy7z++fSPj8t8CPwmBioxyPZ9xOa3MvEh5QxWvagdhmM7nryhy8x13pRKY/NiPb8/WjTEem5
-          PE7civb40jZf/kIY1/aN+Gv7Laif729s5GkY9PHCuRuXRZL8SFkD7fxjXP3h+YbRNLjo8g10urAT
-          hVAGuhVctssh3u5nKZH+/BPzd6yLxX5n3Z9/gv70h0B1TZT/8BMF30NLjO/YSIlMM6SR33mkR7nR
-          //gZcsi2gE790Ul+JwUiti6BAifNuMH7PmHGW7o3+NMzf/waPeHrC7a9PkJGiDfkT9VJezxf6AeA
-          HRvEKtq1XZVVVeVCZ1KkVJFfrJdbXcvpBeboErjJSG60jSC8/faORqe5rO3oPdj9QXLZLs+CJy9p
-          A/v6/fk7LtV2gVbN9ifoVTGK5w+W33BRzQQf6IWC9aTGP9C85ogoXFLS1X5fIvA4bQ4KnpRpyYDv
-          Jdj9PaJruqUtBxVnsDLCK9n5WbzkJ4gBzoQb0eueK7bgBiDUq0+BsgI0I61bmEOJSAsxU5AWs65a
-          oiwGS4u3W6y33Mm4bHCPX6Sciut//Yvdf0Dhk+jj5m1RDyX67JEixxdKH7NdQ5aE264vVXcdj3MO
-          2Y8ZIj2GP0o+ednAyyA26NmDAOAwaXV58lZj14eTtj5f5x88KbVGbn4Tgz+8BGgZuuApiQLAHhdu
-          8mH5Fn/+Tot/fVlDk13fRA0/wbjsfE5+16oTvK93NRYuHzmBz+LD/fNbSeYcHagV7Ib042iMf3gI
-          gaN+iUdsv+DJWxRhEny7QJAPeruewZGH5gsiZKMyoxtHegyl+20l1tfh2gmspIF7/SR3zo3H3V+b
-          gO8yr50PP9tt94vh7pcSjxtHl34e0Q+q/jMgp+3yjInu3zoYfyCLwrXmYioqfQTqtw2Jzblxu3l2
-          3cgrRhdioFBuiQZ9BvxY+EJKJLzAVNydRtr5AVG1o64Jt/mrQgNXGvJ5tnN/r8ehBFYs/pDDnF0X
-          W8XtDVvX6MheL9p17m0T2JZ1I8lderrL4rRv2D6LQ8Ay8B2zn/zjgevNnYmxmL9256MLIF/4QFfu
-          NhQUv0AFbwx/xOLu3+3+OgPtmjrINIv3OB6lkQEC6daAwZNaLPvzwunCnQKRXlt346dfB7wXO6Kd
-          v9F51zN//IMgSbzRib12DpA+U4DKomWLeXMuNaRqFpNzdx/bPz0Hbmg4k8vO3za2uKug8+xzQNWz
-          sfv/Rws+rtcTuZut667VrFh//grxbyhzp5AouQyYMAmOotHEy+M0MEAuLDY4sM0GiKdgFlaz+/mn
-          z+de3d7SK4g3/EuWU0u5YXbgS2c38siSqmXLOff+4df5vp8J2PEU7HybGOXkaIswDR48Wc8DOZ8A
-          ovM3SUX417842SaO6V88sHqR7/ikahzsLEsaxWpC+iajeCnfTx4Kx6nHLNZPLbdkBwxuSv4LoP+p
-          C6GMuA7+xQ/yemXEX2PNoJwXEXJ4K9WGR7ZU8p9f7VIjG7EJpPoff7i+1jBedr9LEmXmSvS6LOm6
-          8vYGlHOQB1t4T7VFVXMId/wOJGW7jWuLwg7o7u5/ScngzmeRYvmClpF4h+GpUeM9Z3A+KRQpa30t
-          lrWQLbDwyQn5sjMW23tlM/jOmQC5dGi05T5PDFwO9ER8XTEKWpzeCgySuifpXp+oAUAAmf+Qdi1d
-          ysJI9AexkHeSJfJ+SRAQcSeIKIjIK0B+/Rz6m+XsZtmHbhpj1a17b4XKG92wJjwabzHd5x06Dz/a
-          69UnHm0OyrCz8Q0H7QXGE5dnKZA05YyTqoD0Jl6OKVSBB8lRkIk2/vl/6/Qgf/2CmL/pbPDvusEe
-          sngbO2cBh/F8xH9+wHQvjgmcM4PD7oS7uD0Pi4h2PjVLi3mKOXZke7hcXRxwctd5dIi3DYr8wAaj
-          Edxz+vyJHRyyV0YCbezAzykvyt64Oe+nEr0o5RHLI+tT3gNKTEb7uds4Qyg/GBKEi55zBTMWgB/N
-          NgA/qxlWHfotvFydCZeW/YyX03hRYeiHFbanMY437eDdIXciKvZY+egtPL66yJK0M9HsR08n7cvK
-          8OQ+HwHXPM2BtV4fH5jV0yNWGLraP33ePeI/f/ijbWngpOD2egvEso2iXra9sbrj/ywHgUa/82lY
-          APOrMpIEMfC29h6zECR2iRUugWApliyBgxPJ8/Z4nOMuQG0LSfu4BZJZexqR5AECXQ5S8tc/Wl3S
-          NnL46blZtj+JNxXHvofGiZ4C/j1v9K9eArgqzAz3fh0psFCA102c/vGD0S2LBAD03rB/PuF4Izn/
-          hue+14M1uE7ectdlFiaKPhD1avkef9KaEDIPD2CNmIy3tKa6IF/obaIWyt2bn3hUIavf7jiRhk+9
-          MQUY/+IlOGg+69H5RQu4+x+73j/mHGLjBhrT/AoQ350G7n5Ed3C9lJ+g71+bRxnBTmD7qQviINce
-          ppV3FuA8gih4np1HTr8HbkGOdef+8cP1e32of37vzu9Fb6qiXwEJUjLifKBS010Pob964s1+T7dG
-          WDqw18tZGMoVzHqc25CDThQIp3dM9/6giXb8x6p2kYfVaW4R2J8XaxzLxuMluI9wOi4XfKqevsZq
-          cVOhqvHgDO3ktfsbfQMPk1oTW9NdsHBHbQTSARXY8aUwZ9ccudLEK2rQ3PujRxc3EtHTVhgchFmc
-          L2FrKui3z5q4aD9pGLVwzuDJ8w+kmFFCJyU+wD+/O/jemUr7xxf/+K3xuj3zyfWX+U8PEuWl1MO8
-          6wk4vH8YB1xQabwsfFy4ilebnBxjqNfT79qAMxMA4tR3dZ+J6LuAK1CF9SAG2lgTq4fs7XkhVhkw
-          Xr+VP1G+irGM/eXu5+PvAGb4thMUICCN+YDl5waexfUasEM6xGuIkgodvV1v7P2eiR9f4h8/mFtw
-          27wJeGwgw6Pgkz+9RBstGOU/fuSY12igByNmoaRGJJDkUAJUPz1b2F0fwh+fB5VfOg0I72uBy+Jq
-          x+ROwzcKz0lBIrp5+Z+fDrcYavjBs622fbtvCD7lPtPnX70a6wgdstklxypJ40VL3z2qn7cD8U+3
-          L/jrrwApM31yVJtloKe7vMjf3/G7+/vKsCzWp/h/dhTI/3tHgewUPTGkx69eXnlVIfI83LDi5/VA
-          oxnw4JycOGJJ74O2qMwSITP7noL2dFrrqXG+EbooUCL5lv7A4NaXBowPaJHicTpTwZnUETqMOOIj
-          GwaUzezjiNIVvbGRPnG+3N4FC4layzPS1NcwvpTTCBPzpQas3p9iCk6PAES1YeHCbVmPOpM6w+zG
-          i7OYM3U+fI9yBrNLhLF2iTVv1Wxmg7fkhrAveN98+yq3O7Qfskacb93Ei1v8XGTVRx0/9CHIecv6
-          lXBrwiVYM6Wpp5PrqXBi4IJd9kzA0ns1A7HbhTOTf7/ecimmGYL2KOLTdvMA7ZUTD5+v0iDmZdgZ
-          gHofoTUXd2xg4gHucT1E8CqHR5LH4pmOL6EK0FEbr6S83KSaFmo0o0JuAhyXF6Bt6VWNUKawLEmJ
-          noBNe8gK0m0bkIudVrvj3DPS77vvgc+Wc81qN61EYdRzxOJUCOaotnmgZSQKeKi2Ht+0g4k2J3qT
-          4zd6DOvG9jLUhabBAV5Gur4kx4fm6xMFo5R3w3o/ZB0ErljgO3e/e/z58LkjU1VzbNoTyal1PIoo
-          l8491r4w0dbbbPHwmTiUaH7wpON0PkBQZFNMQpKfPdofKhFtnXAjen7ugXAJ+xl1RXfHqhsKgLqc
-          FiJXkXmCizKL6RiaEP5KfJ+7U0jrNdFAC/3+4uPsHmYet0xWANnSxsRqfv7u0KMStJfgFLDgoAHB
-          ipMShUMnzYKh9sPi270K/+Lx8k5/A+0Vg4W6Txpy9H4WZe2kapDwVTEJvcQfiFArIpKjkmAs0Q9Y
-          xNxfYL9WNom0is2XuVYWSB6sGiBp0cD6NUUGfeAlxWE+/DTucnmF6Jl9A6J1BT+QptIqdNfGjOi3
-          +llPcFRVRJ7ohnP1Xg7CQ1R6uOcTOQnVe+CY7LtB9hq/Cb7MLKD6qswoUIOaHPVo8Uhiv104ubAk
-          CTlGgGxkTKHoOYQoSvOL5/AThTA668z8fUUsnR74zKPLnGWk7BQhpnVw4YHtlN383i42GD3HFFHp
-          TBA/frdvTDfSpMgwFJ1c3/lT4w9c3SDFCSti9dJEB89cFZR4xkCUSF+0OcFti2J1YPb4QvXCJeLy
-          hxfz4Xk5Ae4v3pLXK8SPt0rirS7fFXrMTUUUMb8AwWWWO2LipSC3HmLKie1dhI4r33F4yztvURw3
-          gQofCvi6f555LS49whc6BtA6dfWmVFqJLjcGBey3l8GGhNBGL1GOgg3AF1j3o7tRwycAK4ftNPDU
-          LivAZFeD+Ly+1NtHRD28nm8etj6boC0voQtk3nYR3vM53uOrgO9Yrojjl5nH+a07AyFq2eC31D9t
-          PYrZDM7Pn0e8iFgDvx1WFRVda+DAUH7DavWfFuKiv83UKTq6MPLbRY+5rYhSbJW3mKmswGfi0Vkm
-          65Svw0lnIKWuiv0BHAc6vdj+H17m3+4G6E0fXYhr7M/cHs/8ob6pf/GJfY+U8RooFxndi7nA/oIk
-          SoRhTkErBU8czlELaBdXJkIoFbB9LY/echHtN9Dn9kn08GUN/NPsWPi4kJ44n7jOhzer9oj9uglW
-          kmTNafpG4f7lQKxvmQsoeKoyXGK2xMnjNWq07i9vmZEDhWhWkGvrb1sX8Cm25ywy12KYIvOkwJPv
-          W9iN7mO+Cb+uRG71mLB5nmLw05pbC39Rx5OrbFUafy/WO/rL7/W9WHS914iX7VS0yU0rabxeCdzx
-          XdTI1ZIqbzGuqQhJdyix3pzxQIxv3sLQ0C0SWFGprf3+jh2rjAvR8X0Z6Gfdp4bJiUH0SbRrnizJ
-          IvPQwbiMnSxnj/dFhbJT9sQW51BbsrFlwVVgCc6GhaEbnz07aFydHCfWS6WUHnkG7vE43/8+X/9W
-          fEgvwoKxbwE6RMUr+sN7crmkhie41kOUX6FWEJ+yOB+BWW1o83IJH7mrUfPu9d3AUYSX+UCMC1he
-          E9Xhvh7EPcm9t+S/pfqHn2b82IZpammAqMBJWKk8b+BSj7WRtZTNzN14K+aUJCtQXtoBzq3nEfDf
-          qGZAkbg9VoxB8bbncIJA9W2d6N/+Dtj22ld/+UusGygAe2O/GTACJ8UOkBXA8f7YwUZNfjh+Fitd
-          cMAoUI1mAbtkPeXcLeciNGjvM8ZcrNXrZqJK+j7XM77r8aotz+MWohcRT8RkErne7sWaoeJJAhK4
-          VhVvvjaUUHvuMza/XxHMumi4UDVcBxe996aCYLU6Eh+eiO2T0Gmj9bwWMBx6iZyGAQAy2EEGD9Jl
-          xjdmvAI2km8MVFmyn89aNANhBVtBciq72KxZ7PHX08OFTLFUJM0Tn3Kviw3hp1ie5HwOjx6nKucU
-          XvwMk3SPV+79OL7RndpHklk/OZ72n0G5OT3R9no+wXuSInk7ttiTzCmn2kkpITotCi7963HgQ2Hk
-          Yf1N8/35BECP9G6jv/oWdTOmW6Z5NrT8zzMQyC3Utoqab/gmR4B1vZ/y5eXIPJSdjxiAibeG1WTv
-          I2Q+y4fEoT/mv+fJviN226d6q+oMJsa8+Ig/OypJVlzVa2tdWKTe5m3mMrLU5HK2WnD53h/4NPzs
-          mPqB1KA/vmHViuKttxmzwF3SZ7B0uhRTJEouPAqcgL0jrsAqLKMM8WUdCb6Mr7h3WOxCLHRfEkgV
-          oy0f/IPw2Xpw7i7zO/8dgssCf9fTc5axlsSjynk2XG+8hJXJW7zxfQYZ3J8Xu2OmeqzMXUVYGqox
-          s6OLciIOxwIeLkEQsMfzkRJdyxPQdMsJR2cFxFv6K1jo4IdGAkNx9vwuNtgufk6S5zp5NDqaJRou
-          1f1fvKztW+zgjY9EbPvWLyZCbYsIMOJMrnNoAIE7fAq4fNjvzC79Ydj5ZSan+Qth29odAZrVPhyj
-          asUqZx1o9/d5H9AJsHW7WR7Fz2sDz5M+4NwVT/m/ehxMwYL9nz56KxvcAllRo08Ar604bLLHBjBn
-          pyOx8p/t8fdj2cL9OjHKDGkrCknwx1+JbS/neHx0HiOtc5QQg0XffB63NpDPqlnu/CHUtnf6YoDx
-          VU4kdVEYs9qQqnDPf2KHv6e39NI5lcPqVM+sffzkq5kdZHn8GTWxgkHQFmNbZ8QIy4y1rde1MS0U
-          Hu58k2ix9qJrWpozMB8OT4r+lOZbIfgVIPLh+68ebAmKIUpT3pwXTkdg8e9sA2/riGdWq9N8c699
-          I9+mkBJ7ZcZ6/ZwJC8uvEszVb2tqEqSyC/d6EjCfhfeIKTkRkM4KIpo3o3zsTtkb3sho4BiPDmAP
-          C6dCdbGvOJU2HazJWjFI8NUjUf1K8sa7+XBhbTwrbO/5u+HPYY/n+UfU2koH+tE1Fm7d7AUfkgbe
-          GvVGKK9SvgQ88paYmqSwQbtPOQ+osMYL1b4M1HK3IXi9XLVNPHUL6Ozg8If/YOlJE8AjjsAcnZU8
-          p+r1XgCXBTbxHKJ5fFkbGai/jx/xeN7XWOv5LOB9Fui8zr9LPYPj24fkwauzdPUtOr90GMKoeO9T
-          Zd0X3VDOJUA14is5jR6TL+7RTECf6zX2Kl7TJrlQ5D98xA6qp31GReZCJt4KbJ3ePl3k+pTAP/76
-          xyeX4J1FULoUGbneAKSro+cyPB9MFh+5m+TRcOnu0NVSm+DjVcyX3+WmQ9n5isHERLdhReHXB0Ee
-          fOfDerl6Y32jDLBDLQ/U7dKBNZi9CM7vgxbQ8a5pdP6hAGJ4Ckl+PLP5ZIT3UkxXPiFBEuL43++T
-          lLdmuuPrV7jFCywD54qz1zf0eChwKgw3PAZr6PTx5C3JiC7feiDBZ1aH6eb0s+zo2oTdb1zWq1zY
-          Ivinx66WTldlyVxUK6KAcfgtvWU1ijfUklggux6gtD90MtjxNaCL/ImJ30shEM2vOS+HbRpWv1VH
-          qNfOvuMqFYbJeH9U2M3rSozysNaEyycfTCPzDRK1OAxEvu78VhaqeQOXKd7e6Y8BxcEWsA2Nzlu3
-          d5hCz9nPSXwWZ7pFtcLK346mRDtBTyP6qoxQt10QcPOpBn/xB1pl2+bFA8c/vjrC+2d+/elBbbGe
-          1xJGWhVi/46WekmbOYI/CXeB5EHVI7veh5kfuURboAjIL+86WFZqjBVOf9DPOu47CG+yvNefU76M
-          ziGDmR+62HlAL19WOm/yfp34u/6fzlh2IY9dkZz697nexuBooqzKB6ILz+OwAdXS4c5HSCptDRhb
-          A7Lw+JmrPzzUph48A2SvYkrU2tA0lhNXBZRNC7HmFDZdGB9u4D2KDU7D5lfTslAYtONXIJ0+AeAi
-          Kmd/fgJ+aoZW/6vXTZNuQUcNLeeH3l4gGOUPPu3x+rceIGM5c2be136gUNMC9HrIevCHT7RI7AwI
-          kpphZ6kdTWh9/Q6T113FriIbwyi/qwK2xiUnefg0a1rOeQV3PMQ+ZBNtktzcl2tL3LD3ST/5mt9q
-          Fj3lwiHGcpDpWl0mHoYOZYhfNT9KnMkd//KFWO/84K2OHovwnT0cnP3xUS9iVRjZpjSLN2uis2uE
-          ChLQ54g1JDuU/PKqR4cyiIie3S/5yhhL8scHiSVej/HyrSwG1tJzw7r1UoGAn88G3rU5mznIvupt
-          LokJe++0YuXUTGD+uPUdXpb7O1jL8FMvpa9HKK6sMoDvpKsX7uul8ud8dWbIJme6bOCbgLa6/gJu
-          j//1/NbHP31KrMMy1tssOyIKt9OITU1X6603P/d9h3dEymw576eu/BYAv0mBn5c5odwpF0doDC2P
-          jTJ7aCvDqi1s1iUiEb7Ow0aWYoEl+ijzQaIG7Xkql+BivN5YozSIuWuLQ+hNKUdMzeS16fmzdbDj
-          /8xyaxr3qiSN8MzrC/Hhpgw8raUEeHwiY79ModZ+CsTIKAXxLLzCY8xvubj88UH853f8448XqWaJ
-          EkyiNv6+Qy8Xuf3Z+YE+rJ/DxYeh0y9Y0y+btv5o2sNPfz3ho+896Op0hQ3Pz8HDCoi8nP6ELoF+
-          AUaM4zIetgt1GFgtH4rd29rk29uHLLj49WNm1u+SD+eWhuhfvZ3qKV4boehBe1t4kg1LSWmQbvY/
-          /SyOAkeXIW7eaHq9zsEnb02NXQYngZE/vIkfnk75r7xGGUR5OmPL9BKwkFvco0TOzH/6VHiIdi+L
-          a2n+41vNF/SL/Kc/jKbt6QpHVUGvVbwSzzs03vrANx4Ol6H7q7f5KitHH/Wq+SS4NdZ4jm/xCE3z
-          q2OjFQ/5yEkPEdZzkuL78x7SpdOGDOzxR06yNGvbH7881ueA5AH+7PjjiHLqZC0OQTTk29aUAVz9
-          0MOe8KjqpZduKczePCLu3PP18sEvBq3f8IZ3Pa5tLrRGlEt2gXOG3MD2pz+x1l+wunzRsA29vUGu
-          a815fgVVLnRxZ4KnqhQ4Uq+uNn/EcEPfrrd3PqxQ6taP5p//5F+SxpuSNLDBZnIh9rOJ1Bv31hdY
-          a6lD9NRltelUXFr4nYIVO3P8pdTHZQGL5xTMEmcw4N/9eejhoLIQW08b+xbR9uQSfP5VXs59JrVH
-          g63LxN+Gj9dXFYXo019O+IhCGm9Ngu7A9LwJ+7GZxhwKiQ/frpxj1WPq/fz50Yc39cYS/Nun8u6z
-          0GCTPwB5pNuvpje9caH8kI1ZNK4GZVd6lJHyFCJ8miSnnv/8F9cS17n/ylSj+TIy4EyHkmgPP6Rr
-          OKYzGAkMsb7n27wkFxOd5eGGcfJx6/lxvo7gZ5g22fPDWwftycJdP85U40i9VcErQwO0/AB+BmXY
-          ztNhhq9UdLDtfVs6ehFUwdn8WQRHFycXfsWpgpZfnmfkbIa3LS3pwa43sWKybt75d9jCg3SdyfEu
-          TcM26TkDpaVoyfGj/LyNp3IB7VVOZ/6KWDBfrl8XXD98S0yqvnLCPNIUlErTk+dlPMbk2RwrmDZ3
-          Fhv1k9J15y/Qd7cLPlGliunDjSOImoRiV5Jh3l5PDxs+zt4RG5f04612lTfAp4FGNGxcYo6aWQdy
-          yS2IkZGwXgs0mpDrGhMHUOA9umhDBbW1cYgCb2pMzUdbQOvybWZmVXKwVf1LR0ccghlYzxflm7bW
-          YX1JInL2hjwnu58u/eWbr6PvsB8ex8DSIZB48+EMRu3mFcAOjme8+9fgT48BXmsS4npjkS9plI/g
-          ejmw+GS9a7C5tA9hQiwU9MN50Prd3wH1klT7+rzy6Xp8u9DOy/Wf3zMPwQThxYIstrbF9kj6Du9/
-          +DzL3Z2raddeK1j6ijKfjVsLhk+aZn/6kPh46sGOJ6w8ve5a8DQu+v7zmd93RPTYWn9rTO2Pa8tN
-          k2x/fsqwWlmugMllyoBV7762hoPEwgO7GrPg3941iSPogrTJ2OBwM3wqWEdHBLveI4HpxIAvnKsL
-          d31JovKj0E3/bP/N78folfFeL3V5/d6+xM8mXO/+yRsyzPIjdsPY2nb61jPsDzqHg8Jy6q1Jsgo8
-          PaXCKiz0WIhmygOOt1nimJyXs9vVDeBRm69/6x2ve78F2FctDbi9HzBFRuWijBXMAO76cZm5PIHQ
-          kin+p9f+9PKff2En36EeDcNPwO4n4ahnFjDK3FMGe/8keK8zm69GeEgBXNIXCcZKqpf8J77RXzwY
-          uz8wqowYwqj8DdjvHtd6TBXHhoopBPj4vGEwZvckhLs+mrezksfbeRJmKKeiS2IryL2lzXAGR8KE
-          xPR1K++WwUn/+Of8e46utzIZWcDjSygxrVaMJ6G8pzB9ogyfrlIVb6uwlIDSc/iPj67eUozwnJjf
-          nd+/h+3bMBX8dp2NH7Z+o4siFC0Mut4iRssPdDmUfQG+rzWf4bXN6vn3/pgwNPAFe7dXOax8Kfjo
-          ReQT0Zf+Wa/cdivgV71sWLEfV28WflUBoZa4u59/GNZ5BpscVvqFqOWnAguEtIXR3VCx4a3PmF7N
-          cEZ/+fOnvxekwwC+X4+AHO+R5C1o+zVQe03XP/4GluxeRLBI7B7fHgxD5+zymCG/TBIxOTaLN/FU
-          bTD7hhkuvbAf9vXaYFbmLxKoG6AUy9sdFnIb4EDdcvCn70D/wDKxJc4Ec3RQKoTEJcCFjr41jYpX
-          CK27ciAYal9tifymQdL9eMQGb56HP/0N/vyP4saaOY15NgNn+ivJcfeTtqj+zFC9xtUsLpya09jB
-          irj7mTPgCJd3TkttoOV2MwuHxR8WCtQS7P0ZErCMHc+6+9zAd+1OWL1Vj3r3U3vwY/sDMcfTIaZn
-          sWPk/bxYgpG/5uNZHxUYG1aIsWZow2Y2gSpHti4F3PdRxfTT+Tq4CgjNHG+uw2ZZrwLCOb3ufswT
-          0FClCjJbZZrpjX7ibazyHhiVV2Hl6rWUMo8yAZsafYP6+rkOS1zeI5hfXBfr7SusSdc+9xmI0ysA
-          n9TI2doOIjg6P3n+ypaiCfW+w2THD3JGYRxvPRttQDAU859fTxiiKygVeUyO8Zur/+ofLA6uQHBR
-          6Lu/zbEwXdkEG76Y5nNb3HQYrL2Pbe6i1+yh7EvoL+AQsGzne1uC8n2mtmYGfXQf47U+Hnlk4q89
-          jzs/67XHpsLv5K84kMq+XkxZDuHP6dd9vVFNsV/ZqHP7M3HUi6PRuxzpaHrV59P2ugqANqcFgkBe
-          S6yIOQcWr/mNcBSZC1FHw6Jk80YdSv7nhc+SBOLtpbMh2tJRJeVT2ep/+nbX37N4DiptdMSKRZrc
-          nrDvsy9K69/SQ4eRR3zc/Y1tDBwdnIzLnfz1F9ajW6dQKsp4pmkxxeweP0gwVJN4E441/nAbxD++
-          PTN7PNEs4G249wdwyG8/rd/x4l9/Vo3mtu4nPYb//OiTqBje7vf1f/otYMtOBdtRLlt4uG0CUeyH
-          oE2L3esQqQ8RH6nVxLX6+yiAunOy88HHMN+kB/zj58SDDyMWllPYQYXMOnb8Qv/XTwHtN0iD/HWY
-          8rW8RnfIrY0X8DWLtfVzePhg54vB5pefel8fHvnCMpJHt210NoQqhOfcfGNXo/MwOg9TgYrMVf/w
-          f0nn+I72ekbMcVPihZ4eJvwwS4eV71vL2faHdPhyih+x9/ys7Y9qA+JkX+J+Y2booYD+r1MPwP/e
-          UXA3BH2G+xn04wT4BYWbB7ElRY+BTrHWgs6YfthocEepzv9cCP2Yn798eqFUto4tAkkjkPhh1bTz
-          pG8Lp5slE7WT2Jo+lFsHYWGx+1TlPt66ktpIm35rIAYHZ6DPJxrlmWY6tpgTm3ds4iQgamoHe8bd
-          0thEYXWoj+SLzXCGOX0+jQa2M6TE9FcIli/52TCO6TGQYH7MuSWsXbSkjzLgnE4faNO7HbA5602w
-          pqN6vj3IDMPgnQfMwWI8Cj9cBa3S7XGQVxWl13bmoVt/o+Bm+kdvG4dw16dnFZvPpzJwXbuZKLwX
-          ED/s4TSs2KczbI6aG2zo3eVCxhUV2Fkk9r/2a1g/UnaHV/aIAkbWm5qu5bKh3po34nG8rNHzSHh4
-          Sz8+PvrOOV4WkYngVcke5EFNUu9/n6GHpRskj657BXv9fIiY53Gm1zOvrfhU8+jojTU2deEB2ukn
-          BpB9Hu/42I2Nx10v3R0Vh9onXnMc8p8T7buRIFsEvw9lYnrbO4z8WW2I2RS3nF7bloUXg/ED3l8L
-          wB3qi4mk8m5h58qQfGH1oYIcV1j4cX/kgIDvYwOzXbPETnnb2z77OX787fwjkXZzNDq9XxlS+yAm
-          dvv9DNwLOSr6HN5SsHqeBOiy4gwa9Xgn5bvo6s3LaAI/xbHC/qnkNbqVSg8Hl7VwPvIvj+0Zy90Z
-          UzBzmyvWBBq/FOqneJoXXT0C9ht+3kgyOUy049MCbBKeZTjbL5bos2LXS+6GIhRGRSZ37ufUXIWf
-          I9zjhRRf4+XR9FtCwL2WEd+txM6FJ4krtMxKSpLmBIY1EZEOrwMjzYe2FrX14/IMMjU3wcboWFT4
-          pb4OscwpxLRZlBNnzUKYRLyAvfiSghH5lQk3ffXxI26SgX0KagaHtGlJ1MCLJ/SXmIFNtuiksOrW
-          oxuT8GhjTyZR+LuWC2laNX/5R9zYNuMxHy4+/DV5HpSyZWisHCwFjIrSIP4jSrRtyVQVQSHd3xGf
-          a20Z83cDq+Y04hOCaU7TtGvQng84SW6o3sCpi9DJ8K/k9lselFueGYvYLd2CV6JgSk7lkUUEJw2x
-          ksnRFnKLArTp1Ce46OyYW54hDz+FVhFruRgeP7t+Cc+cgfEzplG8TIBZ4AX9POw1h/fAzZF2h6g/
-          dAR3RRtvsCx74DG3Fp+ZMooJ570T9L6XKb7Ix6u2MKPu/j3vjKwlrRdl3WZU/xKX4NN2HNi7d1TQ
-          /Uw/WL1apidU5aOEf99f+PucBn5y2RaY7y0knnNt64l39x0atzPc9yA2dAk5k4UfYT3g5HGmdGo8
-          osPJMFNyKlJ3v9+lhJx744hlyG9Aa2CPMNaQRizR1wY21JsZiU8Xz0sEOrCc2XSUKp1u88GTaLzA
-          uXjDl2AfiO9Vn2ENHuuCXGC7RN3G10DZKd2gqaXpjIASDOsU0TeCgmsGK1NG+VbcSQB3PJoX5vwC
-          fPTqVCBfJg17X43S7VYILXoBpcCpKp3yNfS9Fo7n5YCjBnLe9vz6b9hb44atug098vlFb/i3vset
-          Z3PKgyRB7VacyPVbk5oyzSVCU4GuRLucdMB5lWuLM5F7rDm3FlD+HqowwV6OvZeqaZvwPmSQ5LFB
-          9Fn7acvjp7Mgt7frDPNzrdF66EV4nlQPn57ZNqyq85QhzMG2TwWvhr/vH9a3t09uhveLWZzIPURP
-          zQpe8H7U+O5ZR1AhUCdaw6gxiz9Ki9orlshpX6/5cRNMqFerhW+ovIPNcx5vyEm/idhMpGobbroK
-          ZYNKZoZzJtrLFVHhZRJOM/I/fs1xAGbA9cYQRwHGgFsPkwqrMLCxa2QSIPNLiiDTH0v8FCKfrugz
-          FRA+0Rk/PkEVE/6eqfspkM7M3hI2p7n6VuFNBSt2hLqn8wEcZTQ5T4b45icA7GFIVBiE6kiMuE3A
-          2opVBL6tMWGD1UzKGa3mAiR/G2Lo04luzXtYYGt9EfFOSett53dooxNbYBJq7pCPNT1vKADUxY5v
-          nT1B5182isZQw1rhjPUqOo6IIKEY2xgp4L94Cl4z1pOi9OZTa29QU61LAAe1GoRWtH3osF42SzV3
-          AuxvVmSonUGMvdpMcj5Smwy07x7ja/HttJWc8xBKQeNgM9k7tN9vH8DzmYozrwsPupRWmIGL6Qb4
-          mQ0d7cWwW6B9exnBS35tdE0cb883PSVG53fxfIyeKrxzEYOD5SwMI27CErF3BHBI6wcVYie20d3g
-          dIwjLtd+X0FREMv7mKS3evPGWC0S0HsNxkmWPjz2xwgROovMSPJNOIPtIN8WdLls+ykkFwdwsfxI
-          oXdvdZIa4qUela1nwXedDWJP9Snnns9PBw9n9UyCNrUGdszfLYLDtJDcT2dK2XyJkOoutxkdyYcu
-          Tdyq0NSSdL8eAPouuAD+rT+2BRdsKQNmcBWkO8739Rwu4kNBWoAFooKfrG0ZyguIvOsYoCr+1WN8
-          3lz0F09H11Zy/n1JU+hPQoXNz30ZNq87u+AHmJikY3ylq9rlb+QeLgtx2WHLV0n0FmjKXExuRuBQ
-          waJdh2LFaOe4ttl6K62ukPkR6VjteFbbQBBm6KKaBGM2bPJVNtcK7niL1S94aAuR5jscz+iIbSZ6
-          e8shZ3xQHbojue38Yl4yV4XrLyAEByKoKdKPOpqad4hVwkFtdd5R9YcHM2qTe0wmet1ALvpnfKaf
-          wVst/WAD+xwquNjjkZyHwyjv8UzSNZiHzpWvo0SYpgjgXi9XLTNaaUF8gvXwydTLMHYdvK/5K+Au
-          hVrzYXJvYSGEFonwyfL69uW9UVUpJlHO6DwI7zFWEIdeLjZe8VEbH7bYoeP1oZK0zb5gG0a7gW79
-          iYjy6WZvrFpYwHfaSNgc4yvYQh0ugLmfqvkd+STuH7eDDic2DWauTlY6kaORQPdJWGzx8SWnmb7d
-          oeawGN/kN9UmX0nvf3xwll7SM18l+7ZBya36AGqtOXBfYXEhvb4fxP2IvsY9Q4uH0Rhps7DfbwNP
-          c/njr9h5C2m9Nh4x5Z0fEGwd2HzjPniBzNpeSbxwCuU+P7mFYW+KxCq+nbd0s6TA83kV581JJO/X
-          yPYbHJbuNHOfUKlZSVE39MePNEQyuuQyUMEfv9DjW5NPasbNoBIki1z5mMvnY3RVga9aKrHG7DNM
-          22NpUFkGzJxLi7Tzm4KFgA3ofCTD6C1376gif01kYg2PedhGrxkRrLspoK/jm85qVbSwffNnfJq9
-          /VSo+qGDvn6cidndRLAoW8//5Qs2RMXMBee1FQDq6oq1SvMG+oIdA6dbDIkZd29tlrH9/otP/FdP
-          BH7g3/B+GVSiBEqkzVXKtEBnwyp4ed6N0tveQdCzWcPYOiRxz4BNAc9HyBJzw2a8Ou/7G0T20hOf
-          E78xWZ6VCP/iN4NpPPDubSjht7WmYPX2d1wyLnnL9esbkv1UjIGN9rckzpPi4cChCeXAlYh/eijg
-          i7ecT94xqKAlSQ72dryZZ9cv/uodwXWDwdS1mw79M1CIRuTf/k4V7OQyyBVsHTxvWOzBaWEtuTY2
-          LMLHlJ3KBaTH9IuNIVjoZkxvBelHxwu6gs3qZTOBDD80u5L7MZfjKU2zN+R/1jfg06/7Dw9kaokv
-          7Oz6i32xsgL/8NLKx9XbuObawYRPQxLI1sej96YswNkZWeyi2h664NJvf/wvkF7SId6iyN290fUZ
-          qNWx3Plp1YK5wJiomfWMyag/ZCjdk44EYyJ60wn7OnDpKcBpMHk5N+7nNu/8N6i+5ADGlHDwH5/V
-          rCqrR+Lo4R8/wb7xFeoVGucIonRTyZWNmpqIv9ZFMfUBDq7tC9Dn89RCeq0eJJD1Ztjc3zEClpr5
-          WAH3Ip7C/JVBpYh+2FPeC9juQDOhiNMzVvnfy1va1cngTZVW4oxfLSZM84gA+67Y+RBxwJvMT6zC
-          MTBqctrxfTldNQVWgGmxyXmBx/ejOEOOK/ep0d8wpqQ6K/DltW+C76M0tEV/UmB3Kmt8LOE1357X
-          dvzTm1gJmF/+S6SPD2HdT/NWMnr9K8MDAzpJU7ArtY7GDkU0w4BWBrY129O2wlFmZA7ahRj+dwRE
-          /2oJDOrRwreXdMi3/hGJ0B1SP/g1SwM2ZVMXxJ2ZOdg6KRlm7TmW8n7/mWEDwRvZRCqhs8SA+PG9
-          qfcduiYUvYIltxzMYHbdUwcN5XQk+QRzbZlB0MGSAwze9RsdPLKoDOjC6E/v06XZXj0M/DrCZtbl
-          8TLCWYb7/8fBn/4xT20Ami5TiK72ZFj/8kd82hinyJi0VRKpDL10/mJtFI+5sJKvC6T06c3fqTwM
-          awg/M7jdi19wuI+3Ya39ZwDrW+VjhTkfwTI7bAa6Y3LEJeec6PKwlTeyesP740/aql8vBayKpCfl
-          zi/HdJ9ZsdcXjLdr6k2G+cyAAFgRn8efNrDX+TKi9REFxN7yul7fJ5zCoc0acg61JR7VwxKhqBpG
-          YmafO/ir3+CGoIqL4HkCf/wc5llzCZbDNQe/q+k2//LBbfk6XhltVdF0VoVZ+i2IjoqqjH9+QLBq
-          CYinP/xf7gnAuvkl3sZyjxGEksiTqyO84+XNOjL8lsGw71iRvPUNugU+TyVPMMlOdKuMzERKEf5m
-          sVhFbW1vYyfXzhxg73z+1Mu7aUxY2T+Md71ERyLNGfR49xV8dj9p2h5igx4R3+z+T5KPrsO18KLq
-          hKQmPwJKjpcMekzeBjC6jZSOMR2hbqgddl/MpK2kkkxwCW4xSXc8WvBHaZAz6slfPYq3v/wZXy3C
-          tupSSj+imADv/bNJKtmn/wAAAP//LJ3J0qowEIUfiIVMkrAEQWQOMgjsAHHAgUETSJ7+Fv53zYKq
-          hJw+5+uiM9G+BTZ8bxc7lMXCMGYVVibcs2uBLE3lALt9PVvehi8HORc5TKg9DndQ6eoby8N6C4dC
-          IFWaC4iQk2uFJ0adXEHYtRPxDq2cfNFZH9XKWf8RQ1XesdXfwVo3zgit+YGl28CF+WZMUXgck2RR
-          s7WDtmU1Mss0rVnWS6ESz0mPfLzRPP7nR0/B2UHho9p683NHOHDJbgrePsWWsfIZPuA7Nh2SL/XG
-          IOIUVn96YV4jaMzB68qreeMCLNnbyaBo/1DgRi8y4qvUSiR87NJfHvvxmo7uw1cBQyIOIdiy7cSu
-          uKvUuTqBcKNrfj3HVGxhqVINtRb/mXr/HcG/vISO5GZ8nsv2Ckd2REQbX/uaJ44Zw9VvolxXjZr/
-          +P4Md5YC13pLDXqxNF+9XGGEdPmsd0zlJx/i9wYjb33/MFrtek83KYj/dOLuy6WkUtbnxPDOssGO
-          0wYr6cEsUC773UTVr58CdVR7FI7JwWBGcR4hulg3YnqXfUI/3gdDIJsRsZp7VdNsFm2VvIMXMgc7
-          rj+HplbgvnFuIV2wW1M82BZsj3KFDunsTLyfAxecguSAdicnZC9N70fIqXiLAnUjMrrOxYbHrx8Q
-          /7jO4Isn9AA4sr544/NuTWhAsWKoZ0L8PO2mnmXDA9xpViDPuMmMikadwsG/l39+bz4RgYOHKglC
-          1WCSNzDOn6FrqKc1T9lA5D5NCHN7uBPt/dnU5Oh1VzVOiUeMzXRIhJJvUvDz6wYVjx1/ezY+9NzZ
-          Qhc56BgbTtCH55NiEHMyNfBNHObDVx5sSdoksFtKQwlBiW3zv78RXRqqkhY3yOlHgy2SQzWVpykl
-          rpNtk5lZxgzvY5SSZtV/Bt7nGXjIwQjtgeEty+a5U58JrVEoRZduueL6BLZb7kEOQtFO6/p81NgE
-          AVbk/shEiWY+1CWTQ/tncPWWt2TvYLuL21CpTGz8/BFIuXeGdpmde2Ia3SDMhMIi5qyX3UeOIgxX
-          vo270n0kH6+YYujhd7/6+dCTXqX5gjlvqCG3tYJa+vIHGa55E7/gIWM/Xgl+fNsfzaGmh6bj4G70
-          E1KSwWNXmQY2XHkqCgC3Z0tc6a5qOCJCeiO/jDl8yhj+/BJa13t+LssVXuaHjxqVvmp6tqOHKhE5
-          R2hqQyaSQrH/9MTcDxIbFY7eYTWkZ+Kv+i85N6WFb1V8o4NQcNNkJ+YMWmHLYdk1Y0AGC0cQR+aX
-          BCufEAt/SaGsyJik8yYCtC0tDvhCVeKhUvMJ187Mb2/e405c+Fzq/nXMenAtzgEqc44k/e2Zhj8e
-          Fr6FrDVw8VVHGKlzgsrtA4BVbz/wldcAi0uIO5xfDhzQJYsj1l7IvPnHYzOhskJ4On3qmQMDBJsg
-          FZDOnp43P+gwKovgRcQzG8LmZ6zKkL9/RMxEtnhEGtsIYlaZxNGPY72Yh82onDJth8yNnINPkHsa
-          gOQdoJ0wDYzes1MK3Hf8DZdN40zzRtmKcEePObJPYm/MF+saqRu9yvDC+ml6lf469T8CTTguZOxW
-          v/yAv/qMtF3mCRY/i/AjxAVxAySvM7xoDwpihiTNTx1bOGO7A6W1hFi8XLROPJFMgZ0QTxjMTJum
-          732o4NDHcsjNR6lj+ObYMHJ7jphEvyZsfGxn0FbZDXOnp+sJQn0NVdkbRWIp1oMxh2oFRImrh235
-          oAnjIXzBfNOnJCirXS2Ne85W7m5fkN0h6Q163a+/6B2UkFhoune0qd4haGOrQTa+mRN/22uacpof
-          l3D+PDKwzNFkg63K6+Qs7nM2k17eqdyNlkS3/YVdvd4x//yukaJvR6oOp8DY7TO079IjWKa0noF5
-          Pkzr+fkmX8N+019excp924H5M2mFuvaDfudxWsrUduFLFASiazEFKw+lwDLslLQH2rH5V0+iF6pD
-          erLChFqfrQzzImmxQr83Nj3qhlP0fDKJ4663bhf+EEHapCZqn3XE1vV1IX8xKiwkHf+3X3D1K8jp
-          hIDNjvv903906ssCkHZza9S7NQOiO+5g/PgvWII+Rg64jjXNtEYDq77hbaVK0zI9hg+oBBOS3cnB
-          gF4s+QStOsiJpuQ7JuGb7qrOfARoX5Y778dj4C0ZXuF7vEg1FWNdU88Hax9udK6blg1VFMgSYYei
-          ++eV0OA0cvAyGAmxhWpi312hYvA9atLqxy9/3z/UZMMmP/7L3Ll3//Yv3ItZgtui5YEwSISgNW+s
-          fHmG6mneoeM+cDqx2d55WD+uIrHPlDA6hE9fXcpqIaEUXAz8vLERvpRvTJKb2SWUf0rKn354eSzW
-          s7KJHzDpHhu0N79fRsV4seHAB+Wad7pkFtmzgoXCf4nbS2k9R931pOzioQ6XfTB0f/n75yfklc9K
-          rLlW0Ha+B7KziiWh36oqYMOqkXgStac5OSo2fJ8uC/rV64VvihNc+3GhcgN+JyXFzAEeoQZzUnDx
-          lp2TKxA8KgGhYurB/Munv35iEA4qY0srU7gZ9ik6Wt6VLbtC+MBVr7DsGLXHctEc1R/PDNfzhn98
-          WvuYb6RZVuKt/h3D4xJQhADyE8l81SEITdMjZfoPAAD//6RdybKCOBT9IBYiIEmWzLMEwQF3gIiA
-          iEwB8vVd+HrZu15ar54y3Jwpyc3O/JvPA7l/l7D0mojbvJdThkwl8Ta91aTra+8YoFH9C1HuOlfN
-          N6Zvf3n3xJzGiE7TV9LQu+dz7HZKprKXuTFgrO4BPnrvoR9K1hbgrVDu2ym7XTS7rWyguuiNCb39
-          Ih19P2Wg+eKM3/xqRL9GE8DBkAosRYdCXYiMsr988ofvQ/dQWASGr7j5w6/7p4ftoN9tebDU79MV
-          apCY746YMlKjYeyCALmiJ+En5la60GPXQu+yPsg2/1n9xhsC2R76B9uywCLYsgBH+8EQ++f3f/V9
-          s/oSO9t43vIkCT4iqfmNP3e8a+UZcvftlOkt32V/fkg89fUknMV7tbcX4fybr/nha/XDc6iE/AMb
-          9nruh1lgAgE+dyfibn8fefFsodP3OmP9maxgEfm0g6VyyKflyfvu5r8VuLwvITnOmumSvpbbnx72
-          Dxs/0scMZniE5Ozzm18cu/edhUWdTuSPT4STxYLdPbr6iNN5sBj6JwQ/PmG2vIJOil7Cxn0kWO12
-          kbpO9y74zT/5+3dQVO2Wb0NvH99JAC80el1uVgttIIQ47IWsWt5fsYZ/9RQdJPWX96Kx4PZEmgot
-          pT89sM/NwAedAt3VPPoW6D7niUg+Y0ezcJI4mKnMvO3oPad/+vJ/rCiA/72i4JRpDDYKbazW7t6U
-          8KwHIsFS+o0W9g0MOLwziYSTSar1lF07mHLwiV385qo52t9zZCwaJfnBM/t9Zd9WSE/sTJLyJkRU
-          F04BLNpLhXX6Lqt19E4NvD40GauP1wHMmkgH1L24YBLZxUj37DtIxG8091i724Y6X76yAEPe6LD+
-          iXp3Vm/zivQkzIm8pkcwh1zOQuNsDX56XY89f9FeIiTG6ehXZY375UCgD8i0NkS92KU6wBcronng
-          XGwet3Op78a7BajWPGKhdKzmhLoKzAsn9XdtbarLx5tDuKptQSxw8qJpAIMBU5B6fmuCW08ErRLh
-          w/5kE7Wh6/IleRbwMug+vjKslXInx3bgdFE1f35oC50/ni7C56nCxNCvKFqHY84Aac94BIe8TPn1
-          GzZICJmcnG8y7qlo2zNyhHNEYuwc0lkQMwVyiShh9RqM0dpGRYPoadBJvotpOhA7m4WTh78+ZN5l
-          ym3d11AJ2JZcwt6oVjbdM3DNKtEX9yKM5tfM1ggFckH0MthHIw+nDFYxSLBxHS8pd6wUBk0mDPEj
-          rSJAerhYaOwlFyvcI+vn9TlmQGo0ntgHXeuH9fnOoeZtCsvj/Z47smiGvC7XJHztY5d3FidE8Zfw
-          2/tt3LVKIYTG+XQj5r2ZVGr5pQTzIjKmlXqzSuGgTSi4JtZ2zqCRckmXQKB9Tz2W30qvrm/8jqF4
-          ZGzi55nX8+xDVdBIcoG4DOLcVR6KGMZorUi+7OaeFN9TjWJGPpDAvN+itnWtAt7I/kuebQt7ekch
-          C173h4wlsMqAjXdhjZ7MuSOpFu/A8smlBBbtrSIOnVOXrGI3Icm25m0FCVNt9RIgsVcAkUjRRUOg
-          NBmqbfFI5Af9potPfQWaTcvgSG5bwC/x6ECbyXQSrWbUs6wsJUj7Rj1xpHOTLuMBXiF7C4JpeSuu
-          +q3vXSF2XIBIxqav3/gKITMy96myeKfiqLAycKRdQOR4asFc1W8Fncc5IsnLL6Np0ZUaDodZIQr8
-          POjqkrBA7uU14xNDApVN6HtCzwuCRBOYV8XeHg8Jdi8IsC6ydUq0T8UieJU+xE9dSd2j28OA/JiE
-          RMnmW7W/s4GGKu2sENvhUpdnp8hB7iVM/O8iHHuO918Skvvyhp2n9nL5OH2E8BNIATbPlRDNuvvy
-          4HH0PXxywsBdlcDQ4Iu7W/h2roR0yZyWQzjr3lNwoe9odoV3iB6PzvKXDQ84Ig0DvBP7ho+2HvR8
-          N6gTemf93p8E8+Ny4dis0O7SOzk2k6TOrSuV6CSICTYb1Kg0TsTmr75cE3+r6R4XFlrC70hwk+iA
-          3deTIp6C4Iq1YTpT1gy+s2icJRkfk5ymVJpEDjoQ9MSNXrLLzZUG4UF7ogkaHk/nLpANtI1PYs4J
-          06/HsggQL5HSZ3dWWc2u4DRwceR2WkH5cOdnzypowyd8V99Tv2oqV8Lf/Qiv+pNyX07iUN9PD188
-          jrbKGjs2ROve3fnQO/spfd4NDuo5z+Ebe/2667FsQ3ivrsRfrvWnp1xAcxSvN+iPWXvuWeu1CIjI
-          XEGcM+wA2fAUSWLoT2wRf8AgfF0Bftp5h48w1dMZF4Yjgsy646gTTmDm9lEGORiyW5doLZ3s45kB
-          qO3raXGCdzopSCjgLuoDbMrSORpD3vZRch4rLH2v33RxDuIVetXnQux7GFH+PijTrz5Izpxadb1e
-          cAe7u+cS3+wauv1ejvLCSkkylnU/XdUzCzY8xNrOUvrt/lk4eeGZeKPvRJy2tBBSeqyx7N5Bury0
-          pw9H39CIJpp8NX15wxOnWrjjUzxH/T47gCsML9yLyANW1DV7tINIwD7GOa80ET3edRZlMzviB6pB
-          RMuHo8BDhq5YVRYGTNmjnWAuAOevvkal5xTkrk+e4A1vllZ2E0imuSH3o5SlFGE8/D7jRysObrsg
-          3YGSNqlEeS1lP3fMLCCnjj7YeJtVut7eSQwZbrRJfjJ36bp7WMqPn7A/C3q6v8mjBikDXhNv4atK
-          uVcrICPOJ2xt9bPkzMmHTn36YP3hnunU310O1H06EP28N8Efnx3en5TYxevS889YztFdMRqf1RvF
-          5Y/7woJMGWQ4q/afaN69BQ1dbtTDzisY3dU0lwkd3u+UyM45oUP/AiI4oYuJc5ttXaqUUo1+46PW
-          W4+y/I3OiJlMjfiedFZn3+cUKLMnE8t8XvUrN0gD0kuxwtkYPinrfB4WNHPpgAPzNvQLuvEz+vHP
-          VVENd/6IaIDqtZvwcUQvwF8/jgO/Ld5OjbGTrUfSpUb1Lotx/JmslP8mtIbyjr2T2FhfFSeN71Uc
-          mEtCcP86udyxerdQvRYquezpIV0ek5GjnH1b5DGb1XZuaG383u/E9RFWF7xKHQI1sKcaz6rLVY5j
-          wSGD24qpUU/5nXLJRNHXWJyxD08detoMqGvu7bTwuVrxpIYDFLZm20zRCtUaHJgr3EXFG9vDoqT7
-          Vp40ALvzF0sCENz5FEUCqncfleQ2a6lrPXQJarqakDgyKrCO+UdD07JziUOHd0UVGg+wkqyenHLx
-          XE19/HLgNzpcsco7djqTFNfQ9O3aX06vEyXpM8jB2Safac+YoJpSYCXoqA3UL4zglc5zMgggz8zp
-          N/4j6qg3Dna6+ySmerAol1cu/MP7+815RNOX27oMq8aBGDinFWWz4ow+7brDUn1qwcr7XwleoqdD
-          /KtzcWf9cZag7lTsJLxqM12P7H6Gu8H5YkVnFjo333ct3o0v41MBOyrXmWAQvecpJI4IzGhP0mUL
-          WLgLccqbkBKltBqkbD0p7qFsplxsigP8mIk7nd8PK1o3vkSswvjEsqAC9kVxKKD3TNFU930Olt/3
-          +SDJsBFU53ShZymAq3rgJk7oPunEeGENf3zuFcyjIqROJNjIsMe351RHX/sWGmjTj1jfExXwFjpx
-          iJcaHptiyQKSm9wEk3bGxOgj4s62FvqQkZWIaN6+ojNnVB0MdoqHTY2x+0WFVQ6O2kSJ/n146Rox
-          Z0+0bek91Wre9ysDnES8WkJH7vfAVvfQIB14NFXlC+A0pIMUJaLoQWPCzp4zAI2HMQF+uMZEO0Gs
-          8lrrl1DXYeYfuq9LZ5q3Btj4ZmovIaHr7R0mMBqWgmRhIbr0DgULED6QiLfplWXTL7Bm5cVfX3vB
-          JfudVcO14NGmR7N0jsxtKc0za304hjvQ/vimObYtkXvnlQ7fsVjBpmexwXDR9v9VCOcBcnjjc5cP
-          ciOBl0/3wjIzvlOiOw0HeSnzcBwZKuCRP0rAlY8Tcf3cjRazenXox99BIuh0vXeoBX44x5uf+fRt
-          0xYe9D6H27TxS1XxutGA7u67WM6FL/2NJzi7jUJUi++q8UpW/0///uptseJbDkuneWLNGcd0Rq/E
-          F5NzfiTYGwuwvC6fM5yC5ka8uHmDrWu7AcPpE0xi5noV96RLjGhjrRg7byeaqygTwEHzP9hzjwpd
-          Kr4WQVp2D4KbPI2IcL1OcCSfimiRmqSL9nRnGL+tJ1Y74UQXdNvNQAtKiP/qKaHuprdPxsQ9hFdF
-          zWGIYfWqnsQLs7qa632Sw0NcfslWj5TeHN2Bm/4nVnLRe25hHwak1+jhcw9B7tnx87TAcBBHnzf0
-          M1gsQ2Z+eEc2/KWLUX8U+DwlNVF05gSoLgAGinZrTpTTW7DsrTT5w99wjkA/D1STIKm0LznvlZtK
-          6wozUA6FaBJF8Emphb0BpuDu4dOPz4Wnv8K1qD/+QsuD2mriSULasujkuc4hmK8PfoA7/BR8ah0u
-          0boYmIHtXXiQND0IdPRGy/jpSez6WVHN584MoXF2hk3PO9F6ZNEKuKqbsZNa1/7n38S+R9RP5j0A
-          60LXWGRZ7zbx+bcCC/sOYvF5vPDEktqiGsdWLuFizyExPVfq6Qt4CTiOV94X5OVajd6kZ1A5aRU+
-          zvYIZs/e5394d8hv52hOtdSH3PM6TggkokpzrinEYeufsQOKk9JD882hyy0HbDsccJf7/pEBMUx4
-          /6C1dbVMrlZAeizHiY1UMaLCNR/EQ8b2RN8vU7++Pi8O3pRV8Gu9Kd3h6mniH766l3FUB/OkZeis
-          O3g6e4d7ukqO5BxEuzP//By9XLMJ5ibdlm50JCJ+XDkwtm4qsZueq757c57Q5s/JVdFnMP38G+Gr
-          K5EPrUbnSdAacH0PBr5sfoPO5kNBobzsibnpFzZYrgE0myL685/jVyoGUBiQ4ItxxyplxzlHKs0c
-          ordX2eWEpYdwZYUvjibu4+adeVfg5j/IEdVpOpu7UwPukyj73/y+c6n2kR3w7JSWHF1Gd9msiWc4
-          WEDcvu/l/j2/GQsaDsAq0x9+wlYnhj8RqvVLbYUx0DzDJ/awlNGiPZsctNPFwUY6FdE898kVPk6M
-          i+29nbj0c6ESYsrU9EVBQ9U6HK8Miq2L6jPdeq8W+5YYsLufuYnfvbuevsTXAM1cOWCcncV0OLh2
-          CUpndDBmBTul1cSyqBqNBeMmedPxyIgiPAXVQH54SSVpHOC6V8+bfytUyq8HCeTC+sTHTubTRZuZ
-          BvB3I/DZvme2PSYkAatKXWwUhhAt6VcXwO7DfbAyR2m1lodzhxzCHabh9tLVWe2MEBapZWBH3EG6
-          /p43K5R3rC5KpM5ixp6h3d3vRCXgUy2H+Sqg/eXj+mK4fqv5u+NC+POPlzK4RCuHQPHnly/t9aVS
-          C39iqBLL3BJxXC0q7HN41i1MYoDlnpcmkYUBPWOSeecpJZZfKrATxw5n3vXwlyehTb8Th05zRB1l
-          b/0931c8WWA5zLkAn5cd9A/rTgArA5QY6oZpYu/cgZTcESrA8dWJPrc85miW3voAl9zbTcuD2ule
-          lL4T1EH5IUeije6kBIYBjaHTiWmojruUzS4HwVGcsLrpEypotg+vokWxZASvaHmkAoTofCuxGmjy
-          5l98TtzGG8aJ5adzcwsMpJJgxbKqTulyZwMDVvFKsLXx7yI8jRmVc3QlxySzegqhaIH1xpRku/5o
-          /uURhVIqvrjpvz/8K9w02/T6UeUOyikHxrDOE8NEIRhPTTSLS8Nr2FMZ2WVvxOSgQyYfax4/9SN3
-          r0rQUMPG0aJQdZEEvf7zp6/t/axEGqbDO6ucqUQDk86lXBrQgYceK8XkAd5kbh6KYOEQ5bvtKU3o
-          OPz4Abv8dE67RXdq+BHas797urU6S6HgHZb80mH/cXlF6z6ZGzi8xytxnoYLOLFeJrjxIVbLsgbz
-          +LlZ4vvtblv0bueo6lHKgUFUY6I9o487oquYwNvt+MZevi79UPjnM7JuwMRy2CJ3qZOj/9OLfs99
-          KZhc+RgfCoPTCD5yV8D25bWBrwWmv/t31y0fFPEhl/39nqiUTtFQwwcjnoimVUO6fhTqQGI4pb+7
-          y5W6vlqtgNle0rDsZad0SqA+QwbYHLZcvlR/4x0p1tsm3ho5gEJ05mC7liXBr7tIx8l5sWjjH3yR
-          QxXst/EL4dT1xO+mWiXmaGqotoXjtPZArfY7Vy+hkxY8Pp2qrOJYZ2Qhk2OCLeF6VDd9n8FTkMo+
-          XaQ2ne+nYwZ1QzdJcJ40sOTM3T8sDSLTzhsLOluvRUSbf8BXYQjV5RlMMyxq9TWtrmxW+7vx7qBp
-          1x5JMvtFJ01lSgBlPsDeepJcGp7ODnquS4zNe+O7a39XuT9/FR5dC+yFYwTR47FS4onfG12+IePB
-          3XyVJ373uaRLc0wYeHhZE8Yn1lPJ4dNwIBpoMdGrX24duw8CtAYjx0eJVP3aFKgEwXFVsZvu1Wja
-          5ccEqsQxCc7OSTo/7HQFx9HzyCYOwIIMDwIzPEPss/ZC18vqddAYWp2c30/qLk/8SmA29Bei3WRS
-          UaXsQxC58YDxtD6qYctH4Gts4+kwikI08/E1h60+GgSP308146PQHrSF6thslQasbzzGoKGaTZJP
-          rEb7n366G1VE7Ni1qvXGEwHc/LODjY0/Z7mWzpCeUEq8uJ2rqTxkHVQTOPutEPOAKmyuwY+7q/w1
-          qfuKfRtdd/hdf3Q63voxULcdQMFOwB7zVlKWUcUCRLB0iPLzx9hPr6JXZR2+fB9DNFYCbsQJlY9p
-          ESot3YMwK8F3bkzsne7+v3i5NHttEjb/vG5+D/zyQXPTg998d21gVAkGvv/4z0J3FvB3LcDuUtOI
-          3IQwgHXYIqL6O4nyjpJkEC3Rd1re2p5SnqYb3t46IntJ0y/v/nGGG1/iLV+NJid3ElHnbwGW45WJ
-          FnBa2MN6gyWWc/Hcz5t+gcl4+xDnOH5d2ujHGu7w7eiT3/uXzUaBMZorIlfIificU1moG3idEHfS
-          XC4/URbilxkR6aF8VOrV4vDDdz+eUOdOBu95f/pA+Y035muKcMtXib2iSSU7OdCgcjr6f351taG3
-          Ql1nMmw4Nyv6yyd+///Th1zqBY34YITTxpdmv6yr0wJJjA/YZ07Wb7ytELWxj81mFwAqTbEF34fn
-          CTvSdVFnYp9n6JiaQww7/IAlba0CPI83HmsUbV2yv1UDrducYR0pwKV/eu2bvyamriv3p9fRR+jO
-          xNnqcUVg1GDvz1d8294n7fqvBE/BXfaXW7+CCV+DAY5KYvqzlZQVRcWTg5v+JJv+Scfheth6zKQn
-          YmgnCEZdrg0EV07FRrpeAN3mB4AWJCdiZPSlzvI+yNE9GTrs7bqSEpDeAniUkgX72hJGY6+/wm3n
-          tEIe0uKnS0uA98v7sHR8qip/EDgG2J2lkmw3CdH8ndwM3o1XhPVMs6sZjK8VpiB8YZsMcjV11jVG
-          D/tqTKxacOm4+RHQ98ODeO6xBEtosQbY8gmCgfEEK+PtA+B95mH65c+kzM34UPsexmZ4G6IZ+9EV
-          bfmTDwsGVe31wU+/9+PvDsgG/M+/qLzIY1nHYkqX2MlhcC0ofrTCKxq2PB6i2vB8Xj20dPMTCfrl
-          RZfPSUqpo4Q5mgfWJT/+Hr5T48AtP5jeZbBPf/mO2PtAJj5mvi6Nyjr400/h8yNHvH7IO2DuTgo+
-          Cs+Erv13CZDZdMzWo6WIJiIpCgzl4kkesytUo3DkWhB7ZTLNkC97yiwfCN8Wtvx5y+P7DX9/fOff
-          6OVBx2fgaeIpMxi//GoHsArPxYJtIltYf9mdu4DKZqCq+DKWt7yVPCJLhHfi3ibu2jsVq34l4Y8f
-          7VGMIyqfEAtGpVyxPdBRHX/1nKdcjL016ugiSt9BoPblSuybg9JZE+8S/Pm1rR7cbf4mRj/+M8ip
-          jvozhQnE8bnCN88tKipcoALp8Rtjj912KH/DtBC9p/X5+cF+mQPdAD/9oZ/5tV+rNAmBgXdnn2x4
-          P/Nxnok9Vgdy5E61unB3iRXjb/Yi3vOQgR72pYQqyeknis5ONbyedY4iNxkm6IVV//MDsDmx0sR7
-          Rzf96V205Vn4iMqg4lz5mICTd/xiTc0KunzyygA/vWCenwplm9u6Qp47jUQBRZgOHSOI0LqtGZZv
-          I6pWeni00LT5Eh+5aKLrLtdjtPEP0cGhBiTnpgJu83P+chtRT7RrNMBNP2DduNXp8nwxorj5ef9w
-          MHG6NHDM4OMIcmxveRLt/VpCG54Tu/vI7nwPJA2l+UMnG96rq5SKEC5ASf35iZpoNiY5h8PptcPH
-          4BymxKsDCXScFRLdzTRAX+5BhPzdAxMXvL4V50mSgrb5gakKDC9auP56hvOov/wm5rmKbHgDfvXy
-          fLqaO5PUbKA7yTa5vm3fXX95CqM/JBzsW4ku2lNdgVn6J79uSouOL/XOHk5mdyLY+OzV4Te/bLFT
-          9fd86ahkGXh9z1ccfK92tHqSJEFq+z3RN/+5fut5Ap7x9bE8xoI7M9rDF5PR27Iz7+zumS8W4Trc
-          MdbNnFQbvsSQez5jrPq7Avzhb0NvT6IexSKl2OMCaDP1vOWBZ7o0oNVgJ9Y7jFnI0DFngAiFWQ6I
-          bw5LNfzygbd6MLDpWk7PKqZ7hgdr2ZFjWtHNlu9ycNhfdWJv83HUXnH7f1YUoP9eURB/hoS4Miir
-          WT/6A2SkeTeJIfpWg7HTGFgd+YGon/VUtZf+mENBbDKfj4ZGXVr0bVDDrAFxXrwLqBF8IVAmixC9
-          SCa6jnOgQSPrQ39ufZnylNwTaLzbDJvNcwBj0h5ieHAO62gPtevupXVowehc7OlihpJL9mNXQD2P
-          Ngd1IO5cX+YGtdXJJapfmf2+K2mN7HxtidNUr344XKoGhuMR+EzWK9XSW8cGggOTYtxkisrbirGi
-          yRq+WME7rp+X07eDLZmJv7fAGi357cTAOaNb1yXh28/aB07gubd9knazVi2F+U3A46UpEz/Fp3SJ
-          1oMD1+A1TIgLVcDvny0Lj7uaxfcqSgAfeZGCxv0u/D0P0L2tUw3xLWCJIowomtLSFuAlnm2fC2Ad
-          LZMa5nByc4mE2CT0O463AoUEhSSZdSGdD1M5wJO7TFMzaSHdC2mboPWXGNh43rp+8aVYPhgNS+lc
-          Ana/Wjl66IpLpFLv6BTpEoc0vHWN/DRSxRU+qCE+qQq2Db50V3wfCihZeoAjPrZ7bnfcJ3C0Gx1L
-          gYr7xYnhGb4yzcZav2P72RFKBrlHeSbKTTpHM5fkEzTlISLJ9j7YS69ncB6khaR926dsMCcG0jUP
-          4yc5qO7+vCgBVPz6QWJbgxV9nucYHdbBISYJYpVjcB0i02R0nLbnV8qzN9JBaNDrVMZxoS6Xyqrh
-          Nem0ieP2uct/gaVAMQwykrSC3ZPwY0swdZWSxEV1d6mguhAGl/ZM7usRuaN7ETTU2+eORDj1U7LT
-          hQKu3G3Gap1/ejrtQQxVKbuTG4EjmGF4ZOHKXWbi616jktsLJihKt75J4HiJ2BouMZKJ5JLbm97U
-          lfd2Awyboz4tdxa7S3/FIXhOtY6TYvd295/PNEE1qSBxn4AHq04zDtp9cCQWifWeWnehBMYdgGk2
-          1me0dIbgAbhqNsk+AevOuzhwkPXoF3/d6nm4bisSTPYSYj1odUBbz1DQdZwWYnlC3w/ZUziD79s1
-          MG4RUCk37QS41NYT56MtR/tS3CuoM3NCjjxXRGxy/c6wq4oGB58JgvGwhwLEOxZPM7Aq2pfnXISp
-          HF6I9bg60fwNegi510Uglp9ZLlecdhwcXkqH9a5697OSTxZ8wfsT68npBnjcsw6ChckR3cKwIt9Z
-          qWHS5wFOvFYGXMutGiD0ecWq+I7TxQ/DK3KFazUJl++V0svl28DpUtXY2CVCSuejOiAvaT2iKk6e
-          zgNrKkjJD9Rf7ixx2bB4FSCl4OyL+kdT12IPE3hpr6l/s/q2oloYcAjUjxFHZmlTCph4hnO27Im+
-          Zo7LziWpxcF6+vgo3u7RflKTHMbMqcRmaNk9u5xHEfJ73yCZ6aGUyow3g7hG6vY8pYhP+cuKFOvc
-          +7TOzWqR3YqF1GVz4m74wnXdcYXLrRMxVoNVnZ5KuyInaGt8fixtROJccOClNL7EeblXl13GUYHL
-          GWz9oEu34u7LmqOr3AGspMuuom59zuFRoh98PSxGOse7U47qNLNwxjlcuiTy3KDuu8ZkW1nR7/mg
-          gxAX7zN5PLYT5V+05NBqwwr7BzNNlyNQJDhSDLAb3uVoWaGQQH7NCXbowXTXyRla+OMDXBV3QJ5g
-          8SDxtR1xRLEG9BOvZ5SED5NgeCtVImSu8asfrCuNk7LXh7qiYBlckpzuR5UVIzjBPjNkYvrnpCeP
-          z+hDC2a7aeFmr2I//YlBe50zyQUcene4yla89Xig2BbXW7QcMjQDIZUc/+7PTbVY7SBAo7TkTUH1
-          7nJZn91f/drtDNUlUL4reFqzjKPbOQQ8p4kc9Kz+MIn0uEQzqxo56I7DC5+FR9Rv9SWh3fPJYkfC
-          QrX2kjUcBp9IRHbEASyiYs5Qa60Y4+83oBPTVgkapQYSVbD37rzfaxPQ3neZRA4R+uGcry3081jB
-          sZl20fL5NMPf/avXzHPXIXl4sBTXnY/eFo5WxI0arJliTy4Wzvr5vr4dFLW6NB1clklpP0UDFBiD
-          x9Jnyuh8cpoWfb13h3E0zYBGzD2EX8bBWLkEDe3kB81hmuuYYHRm0sWm5Rk+6uYxoXG0e45tgitq
-          KnXevg+CJfTZ+le/+Gx9Ijq/bS1ArsIcfWBbp5RKy8Js58S72JKNsu9fkV3ASfd8fC1Dm46krWP0
-          6d3HxBzREs3vUo2RwGsyCX1su1O5cxgoMtfQv6xZ59KuugawXJQVP1v9HHGIe2towxss0TRI5z5j
-          ZuheFZEcc2qmXHqWvL/7zY257Hk9ajyYPXcOCb1+6Gm33CDSOd7HbuoM/Vwd+QT6JWaJ9DXaak6s
-          1oBv3zQnwUydlK0WCqHDZS2RpJPR7+OxLpE5czwxFKV0yU0EJbwWue3Ta+ap6+6IEmjyxhfbqzyo
-          XatRFsHtyIGorZC73OwPA4906+KWgF1P56+poYJX7elw5w50vrGZBON6p06QTSw64XXrVMUwEZbf
-          TewOypYImHl1xNZxDtPV7VUI/bTrf/ygDvwtyZGsMBfiOZZece9XMKB4hB7R7/eTutYo0376iGBL
-          RhXRwpiDQfnAWBKGolpf62xB915diDJ+HLX93X8r7UqsjsCqFp2XDCS8zQkfyePmUj1qfKB52mua
-          Tx/PZSXzIcJtPE/sarkVt5i+ABlf/uBjHbT97IzvAkqjFWz8O0Xzlx9mqAnPCRtSrlfzM0tqOE+r
-          ha3nOvZU6LwZpulFIBK+HSoyZTCAwszi3/NQ58w4eKLuZDI2rd6qeF8ocjjuyHPi1JVUq84kIdj0
-          EbFdNk+3+m2QfmQy4kT7IOXFJPNgluUEuymnROyNPUuof8cvEsj2MeUPvhyj8wp1omjGmtIxryzI
-          Qx/6HP+K6HIVYxY0e0XFWgjTath/VAnyofDwPau21e4snDMwyFmMc/LWKvb+IRPY3gc+LVHfUxU+
-          RKgWUo+lezH3q/wAORAvMz/xylXq9/L1FYLxkXLEv3QanXkmDYFEE5Xo8cuJaMtoATD5/UjyvTj0
-          U9m4LPzpZcceDtv30Qys0Un1191Xj1ip32aMD0Qi6rEvwLz2nAPNcbaxlGpHdybo0sDhE2YEG7Eb
-          sdlzviKOXl3sc00ZbXySQe5RGERRLwZlA/sUQC6WY3LxiJxyrBKtcMND7Ki3pRrDPYph37hv4qmm
-          CRawyDOql+CIZbYQo2lrJgzvo7F1zbyV6XzOxW2P+8JhOdbkavV9YkAhL95Y3aEKLLP3kmA4ZLwP
-          StdJqdlxOcgPlY6dO/NJ55y3Gghmz8enOdXAvphKH6BTl/7x5coGjwxseIUfuXna6i2sobEjO6Lr
-          U9Uv2V3iIFNcTps+VtzxOEpnuF2vL6BmVenzLMSHptr0r8ejqhNFhYWbvp54jb9Xw025rJBUjoRx
-          QqZ+iY1+gMJ5jbGSH4SfHmehXBafn7/oeff+VMTn3vUnGMqOypbvlwWKw1rgYzO4YDYquwGhlMrk
-          uKxPdfEv1xqykENEUkYbrLN+KMElXm3iiuXYL+fGY2G6uGg7B/YTDTs2KgDwvy2+dnztzo9BtWDJ
-          h7WPUPSq/urHv8QHEjF3M2J930vAum8GItmTWXGXU9SJG/5PzM2XU/47Ow2MXryPPQVvXfYZP4BT
-          KdtY6rq3u77uiwJ74xP5h+VyTem9OBZAH8vJPziQ72faLiJKXakkG3/2A6b5FUpasyc3jb/39BXJ
-          BeTQPiOqQ4RqRdrEweKW8VhdLD5d2rrLoB9bweYP5JTl7WAGP3zW5pq4LX2w2Q9vSXSzZLpkkj1A
-          mAWaTx+PL6V6NPl/+PMQb4dotfPFQaoxudhkx3v0rTlZRD7wEbF9LaALtdUQ5I+nj/3D1nNVUcoV
-          nh+OhC3uzaY0lCwPtjbT+0sau/1+X0SlqLx6l2iScuin7HJp4M4SB+Ibc1nNB/O1FVI4YDV7V+oc
-          HyMf7i/xHcvXYQ/WXpImeMHrjtjno0lncj2vsDwzgPjFHKtLqO4ckTUp8ifm2YBFsVkGRtN162LP
-          tena+UoMpdRFUzfr2zmyycUDL+GgEe3FITofO7wCbUqcrZXJ6M7yI1jhVawGHwaPY7X50Qw8hf6L
-          VfJ8q9S6zwXy80TBJgkEl94PXicyu9rD2N0P6do+Vgh3/e5OHKFbf3rfgbc39yRWJyUqcVHUoIcu
-          ueTKlKd+WZ3VApFVTMRkuFltBi8tgI50CUtuqLmraHkN+vnzkZwMlV/XRyeCz1H74491YE4rqMNs
-          h8MNz3hmWK5o16M7xhI7gcWJ2TM8HEYJh9czUw1RtPfhqash9sjeT/et25/BRY1S7A3jJZ2fapqJ
-          9cW7+vwULymt1Jj5ux/zIYZ9z94+Ldz8FLZe6Y5Scp8VNN+jBJsO1t35fp8kcE1ajTyWdaeuHn9d
-          QX546djclW86S+ZFhGoTPPBjRzTww48/PzPMLy3iZ9Yr//DTX9cond9vs4WGoTnEfTzZftWZMIDv
-          81n0wdez3K/9LlpQFe4VO8uXBdNPD3KsibFZgzn68tHH//kjn+6QCmiB/BqeXDoRafOTVMhUAyUT
-          7HF4yW6AM2Efw5pEh2nd8GxutZODPp+vgo+6FlfLHl4FyO93DsbycKnoz2/7E4l9wZO03/0J8MxR
-          bjqALleHcXFW+BKAhp3vHESjhccZKnVSTqzjzhGRKT5DZ3QsjMXdNaLd5dxAK7GvRBKEpR+3+hCR
-          DkRsevWoztm4Sr/fw67AnNT1+X3V8KaW3MSTIHYphC8BFUUm/8ZzOt/rKofP9hzh7GRXEcUaPSM2
-          KnNi3eQymibYiSIbFTn2onfTr59Ta0EisTfs7T63dCWEq3/1i61OEt2fn4ZbHkHMzd8vyo1pgR2e
-          NKyuUu7ue9HlYLaGLLar3VzNjlrk0Dsnib+64sulP/8mxCHF2IQmGJ7PnIPH9ext/jWO1mMTbzsa
-          zivGtfDu1xYvAWR89eOXp5sSDUEZJ5BLZoINcFHomue1c3B3l7u/0wut+l0/arNBIT//TBuFz6Bz
-          Fa4kzsWGLmypxKi1YY+xUEXuuOURsIhg5l/MsFDXfBhZEDNR6dPNn2z4YonXiqmxueHxfOmPGbyU
-          2hdr2eFV0csyQPF9M044PckXSu3GvYLQvPPYQ7xNOX2OazAWro+Vmajp4iwgB/tYZLFBWpKuq6a0
-          6Po6Hom08BGdf/ph089YOdeLOngPFsJnyL/9eT8L0W+8gOt5bCcOhruIYJNvIDnZI/Hgk4mmxyGx
-          wDvgnlj7VqeUZhlZYY7UAltatETL03l3MJ5IMO1MaNJR3c7xbmLoYWXTG8OmH1H4HQuipMuzX4xT
-          1sG7rIjYeho8WPTB6WDYYH1ayrudsrz3UOCi1Uecc42Ssv7d5sCG73jD+2rKpmyABgMBwe3SuWTa
-          gwRsn7E0RLeIKrNrgB/eHoJ7DPr7Qw7g59xcsJUrkjo4QrmdWuAw2J6sT7U24O1DUgFtYnp9R5dT
-          WgioKuwrtoxK7Xn+sTDIZeoL2fILMEcvTkB3Mxg2f7BGNE2TBAzVLON84w+eeQEB7tJWxQ9dE6r5
-          7g5naGTfcNJ8vnH/+BVrrweJPe8CiJnBFZwrYk97cHBV7ny+iYhmLoultfpWs3TILLjlUdOQ9wf6
-          0yPwZQQ6ORV7j44/PMoN0fAPhjukwwHWCdrGO97yrmrSb/cc7nr/Qo6feu03fvZguxYmuflnsaJc
-          wXmQOxeI2MKVdWclfYkoNwQD42eqpUtuxwEgTl2T4yd+bacizNuK5a3nqnC06bI3sxgo+vvsU7h3
-          wJK0hwS4gGVI0p89ugql2ADr8V2IKzCLO92S0oMcxzXEGmAZLfVDWWGsJwsxr/Lkbv7JAO01C8j1
-          PuKKPu5VCT+ve4CPjcS5XYuXECa285nYjW+3PHCC6Xh2sTfe3Q2/vg3c+Bebqv+JvsXUecBK3CvB
-          oXOO9uJ9KdEBZjey5UHunFiF8ZeP2YbrpftfHn1+WNK0wOMREMaCK9zyDqLbn9b95WE/P7H54U/E
-          24o/i+9P+iG685xou9vFDBy/zg17vIJc+kCtBvpFMMipCY+AnlQt/+lhrB3sbzTpyWUAnujnxLdP
-          oJ8Z/jaDat9qxKlSKyLRO7Eg+GDN58TvUWX9j1ocDmZ6I/Z6ROp85RQRxHq84GDhI7Acsv0M0Msf
-          sXIbHHUx/SCBp9ebTrCJx4i+LM8AnewjHz4oVcdEnmuYrOp9899NvwDWkOAplTpscPtcXfefcQDD
-          FIV/+eZ6AqsPNz/kL922B/xzHyUY3B1Cjuz0SVfrykOYTExPbOWG0un4vWhw2zdBTJe+Ku61Chb0
-          l2zGytU69vuvzWaQ0Mf1T7+tXP3NwJbXkF9etIqiw4JSwTWR+DNLlyt6GLDQEmdiYvfxb5645XXY
-          u/gC/c0vwG08YXzlOpWu21lgkcjeiKS+557+8u7qobUb3+6r6VCbDLxkbxHLbSmlLH8Lc1TU0TqB
-          tgjTtXm+WLTpPewdL+doH37AAG5GeJwWAp5gzp7CFW71gs/2M6nW4iNOMMKv2l+Lk9bPNBF8SBd/
-          T+zbpaCUBbWPNj4i51BjaTfBTvjzc76hDXQdaHaGn4Qzf/oGTOKwnYIwOgExxvihkqq7r5BfM0KS
-          Q3mIBqF6B788clppkLkENkSB0yvU/cXKv/1wv08KxIfmQYwoNisaMacQmWcq+e9Eb8AkndoWcq+b
-          gJ0qbaOaJrMHw0x8E8uTajBfsaOJ/N4zfnlsxZmX7k9vT89a0Kt9/o5auNXTxOxFrx9jo5r+rv+I
-          Irlf24L30HR51eSXl62VI2Xw91l57V9u9dN/eTTKxFFvp341dqQVApDF/hscLylBaBeCuzqf8E1a
-          a3fgmShEl6k5TnJSaPQvX978z5bHvPv1qGQBVBb88FlRnl36y1+28YIlRWbdUZKm5KAny23aiTsu
-          Who2UqBoVABLYSSD/UxfHNrGp38oU1bd1grXYJsPwXh/1VL2lmYiGDXzRtRVYtyO8VoJbvyODcWn
-          6i/fFTLV2g5PVKtqRg83gPs3IRNfNm26NkY5wUvPciQt+UKdLfHOQjcLBmxGg6EO14c6w9yJX1iW
-          3m+XRl6qQJDENvbS3FXnktl2sJnh4EPksurgyYdafOTXbssHl3TxSkUEr+sn9Q/Zqqi/ege754Ml
-          /sEEEcWHPQurfacR87aAqi72pwE9Q+2Fnate0pm8Eg/mxQtihW0/dD4UXwk4p3tLzBoEEe0PJwsR
-          ibv56BPL/W9+Em7PHxuKoqhsdl5FBJ/zE8ePaxdRWJUr3PyNLwxQSZfzzoXgFtwWbFzkO5iV9CvC
-          hlRnv6DpnM6ae7iKvNq/Jp5DfbU+gzITn6HxIpsf2E4BPs1wm+8iylbvQ3XcxVBCz4ZY3j8AAAD/
-          /6RdyZayPLe+IAbSJwzpRUCCgogzQERARJoEyNWfRb3f8J+dIassm7DzdAk7/o3SDxo3uPu3XU8K
-          HukzNIFl+8iY29cLMff9TJCWwEPmxQ8MNi7lARIz2ZBealdv0cO2hJFTY1zWMdtQbCgu1E0+/I9v
-          7lkd/OkZpBGC6QrOZx/+6pwgQ3eZfb2mgcqHug6xjHI1yO7P//HXKT5/wbp8ekf6u98GGy7jar6t
-          ABL+sO49Jb6AMHtPgd2fobwWVIP3hWRR/LvMEesnIEr1kxtCDO495tXSGqnsZ9Wffgokp2IjsuRW
-          ChcEZfQ0xWOzFLEsgw/fN8jqBs3jJ600YSDgmmjF6W7w/VNmYHiVFXJanqjhn2NawF3fYHbnp6Vp
-          tuvf+KK/9afq3c8ZlArR2vPkn7e+3LmHdPU5dKv5q0frOBEBv7Qs8cPn3NDf6bTBE9vOBEU4pPT9
-          NnxlBIcf7u/naH8C77oAG7F3lFURiuadr2EVMQVBN6aJ/ult4nYtfgd0zdf59eRlTYe33V/YIzdN
-          rgPga3uhU/Rewew1e49q6WAgexBMbwWdn/ytF5AYo6+3HA4hVJrwayLX7zUqfNC4/K03E9PPdcDf
-          huum/MMjWGpNv4pDCSFgE+QlZQK4k/+rYHARy6A/5m40Mb/EhmfpphODeXzzHf+n/8+OAo7931sK
-          cg1cg+95zo35UK0YWHfrHAiPLYgEpXq58AujCzESsjTLzRNlMH/VL7LM3qOcpKSuAnHtknMwWDkB
-          r3ICEFcuOgoaS5fjzUrhOeEN5DapYtA4/1RyzXU2cj7p4i3pfKmUiLk05LSA90hfYepAxeYQMrtH
-          7M2OxbIwadoLumAoejS4iy5MvMeRlKB+gkVUJhHyAqZE+0g8WLsb2uTefNzQEU3jSF4XM1AAfwGY
-          SdCtIeSBMIzOfI6OE9hySqSvrEg+iLGUbyCio13LirlNR+I9lh/YDm9/AONmn0l8cudma2mWQJZZ
-          Y2SnV5QPbqdeFSCHB7zEgemt1OtEKFQmg67v2AR8a/CJwhvPnrjjAzc0Hp+ZzETlh2hwp2TbjEKl
-          YDaMjNuyGot/HBawpO8LCa/+vqnq46uKGr81cpO0I2Dzg36FgfcKkX89H5ulXK68gsEkk8dBa70Z
-          pVsKHqrM41h38pH9sEUHjxEcyNU/qd46CZoLG004IVeqdCpIz9qH12cTIOR862jhi7SH10eK0HXI
-          +Hxlx0hXsvfoITdwe2PpzFBUZCv+IG0CPVi/cWUrq8OWSMsXA3wM6nXKx1V64k9l6QmqNonwYWkc
-          ceX6OgqN/nOV7kC0YN3/f+liZwPeSarIvjs435hFLRTOBimGh7cW8dSzNmWq2xzdD1Yy8ok/uvAD
-          Qo4UgZ8YnPB89XDp2zvyt2cw/n0emHW2JiEuXIN/BK8eDsqgkrK49h5+YG+BySvziD88fMD/jT8+
-          PgkpkvAOqMO5AwgXIGHp+D2OAutKKjyl0CU+bNaIvMLUlZUa3pF9CZV8k5TRVsid15EbnX8R/yRp
-          C21SPUly3w9uN1u0QFiPbkCB+jMmcVZ0mBixi153Eo8bEStZAYMqkvD3cICwncQOFkP7I4VqGNF2
-          snoTBtE3DpoYvyP6vqQdrG9vldw7XjPW3yfNlPc74jA9jQod2NzVofsoUxLAmhikazleaU6bh8X+
-          sYzzlyNQTttuRr5YvY1tPqYpbBIjQAYBNuXuBXTAQXREcmu4S8SJyoGHBMkHpJL8NK7p06ng98hZ
-          yNTsEazx2LRwMe8+5udn7C2zWMZQKCyJGGguR+63b6q0iIrQs2J/YAkfUaUgJBVIF9/DKGgrZKCV
-          Gx7R1A43S6nyLORsfEfPHgBAy/DAQiivZxREwc3jP2yQQXJn9YBOst2w6fyoofv7PZBXsPK44JRv
-          FZPnbHSsOTSybcRmilmPGtJEI8+Fv/ezcs0j2nl4R/OxrSq4sEKGTNd/RhtJ4QbnUmxQzNQnb2PH
-          D4T6+fHAtOCiXLDbLAFH1l+R2SUNZfVrWCov0c/R47bIzZw+nRrC4GqQbC00T2gbLVEyLTCwvOPf
-          Pl6dAuTrIfhpnw5sPY9CGfCuiszXRWu4nnOu4KyzBxTUsUMJl/8S5d6rT1RIL7XZjp9vBdUw0wPA
-          qwvgbFNNlK8BjgH3lSJDeFWsrfy9n5dtUjM/yYqhUh6uKO7zm7Gce52BKQQrcrrsADYsFjWUnOVG
-          ro/Q8Fg60EnZ5xNJt6IGi6pFm7LdvgQdg0keMefbvpJ8K4QK9/QB+DowJlji9Ipes5KBrXjFMfR0
-          rSV+GNy9jTmIMVR0cCHeg2beyhV3X7lqoU/iB1Lp1nN5AN9vm8Xrjh984jeuEnWnitwjklP+lPg6
-          /EkOxvefoI3sfbZcRb+kJTHNox8J4S+uJFNeTfIaJNdbqjW/yjapn8iTtDXa4GuvXwItcg3Pb0ql
-          sgzh/SU+kNYIHFjFZ6tD6moAz+FkR5zwvA8wDpoBaV5f5fxZWHXIDh9EMpKpxu8snjdgXZkAGdq3
-          8HA86gV0/JhFRR46lOXAkkrZLXVIBt6bt+oNioHkbDcS1HFPpx//LKGQyyJx84vnreRxE4F0M1Li
-          2Me+oRlzhcrk5hfkrYdXtJyst6i8WnQnZp+2YCXpm1Va95UFQse/PVpvLQ8J63jk+n3cAN/QMwOZ
-          e9bgphZij+9MZoFsXV1xqpnfaKPgyULKRDZeJ+CAFYtapwRUyIPDR3oY9K1Vk+wsq0aOP/lCf+Kc
-          y/JeL+RotI7HEinZlPEkGTveN2BNpkkFh9Tagh+Fbzpq74uqfL72Ekhv7gyEVYJXOJ6wjo4xORur
-          HDg1jLQlQ2l6mT3iCUalVG3dYpGrouYfnj8eqkSKHW+3VfolcOAwQ5yT8/P+Pg/W2hui+wqMSBCe
-          TxbOGfNBnqLXYPsl3xpcPieTPMutp+PzRXul6cwYPRTj4/ElEll4HtHp3/zkjpxZKO/Z4dC1TW8j
-          J+zHGO31TMrpmVNWXUcebOfXGGxtyo2LjJVSSpruEnS6efLYH38r//QJicWqjwijHjKYPd5H4umN
-          aWy3wh5gEjUe0m5rlFNhtiaFa9AbmaYLxy3Osw5qv3OMyiF1R1ocohQ+aQCQEa9rQ5klSsFkh1di
-          Xs+YbkGvuZDVrx5yP/czENi8YIBh8QM6+bOUf1Mlq6AKsUvifX7wDtdsCpM3NdHvQW1Q8tibYKSG
-          sdeb0/zhJejAapDTT0JNqWoPRlqqIkYRMKRord56CM/j+RSM/GGJJiIlC4RZYqEgC7/GjwU8A8N+
-          5VDKPZt8sy0UA8dJ+WDr7rbHKwTacPgNkPi3ixjRAh0y2b5/ciyG5zFfBR9MENbFiRRrSbz17oe+
-          8i00N/h8+RrQruXYv/Ehxpszc5afYAu9RO2RbTkPOq9ZGcPhUktIza/sXvW1Cl9HKUC666JokpT1
-          qryg80CmUAnGGohXB2zcwBErpa+R/uErvXoHYo4PyxPyEvggoEOGXPvENxujtgXkgNOQx+nnjOvp
-          tu5t6HSMmzfX5jQGl16xgsXBjOVszXR1JxPG+rUPRPYde5jzAx8Q0TyRP75fL64LAf9qfsgZ72rD
-          rvI9gONVOyLb9YCHdT3LYCQ7bRBxWtqsTKiU8ICWmfiaRb1lcuwrWH/mZ28TTowlgQ6E+3wkgfQO
-          Rjwf0wx0RioR8xoejKW/nXxAZfeID8HP8FY62gO0qH9DjvS2DKFel02xD1uLnDKadz3X2/DoG9+g
-          254fj+58B7/QDMlzBj4Vdv0KTx90Rnqb3pqN35swHCqVI+j4XY01K1WonO+aSa5LtTc8eRkbPKbH
-          H9JG8RhtbM6b//jhmFNlXMZuhcrH5n94s47Yo7F36qAVbA5R5QZ6m3iOdHgjNoPvP3mlzy8LRLly
-          8iIQ78E4zk8iYbj6S4HBenjl9HXxfXCoEB8UttV62yLFDPzeRQ8lY1qD9cEUGwBnsG8B8JecrI+q
-          hBPDPVGMnxydtc1zINRzg9j9vR43OcgL2IFOR8Ht/KHjeLdZuD2aGYMfeDbLkJx6KPSqSp7h1EU0
-          J3cHXkxZIdbBcBrBZt0ANHx0Q+cFnvO1aQwH3g6OSew+uwLKjZcA3lxRIpedbza+0EsF1uWJ+M9b
-          Zuz+qJazc2giy/i4Of1apQm4gM3IqWTnHF/lDwNtOjWk/KSPZvzE5wokXxISD4NgXBKoqBBF4hvF
-          4O0084eCAp4T1kCXXr7nHM2TBYZOMuPDncTNZpt5CGD9c0kCqi5aH4GJQULpNZDEgxGxt6Lg//Ap
-          4Owf8DY6SDK0BaFGKF9sj//daKHUWgKQ5/w6Oj+YeINlnprIPx4pIOXSMEDtawvD/Zru/gPwTYWJ
-          Mz+0kQ6dPij4+CLkbvciWO9nP4RHNt8wy3we0WY1VFcG67Ei9ffowfRtV1Px+HCPNHiQ0/WXJvDB
-          uRqxp7zNt6SIVMWrzYGEbfIdV4gU9U+fB9JPIiOudUWGqlMcSPRkN0r1TUuUFEorcre8ajbxyTDw
-          dWdiouP8S7dqfZSweTIhZrnXF9CmZiH4lMwU/PEp/dlJBbfzc8TMlHkjPwhHDOu51DGvWadmr3cG
-          utL0IvfdX68sYPazVz0D4/YhG4RRvQRWavxCdhkv3naVZwb86WlLMVNjlZTQhd5ZbEmE5rJZQPAe
-          wFR+kkDw+h+g6qrKSlr1BF2ANubbmNB//gs/uQNutnf97KE4jWdiN/Hb2MpwksF2SibMvuILHfbv
-          K3WHWQt4WpjN+qcfVTi5KAv9k8E9lH6C1bJ+kW07A/jTW5DLrSdeC/qJfrfpNQAOuA3xNqh6XPHS
-          MgVNaktOvRz+6WkGyGuPyYvRsbcJZ7sHe70HUUa3fO2t6QrGC2sHhxflc3xYYP2PvwxZ24+pW7EP
-          7uExR9p5Xsfl/nyGkHibiiLFkiP8PLC1Mtx+IrIWxfa4P3+y+0lkHEwXbLueUx6co/3lER7d9i2w
-          rB56KLyFlM5/45vNlYec/HrOl/AnlHBVQUG0ko0bSpxxgZ+rOKLjkyubBRJFhsn4C7C0/771Nt17
-          2F6/JUGDNHjzW086+DWkIzL58pj/wzfuZeT/8oP1w34r5UwdH1eu8462l8rpSnPZdxTvfLd0sbqB
-          3sxvAZtdlmZ2LGeB8uXTB1JGc7D+ODaBJXoq6NyL3rheXUGFVVu16KmYCyXhox1gxgkhcaWqpvNt
-          evUyejE1OfFKHa1j0jNwYvriDx9Gnp+6DAiI54k3yoGxnu8rA1o7zpG6VhldpPMvkJNxDPY8wDE2
-          q/3a8KS4iNjdTWvYs4j3JlSmTR4Hk43oRfZ5KLL0EsiafctHlzf2JZZnibvLeTY2l1ucv/wEqetz
-          aKgRPQfYdl0dyNxrjehLra7ynp+g5wwmul4eMgv5PBuQ2ibfhkrPIYDXuYuDteJvI//erErZPw9D
-          3+vA4tp1Igf39xXFpOCi5aDKPTQx0Yn7yRhj/Fn13vY1T9DpTjxjuEPgw7mxf8Hhsd7GyflwE6Td
-          yUBWsaqAe6rPSRL7am+D3VTNol/DAs6j1wXSrg/4HIeusvtzZL9vYbOuj/oKezVMiJtGwMMNtRj4
-          adwNGaN8izDrZS6EWWwRJBrPccrwhQHgLC3kfBk/YMiCVpfF6XcOxPRy9uhdsV1ojfYlYOQKGzjK
-          Ax/E28ZjCm0ZECx+C+V1xGfi+dPdoAXyYuhZh5ocl0Ns0GNcF4o0wQydFONj4J5HV3hc9i0OsyLT
-          5Wj+SvhtmDNe6tge2bIaIBjmTkau/NJzns0Z8w+P0fFg8Q1lwckEf3hwJ1IdUfI4YlAaVw0ho63o
-          csurRTldDCcwnF9gYG7cYqBQv0QqX41g+0RbqyTfGuFBs7m8tVsuhFHnVeh4WwtAdr38V1/IbW4T
-          6JtaDOA9Z/JA2v3G9t6sGuLvzP/LD7ZLdl7kPz1pyeoxX9/aUMOms2N0PA8u5XE/DfAYqhOx1G8W
-          vY+fuFNu1DSQS59LRE4WVRVOCD7oWLFqtPjHeoP7+CNtlp2cPWgdhqGZEaIpNhw3M2I2GGlbhvT3
-          /WZMMtYyqIapTs7HnuaTXF4qpTRCjRTsC42/Z3hu5bO7aQGX0kODR75Soek+BbximBpbxqgqcJ/f
-          +7/x/coBTmGffjN05hlsbHveCO/LrcZ8l65//LoonL48UGFbpsf7/NIpItuF//za5tqxo7TXT4n/
-          +JuvdY1XoEzPyNz9N1sc+BjyKOnRn//cgp7zYVqUv3/zh04CUSHXqBV55UvnEQEqrfxXr0V2OUXr
-          0WQzRf/Qdtd334gEjhP+zR9ybIRDtCUTw/zLz/707ODzYgdZposC/+zaOfenb3HTCciheZ3P8/2d
-          wZexIaKtjOvxJIULpMzFJscXS7ymv2Ux/LJBjJn37ZSvkMSb8lAeh93/6WDB4oDBbG/VvlHf8rZv
-          DFRoDjODnAvKPTrdIx/Edv5F3oWk0eJwnw7GetiTlJbXZs8vEjm/yQnRMBQNvPSrClmmjUig1JWx
-          /Y3vC7oP4q2HQ77I5aOWlqm+kXO+On95nQ32/Bg9X2zabAnEvFwecxW/bvgT0Z7rXOg+swQfaq6P
-          FvWS6nDP39FpYS5gZdRLBbIVX0hwO1uUl5iXC5XD+YHQpox0MPYtD4fqzO/6mxu344dUf3kOUeFH
-          MeZofAVwvVlvov4KyaNPMlac9DHLYPNOmHLaZrjggLaZBCT36LJlQglexTFC6Dw04zofpRrYFDfE
-          qFkDLMLTx//mn99nRb65duGA7Need30v7frkw8g7P2G+id/eBl7GAnc/vfOLHvGWGcjyqzJTkrLv
-          2KB+vy+ZiTRFZnm75tOaeSm08+aBTubv3mz+V52A9D5mwXtWxGb//jX8HAKKe+Se6KhrUggVYGfE
-          GMTPiCO3DmGmPUzi21ZrUOP6buW3qUzIfcUgIs8XGGDi+TZRVYt49OpqDjyId4mcZC3aH/HqIdjz
-          MWzVVBuFIck22UpkH7PLK6HCn19uedoE+3wDFKKOhe+Ht2Im8lq6GjqooFE7lMS+h73FbFge7Hka
-          Mn0v8PD7/eugja4JZtJL0kzzV07BzhfofNl3uYvnXP93fwPfbfcte14iL2kSYnH303sekMn73/+N
-          73JkpQpYLuui9IKAt+dD8r+83QhnBJaaogU0j4OPTN3yIi4/PANgE1Ji9uhMI6VDuv3lCejRy4ux
-          550L2PkVM+tzAksbqTE8TvvT6Ts+k6urucrlZEBi36+l1/+tDzBk/gbcY6ki6na5D/l7MWF+e1rG
-          v+vzITgHh3RLvH/1/Tc+mVC+KeWhVsHuVcjkJB6MfJ2+jyt0xMlABVu/6DYJhQ2fV4YNpFGsjH/+
-          d3X4MqC7HsIcqIM/PEJ/+dDyVz/yp5uI+Yz0UYAHXhYXNQ8CqWTnaM/fN7DnR3iJgitddn0C+3Ys
-          iBN4Qz7b0RKApy1uwUQKbMzVm4mh2bWHgFmfPuU708V/3x8L52Gg691PA9huQ0tUXAzexo1bAuNt
-          4ZF5Dxdve2/nCkxd9UDX7HqlWEHVVZm/+jdYz8M7HxOfsSV6YRi8sG+Qz1c5V6FFgxs6z8AHcygH
-          AywP3QkFrtOMG3h5i3DfH3LZ8y1KIckYOMwr+HsEC3BtjBLIN/sS+fE4RAM5mSIskqzFSkj4kSIx
-          jIHmA4+Yc27v+a/LgGUiCbLerG7QdTAzuF8HNKNbtN5hFSo7ngXKjRxHCoJtUyTu8iJ5xU3ezncL
-          tO/fHENcPI1t51P5W+kI+bhQvHW+/zLgLK2KSumNG/xLSAV2/sWwvo3efBmSDQTeM/zTT+NqUKMF
-          f/x79EbZm4Xna4CXFhrERgGMlnLJWICPV7zf37pZojzOQHOW3D89RgkgQfvnHxDimCmiFync4HAr
-          BHJ60QlsrJc50m301AA2dxlQXZ9jqF8+L+SyT63hJaWxlWWaE5TC2mq4QLy6ivK1EuSJ+grmSeAC
-          6X3TN3RmjCXf/bQO9/kQSJ1IvL0JDoZpmPa4iXBoLHdfSsAhYihByXIzlj1Ph0U9fvHIKdW4ne8p
-          +8/v+3M2R/t8rmGJXgoxHqs+8q/loipQD3RkuqezR/XtFMPu2zLouhZOxIePB4Z7/oOOBQ3ANvK9
-          DvVn+Qq2v3z7z198c/ZEPJbRvG8kXwKl/BY2cufyQJf8VTpwe7xnFLxvp4jlJ7VSdj+JaTBiOr7C
-          0FFukaogfStFD+cMa0LE3n5IvaIoX6PxHkBXGZ7IZlrBI/ep9pX6/dHRE+qa0Q181Cu38aQSzR/f
-          dEE/LwXm9bMf+7FU0Tr35wTYXpyi4wINSm0T+/L7cVqRcZ5QvjXXzYWodfs/PByXoZtbeNWuPtnz
-          0mY1DNLBHc8IKteuWbPgYf75N3J7LGrELqfIhuNp0tH9xSKP2ix/hR/30BM78Kdx5TwaKibfWyiU
-          G2j0X7Zi4OofnoHwE97Nurm4hPWjZDBAsxBt8chXyu6/A+mgtcYanvQKLlN1Q68P39M19oTiL4/F
-          BI0b3b7tIIrf/ugjS/3KOd71F1Rq5o60j5QA4RNbNfjTiyZ92v/qE96PS4KKIhpy2n/cAL4lXyCv
-          02/z6OH9auGghdu/9eP1yIapEicjxjJ9hrv+uNQK6s2OOJ809Jb3hVZ/66nBp+L0ced7R+nT9EFs
-          seajP76HdjM+ya7no3948P/ZUsD97y0FV+5oEdfjq2ibkT6Ag5i5xEeRZbCv7RHA0nxdiT0WqbHF
-          2JNh2r1uATN0RiQ4ud0pL9fBBB0qaRwNyRIB1Yw0QNlBo9x28kRgqtuI0C+9G6zXOyG0WfOE3P00
-          VRwI4wJ5/5JjXi4mQML3WsFvXp6wEDMgIqIztRB8iI28upzHxU7TDb5I+CGpZy3jGp/dBe4tgomn
-          JUm+HUKkg1Bbv+jotGRcb7LUQfEJXXJmzi1dEvVtw7A1F3S7oznCz1nslEeQy8FTP3bG0r7aTXmk
-          W4mcXqbjeu7pBsvf50Se2ng2aOyKJYxvDRdsg/XyVtFlZBiI/hUFsDmM9AckBgzF7Ycurhd7wjBz
-          paI+W42UQ9dEG5GPPPwurYZQVV+jtRiiAjJtcUXnpiJ0Od2VUmKrlCH5XX+BDXPAV9wFQnJ1Ji5f
-          T+wbgrOKZnS8ZGy03vqqVL5NRsg5sFLwCTIhBN79pATK0zqPwu+R6MorzV1ymj/tSO+2U8PNXn/I
-          PiU65YitdHKaT1og6xznTVKIY6Bxjw2h7aHnwsUZXMh6kY2MOjEjyh9eEL6CWcU8jqJxccNDAKMz
-          NonGCV9jrW89Vj4zfyR+eywAX3hFD3lZvRJXNNORD4RxUy6ykARQ5Jx84fDcw0AMrlg6eigXuisw
-          4aBKECvkdzL45nktlSQzrujc3b1cUM6nYD+5SyIXVUqi9dyDBW4GBAE7MwXgmmMbwMdD6In9ikq6
-          qJmcwjV6JOSSEgJmtw5ZeEglQk5o0UYhRmcTRrC2SHmSLWMDP2mCqj11CGWNA9hCL0zYNu1jHoG2
-          jfM6Ywc8BZgRsywMj3U+EqMAJrbQ7T52ERdkhxCmD/gkpVA4BivLaqXEJ1UJVt99jovxdDBcb46M
-          MmbQcgGx7gArBxKS468NBPdZyODj4JCoeI2a+Y4eNuQ6M0PmkHF0/XyTGj7dc0Oy7e1428HbMsWq
-          mQNmc172cOEVw9/rsehthbF+Gk1WDsAXyVPY9LwvazuFI/NWkIakAWDrfYkVWex/6Hm/Jw2Hv0YA
-          eyBLxPGssOFKAQ5QQq2EtL2ex8z6+PDhVjrKtndvdIfwqMNbOT3IfXhs0fplhQnSjLyxZJcG4H7h
-          s4TfB/4gI5GTZikTUMG26R64s0GVCw4Ck9wnVCdBJOFo6wbaA0GSnigyk9zb7ChmofShAKnyWct5
-          /76Wijw+7wgZJycS9sU+KK3nH/JLh+ZLG19E5emiBr/9m+lx4HwOoCCBZ0AXufI4HMu23LBLGiQ2
-          UPMu/p4H6MhERmoW4GaJJ61TIuWL0HkzxwZjx0+hYtkTphpIRzbATQCL+u9k91NEhVzOrmDRkI6Z
-          WfuAJat/GB7cX0RCVznm7LgOHfR128bLwxoMWt1irPh80+DqEzT5Om+JCebCO+z1/jPW65xm4K9e
-          luOjypcAj4FieeoD5Vh85Gvmvh1YpB0bMO53GflEbzLFp5hDtnBKIy58njuIkxyTc9c/KX0vWwb3
-          +Y9e3uHs/TgzZ+BTMlqkG5mQ0yfsOyi+jJjoO74s6gup8PuYPuRyIsDbltunU1I6fAPZuATeLLqM
-          CGOpq5EhP5dou53ePeiEvEC6MnTe8k0qHwqnu4iZJA7BlEnHAo6ivNeD7gL+a19SZeuuJrmyqRPN
-          Wt26ClXUgWiqlORLZF10xcDPN3n54YtypwPLytJbNjBzfNUjzbSwU2LsssSGOR9tgsRWfNRINrlb
-          c2xstQY3+GleLVLHn9WwcxoFcjOaAfFOoG9ejCzqgNhKFFSu8s3Xy2XDyvOYcuRhfJ2IvdaK/u/3
-          nJ5imwt3W62VlydfyTV+cA3RqgsPwSanyOlLedyvWeWhIIBMkVmbdccHqHu+QS4irscVjboNhQrF
-          AZfUb7rdS81WdC8wiE7oDazjo9YBJ3gt0WOnMVb2fYBQbxENXgzFxjKuQwtH166IqqFlpPeD60Ng
-          nd7EaFcbUPyUe0DfrULCu62OfFXrmfKQEgMvjRnlAv5ZFTS7w4R56BqUrzW4/ONDfVU+YOGLooY+
-          Sw+Yj26TR+g8OPBY6Bh/wnTzFle+pNA9nO/BQAZ/5A68aSvd9fLZz5XQ8u1vftDOQwRtQwyG3iAL
-          lHnJCuSgvuVsk4euQrP5jSleFoOq0i+AaPoRlB/XtCFw+PjKPb5oSH1sXEN3fFFGSRnJM4RxtFjD
-          W1YcaVjI6evqHt96WqxQxmLRZRCuESd3RQuTur+iy0lHI31V3xb+9OVGXlnTg+1S1ZOCwfuDrjtf
-          Cw6iEyy/wofseJ0vw/u+KIZ4GtHVzGuwihfOAZ+q50n+kPKIP1zFARakRBgk6znnzyd2780Q+0RT
-          fRfweULqv3oj5yLVDdLfThl8MMoLL2F1pzPiCwb6i3xEL3R9j9Nq4QWe+mJFZ8Ws8s2/SyU0Io1F
-          WfxmwB/+/fEjpu1qU3ZwZRVyKtME4v7zqPViavjjnR734xoA7o1F+58+0tIoAbTJDz1ExqCSzOPV
-          iNPzj6zgcLvilRbm+Lv1fQGb6wsRNKGtWffxkyrql8Tr9y4Dr4p08ILjEt1/4ZnS03hJlfkJmqAy
-          bd+j3ZWa0NgqhF6t/ch/fflgIGtPLHoxdj2um34eIPN5F0gNkZULqegx8HmxD5jlG2Ks8eWewie5
-          Z8QCwBvpdMp9qBDhQHTzb9dthVKogGgOeF78gC0RwkIJyPeKBfs0e0si1Qt8NayIPPzt6OSIngl/
-          vNsju+5oPu34Bb5NSpDvz0W0NPP7qhxXt0PeuIzRJkOdBbHHjUQ7xuI4t4PjKjAJR/K0EOutZiuX
-          0qACGNDXAqINlUUKkdGr5LmtgreaT9ZXNC7fMG0MAtYTqHzlapQJMvxAo1uavwqw4z06vo/xOPfl
-          BUK7hBdyDt4roMOsFKCkUYTUZwIb2t+0TNFj+Yz8w35OE7d8GQjb95E4Wl3lsx5GOoSrccDK3TY9
-          etKEAA7f9U6eL9jTxc+MBW4/eyNuqAgAy5zfiqp4vxHrrJBm3vUgLLstx4nwueTLtztieDk+DVRK
-          +dpsv2zhFbMIUmRK49PbQlXMoI0vAgk+n9Tb6vNx2B8ke5D7xMaUPgZUwumLa+TqHGeQ4dez0s28
-          80RNhmOzHE6LqWhFppMU924uGN7LBKxTpsgxg1uEFebiKCfLLolj3SWwSP5lUhB9e8i0npbH+pm3
-          AFa4Poj2JW2+7vgPJ3Gd8NdxkmZ9f1cejHNBkdP+NIPeOG6AgniSgkPJ5vlWpgoP9fZMycn49vn2
-          TYcayqx5R7F4sT0WH4wNFo+RIaf6lFCqP/IO/q4WQLFaPCh9C0sPT4N/I+U7/ox0On9ZWKfnKhCm
-          ps63j2mYinV/XwKZ+/EGZS5PVt7TDmJpTUPp6zuz4I//+PTrGjwndAEokXhAFnht435/IfzTO7rO
-          MPlsqaEPfvO7J1meZt5iJG4JmTD/kNO1SJp1i5MNKNHdCYzk1xrLITAxlDz9S4JHUdJx5yf4/TgR
-          ebyLMVq/gGNAeONj4orVOk6nwW4hMomLvNf3aryds5RC5fJtkcXeTSCUKhdCyVO/WNz8NVqX36MC
-          TLY45H7h52jtdHgFDOPmxD72HuV/84mBcsm8MPC31qObFwRQjqoN6VZ7zPnuVS2KNzEBMm+5QOfp
-          mMrQZZwzOnKc6lE4fAJ42+yZ6Gza50uhxyaM42IMlN2/UOWsBQqwvDe5H9IwX41RnGC+kgex9JDk
-          ZNczihx/INKfmQo2DNdCIbmuojhSbzmfX68yDN+hTs7aOHvbn95QuepLXsMbNO3ipCkcAuuM9AtD
-          jeVSVpnCrbVDrpXNUfopLUZ55ycD+b9WMwTKVtl//reN3Rw/FJABPXJ/eIjDKd/Ugp2gyHo++ucn
-          Fr6b4MbWFVYUU82FK8kw8EAwBAwfGQ21beLC9y9/EU083Lz55zsstN7nHB0puntzFigbhKmO0J9f
-          WkJopjCbcELMcr4161f4JfDj1cw/PbP+8cs66bd/+EFVReGhVbkX5Mg3d1xOQ9D+06N68xiMtf9N
-          rDw6MSSucQmMxYRAlXe/TM67/52VZ6hCKymPRA+5oNlC66yDd7MYqCjbKVolcZ6gDo4BMt/t6i0e
-          jVwgUlvH8l5v31eRQcg9Xxk5n9kuX3ol3JSXJ17Ji6GBQdJptZVrx/7IhY9Uj83DtIXMs+UxIxSO
-          t/06A8NTWPskgM1rXMXDyCh8LOfEYTrFWOZLDKFhTAEKcaV5SwGeC9j9HcqBtjU911w35XjernjZ
-          9TdvCQcTRpkP0T2eezDlp2sCn7zPB2vejR5lZFGVRyeB6MzoZbQ8F1eGrXKtMeN+w50fVFcJD/WT
-          uNuWNViQYA0qclqQVok/Y8XxZiu73wkOknXP6ev7YSG3Vg558EKdL918ZkB97UwMonOXr4ypDvBR
-          ERN5yY3NN24hEO76NejPSG0Ec6a6UgdNgo7TcT936MNmf/wZdAvmDeKb5wqMovhGrlhdxpE4B1FG
-          3JXgmmXf0XR7LKV01g4XZH4PUT5P3lxDkZo6ia3nx9j2vAV8O/Qkx2rMm+1UvDPlzw+Y3+sR8Gl+
-          LyBgEoscUT5QWpe/K3R4XkfmtYX5D493HhZ5XiM/Pg3GdDycEsg8Ox7T449EmylXExhJcCeGa650
-          FQ8NA5ZFM5AtFxP9Hdu7A++TrAYEXbWGm67Noux4i2L+yBsrpMoGPbLBoIV12tAL6jPo4LOOXrFj
-          N6w3FSIkrLHsfO5FfMXmlXiRrA0vyurkwp9eucNzFxz4IAfTXp/wXYhflM/l3yMTRfWXFxCEmtib
-          D9pdhd0gtoFYvy7GpsOvCu2SuWC2LBqPEimN4Yo5A3mfLh+X0NAnuIVPSjx/Mw1u7YgI3c87wGKW
-          3AGf3l0etobOEj0PQzpl7MbDItZrTIUuBp0lCDYotaFFtg2qiP705QrLp30J6k65GGMTtgus5epH
-          TgP+eeOkxxDUxAbIe6+6t/351VYJa6QqNxLNOBdjKD0iA0PauDkdVLtXngKTYZHcbpQKs83CnqQa
-          ud/b1tiO3LIo54Veg/K4cAY9aYcA0ptpkdOYs83GMm355weCcde7gnUKN/Cnz28XSxnp3IriXz0i
-          Xb9cwRLgxpf/5kPZepqxHnjThJ8Tzf7Vxza98wUu/kCCQ2tLEc6f8wIsT3+g03waxw1z1FemKWQD
-          3BZRQ4Eri/A0BDek8e21WSs/dqA/qGOQ3IzGWFNe27vK8QFCv2Pb0CooS/ldyF/M9UPb/LIbq0vH
-          R6AS98ydwKZ32lWxx9cJmW8nNdg0cK9wTD4pOW/r3VujWNOV/fXBgXvvfrY6pgpt4jc6Xvg5xzDm
-          GVjxmv+PL9kyoTUExkMh3rGyPf5UMTrg2IRH7q7fpyf7qsHyYDWkKmsf0brmRCXvoEUyvdVHGly6
-          CeKj7eCP6JfRNieOLP9dVwmQjaVNuwqamsCh4B1bDfnL28o8npF2Kp1xymbYQa6zMwz+6vHoxuG/
-          /Cb0ggcQ/vhxzm2OHPVP52GPexawuWRWoHS9QpfbQyzhGrY9ulrtMeLYZDahzAMLWXt+tXJmDoFV
-          ORfkX/B33H6zBqEVTw1JiHrwhkIyChBXh4Zo0XOla9gavCLq1RNddYaJ5nMy9LBdeQ6z+efYcGNV
-          XZVdzxNbLny6Pi3JB/QgpMHZVQZAccJm8DaiC/rTq+sfvqD++CS2FDk598aLqcBVOxC9e6cG5fvN
-          h+9WOGMxAly+wMcmK395r30zDGO09GsNilityfE3dYCelKGCgmCSAO34vOcbOlBsPgwmykIwe6+v
-          D5/O67BvAX2MtDtWPuSyYSLq+bf7hzx1/t1vvXm4xiIzNxcuD14jdnlTgeASrALLtxriNZ7k4e0j
-          pvDYHGLkrIsYTVUbY+h4khQ02uvQ0ONTDOD2SvsA5nNIt6UYecDkwnX30++8VzM5AxXxFnJ8H9lm
-          aebfFX5uyCFmfgqajY8MKO7+EFl3NOdbN4AezqDPyN1p0d/3D/auvV/ieDuQ3oZhAVpn+4EISmJw
-          YeuxoFV9mxTVAdMJH4xF+sOjP/3GimXNAgYlGTE5vmmoRJ4ulOrDGWnvDztOd0XI5PPn3hJ1AcbI
-          fZqTLINChSRI9S+g9X1UoX7HAEso7ht6g0AGChVncrfUvpn3fAP86VHfxlewHTw5hRryF3RtuMaY
-          lpJ3QG4OMMAvl2u2UR5deCCrS864apv9edcW8o0rEP/LD95irb0Dj/eiIHrG3il9fDYW7usXyHiB
-          N93Qj1z/+Beh2LHH5RRVLOwGuSXmwOP//BiTc1dinrAwLoo0+wA+jg+8CpfS4NDmxFA//Krg+615
-          sDjT4sORaRRikW2LtuFlFfB3UVdyKbHkrTd5beEjXcqACjc9+tPv8JPAaV/vCPJt92cwOiQUHa1C
-          a7g/PfddOg0Ze32swFk6eEgBwXPxY0d6L7ormNF4DKQ9v1+Oh1MMvAn+p3+E+j7q0O77nBSSnYFZ
-          +w0pFMZRQ2qfrNG6nQwZPqTYCH7SYclnX/cCGMnLG12z4xSt4kVx4J4XkROK+3Ez5R4DMyvZv7yp
-          Wd4PjVeEmHbE826vpu35awVdmP0Q8sMDmPf5/JcPIS8629EnTBeoRN3MYjwIHNgU/cPD8st9SGrd
-          JbqdKkaV/urvvK9PrJSXMSTs5KJ417dz8q4DOBxmSNSuz6LFEQ0b2lv8QUban5sNRacaPu66jEz/
-          JBtUSN6mcgyEghhLefT4rlFSkBbUC7bYMYw9P9GBuOU9MXywnxNtfDeoGdQg+l5vs63gGnTX6BO0
-          OZ95y/m49cozAvOe12mU2/UdrMz0GUCOmUDriJ4NPH5IdzybwMpv+gYm6oQI5SGbz2MLReD2d4js
-          o8REfe1ZieJfHlMgfYVzQ+Ew++L8/czE0ZEHcATVVPn7PUZ/O4NtJOn1734gT1ELbzvcrVC+GkWC
-          GWGWmy30vSu8Zc4RhePPGvk0f5XwdY9V9HhpPV3ipfSFtsuPeGFfw7ht3EsE+/iS5ICH5rfXG0i1
-          e4lQp7DNqt0fV3g+GDPSy7Tc8+L/AwAA//+kXcmWqrwWfiAG0knCkF6kCwIizmgUBREBEyBPfxfW
-          Gf6zO6xVZ50CknzdTrKlDPb2s8PQ+5YFLaRo+3vOgZxDdnS71j0mP79DnOmSxfNLs2dwv6RKIBuM
-          Ov7VDwJZwEgRUpN+4uSoAGaWHaTg7Qzy2+A7AEH+JKpSVQW94ckCl9MtDgQrv49081ty9BkvWDgd
-          VSpMSWeBuX2dN/850kW0hlDa6k/IBWo0TvWYM8CNIxMZQpg368E+MhBZJcb0Dh4A5+2kQCc/1MGy
-          LwKXj9pdAsFis0j56HD7/56R/PP/p4Nu6UJ36AP4y480gJrme2QfDLytwREdI7d1V+eAB9idggJZ
-          0Z66i2baDtzqBcH3BaJ4UQB9/vJIcqnzj/7xeDGER574uM9TAXzLpvLgVh/F2/dquIv3vsGnfLkQ
-          +7WaIyvAgIX6iRkCOXic6Lc6eQZ8yucLimrrTPl+V1hQxjOHjkWlAqrv/b96EZ6+dq/j4W7e4KB9
-          6mB5vNiGeDIKILpdY2Q72UtfPOCJkFtfEXLWh62zxwtXyrci/RLHZm/NMl4HBfZC8sb1685tR7T9
-          DGx5MHL5WXE3fAzgIkULMgo+1/GWp8mPoHfRnVo+xeK7yeRf/lp6SkjXVMjKv/xqm7/j9HT9FARn
-          cAy4Dj3dX/4Gfvo/ZPVTQyM+g3/+ytny48WMPskv3yE6fCsxb1PkQNvhCfH43KTcL0+MYW2SkrKQ
-          zvuwS8BvffzqWZNGcSnlTuETr46a4qe/5HcQCihs6imejhfuBrnxywdATdN4iQe8wlh5UoIspaFL
-          2Zw9mY0lK1gGx6KzeBtYAK5egK5KwurLVm/Zv8tHTFA3aTqbsysL+Qs5IiWEIVhPjIXBL59Wi851
-          Bc4va3B7aULAO/MVrAAHCdwzuYSObdyB5YuPN5iTh0GyLf/7/R5MwyPE/KPCdE2eiyJv+ph4VfEG
-          NFt8DDd/QGzcDwWdx3jzZ9cncuduLqht1KyccEqIZXdIY8oWJw36iv8lhtUdm4G/3Xjp/9lSwP/3
-          lgL9tErksHQWFfLukkn7sDZRoC9aw3KPHsOqXhtir7Wmf4OxcmAdL4D4kb+LaTaeIigZbod393NC
-          51skZbDpBwuPeU/0FaYggfCt5ugo3nl9HYSpBd5QBkQNVA+s4oO2cg2VmNh0/3RnwRRbaPHzh2iy
-          ZIwTdlUPcraSoQh35sjlhRdJXCldiN9To+AChjPgA7OIZLFW0fm+lyIYkMcH2YyXjx+W01Lo+4xL
-          vEb9NMvudHrKl+/KBCBi0EjP1ekmG6JFUdCUtd4vacjISlHYCHW7bTP2sU8B6nKFxOVFi5fPjhpA
-          8cxj8GhZ1yU15CXgFQ7CtLwWxTK2DoSRcD6ji2jcR/b0Ail8ZupCMsbLG9pLt2Tfju8Os1ko6DMz
-          30Jo7laMgkhSi/mcuRpc4mkgmbxTC/5o9bNMriIl2SE1C24vba17QzRiEY/PeA6MuJM5UJ/JyfG+
-          7nSDQwpXz+hIfHkUgKuXcYWoqjoU6BSMEw4uT/ho8xAdK0HU53G3z2Dw2X0CmCZOM+y5ogZS5inI
-          fnu6y0WHkoep/7Xw49a447KGay3f02aHDqKxG5dTnUiQuyQXpL95NV620pFsquxK8mmKCs6gFQt5
-          x46R9th3I3s2RksWv9oh6ITYLaiUygpkDjTG3OKGxVwseQv3r6ZHxvQ+NxyTqJ28l65vVMmOBvjp
-          1j7h5Tsz5GKVjr4mPR/Ck3a9IR0NjcuBu8pD53GXsXhcX+60xrUon8aqIMkynF3MXmMR6qdZwkKV
-          280yCU9RfurHldw+UkZnhloJ5O9KhC5H3XF51V0yOTtFEE/G1aVrK5wtEORPi+hRDN1ZnMSbfLiz
-          HSq3+cUxyqeHLIY5OQWqR9lhOpWyRm2IAn7N9OUlxDkcx6OJYrVtXb4S/A7yax2SiLJmIZTsPQDc
-          uwLktrRigbOttdd0PWjo6MdKzH1ZiYfITzQSm9cG4NUtLXgQxRYp8h3rU5BvrbHAVOMhMti43w9K
-          Liv7i4kFUeua2X2CGX7KO8Jd8MIxrvyqh4uRjSg7rqYunJgHluvndU+ShqQ6u6QZA7v0cMOMg4pi
-          brCkSfl+mpEFbNeloXri5Z1wdEg+hFEz3z5aBu2OfZBq/y3H5Tm5w9/7Ru/jPcbWe1Fk7/WlyJ6t
-          YuTxNKXAkbKURO+DFbPHvpBAOj90dGsHxWU/650FzyjxUdwohUvNYB3gZe0Z5Cmtp/NlaGK57fQu
-          kJPP3Z0/xaOVz+eLg7TPa+eurK5LMhoZBdMnfQBO1q8TfB5Enxy9w7VZrdNDk7tXUSF7VDU6w7vt
-          QP8yOagQhInOOLjX8M0LlyCWdrXLh3Pvgch4XJHJKbNO92Vlw0xK97jmODzSx77u5LL1enLZ1sfa
-          CtvFZxFMka7Jhs4NcWzBxjx9f+Pl9rmTT2I1Bs/goyNJn+Uys+A1zgVkr/XTXW/s+Snbj/CBsiUp
-          Yno/CxrgJi8lV/l7BkJJRQZOlazij23WDR+poSP7+tMj5jtcYlr4l/QPXxJvEehMuHsOc2WgyN69
-          pHFxvFcJy3vaE8WsVpe+u8yG7jeIyLb+xtnkjUA+tLhD6hgOI+UlOYHzgZ+RfmOeOi1rIYLXiuWR
-          hWjg8k149sDVsESifGzUkFOyzJKlwy8xm+I04qwwIXSKNSPHohCLBs0PSd4Z5wD5SngYeUefsYz5
-          jCeZCB9UaAslE5uktElwZ6/jauifAAZtaJLs0z8pfzxmPJSvXErOTbGM6yKwFgAtuKONXyjXxHoH
-          csujSMm6EMw7nQ+hv5RHVBUmM86f4tPBi/FViX+0LSo4qpPCBnvWNv9ql48jRfvNL+KveuVSt3cC
-          GCpxg+zzTW7IiXlMcmD2B5TKrUPHZ8PzwKn7jITu/jSyLopZOGqmjcpX9Ylp/bkxkNxLSqLf/PRf
-          WQkl2DyJETjCFtG5Ney8850g+8I3y7WCCRxPxYG4Pz51DkiBbdrqxH7R7aZLrvL+fq8fCwhWWuqM
-          PD0FFR0q1BdL+TEMyLbPgrgRQ5rlJrgiVCTdCrjFnYu1zr45DD68GXBrGxa8OR4S2KbkjYLOvLpC
-          utx7SGPjQdSP4IycWZm8vM0nzHyJOLb3Lk/ArIUz0u6VBeg7VGzZuJYcOuJrCtgXXHi5umgtCrTs
-          OX4l7sQD9XnV0aVCdkxrmPXy0dXuOOHuCWWZkk9k9znHpJIvUCfn4/KUe+YR4SWsvJE/5ZkHXRSd
-          USz6is5JoirKN5TXKDh0WsFLllvCex0zRGd4EczSO0xkpEk6KvPjUPBC2wwQMGNFYlLXYE2iLpR7
-          1oxRFADb5aW1uUl3FxYkPEYnyo1ho8Binkak40Ry16H1BijM74VYuoXcjY8Z+KwZl6heP4HVzTsG
-          tIfIIEeVVV0i3v0AbOOFqiC1muXcs61cB3uA7KvVuitOy0Rkau6EsnrWR8q3r0RuBGgSj+y3Wy/E
-          G4SBVU3EuN54ukqWftsusr0iHx7Lgke8ksqPT4SQel/cYpakOZQ/vB+Quwl4SlvbwnKJhzexI85o
-          1qvmi/CheBGx06GNV4e+FZgdahmf7YoW1BmkCFi9PaJLTu2GT3nGgKbKr0ivPc39MMqjh1aFR+Q/
-          v6a+KoBgGBSXEenJd+MWbjulx48k2HeSMM4OADx0r0qC9KF6FFNcFx50xbuAgf0NCzqwvAcj53DF
-          TGKrBWVcOMNcTVnkq7d3Mbe7hyjzyVcIxHSSizW61DZ8qLqAgfec3WnjBxjv1DAQr5bhEnbwb+AK
-          Sxa5UPIaPD7KDkL+cg3A4R7GdH68VkiP6ZNoHbR0onCjJn/s04kEblW4QmqcMlDzZUwOxHnr6815
-          tfJP797GOAPLUysmuN/FHXErIwBrGfoT5GxOxvsAX4vBOn0UmO2kPVJ376NObWu4QVtPWvwVjGb8
-          OE+uhGl5MRDayytdmVMR/PFNvO9qlxqMgyGF9m2bTx4YgB7nYMH2B/kbv1GgvDR4sIyeJGdWiOdB
-          ezwl3gwTcuH63XaftZHI9B3df3jhLs9J7yF9ZjVCVKDjyg/6DOsqO6OTIB8bQXjIEdyrnkWSKk5H
-          7OQKhtZXG4j9HQCgG57Bs7N3yGm2B32+7DRJBrSpAoFnu4Zk2b3dz+VEAtBnPsBMS20oO48Licut
-          tT23RiGsAwDwza2AS66vvQ2503whly46ACFp7ixsPd4IxFBvR7rpH+gvry8JCrkGq3khGRSF2wGL
-          30prfu8LJwPpm59IXUziNoK3uLsHMjc245oVJgM1Z92hAHDCuOnbHO6esofM8BABYeNDQKftInYt
-          NimNX48eSlmgoKxHT7quWwM//3YgxH3BwF06h+YwCnYZll9uBajv5QYozXZA9pKAeGb0UJH33n0g
-          1tBN8XRCTSK7Qh9t8wnoa/+ZVnBnk4zYYWAAmnifFjr1kOE+1mRASyrCvcM9L8TRZq6YB3PHQ8mW
-          VOKVL0dnpbYOYSPXJnF4eWkI8/RWgCNxT26O9nTnRb9E0L21CTlJ7bGZh+t9Bezg7zALQFJQSxxa
-          aBgtj67lRywmvlS6P32tvCU0LndnFOG14nlkSN+kmdvpCMEPf9Moxzr9tf7kzSghhy88xPxpsUrw
-          rpZr0EecMbJdeGXhmRzsYAmraVw1HrXwyktsADv+E9/OiPFAg1qb+FPf6hN7uluSd24RuVyOH3cl
-          8RSBfK+RgH3YaTz//BdvHfYkCwOD9iUeaogjaU8Q/lCw2p4iyXc2zYL9QC4FeZm3HjbiMqBkvJ5H
-          NvJoCELz+CTa8dsX67GcZ1k7eztyOvHT+Nn8C+SZivv5l2bG7OLIs5/NJAON2sz7i4XlO7Ne8NL5
-          o06I9OJhQ/SRWBX2XcrJrxzIzfuMNFYwXRZdQA2qLmlQ9hvf+87Gktjvvsj35G5cG//rgJ9fY20o
-          umQW5BXUVX4mhwM3uN/mU7SQi8psm58rHUxG1YDCVhbRu14d58EUePA21IIc6TttVuN7cuAFCSrR
-          huVFV638RJL6+GpIu9B3QXNvb0Hxi+1AdpodWBgViUDwTUoM5pnE23oJ4emwuwV1CJd4QdDXoKdm
-          J7wyBxm88OrM0s/fePtdr8/PIjfAvEVKm14eP0nmTdDczRjPZTgXf/oX8Y+WqAbzLeiBOiV8StOR
-          qN9pKNZP2OI//6zWO7eYGTJrsqMJO2I3F71g+VJpATdn53/82RUTI7mGcw7EToExwXIlyfmQL8Hu
-          UyQ67w5iB6X3p0H6yYTF9H6o2ynK3Q7pKjFififGluw8KjmAw3h0OQ1ZCrzdpjuKjl+7IL4XWb+/
-          j3z5UrqEtGcWti9JQ5qm22AbL1v68TO68NxI5BbyMB64KFj27nGcFKvtQTvfO+Rgc43pG25bU6vl
-          ijRwMwB9OlkNiFgv5BQbU4z16+LB+eIQZN0PX5cUeyEFehnst599l40d3ZM3/A/W93EXz1SqVzmg
-          nYH3B+MKljXQbpB8vBWZCpP8W4/Q35XB+8ANOoYpTeDeskyis1kdr/bWeHHvVQPGbqjSJTibDpwP
-          7BzsJJvqy/XQJrB0yg+JT2ZZDBfW0mSjX1nkF/prpO/9bMhP0GDi7h5TvCo9WkF70vSAyVuFTmN8
-          /csL0GHdba2/2a8GvcOBQ4qSr81XQ5YG+u3WUPmnp0NDlaCIfIx8CZQxPtbIgH4VOsQU1lDnoTIm
-          cGgqB1nKHcVr+zAt+NMPtLw8izXwwgxK/qyg+wmdKfcw9fn3PuimL8+GavogARmTMzG3/IDTXqki
-          s4NRELXpXHdNeiaC7a5X0EVRhmZKzxED3s/PFQvJ5eOSTz1rAFZxgRlTLcbFzKsZbPkFFli/A2va
-          3GxI2oRiKRyTYj4yowhzpiWYdpxEV/PyzqD7lWKk392iYNk2W4GyP5vEOl6CZsrrAUP5454xPFkD
-          WJ05xLChaEBW8AriRZSfGuQWX0JKbc7ueOZ3M9jGP1iNwQZj/Fw6mJ+fKzLP53VcbW7m5dNDLdFB
-          jnYF3t/dWjrKfBHs21gA05oKnvzRyAnD43ts6EX9JGDTS2RrkaXT70eG8HjtK+J9oo9Oj5colX/f
-          71CqdjOvqRAA6XloSNDpistjjc2guugqsrLRHzm1tyy4bI0TzmmzxvTpdzz4zc8bMtVGCL97FjZK
-          gfHG583yLpoQir38Jc7Dp+4XvTgF8g6boMyADKDqukLZ96FLqkuxc/GXlJm05Q84IQEXr3p1qGFx
-          Gb4Bx19DHQvt2MN2N/zDE6quEgPF5ICDDocjnam8s8DZAQ7yX64MNv7rIH8tI3SabkMzU6lfYWub
-          H/TTR+SVMd0Pz5AtYJXOxWHoZbuynnjy9LGYpvCTwjKRPpjeXVDMt4+Tw4BRdOJzoz5y72KMYJ09
-          U+TEVRKT2a1WID47m+jkm7uLZfgzJIyxkEOV9+O8/6YYBB/HIM40RfEaUqYGrFtWxBQ5Qufaj0Lo
-          7VIfHRclpWvw7DOoz5mIpWla45kpmVTqLHrf9CcGJN6rrQwjt0ABMtWRD5B9g9ZXGUiVlC9A2ixM
-          5O95kEnAKqJOazT3MDSFBrlg8hpulxAWWu9yQD7TaFToUquGOgdFFJym35GbnQhv36QiVeB28XeI
-          YwN+rZQh9i2FW4nwYMkfOz6hwFSL5udnYIMDi2hrYcbCS8MdlOL7ixwfk0A/557tQAJj75dXjPSu
-          3S0YiuMdHdr9gdKsfwSw4voniaLXYWR79w7hyeMq3Jlf5GJauhA82IbFwzStxeIdzFR+MHYQiPXc
-          jLP0zhKYPDifGLF/KeZOUmd542NcNHUQL0lmYJDEPkIHM3C2n71JAsynIhU3Ns03YGQLjKl8+MP7
-          VVqbEnw6gDCd5z0l9509QelJW8xomTbyPz7OpGSP+/PJiDe+m6C1iAPy5XOjr1f8a4RZzQERCdVp
-          Nu6esEleE1KO0UK/ah9YYNPD6NaUtUvP9HMDUSxkxNj8sFAdhOCXRxAt4w9uXxROB7397CPreMHj
-          TOyrLQmHDJFznq3ud07zAQp16f/lx/Ma9xJoT4pOPKWddKE59CXQxAki9xYFDWZJI8G7WpfoqiPJ
-          ne3RhXDqC5Woc+zEC687AXS6C0THk7N3Z8bCKdS1EhCzWFZ3nstLCmVGT4J5y7Nm5zVuea8T/55X
-          Z9PxCYF8L9INb4760mtjALxaQSi60EM83NjqKd2H7hrs1yDdSqCeBsyTs+BpGD86vZ93GrxzRxlT
-          jrk2j7EDym/8Nz0Yg1l/JQmUokEh2pYvkr7pDTh3JUZJl6b6+N6LBhBhYqHT8nBcEoFlgD9/fT6f
-          o2baD3b28xN4/WYvMLeTCqFQhsEffvP5rpr/9NE78+ztPP/hCe4uU5Dg/q7jWZmTFgZ4MkhmEHNc
-          eq3xwPt7yIL8cqCgZvRMg3HxiJEqyMdxMSufhYD1b0if9qq+pNHUQVE+jiioOSXm1fsp/cuXfcHQ
-          R6GG4QCFovOJfpWLkZogGoBnhSHKPr0G5t69M5B3+IToGSlA/xYECDf+QPaD0WPB8rQOtmytEN1W
-          dDrLHJ+LHtN7KIRWoNOvYinQYz0faeKecbstT/jpUeKHaxd/54/U/eGtB/w3+Dax28JtvqEUMTje
-          rs2v4bb+g4/5qcGUoZWVr8deJA5UTsVsf6EGT5dTRS6lujWW3S7ahpdvgZyqice5rqVA+j1PLFyu
-          DX2y9g2+l/iCdzt3bRawPla44SkKHO2pT7KlGDB0yBM5R++sr+g+JzIg5xuxRtZzufPOa2GdjVLQ
-          /97/ps48TPivR4JnsH1f7pOA+vvYE/eKkDv98t4tHw52Uc+B+WlqhmwDXAf43EjxMivOAB/5whHN
-          TuwYi0HRQw48zyTYDUKxdLc4hbrfAqKeeK/BR2qvcL+zNZSu5jguGug0+Vf/+c3PZU6jAT6j1A9A
-          opICnzvLg8ua9HjtL/diJg/h9qu3oWMM9+6cdRoPo/57IfapbJvv73uwO5NBh5SxgVAeMga+M3RF
-          x/qW6bQPcQ6OTnpDx6jZu4P8bjx4fU48yeyOKbBDiQako/7a8utQ555ajCG7z3J0pW++matxLWVR
-          eI8YSPE0LrvkzcOGOQ/kxl1FF5sjSsD6clqiI04vyG6FvRTcczOA2Tz86jkaeJ56D+XjZ22+F9+4
-          gW39IHVbf8IxdzK4Dz8fvHi9R//qGxveB/DVH5tlywflQFltEvBvgVKXyT2QnUJINOZQgVU46RMI
-          PHX680/k2fAs2PJfzLTCPv7ptz195jXRyrEDazjXniR0io+84ukU7FZth41pH4Plvowx5d7nXN6h
-          KkA3X1Apl+ASSxu+Bxs/jn/1uq2ehHx6KgF1mSiQh+buIP2RP8BqGKiEu6vvY3ya1EawxGcL00tl
-          ktP2frQ6CNvC0Y6YN58BmDk3nOSUN3XizdBu+DiZLVm97PRtfTaAlocMQmH8LkSVlU9Mt3wFbn4c
-          pbXjUeFZ90/Yf8YYHcGJd396HZ6Z6yFgZP3SLD/83/wBlhXzHc/y8ZZD/4hm5AmZrdOtvguV1/GA
-          3CMIdWF7Pwind04MqnPuplcnWFTBF0sez47Lvv1A+MiHEzrq/sNdleBx+6vHWh/fi//y6ruVFaji
-          9/qm330DcvqOwSxzSfR5tncZPDrJDdPrrBVrAK0JXNaBQYHNFwXWZzWAR5ktUPWejeb3fGB6cio6
-          SPxupHl3z2AdWClBoZrq3HsBEaTv8E5+9YxFzg8QqvCokrx04oYKohlCh+VT5HLhV5+v/Ir3F3zb
-          bXw2FLOw8qzMy32IrhtebXohhU//5gePO3Qb+qtHFJHxxfSSPJvmLVvZLw9APnlXv/qLArfxQjZ/
-          ftBenMQSpJVuEuU5ETqn5xz+6q/Batz75mNHZgs/Iw2C5Vx7Lu/M2SSl2ttCvtPcwfK29QRu+ixY
-          YmMq8K8eFe+rkFzFO++u+Siz0D3iM2bCWqJzdd4aPd4zk8TDPS+2vN2RQajXRAfevsDq8z0AVz7z
-          SD8WJaAvDbdw88+43/KJ9Z2xEjy7b5f4E1fFc387Dfv/Z0uB8N9bCqqDIJFD5rHNYhZ2Ks3ukuBc
-          vh7j9WYwAUyM74G4n9s3no/RK5KdYWQDtmu5ZuF6xMM8N3Si1o8K0K52ejjYcCGHYrJidk7SBHhz
-          5qLMxZeGa8JjC/g4TYgB+5O7GrViyKYGJ6RKpuuuLD+VMM0sFMy7U+Mu1OpLGBr7FJ35vejSKPdT
-          +CizA7lWq+byD8lQZAF0KnGDJCwI6jIe2oMSotRldy6WQbZCUcQJvkTKKZ6lodHk21ryyEsqQj8H
-          a0mh9HJkvHrHrpkL5sTIo7D2uHFkXCyxrSVwCp8cSabQodMrljTINSxFwUd8AgKVsIOX4lXjBekc
-          WLj+wEL/e47Qyczjkd1pC5btD2eSVHVrii0tCeBn6z3tN0dElzpHCgzH+x1ZhfcpSKVKN7h/sW+S
-          wLEBM8+dbdnQK5fEUwjA2r5kBVxM9YKOb3HXrGeRanJZhQrJ7aAB34UtUkD3nY+fr7YYWfbYBjI+
-          XVosRRfWnbliiSDvrQ4yVfdE2dPl2sJReFXITI/jSLz5FgLro6TI1sxzwUJv50BbjQx0sPDUzIgN
-          Z9mevBhds48P+EtZP+UzSTVcMvyt+X7mGMvepaTkfGoyl33cQ1ae99YacKp9d1lpGDXZOPNXLMnX
-          TzFPQTlDZ5eFRB2oVbDnKrbhNp7I6g9hwz+9qwObKlNRxN2tkf1cAxECJcxIwVyTeLZu4gxhzLDE
-          0+9nIATldpNon0hIy7nL+H0NuQ2lkNFImBDo4nvdO5IUOyZBxu6rL+52kftO9gpyzXQpXtrHB0IO
-          xQ+keUdrFPaug4Gz0yXi9BkEU+tUNuRPzokou5PucsQecnkbLmRX3r0RHgeUQAmIJ3JWA6PhKAp7
-          uaSsEwjv21Vf0u1iTQP6F1S+17fLBwf7CfP6rhF0fCkj25/cHJYkFEl2kiSwzLkzwygLb8jTdm9K
-          TzuI4f7N3clpoFY8H+MrA8cbFyJt9Cu6gv6iQUaWTOS3aBnXcFAjuabxC/vsMDSz6QoKPJzlG5aT
-          y7VYhzCc5WzrRReRWI+FZmDX3/OSKmazmNshyMBHqS7IkjVhJL7TibDKdTmA7Ky6S74Yw7ZL1STp
-          2KJmftwmGwLgvYkBXporHM1ckXT0bpA2hnMzX3bEg/z+G+HubTxG7rWeBxgsOiFKGCmUXmraQrF/
-          HNGdFMHIIU2ZoT8mObrhohlpC88ObMvpgpQh/sR0Nx2fMNgpPNGBaerLZ44nucrzGBPHVgse3Dte
-          ftX+J3h3jOsKuMgNab07JbH6w9zMB4EwcHxdc6R5Vkwpmfyt93EQIwVZ+4be26CGjgp6crRZu+Di
-          XW7B4X4h6CC/DjF7jK8QXjHTI1f87grqO50Exf7CEKea3JErE/8G9R5byAB32NDufoa/9UTcq3XQ
-          J+gJNjTUk40/gi/oxPjOmqxGuxkFxn125z5MsZwtHxcV+9Qcl1XWazBH5xtRh/A1Ck7qYyh2Z4s4
-          Nj5RbtB6HrZ2egqA8NV1ASpZC7/i+kTXizLGXR5WGpzCmkNpZL0BBUKdwEGMDwSxbePy83ud5NNo
-          u6Rqpnmk0E4s2T0vCOlk9se5qjsDgugUoiT1SDzvpONNegmOjXQFZyOv+1YK2puYo4Pau+4XOSKE
-          dUkKEsxG59LDqGby9n2JQ/3OXWj6vskH9Z0FwB6xu8BWHCD27Z5kQlrHLFfsI0hfwYEoV6C7y8Ip
-          mpTs2AeJT8dTI8xMxIp38rJJyqdWQWcmZ6GepArycZE0QpRL2u95g51RTvGillULiXANAnlbH8vd
-          UTQ5yG7fQBDSUyxs3x+WubRu438dOVjaJQT+/khueqm6k2Fsu1gr5Y1uRlnFK/zMmVyWYEB+LLj6
-          8sWnFWJvVxGN1w4Fv7ccBvAB1ZCRrxqltzjiZSVVenKnDCn+8DjGiU3OojY106scPPio1yDY6fsx
-          XjnlmEFVoV+CrrIFaDx+J9jY/Bx8nucnHZSno8DYRB3RR0dyF+EcBfL294J5oJeCxf03gU2Vq3ie
-          717DX/m9ASm4SuT4UVl3TRW+Ba+CexL7JH/Hmf9ety0EmoG03fgZZ4GRFJib9UoOy6wCgQ7zU37W
-          VouOT7EZ5/nsKhBEZYEq9n2ha8CEofyo3zoy4KgDQXoVEdz1/Bkp+nbK+RZ1Nnwd2QGlbxVTCnjQ
-          y3kZWRisbUB5430T4XCfzA2P82K1tDL4w1vNsyhY4Vq3csZqV1QxvN/wq9t78FVnLLquTh//8a8s
-          KTdyD4Pd1tt4x0LWXUMU3tV25H58gMYdQ3Jh7gA1d1kmv9hZQvF5WuMBrn0LbNXYk/Pj1uh8d3zW
-          UOxcELBFFAJBIQsDYZdoRD8dl3H98fmjzA9Ey8WhGHdjycPpSGtiupeErnBBARTa8Yk8Jp3HFfuD
-          DRtXo8Hq6gYQgvhiwJqrepQ9M2vkloEv5TInb/LHt49ba8vpKiTEbyMjHvR0DKDCvD9IR802HivU
-          IB/uLBQkl37j02SW+1pCJLpnWUzp5+LIGx4Qj7k28bomYSZzUnfAc54x8WpXp1LeIUkOmrvIxd/H
-          lFhSYu0ddHq9dzFu8deD2feGcP89DOOiuB8RJMWtQE6if+Kpe+U5LB8lREGL6nEGRsBAZL6VYC/v
-          fMDdbTxAGg48Fr+HoSGcv08g677NgOXLL3hT2k+wzL9vYvuX3F3HC2nh54kq5O3nSn89OSOU7fPg
-          IFPUvJFzFq6FNFYMpFxBoy8Xy86gHMcOhvF2atly7hhYu+4efPW7AQi0SwsU3/mOPHZWdd7nagsm
-          33tMrMI7xvSqw/w3/4Kze0kAeyG1B680JcGOFLihVmjYcsl9BBLnawyWA6MasvR+ulhOcs/900Oh
-          aTxQYDyflESnypGk/UNHnv60m76342DfuAoNQKK8xmXbDQD5OEn+5v96WpebjHlDQSFn1u66PZ88
-          P5+fgHJwBvgSzylUUHn76dNxdfOBh/blo5FbS3Azw4s9wbl9PIj1yKbmS9N3CcvK4ojeGVmzlmMe
-          gfEmhFiMqmKcR/mjwHQqz6iMc4Wy8f41QFV1dHLU+pf+aQ6nXO450yEGfy3cRcPPAY53pSB5NY3N
-          DOBsyb/1rC/X0F2Z8yMDG98GchDYdAZmn0A0ygyJXlsvvFRhOmBpsRYIABpg+jx2BsT3dCSWr1gF
-          /4XqJO+SEhE0P+Ziro24kyLLKvAuDO4uJdNaQhrc1iDT+nTrTemncJfEBrI2ff+lvhBCj/EO+KP1
-          L3d5spcOdO4zQG4kNWAxou2i+JnckQ13GmAb7dRD83MeiHPz9g3OJvIEG/+iVLWxvpy+vgIHDuyJ
-          +nrvCrwmWQ5zTokDyPp1MQuW3MLXc9XQUdx7MZHK0YACQCPx7N6J+TO52VDwrJr4n8yk3FQuOTTl
-          70zcOF/ANxBHVtr0BH5tfmZZxtGGlmLsSFwWfLE6u9cTflz+RpT6YMUbfykwX/kQp5YLKX660w0Y
-          cZ+SYn0/xlnNnOdPnxHXcBV9EbynAu25zfE8xMdCCOK7AS+1vaKzVx/jxb68tB+foXJPEdj8XgKT
-          VXkhbYp3zbzpb0BfCiShcuF+eO3AZ0XOGDI806yHmO2gPWghUnlNLObSqmowVKmAvC8QxvVdOCJU
-          YLcgq0RiMf30IN37QiBGdRBTVXlhaI16sPm/xV137+MKzYv7DqS+QOPnyc2MLCmrQEomlgFWh9WD
-          XOep5PQybs0KslCT7QvrY/Y8qTE9PKYautM5Q3HOWSMbS3kKhI57EUfeCwUeLCDC2nlk5KyOO4BF
-          KrPwVkgTMru2bZbHcxdATatO5FpNsKEnu2Xkt+hoSN3hbzxbHYzgq0Yf4voLauiQjCwYuP0e85YL
-          AT082hrmOGRQwu2/+nQ9oV6SgHRCumdpgB1Yz4GLG3LBLh5v7sj53gCv+fVCDG7/dRfFkT1gMoYV
-          vK1xTwlutQiO3scnVmQdAP+QxEm6mMGTuKx/imfFlG9wqNyJIM446ZQ1j85PP/706khFLX/Cji9v
-          yGUdZqRnq9GkjW+I4wudviycrYENv8jBZbG+yiBc//S68waPeHnKhxWcvmlATNgr7korpYPxV16R
-          YS58MV4+4wyzq3UMmHhhR3qf+Alq2lAG4FNI47qlxtA4s1fi9KDa9IOZwopLG6JgftA7u1MHed8r
-          AykbZBVrNr2ff/hvelU6Lry+sNI5nxf0e77vprflNA+jTb95sXB51OGf/5Ab7znOmRIz4HXTXwhh
-          0ae8uW8Y2TiPAWZ0iFz+w3YznHmRoBiYL33iv9cBoilNkK6e9sV8vQos/OkVjbtbGz8Z/d98OYrp
-          OK4+S/M/f6Ztfnrx324phbO2oA0fYir0RwkyoBaI6igHd9n0IniwuyMx3fzdfKlVl2A9uefgcWdK
-          QBXiR/uNz5FZrF+wWq6RySJiDsH6ML7N4uo3DX7EpQu+LVqa6Rwz0f5182bMpdNUTNLa1fBavnLk
-          H012nM4xH4INz9Bhevvjaoq7ENbcvUc24IP4c6yrAKILExBfv7d0bbsPBn4KE3KtjgVY34Umydt4
-          obPqLmDVz0wNidBeyfEZhTpl8ofz40vkRew3/hT6aYZvbmUD6S0GI31mBENLMwWEkF0VMwzCGvDv
-          KSenewH1+TkvAdRA1uM9o9XN8lmUTN78Hdn8UrxaKqlhvGKE7Oxz23o3TwFsm+iB6ejkLtb9IAHb
-          +6Ozmx8a1j+Fivyb3zlWd/p65RcLOoY5ElV67d2lAyYry+8xxPDatfqfH/v5T8tidX251iEEaJ4K
-          4rHOd5wM3eIhMq8c0cz8OeKfflXjxCJ+fLqMf35h8/8BI4Ufd62gCaG521PiivOq44c0T/LGd+gg
-          3k3A1TnAEF2oTpz1HTXzt/7w4HHz2Z8fb4giDg6QlFlAxkzHZuMbDdhMniL70sj6QO5aDUenPyNr
-          fGvFtj5nmY3SHCF978afFI43sArfG/J6sQfrrjsbsHycFmJHjerOhk5rWQCtSnz28AFTbRQt3Pwk
-          OXjys1nzQ8vKyCy4ALNvi85RvzDy5pfIYcvHlp8fBx3/IPp5igr+yl0mODwaipy13YG1Yp0SZOR+
-          QJuf01c3f/I//YvCG9jpy/GzGtDV5wsyTh/crPPbUqAY70TMqz10pxGiHvi7xxkdf/rwpjS2jC4v
-          8eeX4iWMYAmDiaVIRZ9jvHK7soe+VuhI2/UpnSd0LH/8T0z4+sb4Et22U4SDGmD9iuPl7UcMjEdG
-          QggFT/D7vtB4cydUNUSPl0ebdfCX710rL2gWDQ8DhDH9INtdnHhdNFGRN3xHVsm54F0LSgRPRqch
-          n9krBZu46wwzKgREuzMu4CRVD+Dt8OzQ0U/SYmVxz4BNPwT7jU+64up7cHROO+IKb3lcrKmfoASI
-          hedzVNLPT68l3yoOdlj80uXTZx40P5OCfIZKdNo/LBsOrs0S+2KplGWjJIMPdmGI7uTfkbayyspQ
-          esq/vGzkzEJJZU8DT3K4WCsYNBi1shRBiahKsuoUYNGB9+pqopi/nxrBpp4I+b1YoQ2PY659PCBk
-          ANYChj5e7joc6lb+5Rnb+nEX+VuzUB/EIzlbbKIvAbEYuSiuJ3SZBiVesSeXsM5xQ37jRUfysOSa
-          6ynJwP7h0prrRKAqUCTqbqndFX5cDC+rOyDPMI/NVGXPDqqAQOLFl8AVQmHx5C0/wL88Y/0aCwQ7
-          oGWYNdmjK9R83MnDHZu4BzyOyXCzN3xjti575fpvPXzvghMIPA3Hvzy4xuAWzPdH1CyvIbLhh58x
-          KdkvjqcfX++6IQqAvLdiTG5nFugDZvD0ZYZ4feo1I8H+SEhg3EOXmrY9wHuR+8QPlk/zsaYaw/jS
-          sX953XoCj+GHp3h9HJSYE7xBg67aJXgdfRnUF9IHf/pUO682mLPjvOU3yYmk3S3Q+Q+jD7B5rDby
-          rnoTr/VS8HBw5A8KcDaDby59anjYWQpR70WpkzUbb/DheAa5Mlo9rvYrgdDXugCX+DHqvBRYJfy0
-          3Uic7+0WU4Bn5y9fv/qC4NJbkSYw1gOMlztT0snlsAa7x95HOjBf7qSiPN1OUg2YPzVvd1aswYGn
-          dDX/+TNe923oX+0L2fRHM4+QTcH2/D98bdbipE7gyZYDusivdzHBM0m3wgcM5jCq6Vqcjhj8vu9R
-          3n3pxqc3+HbYFmmoiei05b8gW9j0L4+hGZhZmBjkQFA7OfTPn2LpMCPrckwbWqlSCXouPCNz+nIu
-          3vJkkFhkR7yk3m6R3PIpWT0c/uoHs3gYQqgMOSYHa7xSXgqCEpra8kDKXTVG/srdMTjKeMJcmanN
-          +pFvjCR29YlcNHcEmz5VoNVzLhYvn2dDb3HOghe/rMTzx6tL7xMzwY9IO+Rfmzr+y2uWO2ttetRo
-          2PSV9lDp9RrT0/E0Us43hl8eFeyfqIp/fP3nZxNozsXCXkoHagl/Jzq3v7nzI+17eDb0CulnYuqC
-          79kRKAi9EO+T8ZSIotxBInRXFMTsO16ySz3/8ChYlEtLF8xXobT5Z2LD3ZOSjz47csffbsTY6hGL
-          2bgs/DzzHhlLOsXLKs+e/KmENGDdXGy+x3o74tLkTyx9b0zxZTObh52944i//+7jP/zY8ABLW71i
-          lt8Ig3ANWOTvv9e/94O6bHnkcEhqfeGlUvrpF7x/ciFYy+dggTmwnyjMbw39y+u3/Im42t0BfPg4
-          GYB62736nKnogjBL1i9PI846mDHPeIsH7wXSMFeeAzDJdyPYb/kiMksRjcuTm6FsKq8bslRXoVzh
-          niMoxycH75ruqC/+e97w4tUQ207GYhrZmYfvu34P+M3P4+vuEQI0uUek52tM17BucxlKlxWvuejE
-          a3H1A1iVV5f88sXlYinZ3/yyHuBb0A2//vDwly+Rq9d2EN+TkWz4QFsxjG4Qnt4d+uXrfHFSsYzM
-          1SXHTV/N+BNp8HXo38jxT9G4XPuyBDvFOhKrII47f5qwhHvfyNH9Vr+bJcx3DOiqxkIHL1cpm9n3
-          Eh4gXoPxduTANgkHOFTHiZzhUxqXOnJr+Cymihyzz5fil3DO4HCvJRIYpVysW74M1Di10Oavx+kw
-          n5xfnhPwoHyD+ZwMEFIhcwKwjddcG0X3q3dhqiaT2/70oilvM3Krly1Tuc8B722tsgc/AhN7bL19
-          hMcy4F2pbah88sK9HAA9OHHPIV5pZXfw8qUNZhKBLwgLQxG+eEZG2r0Y3VXIT1C20tHGs2QHlGfO
-          j1xWI3n+0y+U3W6x2PLcgPmcNbff8Ak2VcBv+HF26eMAtZ8/RwgDZeSMU1NC7G8X9d8tz32aS13C
-          1YNXEqHu1VAySeXWCCInx61euZTVM5Hbeh/jTT83QpDMNzkrTBWL9kkccWRcxV+eEOSHUfuXp75F
-          W0Mh9wp0DkRNJ38rPkLIMfFIpVcR/uUpRs10+oz7bwpvx8ct2CXLnZKr8bDkRXq7KMCMES91JHpg
-          89toy1/iv/rpr556m9U4/ulj2c3sihztgAXz/iGxkDg3CwX7r12wbFRmEuszMfIxc44xQoYCpyMz
-          Ig0GzjjdQbMdaVFZvIafZFzrZTtS57EZQbf6PU5m8vbAs5JqdPBzmX7amgvh2fB6VLJD0Xy/36/x
-          93x/eZCK8gQag5oQK9LPOg8edvCr/5DiUhmNsI8DD2zfhyjcC+s0io48eDt8izTlFMRr9k0CuEtu
-          iFi90xSDnjaevEtOBsp7Zo6/TXjs4GnBZ1xLNqZLsA+fcjw/PXLM9Dymzfgo5V9+uW71qzXbbr2w
-          mSxF2cik+hz6cAJYAN+/ejc/DY8EKugUb/nU3V3VlZsk2DkCUg3JHDmuZ2tpy4f/9O/qHJ0SWjI6
-          IQO+bvG68b0s+ldMbnFAwaSYcgkZpe3//Pv6kDwFwtduwvv3+UY/P/6Iv7v1b/2R3/zFnlwRRbID
-          IFy9toWV7YV//LDi82uAj6OAtzy5jedjcb79X40PxP/eUrDYaCb6QxwolYIZymGfnJD1COKYjqhh
-          YH4Pe1x3sUK5c60/ZV4pFZS4oUvpsU1z2cD4QNTSfrj4abdQlhbri9wMabGQhfsauuO1QTp3LHXu
-          dNYzqHmwIFa0t13+thY94PUwQj7H9O7iEqKAB+YJFjy1AfjpVSl4HQweXbuJjHT+fDvoMmFAXOP+
-          BsvZJAn09vcDsQv4jeeLpkTyXf9gZOfGo5h156XJd2PwkT83GRViJuvkk2T7JD7tA4DVxymVy4Z/
-          I79x4qb9/XuhGMSA7q8N4LxJ8aB94AC5HbHW9BdTHGB+S59IU2WGLrlkOpA/aENQ8eM0LtpVTGD6
-          1p7I9+zWZcUdY8GG9xVyu4mWu4JAxPAVvLpAdkWoDxN3KEE5ni/IG6tDTKu1kKCR71NS7td1pLF2
-          VORwCCqS3xypoAEvhCD8KoR4O1+l3N00U3kbH3JPvlYhPDgaQYdELa7peB9ZW6OR7H5jQg5OlxRf
-          0T1Z8qUDe3RA61On0naxiCejO7Ll9TQKx5Ib4N2zZaSpdwDW21t7QqPgSmJ+ZnucCr6s4R4nw/9I
-          upJlVWEo+EEsZJKcLJlUBCQIiLgDVARUZAqQr3/FfUsWoZJKn+4+VUmauPH8QSteOJxFPKZa9pOq
-          2bWdALrTWaKPnnqZJDyMHF8Fc0+uE9KQ9GnJB7j8uCN7MRHRQk6lDE9PO9LMqbRu2sYvwO/TryCn
-          l/ExhNqqR3j1UJFHYpyQwA7bHsltU9L0hCtjfvpxDt2enjwB+yhbotNZAem29KN4xqVR89D4UDyF
-          HU2Paoym6+mcQF5oIrXJ51QtKPiZWEEPi8Zncg1H8eGsKfW5SuIXuiKeV5MeNtX7483pMHdfs7lO
-          cLkhcxQaBtlkSe0eSztNILbn1hVv2IMK5G5uaBxsG4N9tmkDZfXgiCowVs3uuVfRfnOPSH5+3xwB
-          vCkHXpECejyXrjFLRMvx5JozzfXFZnONsxKFDnekbmyW2RDvZAXmm0vo3ZHB+Y+/XSvtiSsfG2dg
-          h3lEJ009k1CQPmiKF2yDAhdhlH+XoJrT6Jcgw2//46Oi4Vkv8Of+9sljU0I4nxV5xNVWC2iufatQ
-          8hI+gPK3K4ldxY0zpEufokvCcSP3SA1HiHW1xN9GXLMa0QtN8bfXUXtKCuoqT8mZYep4JdTqkTwe
-          aWUs9onJmIoZP77oLXB4eXA9mM1nQknT6+Gy6w8KWqKbRZLU1DL+/fR9ULZVRY6CprAhO60Ph5JY
-          HpGpmUw4n0Mdm4/fuNa3jKgqMR7bpSl5EJfvrC+0aMT2GD8o+dILEqOk0tElUWW6ryc/o5z90fHS
-          mjvyrHOlmqyCC6Csco4ETeCGS7TgBu7c4Ua1FEKHN7xLitjPiImVzjvGX6PFgiKJL8SowUTSrp44
-          8OuqoOZcM2f2a1NBZVnI9G/+jN9wIlzON4HoG/Z2luozl3DajZRamoCr3uhfHC6ntidEa/1sGYSz
-          jvklyUk22HXYZyObIDrRnqplWWbSvWXrcahlR7WDm1T8W1j+74eHIg+MWYpbF7LCfpL7Kzh0omjI
-          H9jK/Z7khn9C7JU996DAVaC7zj534u2bFRC/D5gawnfjLP3VfOBt4UmEhJssY0KardI4u+R6/7XZ
-          ZNSbFoaXQsi+3sxoCtZb3P43deiO6+xuCvJOhrij6ijgwnAEKU1rKBF3oqZdfbNJfncPwGQzkMOU
-          y91in/sA1I2I6Q5rNyYcXtYexJ39IF52+HZT3EUpREFe08PzFYfCM9rzaE78I73EhHR0gpcOWZer
-          XmO2sSFNh8XD3Es0iHtVrZBRlcgg9+U48s+hRUuVFSnurPfNE65yakjH4qNDieBE0sctQ2JVadaW
-          VclCH2v27YSnrkY/9ERk9wt2FTM3Tg1gVSdyOLZeVt6iVwvTrjyM8u6rZwI4qolZi0pyVE5Wtczm
-          8QMmbU9Udd3Z6HlT28MVF0ADDT/YxN78A/ip9L2llQ1DgPjmggKpR63+1ndMcM41ILvUqUd3OBtc
-          2/GR+fYXariEc1b+5PDKv55wUMxMnA9bgCFg4Il1cAqlcgwKzAbKe8rEeWiWglmE5XL1PGpUZbfM
-          vyFWwv3Z9T5nrDt/+ofLVmuoZ5ozmheVpADq/urhaSegVlRVF9/Fz494dyiMcf4NEbyu4Wek17lw
-          pO7jeUBSfyDXpxGG80LSHJX1zSRnW5uNRb1fEzyWQ+9JOdpUS2/ZAHa5l6gT9U+HdT/xA/avVcft
-          1w665TFWATqpe4dEdnUIxb681/BjG4fYSyKh/v2ZdTB/G58SLrKQcGAz4NYpI5Kk8xvxgTI9oEi9
-          F40eZpGJkSa5mGuSgTyG9RZa2yYpuJtDt/KZxvj3tpSxcefPxGok0k0FrgPotP3Pmyp1l4nonQLM
-          SXAc8bS7sFbjDRel5vZEdT23DGaz12M9KTwSa8UD7yobGbY0mAhRtR2atvGP2z508Uwiyw0cMarA
-          wpfHt/zb72r2EzbhG4lkul/3ozue+xKi4yYixjv9IjY5bY3IuyYkQAuH+iCvZLx53xcao9rs1qxj
-          G3eGjKhdGC5bHltIABM8eNvP3XPYJ/NMyPLL4DXcMzbYJFgBqviIJ0kGp7AxXz6P1S1RxqJIvY6q
-          N5mH1R+R0xRJ4XzWzwXqz/Ak0fH7ytb/p9A/D4y42rfKlvmw5cDZeBE9GsouHJrrxVNuJr7T3SHX
-          sukVJTZy4Xxb+aLtptMlt6Awq5qQn3RkPRX8Aq/6MArxEHS8UW8auAr7/R8+jEW9cA00Svsgdh6M
-          2bQNth+4K1eNmE91Hy7V019Q8VA2VAeuZ8tFtxYokuhCU2/7c0a7jm18FVXXk3Hkdf/1HvtwIgf7
-          urCFXIwGN7WW0esz4BFd6wci8pVH+TBSNP/MKoLu+HrSP7+yKAcvQW9roxPDxxaaU4wLtK6fHFz/
-          nc3PLW8pf3g8OKa0vvJl2oiVeU/Cs9tmyy23U9hSfyLe0hTd0nyKEvsDtwZt/bROemXXPfSB/KbZ
-          f36fgFc6HQQafG43to5PIP6qJb2sfmge1RcHQz2+Vr45O/yxlt2/b2JIdGHs/mx1GDnjRFXe/jnT
-          9lA8sPnxZKr/4Yk6igh2hY0//1exKTRiXLabkZyKd+U0XNDUsOP9nD6/9MLG23OZ8F7UNtQ/ixoS
-          Z7h6cEmA8wB1JptO3MxjRX7faeiPh0zUYtWFtf6o3kw0Y0dO+aAuX4PXvsjO5j+9aEnQE3MstkY3
-          zV4MxVPakd1gmM7s57UFzZexkT9asjO1wquACkk10WjDuuk3SN6fvpKTQPedeKs6G42KL1Bz5MD5
-          3ltmwV3OBXJDiW8wT9z4MOLDgVq7REX8un5USc7e6wbbzGj3GHLgudYYlalw2Kxv8j1i5Wp2j0/N
-          mJ+7UwTFNSGjUr2PlRiJH3XrUAd5r3U+I7orLhRBl4wb//7oVv010arv9Jxfy275luMHsuz2pWpF
-          9kw4K9OIMuol5OR+vkaPZTOGZhMBfezILfz0j7cHw0sm9ClnYrXMb55D753XUnWtV2kSVB+naW2T
-          Jxc1bLz78QidrlNiHfye0b/+7MewQ6zDcwyZ4Nxq1Nq1RwPlK4XzcDW97Su+FnT/5W2Dt86wAJ1m
-          n9zUxOokIQ336BN5EbFFSXOkCV4qBo2ryP57eHd01/wiWAb+Ti6wyNn8jPwUePqaqG4c+46pb437
-          jy89Mkc2xd9ax9FjdGg07yenMTYZD3/9qbVZX3kzb48AttWGjMO0KGH//tAcNuP2QCLW75EwCv4E
-          q75QQ6ydjH+bYMJvtrZ/9RCOvVq6+PiMeeIqvlAx5a484FX0QM+zYRiSen952G/iswdxaLBR4x0P
-          roI2jPIuKdhcL3YCZs5KD8LBRbPkSTLyovFGSH26OZMMpx7B8Rl6lVTM2cygKDGO65nuNmLgjIs8
-          pZhf0nz1V09nroVlwkuvXD0IXk+D7ZpfjPrr2xj/+o0/PkBXwRioRqWLMV8KpwQpsix6Im3B2Pbg
-          tQpDikZI+JIzdjyBDVfUTITogoTYFuEcpSE+UNOO3myizsKDmaJ4ZBIN2N/60VNjJd39zptwGEn1
-          UO6MctStG2Kw4yWXQZ2YSbyejuGEb9IEq98gbi1N3XQ9axM2Lhd95G6ekM1LzFzsMJlR8tXnbF4e
-          GQfaMQipvqmrjiqpvseg6Xeq5Ve9k3KLcxX+leGxvJA4m5/+mg16PLZkXzZ52Kx+Eot65I7T2j9R
-          m273aCNHd09oWJ7NI+lyRdjcC3IdEq6auKD4/PcTH7mwkeB/Bf//fLNDuGX09Q8AAP//pH1L74JM
-          l+f+/RSdZ0veyE2q6B0CIgJWcVHEZDIBRQRE5FYFlcx3n+C/04vO7GZpjIrUqfO7nOKcle9qwqPE
-          kcTx2TLDqwcb/1Zhd7K6jPUuWLsk7jZoAWcAJvAqGzWrFIkcGzRUM9MSCKf6LWPb9CJjjXddLQIS
-          Iu4BNZds5ZMJS8XckfA8s4h9SpoAa8tTfPxIkttZIOVACeUII3MWs6mddr7K3d096kepZdSopQ5S
-          O74RtCnv2Q8/4H5JfKwV7miMh5dm/fQ4Ele92gnGR1MmOf1OQG0eBgv361PH8+WFndxJo198ABTR
-          CL3V4NW333HjKeN0eJCd8zy77JHsQnUs4PF3/3qWnL/yn744BlFXzZC/Iyi/ORMfnIaP6Inb8nDz
-          uXIYf1s7WqqrrEPZDw6rPg6q78KLCL5395CExof05LUz8x8/wg7aHo02VOQ7dBf9g5hwhMbi3IVE
-          XfUGRs7WMcjDLwrlrT3ZRMOqyVZ8jcExPB5w8JjdSmxL2YYQeQ7Ox9MczXn2aZRJulfEPcx3QL1e
-          TNRo10xkNwFsMEe6iTCeqnCifDsYs8daCMdCMwiaPmnWc4+nLFvKIk7S/QGr5rzbeMBd4EgOi/XN
-          5jS50R//xrtH6bgLTqAGHtz+RszHWzKo8chl+DzGPXYe6JzxZC4cVb0WEsH+g6sGu20bWBwuIt7l
-          2skddd7htsqiM7w/x4+IBv2YKn/49syP2bhoh0SZd1+LHO6Pez8/lyyBaz6aJEcQAI2ViwJ+etfy
-          ozmbxDTkYfzh2gleczf76T21uJsjSTj/3M8Lz3lb/RDlBGv+J6JbfPRgyyUi3tF3YdD7UwlhhIcX
-          2aXjXHVW+7JgTJ4JManjuOI56TV4jKU33hGH9jN8h9O28+4pDq9N33crH4eqrx/w41SOxrIxTjrc
-          SlRbGwO3xqy3Wx6ufAeRr7PPFvHQwl8+xafVv2A8Vc+wO6UFOeRvPyJ9+eBBSwMD7+N8NiiD8QA/
-          l42EndNTM1Y/IIQrXyb2WWbVtPoF8DXoD0Sr8cgEnXcgFI88xA+BrEcw2Qup8fC1cLRt39HSbsNU
-          TW+tufIrrRKH5sRBHdMRo8QYGd2Gc63+/MUfP1oetsDB7GW9ES2+1F1uR78EZeveiZkaist6M+Qh
-          P+fxer2B25rvdTBnkw3YHSUbMOWx3NXIU/CkYlfql/La5PC3Pj+/damabfHjr+TwcA/Rqt8LuBlM
-          AbvfI2eM+o2e1VWPYTstzUhK9kUCN3y9x0bgOREz0d5Wf3rKtewwkxqhDFV7vyHkVA60X/BHKNRg
-          +Zio7GPLECZbn4AUwxQ/ysGvBNkNTFW9tD12h2EXib/1+fHjHZEEd6kWvVUDYEYkf0DNmHVnDJV7
-          020nxTN9xnho5eqKR/gJplUfxsyDaijeiMPOD0aNcL9AhHcZsb/sUlE5d6Y//GTVRs4YsvQU4sxj
-          +LeeI3M7C45tvkXL1bZ78XbSZLX2eIPc6lirqF03rUI9a175ZA3+/GeB4e10Uf0sWhZZTqDf2xVx
-          zembLXm3yOoab0itbhdj2hh7HURa204t5/NVe0+CAmBfSEhcGy546M579S8jCd1riVbDNE4eNG77
-          lLgJbNgcHBUfrOtDji+dY1+4GWPwmIWFmE98AOy+5VvwygwODT3aM5opngLLspSRitObQe3hkIDi
-          Uho/vyCinefySisv4TpbeR/NkFYiFLVcI8f3pjcYxxvTD6/RdssdjF4JvzFczpmNFvWEMprtYxPe
-          TONKDsfMA8S8vBWYhhYgBmpYRH/8B/lwR9JSfhtS+bxpv+v545cD0V7TH18+brmDy7+9O//nV2uV
-          9s5oqN9C8HbnVW9+jxENuVYEK15NaPWzmYlOtrKXzAmjjB/c5Vy4FLKRnn5+j/Hjh/C6yWziHhbD
-          mNPebRVk32d8UYNdP8eBowNzXHZY3+1kNu0HrPz8MWL402fdr2kKXWHZIH7SY3expLKGw3mZyUk6
-          kp7sH2vXQpjq03b116YPhgPkXryB8/G+gNHnrgW8bpf6d/3ZGOhBCbdtPxJkuc+Ilc2WQuliIaKJ
-          ReLOS2mewZrPyMkT7EoSDMqr3GPvIrkjWS/Uz6QENZrw3+fHupl9oHB+SpKd8OhpKGkWXPkvAgMo
-          2ej3YwdXv2gSVL1wlxkWFL71JMFZMr6jZdUH8HxwdcSt3zd/O7+FTRSiiadO586H22OB4t7OJ6F3
-          gmrJ1WFR9F2n4PX6einZtylc44nordGxnsKvBj/R8Y2YIFlMUh5LDtZ6AILXvM86JXwWYPwUDP/8
-          pPH5fBXKlXvq08o/XXrSuxDeZ+1CklYiFb1XZad6yu30w8d+rAWF/vw1nOfuq6KBfhhU/aS81nic
-          IvZpCgr3kjURr+E4g5WLroC4owoJZnqoJBNkHsRpOGLvIhSAZuK52NaeaCAWbCe2PK5SDv2vfyWR
-          r7Zs2T9QCgV6+xAnd5Ro4NLjGfb3xMFXqoBqwQmvwW25N6ZGJl7Gx8RcoKTjCm2Ve2fMPvcsFP0k
-          vyYueflsuSWhDnnAEmIggQeUx6GnNKYLsK4t0q9+IKvlZnpP8urHUi72OXgvDJH88GFOz1deOZ53
-          EXrfomPPXwsj3L5G5Tsp0sV0uzk/IziJxQM/ioPmUvnFLLUFTU2c0/SqfusLfv6MQ4sefJ/nJNlC
-          hBwEV73LPt3NgUWcHzE2ltodsiPKwc/vNwKvi1Y/WIOEfUpsBN0pk7RYy9VVf5CDrT2q336C3bG1
-          cVyHp2wRw8WEc2Zl5LjVdwZd9briD+YXBwtyquUXD2u9aRr06hjxq35W00dyx/uBL/tFCa8lRE6s
-          Eg0tBzCOKayh7lqY2Gv+JJ+npkFeVteuZnvkrnpI+fHflS/4FbXbooFcMHdTqSQCWObZMuExWRxE
-          ve2lGr/znMIVjyahG1BF7cdeged8cAkaOMLmQ9vyYPVPVz079cS8jDIobn6GPWPnReN3OmrKWn9B
-          D7qk0bIImg4f/PAg7t4bonnVg+ptd9qjBRZfQE9SYK792wxUrHxB3G3uMQzBluJb80CGFHJYVyT0
-          dJB8eKJ1PyYdlM/GE3ub0wsQJGmiuuT0V48aATtezjIw63jAXreps29d2QocK5qRWJEKg76ngMKV
-          f5O9/ry4Q4d5T33d3IUcjZPGeBXACa71JqTMPYuWPDlq8NWGe/Tzp78Humvg8LQ9HCuSZgjxeUlV
-          C351/IzoFjCHzKb61JcdQWYURdP+luswVBlECtVtxoRQiYFHNzL2gNJEP/8a1ri7YB1yHhNFKphq
-          Zw86eepNnFH7cZJhdBIviFVUBPS6l1v43DcR3s1h3S/zYe2SfkwkfEIwdgn3uMq/fIo16WFGUrno
-          sjrvs5zoF5iyVc9pcMUHxLn22uV85Qc/P/5A+9SYNHSRQboXIhyHntHP2Ulr1UuZnqfNxT1monPb
-          a2ouvBRyOHpuJL3KjQhtr15ICPZKxIhMTCC3S0x+eCpU1wltX50TYG+XrINvLm4LnL70iLddlor8
-          8vnqN+LwOwjZYEjMgp/w9MRasHadObVIh+63uq71Uj6r35XWqiu/nLiBw+DnX8ErrEVsyfoHDLIx
-          FWAZ5Cu5HrspooGOB3jlHjraBubR/csfa/6aZCj2PWW8Cv/8STRJuvvHb09WKxA7y52I/eo5V0lO
-          yE3YpaDfP3wPvtGnQaRjhLFov00ga/c++fHL4fV8harucdk60bN25yWPOCh6PZpgUwCXdrkwQPH0
-          OU/86i/Pz8RelJsl6+RXT5/KhsD/r8EH2//3kYLh0ITE3L3fEd1fhER9LE+KXVvoGbuH2gLzzof4
-          MOgLI2XnF3BWd4+pIbFmiNouatSxTvfE1sC4jmRhytZloTwtnhVGUuZvauhb5wvOvE/ARP1ygarP
-          v/fY7sZXL5xFQ4O2/YGIa2vPpXmmWvCwkAnJ3fiq5nmRU3g48Ft8PQWOy8cHgwdlxdSpFvcfd9k/
-          diI4ctAlPhf1GT1na2MQGO/QJiKbjI79K4XXujWw55dOxJNNtagk9nLicB3XU/fo1RClewkbb1qB
-          5eEWnbrnbjviCR8ho00sO3BOdIno+2TvsuOFcDDPiyf2t31TzbJa3NVPuyRIZLblitbH0KAcbxWM
-          qnOZSVoAFXho1GD6COE9o9tv2wJJ/jpTAjcKYxV/GyB6LHt8WCG10j/2HQTPx5Yk8jV2Wdgashp9
-          7jy5hmTs5+urOYO+jG7TZKt1xVJLTFWOa2LiJAff6HVHUqDkh1/sbZYyE/rbM4cBp9Tk4GVuxvh3
-          4ajnXtljzb5YxpxsMYKH850j7s7YA54TJx1ei42A9ScpAW2Ofq76qoCm+tJqjBe0cIAVt79jb3vL
-          MrZ9n0Qo4F2NHs/bkdHJcGLwnIYTedRf1jNLqs8Q0/sBe5WqAZHB86Qu3lGYOOeru9K1XSh0mS+T
-          c0sTtohPTVOTY/fFeC8FhjgdTw0st1yOXRThTEyIcobumHxJAHtQzdx4c+CLaQrWx9s1k8rdnEN+
-          ryfEgdnLGE/RYsLyASbEbY4VI9AMW7Wymoq4ia9XoihcB/Uyfwk57WUHCDRFMeRhFeBTIY39LFON
-          qst8J1g/DJ+MbtihgLt5bUz1GIdojq9vR70JNY/j86aMhGSgE/TOoCJmVRCDkHoO4aIMMjbffNBT
-          ltMabsO9jh93bAOBfhodTsnLIPcEXlx2g6q9dXb5k1xPQWfQ3UgbyNpNi+Tt6Qyowr68rHPgRPSh
-          SnoWb2UIMWQyNnzpWIn9owuV337TeDuPpr945gN/WgxTr1j2LHLwuoYPjDzSVeQ93gdYWXWFQ42u
-          gxVeRqxG2ArJbZCuTLA2NoVrPGI3S41+Jo9BVMRMuGJU3jgwkT2x4HP2XuQm1LuInnw5h9/zVBHH
-          1UA/vbvMgkF8ojj63HeRAJCXwM+zRPgU0ZsrlbttDvubgSfpsFJUNiQhDDi5xvnHsjNes2Ud0NHR
-          8VM2aMWmVDZBfqVfYsgyYMN41kpV5E9b4gZY7ClpdVk9zSTG2kvcu/z2W3TQ8s8auS+PsyuKwnOA
-          TX09TJvOekTU2AEdXO1PhE830ETf69nq4DXbNzi48l3EdL6wVe0VJ0Sf3UcmkshaoJuHwbRIz2f1
-          3j71Wl3jEbvLR8skE+gifG+PGxIcqjOYFbHPoeT7X2yK+4+xxm/8l4/cyjtWbJdxJsy7EBLnveEi
-          0jbXQSm4T4s49WFlfM2bokr82Ee8/Rr64ff/DuL+S6ywyVwxaItFLdX4RpznVmB0215zGGiXEpuZ
-          VUcLHwkdvM/DiG90d3OlMB0Q9O9ehZNbc4rEzSVKVXT41IiBNHFZTYEHQ5MygpvzNfueXrdFlR+c
-          gRS5jjMafTY6ePbnEOf6cs3o8VPmaqmeb/gxcjsgFcq+UEd39Mm1dMxsPsTftYWWnUxFkx8ZDSLY
-          QtdODIzX9eyG1F1g2b0YjkXFc4Vq3IdQSJodyWV6Z3z02Wjb8DvfyOn80BnfdkoB9a+2wTbXJ5G4
-          c+8LvABWEDTEobuIwnOC1rc+4sP1GkbSsU5EdXjcC6LdNkq2zG14VskMAXZv1i6T4LAdgGunBjG5
-          Z8DY+LFq6F40Gycn9d6LuBMttfLyGZ9Od5FNvlEgVTn5R5zu6GKQwDDOUPt4zcTOOwew83HSYfbU
-          ejTzAQXsGM+2en7bGOdc9HZbFiQKRGnikHTJT4D/fu46ANWO/X4fLPArm7/7hY9RHVa8gdMBnrk0
-          Ifg1LD1TYDPAPC+f6/0oGTleCISx67kkj7ZuLylwmiAa1qdOvien55N2lGEYbV3sukdkCP3telcR
-          1AXsSHDOlqqSkz88LZTXPhOjs2dCQfEGJK14K86Coiu7fHySo926Pf8ptjFQ5szGLp5DNgjvyIff
-          SgT4kEbbat6EtwXsY60gPm9z2SwuufW7H/jOz3ElxcMt3Ep3ecR5eELu8lXW0fFlMeH8kN6N+Rw4
-          d5iyEJNjfq2ryZuBBZ2HBzD+4bkLahHOYvDBaXKgxtyCSITTxbXxCYi+O35sUENvcgeiJ8XDFUwk
-          8OrW8xhOhVRwxRcaRLhbG3fGB/kEBKveJar6dp6TVDpmNJH9x1JR4OxIWG+8fgbZYMKgcmy8k40Y
-          CKkGHLi03ps45SnJ2I087vBVFB8Snzd6xB+iwlSfU6YgbuaFirXpEcH2RDm8W48Fze6MKFj07TCp
-          Av0A6l8jW5082mD3Ln9WfG5tVVatjBzq3GLCivfqy6QORjK9A8bHUqj+8MTL06Tn9/ZlguZ2Q7GZ
-          y102R3lpqw2OCxIFo52xLu4mNbTtieykbAZLbjs2lBfY/fAsehvSWVZXvMaH+6AxPmWyDx6HGuCH
-          U+BK2KubGp58ax10cJldioRZh6QdZux8aFEtD+eZA8q0AOebYwUW6XtEcHjkBd4R6Q2WzSVLwTh2
-          iLhU22QEBstZJYM5EPta1IzCsuogbkCH5o8DswVqBx7aoyz88UOq74NCDSrbJvuhDzImDjoPu6em
-          E98nPqPNeOBhKzQuxmGs9Xxviw68xItCjopnGcszNM6wn7N5Kp6lXvVf71Wq87mF5LZcsmh+sw+n
-          3q6xiW0/rgCvoJevFqHgkZX/sEVbZ4Ovn//LP4ucuBq8uhtEzG8j9VR0Ihm+WW1ib0+SbOYuNFTB
-          kueT9Gk/jApaOClLXxfk6KdBRUnrKED4rCXIZ+IbVNkdG5BNRwNnRXeKJE+/5z/+MDVrfC527Ikw
-          DswUrfjYz6gPOWXNvyR6vuOK16XtAuZxu8fuet6Y8fHGh2N+qsiK/9lMHjUP349biS4PP+2FvrVj
-          gA7vGuv1cWT0E906KFBYEfcGNZe/D8AH1v07E68zjhnl2+0CR+qUEyhjh4md/OrUcvJ2+Hbfvgxq
-          7Jj2h187ax9kBFwSHqx8iTz7JWXzLdjG0FYKBeUrH57tfV/ANupPxPIVtZ8TxU9USe4d1CH86ult
-          5+tqpx/6afEnyWX4Wmjqcs9tfEAHGzAU5yJUZIdb+XyXURFCGyr35IXgZtGz5bU7OYAYg4tjX3n0
-          TMorR42hKKIlN/YZDSK+A13reyTxg40xpTm0wXfZRBiRIjKWQ1UX8OR4d/JQzxGYiX9ah6sfjth8
-          pBZYmh2vK5PeCQgcrq3BvpvjAkBlsOmTgzxbr1+D9PvgsR7UAmOh0NegwecC74un1gvRcOTBZnO5
-          EWS0ort43lLC/Lp8yREvXs8DtdK2mo407FyfriGaRXGGZo4jJC7hJ5rfj8SHeWbX5Mzfjv285nNo
-          FKI3cUKTVTS9kgGqYeet+muJxiAWfVh+7i6+fT6TMdCrOSkH1BBsqdq5EtUmqNWR2iW+S0CqmKzt
-          QxBeIwvb2n6KljhtKFCTIiWXqoaMCu8oVI+yd8Gm25/6hY/UFhyCg4FdnRiZuCxRocru8YGkN3Sy
-          Qdkda/ggm/vEJVHcszNEPviOB3PiDil0iRbwCjCujU6M5Ev6Yey/KdR225EkZ86uhl98r3rkl8+N
-          OQCiqPZPtseH1h8YJZueghM56Ti6uj2YC97x4fhZNIwFgly+vu3valNsPHLSrRpMdJsU0D9ZLnGz
-          C8eYM6U6XM9Erflcjlh5pyIEgnhDdOmUflj5FOCc436C2xvIFkEWvT999NML4mE2PfiISYR39b4w
-          yLVVFkgP0hHrXNH3S05fHjSxCabl6LgGnfLjBLlhHbkSbWewVHNiwZy6LsFlGUTDp5hjyBG3J3i8
-          GJkU59sGqkmZEuvExdGCO84EHrBMYleDUM3P7mjBVLxIGE06rubq+a63g+ldccL1cvZbfxXAGpPL
-          Jq+MJb+8eNi48I0v3dtmguRqCIoX9Yg2/HbtOsVwse2VXiF7eNbdeXy5ZzBdjvaq90DWZuVigjV+
-          0NYqK3dWxOq+Dh464L0AFLDs3DuFxJhcJM2FmM3GV0Ww0Yc7cRuOVBQkXQ7dT5ERH2UoYyqKeHiL
-          4JE80vltMHHblfB0nwhablnx48eWStPrnRirfqCj1Yrwxx80LuqjAS9p+ae3U0VOKoHbflolfWwu
-          6yCPHrB+PxQwjICL7VU/zUc8DOBb8QAHvlxlsx0ZCBpj7mMrfewB3VcXCwZDGuAdet6NQQ3GHCbP
-          UiIrv4++h/g1QLMgH6zdjqCaBq+21P0tzf74Bzns3jVY9Rrx3LComNzWZ3D7iles15uhmndVWCjn
-          8Pr9xS+gbPB91XoDMKnW9K2W4m6W6uoHTA0oKRixcHfAqu/RNg83bHHwQYTJ7FZIeKum+7ufQNG6
-          PfryX5XN5aVOwWZzvU3i7fMxxmSQJ9BcPLIOOrIM1heVDG9keGNUOLts6axjqqbuQ1/3/7WnU74b
-          1IsmD2t++Vb0BXaiGi9kj42xPIIF3v1EdciIMM4ax2Xa896A9TXxum1rjBrvT1ApQzptv/uAzcLp
-          USrF+/pc8bfsWX1hPIwmr0N0fcqTukezVtUXrfElt24RuViHEipHGZGLHkiAPQ9jCC3h3OAHts/R
-          rPHJAM1OSlD3RWYkTFHfKVRMPOJreRbRy/s1qHj8FutTlin4vty1qxPNKV7zS8Wa8HEG78PrTPYn
-          xa4W2347gN91D8SfjkFEh1fhqT/8My+tBnhcKzwMQPlF0uqf0NtCTLjyf3J8Rz1jo440ZXPgC2wj
-          lmViRNwcXLNDQ3aoLti8b/IJqpZ2QPOh4hnFpxsnc8L5hDN+XxvTsmEIfsKPhA09DCp6zopC3Xwf
-          LtktbHQn+KWWsgGbHBu8f3WF8jIksI5pgF0KdCBh2Dawf8779RlpoxIfnC2rKeJVEo/vB5uq2bfU
-          87ycifbSmqjFG8UDq3+ALc2yDYraIAar30fQ6PlsfnFuDpLYn8hjIhmgzzfnw8upoDi0z160yO9u
-          +uN7iWJE/ZiHBVLNz+OBL63hVfxiWSWABUFIWuKLSw2cTtDyjXGC0eBmDBNNBjZjJl75H1ggkgfo
-          31G1rpcCiMF9z/AG4/UpwbRj65gIDdKDcMT78UUNNj/4EIiKSibYbW2XFlXLQccZ2inty6YSrq/p
-          /Mcfr4pQV1MVJxNY4wnbLXxGNHUDpErkpk1lUqgGe6AyhJiJOnaF6tB/g2Ncq8tlCn/8bX2Kde/8
-          /ELsmdtbL57zUwgew/tFjl9Ae/r43pc/f0fPjXc0resPg8Uv8HHF23b1U+HpduYxsg9v8Ld/Rc0C
-          eO/YesX098YHbtCeyO2+3Rn8L76GZujweX7oEQvTwQOrf4E43pYNqh0aEXqvqp0Uv6rdKbyWtSpQ
-          rsKmCPaAR+ncqWt+I5fWGP78QrXsRIs4iTqxZfUngXc8WtjbqWEkNJxyhvf7ZGEvGvpouR2yEKT2
-          /oBk77l1GXbsMxhd4hPPyC6Ami5E4HFoAEpvbKzmoX6JavnJXYwa/9t/maXwQOGZQXDWdMZ8mPQa
-          xIGVEldxh6xfdF6BcdWM2NjdQ8amlFqqx8cI60mhup3M9FRdPFeYhPdhn0mv7NPCqv8cJ9aaTfTD
-          S+BRGP34syt9dk8Hwnp/XP2TPCO5OmjQt+IL1rZ909Mmpra64hHGWL1Fi9P7iVrVUjR9YGRHP72t
-          NEnTEVvZLtlcfycKdvrGwAfL54yRu8g+/OqeT06ne8wWIFo1/OXTi7rs+1k4XUr4Fa+IOGC6rvy2
-          KkFWywrekd3BXRsT6yrH1fGf38rTqzeAz+BE06ESd4AK7yxUbqf5gw/jRLNlWbISavugRZtCjJgA
-          tGEA/LXosbbiJUm3eQdnREf80x9z9mYTLLcwx6eNNkdkeJxSmGUdJd5tYWw2vgJSygntyMFlTSRY
-          XltA0Zs2xJqKVzZ+KgUC7+haqOitqV9+/thmkKw/v5guj0+qSDsnQMs+eRus+XLoxzcQGM3EnVx9
-          G0OzGD9r/pr7dX/60Aw+eOJO729EX/a3ASs/J0aPPlEn+E9H6a6bLbYbLTGWFT/hJUo2CGSXnNFA
-          FjwoV2KIhOJZVEv3sik8WeJrUhT9xuaiKWvYJHWHrSvagGWPO3vbKcsTG3Q90nMfmK+ufgiO5jNm
-          czTseCgfWImt+d72f35qWgk1WfMno189u8Pc0z3ifU9dzwK/0oEQiC72dgdWzYsTnNX1/09yfU57
-          aejeFlz942mrIY2JMnMSQDRDJcdKsN3F6ZME1KcTT0z75fVkh+IURJ1I8O99yZLqGNpZWiJxc46Z
-          FPi9Blf9Nwl10FSzd01W/0m3sbm/9tkCER1+/gFx4XNi3cUfG3A2TgHa5te6Z0AJUljypk+i1Q8d
-          13gAwRIWROvhGC289eDg6jcQdGRGNNXFTYH1xlnwUS7sfmnteICmEhqTTFlnvE7T9g7b5liQ4LhL
-          +tVPcFS89zcYl+WcTXqptiBTgwzfVM4zFosvB5V+6WEST5wY0dX/AOW5fxPjKKp9G5NOAdEn58nR
-          FN/9lNu6rapUhhg3Zykbzsjmoee9dlOz8tu5aLrm56eR55o/2NxnEJyJrxEdtaii9GoO8KsjH++O
-          khT96d3L/Wujao2/+ds9Ong6NVfsSvht0Ot2MP/83uCbc2x+3l8pTEKF/fQ8mNuQX+Cu/kxY27sm
-          oO0j5FXxeDUn6NH69/0hhDlSkNTV+IeXZ6g+fZ+crOnbL1L2LaA6pS2SXud3Ty+q3ii2oe8m8WRf
-          K/LLT28HYOK94w1b718HljC+IgHRQ7T6hRpc9RCx13y7LKeUg+X+dcKmI0xs9TtEeLrF/Dp4Ycim
-          i3GSf/GGPTXXXFG7pxzsDXc38Wv9hAlOQZXc7gK8M28RW/U7hVt6XcihWY7GdGmtO1jxD6d+ZRrC
-          XpUahZMnGykXt2XzEdeD+nXzG9FjT6m6NDjnahOIPNlpn5EtVexPkMIwJZ7dP6sh8jbddt2v0+Jo
-          R0b3igDhXupdfMo+BqNn7evAyzc6ocUsrUxqTKEFviqhSeZ6OWLi+ctBbYT31X8qGEXCVocYVTZB
-          XPQ2qEl5Cra8phGneKWA1VyqwIWzGNbkU2Usb0mhyqelCXkGxY6J4bVsIBqUBxKOocFYeo5NCIxt
-          hg8wajO6A77/Wy/s2f2mYvfvq1PvgtniZHoeKmHYrE7CgS/+9Mfk9H6qjnWyx8+geDGqBu9cvcFz
-          h6Nm7cpueEcTUvOB8GnkXozypyWF7627IbuPc4/IWm+D4/mwID7139VYVqaprn4pOd7Xrj2tYDfQ
-          QLs3sfS0i9jXexXq6r8ShIMGzFdrgjBuuBafzklu8Pdn1sDhUIf4vHu/M8In9v0vX58wx4w5mJ4O
-          nF7zibh4XthQht0CReMu4l0wlcaq733oJUGOTVpyjH4/dw10WXsgd2w2/XIrw1Ydr48Rm14U9rTw
-          bBP+/HK07YtsOkqNAi1te8HOyj/myzIUsD4VBxz++JT+0e7qr/621qeqxTdWPj1AE59u+0M/2I+X
-          D5+ldZzA5kvYEsHegSfukuLb7vOsvkZ4P8M8kLOpvkp59vOX1Fckl2T1F9nyq3f88FYfb1LEsmeb
-          ryPJALmH6MD4rFQs6EnTEevVa43nbjChflM+ZGe7W0BW/aMq77bHAX12gLV6Ef/0BdmhJzQoObkl
-          FHrz/NMT2cS5oacSizG02GcvGy+at/opuYsPpVNn4widATyt3W3KH3NqLLdD5MO9R2Oia7TIJrNo
-          Y4AD446Gnx//AXT45Su8q8XUmIfN14OtczOI987qfuaNewzPQH8Q91V4xqLTyxlGR2Ti3R4v1YyF
-          uw2vShEQXUcGW5o7SH/1gWnL8Q4QV/8V7C6PGHuP+gaWtewCV32J+h++/uon+7B9kxu58e6k7gIN
-          /teRgn/9x3/8r/WAwD9N+8jf68GAMZ/Hf//3UYF/S/8emvT9/h0s+Gca0iL/5z//6wjCP9++bb7j
-          /x7bOv8M61kDVZGFv+MG/4ztmL7/x1v/Wn/w//zr/wIAAP//AwCRCWxVugUCAA==
+          H4sIAAAAAAAAAwAAAP//pHtJs4JMl+a+fsUb39aOkEnyZO2YZ0gEVOyVKCqgogwJZEX/9w691d3R
+          EbXq3twIlatJnuEZTvIf//bPP/9qi7o8D//693/+9aj64V//7fve5TSc/vXv//z3f/vnn3/++Y/f
+          3//ryvJZlJdL9br9Lv99WL0u5fyvf/+H+9/v/J+L/v2ff3FtrUQssx6IhfXWQuNeuBNt5TqpwO/D
+          Ctp1v6Un0wxTtln7JzhEw4aEnhPrs4TdCuM0K+lhmy3FkhZVAiBilRo7c6cv1iI30ArzioSHl69z
+          qT7nUL7VghgG99Q/YR/0YDfFYVxdU7UT2RUMcFf2bmTSvmY0dtQEPbYlR/TEyPzlPq1k6JTe+62n
+          mAXSTYCEtxRNT+/t9zYflGC8P2dCgtWhm2J/jOB0VxC16sdGn9FMLKjeQMjOb1k3p+Ik4x4tu7FR
+          sKBPw2uRsT8Qg1jNpfIXDUOA0uGc0Pxddv7wvKYVKNHqESGJG7tFVVUL20fRHp+WphV8Y8YLCuw5
+          I4edvfanZuIWXJP3SK14u2YTH30EUHTNpQqgQOfMoZMh3jCD2Ba3Smeev63wbq16NKGCXQ/l0VHw
+          5lRNVHEPWUelRdOAx9aV+HdBSueG1yqsroyMHrND6I+cG3jyLfc94sSV1QlRK8rY7wuRWga3L6Zj
+          Ilfo+5pYVavXIpGJAqCpOxIa6czmib0DCNeaS4JDXvpc+2Ae3tZTRnbKmqBJlYYb5AoXEz+iic+6
+          460BOe3Z2KX1oWY9v97DMW9Nej28ikLkT0UMjakUNM+DuhDR5vPEOQ+UBI9uhZimaQq+zm1Jt2X1
+          9Lst9wyAXTd4XPyjkE4XsfrgksslUsjCzedv19cC4+iciLFypXpustFB12eqUqMjj45zzGTEvB/s
+          6UnAQjrIRuWAdNl69LTbQ7GUfe3gYycXNBBCo5jXu3UMv/2+4o3vD7B2Y/BQhoiG0q2/bIrEwGLY
+          FPTQbUk3qvG+lG+Z+aKuQiK9l2/Sgrm+qclVHe+p+Pj4CqwO74kaOlz0uduaBu7FfqFGcxy7Jb9U
+          Gn7bik/i6930xc5NTtifrgW1nxfP5zb6PcI40JoRzCRG84x3y2YmTkIVnbrp0Jj5hKwXrOhlaJzu
+          L1/fl0qkej5I+ny0ZAtNl9ilJ/1S6UsbLgleLeeQ5mRUUH9wQgd995tYnmCjyZa2MU603CbbjcWz
+          KbBPJVQP80p1bTrVopbkAWJXhIl99LViOSe3HC/XT0SCdXvWWb69P/E2P4nUWjZJLcjdLcbCx5So
+          44sx4qzCHfGOCodo4kut43C7zfDC70binfW6E8pNX4EeNTFV3oqScu80BzCfb55cq0zrpokWZzDK
+          eE32n0QpxN1GXvBGD8RRZGDq4lBfPMAkkUmk3UJ/SbvpjL/1PwrNRfM5LhYE7IXVjqjg3grBjr0I
+          zcNkEj9Efb10nHCCqT8DUcLIqLmrFlv4SbZHkt2ed8ZO0qOHgMincT5+LgVPqr2D6Gt/IYaC9zo3
+          HPoR/P4oEiWT1ILx4VaGyyUnNDXZw58cLZTBqUqPBo7T6lOwfwm4mOiHhiE6o+buiBzUm3Y1vmxe
+          1Of8/TkjGPKZeKU6o6XXpA/OAsskSaozfz77pgAPz5NHSXrMhfjiphOGXdpFwnUJUq7MlgQfhkNP
+          nNPqoPeZ0j/RgdQWiTcG+NOJPEY4SRlHzlNu+QJWtASjGI7UE1fEFzpulaMuaJ/0gItbOqR2tIdY
+          WovRivhVt1ytAWB9Uc4kJUXss3VkPGFqGp2cJl+r54b3bujbP2mglH7RL8f+LN/Gt0CtNPP8xXtN
+          gANf/VC725J6Hoql/OEHyX71YUjpCcOtc6j9KjXGk6XfoxO33VKDDS+0aGX8wd/+S9VOcdJpWt8/
+          2ABpoXvIQ7bwxyRDWeovRP/iRz/u8hFssRui585ta4b1QwCVde0jfE5uaGICp2FNrgbqPbqxZk5p
+          VGDFnUlUfbgi3h8OC3zmrUcz7h7WkzxZDqDwyn/zV++Wo7xpYCl5leykfaYLusd/EETLapzu+7qb
+          B+daQqwoGvH0i6aL56de4diWEbWESesoqxVJ3tZLRvWVX+n9Y3gBxD6U5LCgoJjtHHOIx8Z1bHS4
+          +FPNuQ20oYKI9vQ2BSWf+SOHaTvQJDZ6fxpNrcLN69KM8vSN1xvFCr40o0uju6PoQl0QCbpn8Y7g
+          SkafhWRlQY+mHbULLOjLK45i5K1PJXX51ctng75pYCZeQkOQKl88C8cR6uepJTYulEIcYxnA2Sx6
+          hB3vzibB3ys4m+cHCYOVXfCfbCvjU3Roolm4zIh+8QZpeduQ4/tmoEninwD6rTxRck5mnbkoMzCn
+          KRk9NIVUz3M8CZg7ZmHEkcsbLfQNE1hXXiQHOL1q7n60FPzDQ82or8USZ+oIj8l40a2c3Vk/Q8bh
+          89m0yKWr9ojjmqMA9w6daU4twZ+2xzbHTnX2SGns/E7UkjhCPX10NNsmTsobWWX91avzegedeEhx
+          g8L4dKC/euc/9Sv56++XLx9aWMk8fIvjLT0i9kinzy3ZY88sCAna8Z22TnI24HmHiVxBWgr2PLTC
+          jy+Rs5lMaAqbSsGmkxDipPewnu+XYpLDcYboc1GOOsvOdYSVocmJfyFPn5nWZUGCnJk0QurJX8zH
+          JoCBW+U0W3tzOnX4LeNpWb+pe9Gqmr2vao4lT+FocNYZmwZJNXCcJjkNxNU1ndPDuYKDxi7jrA9X
+          Nl0Xl4NvPxqnvJRZt915H0jQASJutznVn8eDT3B5kXxSTO7Rnx5ju8BhU+2JIy1N3b2vbr7RxSml
+          QQBhzY8t4sC3+pw4qJGKaZ1UGf72b+pNflWzK15ifERmFN3BVdJhV6kSbDZOTb3N81QwUyA3GPlz
+          RjT/KBSLmG5usDfvItEGTtD7Wr8qaApFd3yYjE9n+6x94CYLt3F6+SrjZ+XTgqIrLnHu3Fj8+CF6
+          7G4HesAH7DNu8mLQo2dMIt9ep71xq24YhaNGwyWPfC43ZweTaJSo9lambtro9wDs9cKIW4SGP7j7
+          NweE+zyI03zMbk5O9wypOLWpYxhNwT69N0GjgERKLM7pXJ2iFYwfQxwhxFYh3uxRBuN+35PoAWYn
+          FFv6RKvOS0dOXY2ITUW7h9dS9jT6TJ3OOCWv0Lc/j+tL6jJ6ag+3DX+nOrXkmhaDkEwNCB9bopHj
+          5MV8j08nUJIAqC8Lii6qjXyCrah9Ri5J7/70iIwTXMzUjTY3nhZDTNhq8wbnQ7dr55XOiVd4MM/a
+          mRjf+E/X8inIt3M00LCMic7YlTMg9/GVllO+L7iEX3vSj8/BWNs1s6exh6nZCfTLd4uPmM43PEjB
+          OHKXWizoqDQTJMdeob96+MhtcYYDuVvjys6Wmm0fsQDOZtKpq5zVgp2ffgWNDHbEux+zmMXJ8eBX
+          bwYVRNQfa2UE4XpvyT7NPv7crz7cT99QZ1206VLCvMeGtqlJLmBL77/8EV1QwNPtCmU+cztTAskT
+          1JE/weB/jqddAMUeXsSLBimdx5ZxUD4ikwZW+u46nD5OkB3CHTWaoEhnkpAV6l0pILvHWfX53ZIb
+          gMvrEInyw9WHR3XLQFx1B2pep6SbP83NAp7DR+oZW5uJ7mndw5b2U8Sx4cUWuWsTRJVHQ/ykWrFR
+          k2mOLDNtqd9eofjxb/SLt9qu9mhS20MG791xJt7ka91cqMcYLJYV5HCeEfrygR5+/IHbmbw+92PW
+          wqAbhF4z/lUzf+NYMKnyQI1Bcmter8US2lBDIy/c+W52n5OGy8FaRYu1GL7wi9cXnwnJ+Fc3Wfaz
+          R+diu4nw4suolUw+2PS30+a7P6iYEis8wQOFQrQ5q2/UcnMRAeajF82eheILyRic0V4Kw1FKMzml
+          Gy1OMKdpWSQf8gFNJGKKPERhSa6SESD+q0eh3M3m+B5t0W8rq+bwB4dPGrzkupvF6wagmhOXBL5N
+          9fmmHzX48YfS2HX1vBHbEa/FpqfW6/zWWT6HHmRFIlFNvDXp4nPyHnQ2ThF3rlYd/eKnfKMXnbjo
+          mfzwsQdH5w6UEFx2UxdOnhzrYk6dsCcFs/b1CvETEqg56ixlZ2HbAyeCQtL5fa6H9X4yYFnFKjlF
+          g1RM96OlQV0IObXaOGLLpItPkKlB6AE972zQ7VsP6CzxxAk64ceXSyQGLCJKba+60TPsCca08f+z
+          vxbvPIPDyd5TLTrf2Dc/9lBtRfe3v8XMP04LfPUCCcY9lzYkYhoSW6+L4C4l+vLlh3BUzTXRevHW
+          zdlOqRBL1/eROt6i9ycyjOhQRRfqLK7SsV+8jpTONLCzqhi2g179+Cq17q8PYvb07EE3O0x197lF
+          /X1T5XBBEU8C170hdonPFX4OKKAqiYKaV+gswbwvA6Jq06cYkdlZ8rd/kkN8C9gcOukJf+MRTUZN
+          fbbR8hiU940npTbc0nm/PpawecghtQ3N6yZ3mc6yvNpqxNiZvM+wEWv4Fz+HjDe01P5yA71Jp7/1
+          t6tMyOCrz4irRzt95qtwBV3tpxGj0r2Y200ZI9MIenoyDK6mLjobyB3limibnYYYOlqRnE79m8SZ
+          dC8WOTdjfLDbcBQ/scBYSs0TjhVNo0GInykbWL2H5TSXNPK8N5qkvZ/DV4+S811K/HncaBrcNe9E
+          bPnx9pfVcFPwdIyexB7tJ5vvYfiEutB9aqPnHc2UxwFIKhqpVSBI57cdV/ju6SlxJpR14hqv9ojd
+          HJ8qRk302dUjAbp0dSN+Xii1+MUbmJPTmxgjv9anZX/9wJXvVaIM3IMt5aa/QSdfA+o+L3I6TX30
+          QUweT8QQrgX76uUJqrJZqIpllC7kWUt4qcycOCdvrpdNFI8QvfCa6ofNkrImQGf08xuyX39qsqeD
+          yJB1JKXSPV12/EMDXjdjqq4hQFyB/TP0rhzQn14Vn08hQlXS7sn2hTH704cSkjF1z+qbTc4stvC0
+          8xsJKtMs2E+///ik+sIYzWa5b2G29YDoy4f/+Sc3VDuNNvauiesJfHUCfDipY1dLLvr6KRpshtmi
+          ur56dn15R62sr7sgWvn2uvh0pvH588NMavQ6M4/SgsyKu0b57LzY+Jo+Lfz8kWg+Un8mmtygOwv3
+          xBbplk3XchSgW90osb56Z46nPEMf/dx8/Zstm2iiSxin+3J8LahPp/aBHIjQuaKlb19Tpl+3DT5F
+          u4a48Op0+oF4hbXOHkiExW365RsN+uZHJI97w198lS0gK+cL8T/j3f/quwpdD0ZEVLuhaJRzM4FH
+          L8Qk6UWlm2FEBgiHlo60smm6SJfRgPpqjYRwW8XvRYlNAB4tiKkpbTeLwR7wWA6nUdw0cboMNyUA
+          aQ03UtSGUUz3dxZAUWUGNdMkTpftvI3Rd78j2UyUWtgUiSUj8WpSd/bUjivCbY8+JaeMPUFZt/Dv
+          WwWObQ1EH0WnFg5TEIF8WPbRY5tV+tLkTP71p2ElhFw9r623hC95fSXa3Fk+fWWZBqNd3Ekgc/tu
+          OSinEj7li43iq6yQ8KzvCYRrxSVb+pI7Vt6qEfpbvokQr97ZZExTj4Ebu2//1eqlyCBHXcEfoymM
+          mnp5CPr5p/eokYOpL3F9sFD1kHvipNmpaHer+Ia/eoz43/yaDnmxh/PZtqj51k71Ig1cjJ1JnP4+
+          5yp3+mDj3Z6pttt8iuEsrxvI8Coi1jaR6nnceBrgGbbk/OVj/KdpDdhlJRqby/Sp580hyH71MIpf
+          fkCJbCswLfhNMxzo/vRdD+TzJqDeVfvU088f7K4SkIPNH/QhlEIHjqZfjdPdjVPBOa4WOJRg0Wv9
+          9T/m8yZBybq4RrInd93jUd32WGsUlRTyAoxqmqeBebhq1IpuTTffL+mEv9dTzVoMne0PWo5fQbUe
+          US29GRWNJsGp9Koi+fU+o/52PglIal4GjVx3688fhs9Q105C9Pd6jeZDkEg//fenn5ercBgBv0WP
+          uoNVFjOpSgc5CcE//85f0G4Tw9f/iuAl1/Ws1+sSNjiyqNNnVbFsY5B+9f31c4CNYtz38MVLYnRa
+          8udfwi6vymi1nF7sl1+b1KLHiLOWLO1BRhV6Stcz0c9N6y8Pa9vDbaEnqn1UNxWPq6iCx+39+vNv
+          +0OeZqAPkU5CrWTFnMiBBHiQB2pOvZj2lj2OYBAroGpZWT6v9oOBXkZrEd9O2lq8RkoF9zB9f/Vs
+          X48PwT/LblGsf/29E7Xt4QQ/P+h42HN6b9w+FRxGZ0MS9ubS8Spce8D7J0etbZJ3fHdPE9QEm4w6
+          nyTuFqM1Ayjs+4tqsdHr0zi/SmQ6MaHXtXhMR3WSBTS57Emia3rvvvqXg/XWUMi26CO0gFbGG7GP
+          oujzGhSda1RdAYWTEmosPdWXPdsq8OH3A1XdY1QwkdsrcFx784jiuWbdVx/Jv3x3bdPU+Rd0FvrW
+          c8RFBKP5p4e8dV7Sr74rhu5exPDVL+OX3xXTW/cWpFG4jD++89P7MAbxlrh0p/tceGjbPz+7fMCj
+          /uJfC6J7k6gjYMtnH/+Zw3mDehotJ5tNDyx4v/+ndlPk9YRtAPjy92iOtSWdwqwQ4F1mBjnu8gdb
+          6q605GG/f4yjfTeRKPf3CL75QRXJfnYD7p4J5IoQRxDiZ8EOKX6Cm94tYjHZ//rbnQM/PCg9782W
+          FeCT9HXtqR/PNaJVkiXgcJuUnL764q8/pnvvGHHa2u6Wrz8KOFAaor/Cjo2rUP9gQckEetU6AXXt
+          1g7ALKoHMS1u9Hu79Q35/uI3X3xQ9b7ocQwPLVFJtNE7xLy8G2Gxa+d7/1L3nupC+8Wf2u5n11Vf
+          vwa7xj4l9rff9JqwvcHhJNVU+cwo7dX2ugexdToSnpMbm175ZkElbGfy42eDbrc9vKcgovtvfcyJ
+          bEiwOzcqIedrk06b2ligUVYSDYfG+fOT/uZBSvHh9eHn7+0vH0x+8V9ind1wr84W1er8kY6u987R
+          dz/GsX+8u3kbUA/9+JnVER795gnw5fPRqJRFt3iHakJtyDnkwM96vSj6zsLfeRVR9bn2e3SpV/Dz
+          Q/QmMhC/27IeS/0FE40Kr24yJmmE1BqOf3xu4tX3GaWJpBPLF5Wat3MsgCDEFdnHu9r/m289LkX8
+          nYfkNSOXKUHH/GPSaKuoPge+u6CNHZ+oeXCVgr0jFKAuV3RS1taY/viLfOBLNq4kLdVnkioT/vKN
+          P/47K4riwPd+6AW6iE3guxNcN/cDVVK9QlPwFBTUtnkezT+9HNZHQ9bFJSWeJ/v1+OVzCLyhGOWv
+          f/MJVk8J7It5Gd9f/48/xeMHXUDRqKmsKRqjTWSgXmVWJOvxyOb9entGl/NL++ZHwWY5mzL47Vcg
+          aRqb/a2roHn7OBP9duBYvy9aB/3mJebBxP4kIHuCnz5ZX7eSP+fv6ow3+2s4oq//z9JSjECRhpbs
+          8fFZT8GeCtDXb0KsZbN0dLXdnuCrX4mrxhbq9fTtwb0f3yO3cqVu1jZjCXKn9NQaDC6deq/O4IcH
+          /mNo0+43zyyPQhtBO7rF8tPHgpBU3+uztP/6J7BdXJkGb6r53CMKcmjP4Uh07poiKnzeZ/jiEQlE
+          1WXiYs57KHfMjJ6oaXWaaG0JX783kpIbx+ZlVE5YXVkZdUf74Lef22kPpVZR4v76SzVWEl6eK4fY
+          V0Wpua8exy+lqUekxgd9OX7qHEeorIjltoi1Jl4MuPKjSn58gjX9VUNsVU9R4VhjOrhW0cLkzk+S
+          gjvV7KDTHHbMGqlTtSlb7FucY/nakW98CWq//hkkaAfR56tHh2/+Q5GYIiF8xNdMPqeKnHf5lSgY
+          GvT1X2Rwr9eKGjYbdGE8jk9YXRdhXJv3UZ+L2VeASKNDPe3spqy3wgZ956Fku0nf/jAU8hk92mWg
+          TnLL0Ozu7xy2txCRw9GvUmozVsHanSxi7+y1/jbnzQc0NeuoFpGhG2q7mbAEo0d0v9WLydm1ElxP
+          m4aYl8VIZ4uoCQDblNRw267ula7x4H+t30D8oVudUbSuTKLVuVkIr1tWQiuwVXRN2I61BdbP+KvP
+          SSpni7/ct+keo3Xj02vj3v2ZpM4EWzcy/uZt8+ZgZKDeN12E5AXQHL5WK/j6VcQSpqobVdW15Dzd
+          HkmQPb1C1GY2YetmmNR5b0U2bM94jy75/UpdBrtiOYn9BP/6nQr4H//t/+FEAf9fnyj4OL1GlErS
+          Gc9okMBOWJfEuWwMXVxkWqJr8LpQT/oAmq6rZYJ4Oa3GJtpUbEb6e8TimGf01DpKIZoX5Ynv3SWj
+          3gWFtfiChQOyebGxKmJV56h4FFD30CKqM9Mo5leWAR5UNFDzsByQcOzHGExhTKLUNVhKF/l1Bp97
+          x6RI5rFb7HKrbYTDvaeGWXRokXdyDCzMeKpPy+zPew4JqHVKoORdnP1x6zQttJKxoYltRYi5xe2E
+          x9l/EVuWVcTSTyuAVCpbelI2nd5LL+8EoZK7ZJeJr4J68pSgbXryRt7t9W5kob9Ay6I3VSbFLthW
+          XSdwHZo3LcOL4XPtbsiRu1xasuN2ic9Ph2cFKy2iNLpmGhNvhXWC8dZkNOHXcSoWeN/CSF42sVG1
+          1pcxc2Tg+6Cnx0R5d/NKek04Hr4Tr3XfdVRNNjIMWeyQ83NfMpbIpoWrt0eppnyeiG6fvQUfOb5T
+          4vu7gvNus4zTRcyochU4NsFGDyAWtxEJbKynokvVHCLaapHc6RPqok1s4FMkUuJzVY/mqV32OOjs
+          NdFyHLJlPiUc1tqojJYcD4g931sOujAgkXDZMvZMsrTCia7ZNCi0ZyFe8znA7/c7p8l3fZxdnSy8
+          bcUheg2IFXMQ4DM8qzyONtu7qnPK43SGdXmTiX1J8pptVTHBwfkSkeS+57vJODseWM+MRGwWbylT
+          1acBx+hzpUYmVB3XpnGEJ3WZqH19dOiXX/CsTjHNw/hdsFDVYijEWKSeXNr1IKihA9s09+jp+3qK
+          swMHfhTdiHag724OSjXG8yFxafQ+HYpxTm79Ju30HQ1T6VUsnP0UMB/bM7HfwYEJ3e5u4PAOLS2k
+          YKOPREt6XGZBQNV2jf2+tPoV7LeDTw7b/FiIv3wsV/Waek92QfzFPFfwhodFL/mWFksWCU+QQ7Ek
+          XgxrfYnGuQWBtS09uOPg08MBBLifJJXunqOlc1UfLaBqg/mtryZt6yb7YE+5W3TXdw+2dG/PgBex
+          SqqUblu8+Tsf4KI6vYibKO+aDWdxgfPFSelBlBLG+94thtBMjkQbub5j6Z04IMjXEzEz0U5F/77s
+          8cub79S8xETnEgM1MD7KhlpqdutEdZ0LuJfNkYbE9ZHorB4tFNPZJPFLzTvB3vQOel0GQohqqrqo
+          fBVdF0aEHDXrU3zrXYHo3Gdkm6zdgrs9QMBvyzmQTHgpNW+G7Qdc/DgQI+TSTkR2soC32A+ir2LH
+          F6yn7uFwuaNxeb1bJEgfdAb7/ibUrd5NOnWD78H5KXHEc4yRsXhdaRjVmUSsQR7rqRt0B9K3wYjW
+          34+FmItBL2+MqxpBKlY1Z32MCEuJVI0ghRgth7A8wS/eZfiUEIsrupf3W+pTv5du9XQPVjIM0B3H
+          Qg91fZ7vSgB7Gz4joXlRL9lNO20S5d1TS/DUjon5IcKakt/I4ejWbF4cL/ldT5T0MvhCGrxP4Gp8
+          R/RZVFJu3WgaLlf39ShlRegzdhRPUKhFQeIcGang9hcL+LV3IpFV7PzJpHcPX8z7I2ouY4LE4KQE
+          ODx5b2qrpazTSQlKWNdHn1q5TtggZqMDkuupIwiCgYZyq5/QjX9JJAryjLWhL/VwinhK/bhE6euD
+          Th8Y1fozTsWg+NzLXd3gceUaYh92EluUgzjixJYO0QtVV386CEaCH4Ok0QufbhmngRTA9moHNISd
+          gpbbnS24I95M1ZlMOhtvXIV/67veZ8LG+0ltIbNpQ8Looxfzed9m4A0ki0QOV+kyu0SBV7cMo3BL
+          KzSL2dODatTW1K83WzTzJ99CltsqJDGy1OciK12gXrUOTZlppNxbPgWwji8XYp41X1/CWpvwu5kd
+          cuiuvs4/rFwCy/0o4/rQXPy5UK8L0MRhJPCOfr1oov3EyL9t6AHzScdvel0C5RzvqO0Xx24YqluO
+          ZWcdEWc/KgUvPDYAU3GPabh9ZD51qZvLP3yy0s8LTVqu5zjehBO9pkVbLMk5/KDhuvaIdcgyfzF5
+          mgPGWUl3cjJ3w/W8xDicFItkr+CNps+xctB05UKy5Sa7m97vKEeW31TUfNw0nzeMCP7wRum2Xcev
+          J/6DJYs9qSdEXD2E+Oihtjh9iPsw+5qNi5Ph1nDXxHpFacHbyvaJ5ZiF0fPE+93ibtUVrNc367df
+          vlC0kgSi9dyPKOCILjpkc8N6XZfU8EXGZm817+FTEJEYcYzQvB/UHj8LtyGnc3TXxXd5BWBjppL0
+          +SiLSSzSCn3jT521URSLwZIA26lS/vK/EMTDqQQLRVd6bGnO+uHKGTi/WApJrdb3v8f/FlSs7K/j
+          Iui1qNi+Boa8PlJ/nzz9mbu2CljlLabHN+f4s100Ix7XoUF3T2eddlczjrB1Iil1b/utP+m1nAAf
+          m/MIycLSx02HDPa6cSJJVTRo2iybBI7mJSG7tRr6f/g7k9Kj/hMpbNGUdwbNw5JIsDWfjGkbJv/i
+          O0qnvO+W9vNJUJJfPyPqkyNa1hF2wLStHS3tm5EKmvLeY2eJc5r+8LI+PgMszplOj6vnoZ4V/QgQ
+          tuxEo0OdF+z0upVYR4dzNKanup6mp6Xg9W43EHvXBikLGlKhFb6PRNlOt7qXtN0Tlw27Es2bCv8X
+          X8RzSkdNA7f+9N1PcE/Gnlr0VejiTjkYQNmQEP07LZj0eBtj0Wr2NP8kHZvFTQLYClyeBIYlFawV
+          dzLYrhjTUJsaf96FGw661TRGD+nx8KfTMQdw79KOGP3nrjOtvE3g7J5udHT0jz77D3aCSzQuxOCy
+          MGWJnwM8tswe1+tERgvHmQsOA9Onl8iAjrHOjxChfEIUAruOHmKtx9NVCCmhxELiKux7CHbtmfpV
+          F6Llsa7P6NGMfMSCh4cmMX3ncPU+KNrNG5ctE4snYB/VoEazEeq+f40BfONLlK13RPNS1MkfX3C6
+          G0H86XQZYT4GHYl4dvXpc9YBzJnTqHl8Vmj+NMoTb5LqGkniiepDrokrQH61iWYU0G6sqMiBWCeP
+          cWOqhb5EnsFBbsYaTa2286lD5goGK/hE462ydfHzmj9wHV2Ibu0qL6YdekfAH5PV2LNdWDCRP0qy
+          jY6Upvvqk7JtqlvYXa7tyF/ooV42widDxVDto+dxgvSHV3KSXz40PtibYkHvIoZdl3xnf++ueL8e
+          twlkzTuPsyla+vvWvh348WU9Xy0dfZcHgGRTCiPYWC8WTSQN+vLzEZa41Ce+/TwRz2kdJTRHdbu5
+          xgkuTs+WmPPmjXor2lgADFSyn/Z7RM2L8wR+7ZzIdmQ2EtfH4glyIQF1tcfCGGHIg5OjYHKdA6te
+          9gnbQ9jOJ5pIRd6xEq3OkIrelpqLture24PhQLI5CyQcnUgXnn6SYyY4yl/+Ttx07jcrgILY14fP
+          hnS8SxCddhvy4+9LYzpnMA0qj+iwXurmRNYxsk5hSpWGr7/5p0ewIc30x0/F/dtpwbWdlvgL2RT9
+          5/hx0JK8XBLk8acesbKxIE6XkPq7cI/6l3ApN29VPFGtlMOCseM6B/3+BGJ1OEJLlfU9lIW2IkHr
+          uYibPFmS+2Q7UXVkNvvyAQeb3EGkQeu9Gbs/cQJctszU/PHL5vypkLfJ4i+/9hG/H9QRtZK1oare
+          LN0ELzEHeIkmUX3O7Cb0qCJIc/P111/6/vWMYJLI9Re/bko7/PzhacTltxGx+MRHIFnzk2br64TY
+          5enEoCvad0Ldpd2CABJI6LijX77asQO8W3iCM5HDQ9d98Z3FH3wXn3HEodUx7TOxOeMzBwndeX6v
+          z6r5ctCzWtckkFhTLym6VGCea4naIWm6hZnHD87FI/dXP199UoJayAlVjHnRB71Za/AczYSoaTmj
+          5csvERiZQbbrs9bxu3vCwf06bamqvCt/klc3BwzTxmNlWyNrndEBaJPwEknbdcNmed1I8NYZEOVX
+          fxrbV/JXT/x9/t7dEwHk9RGP624k6fJ8DhUEG/9CrUqh6XBQqgnKSS6Iss6u3VyI8QmOL32Olm2q
+          dvy7W/pf/41WX743vh9mhr74T/dTqOnCfoxKsPHiEaW/EsT6RCoBP6obsdbalk3yke1B78eOGCLf
+          1ENSMQVHlS9Fm2cad2x/fa6guUUt1bbpvZveIUrQcZMHVD1flG5Gsr2Cd/GSqXlfe/V87O4VDk/O
+          m6bt0S3mRVjlcKBriXpnoSy+/XmF2I0Xib18RsQQ24xYx+lq/GxiveZbQxqhh7NLrvgxFuzbn1Gq
+          hwdylYQEjZ2xSzBeb5VoWUoJLfxWuWH50V2Ie5EdND6aeI8P24NDyUet0Kw3awW9Le8Qya9j07Ey
+          9DI0vGlHvdXrXCzvbbOXv3o4Gh5S4c9UlfYy+J5A3HX3ZOwhCRa0+9tp5G6TihhWMYcqMC70JPlc
+          MS39dYWwYDPin4oTm9RZ3YPEC/Efn5qM5HnGu+fJ/8NTMV8JAg6WrIvQE90YG3bm58cPiWqbnj+F
+          /tTD55juiP7Ad318jIcPLHMoE8fj7t8Tw+sYvno9WtvarZvwLUxgjd2O6usy0GncuwoI43SgOZsq
+          9sM72Nz9mCindkJffSKhH79Ur/msfwbnpP30RjR1U6ZPGZUzJLMREe3l3dLpHbIEmyGyiS8FR30q
+          u+OCogNnUK+WQp1b+sMKBt+QSVD2Sz1e5U0PQjj7v3quJ/Vyy/AQqufo9Q5ENr3cVYXkkC9pWKMi
+          nRp/30O590ZidvDR53N5OkPyEUwS5JTpjK36CO7dNaPWbfEL/sGbCuI/sUeDWt5381pTK5Q2UU1K
+          h9z05X5yW9Sr4yWa9Wap5y+/luvsSiPOehM220Xfw2vD+WRHL2onTkpwBiPaeiS8Ls/uT49mbT8R
+          5WBv0mVRMwl4axdTkhZOIfSctAJ2E0Xyy5+u3Ly//M+eR3xqqm7x0quH2MPhyNa9L8X46LUSr1rD
+          p/olMuv5yNIWjrxNiFuUSjfhPPXAm3z9x/+++cULuMs3SSQ/2YUN+HiV0fpevom53RM2dxthwvwn
+          8Uj02Fb1JI7cDTi5DelOejz0pY8NBwfNNoxG2Jv15Ia5/Bd/koRGys0zJyM9AX9kqyOgQf0Uqw3G
+          +5Ko3PTqJu7MRsDxJ47K/S4r5tH9nFHyvDukcExWf/MdgB50feSUfkq/eN2gIUucSMh1ghi/sRP0
+          9ZvI9XIHNLXPTYAHdTNQI7Zf/ps/6QY8Blkjiinc06GR34CGSt788IbR8+mYI5p4jHhXd66pS9UT
+          HPNCGY+a5RXi1MwJODi3v/fTFBM2Fg84X4uJc9sFKXffx0/IPSeJVkrJd+3mPpRweZdX6jxnLh32
+          Tibh/V5ziG96GuKvU5H8/JdvP8nRUrfhCMuxQOP6PhMk9slU4my5bol2fsn+LKcmwMFsTeKcuYy9
+          SN4qcLCfr6ivQs//+UlwyS8p9cz2WU9RpzTYNrBCjI+w95mazBJSDifz269aNm/ySsKsOKVRr34m
+          f0aT8wE/Cm40HZyG0Z3USFiWHj7JN0PVLawtZWBx0oyr90lMR+88cZgj8o1Go5swxgmLBegQAnHA
+          9Zhwhq0EgnSTSMRnQj3iIVbgy++i6ns/Q1yuYoimcE+z3pALOkunGK7B4xKt8WP86mlb+/Vbkr9D
+          qJe3KwfAjfV2lDs9ZlxzriosCUVN1SgeWH97cBw6vtR5xPmWpMtmiyu4Ofs9MZTQZuPydEfIL4ZC
+          D/P5g9iwC9u//W+/+M90s+LwspctGl1uuT953EkGsPgN0Z7NpqM/vrrP7zhat5B283M4SShypycx
+          2GPu2NN4aGB8cuu33m58rLsSsFFn1IjjAjFqvnN4sMOT/vCYedrnLLvl6kn01RHYPLXyHm6X6EmN
+          KAV/eS3ZCSl350nJWlJrPryVZ3So9tmPj7KpEfkPrO/nNzWE161mlVwmwARPGXVyeHbDoeBl8J/a
+          lSg3q/XnlUQnRBonpmmUiOnUZNIJBEIHasVS738OZRXB6a250WZad2hOKqTA16+gv/3v9Xj7n/3m
+          s+79joPLZPz1M91zq5o/du8bcuxbR40S9mwm/iVBT6/ZkmyWTF9cDp9IPunOQLPqMqZj6ZXan/+V
+          xzJf900m5eBj7kJPv+9vn3MAwqHuo4MS7316LK0Jvng2PtXLLp2HY9SDVqM7ddlsd6/PfhDQc7QT
+          op4vt469ckuBV6KFxMgErV4ujiKA8tSEkfvqpeXk3jUk+VYd0UOD9Xnm7gGg8FJQO95MaBADUODe
+          tNW3HinrCyf00PYUACH8Dn2f0NtH8K0fEriy7k/iCDek4nKm+jaz0la74u9zJTsyCmtti9igNp/f
+          eiNwZV3nHLVSoFCPBfn6EYiVSDgjbr4otOzPz5S1znuCL5/805Nd8eZbxCfngERoWtU/fQ6qrWkk
+          MH27Zh1+cWDtDjqx5uKp9/Lq5uHzMDyJtr+btWh9LQh3x47RoR9IwW3ehydMVqdQc60O/tjYuxY+
+          fL4lF+J2aDhI0xOL814nXkqWYnBkaFCWcjeSfMxZn+jaB3S7BE9qGPtzN7tMKyHY7q4kniM/ZfJe
+          1KB5GFKErt4+bQejXcHX/4ig3zyKoeSlCH7+pXoxL4z57+aM2MHek+BVjH5Hj9oJ/fwYU4mCjv38
+          mPHyeRB7b3OIdbb6xHBXUno8SUc0f/ksavTbkV4OFet6D0wN709STs5fvcesly0h21gr1LmkOGUj
+          kA9UE4xEkaRDvdwdLKC9QtfRvJ3Ngp1ds4EJbuevP9cUi9d8AN6TZNAvftY83j4WmEsD0WN0TXQa
+          ZXX8p9+JGfFssOpSkPuixMS8+WvUqeuYw/dLIUTCLdW+85XQQPfVURvZLCoFn2siAA77O7WPbo36
+          b3/780MFNVPquUUloOO2ichVOH3q8fl8VHheGzZJuERny7sXS5BZj0aJtwskjK+3BPZjFZDffG3W
+          qrsAn1oXfn5uwRfepoSvHxvJbz5gi0YvJYzkYZPC9Cq0mF1rwVtN7D8/fpG1bQKHlbofOS4bCqaO
+          oYKu4sGjNjp+ionbuQ3oeLsiv34l8hfwoBRp+PVvOTZtrnkM3/nMKMryHS3X9yCLOepnWoovF02w
+          8QOkKO7pe4LIR9PtwQnycmtfxH0uR39AMlmh7i2VxFuGi89i+xyhAnw7Yp1863p3pyvyr19q33kO
+          u61RBMn2I0Uv6TXW466bE4zK8yv6nhNLv/oBoL0lKFouvZtST6tK2Efz+ft9SrdkN++EtMEao6+/
+          kM6+ajWQTy+N2t2KdfOXD8Iy9gZRN9fan4IRA9yea3ucctQUk6TtGnk3OT3J4hgxtmk9Ae2iJiBB
+          oVmF+NOXR94kxOZwVczearOH0+ZxHiWXzuhbPwrIttb/5glFJVlWC9OrDWg5fz7pd14nIbJ5sJHr
+          kFn08SZoIZnS97iRoUuHuBQSoA5+0CjqHF/8zbOSiLojSuao7sUiveGP92m+9Z2kwkqPErQW9ptI
+          yrubP7YmLoEn1YX4SvFK2Vs3BNiqKh01eezQOCft9wkwziNpRkR9ri80Rt/fj5ig8Po8fJLbzz8e
+          D9mpTSfZSjLcdhiP3BfvZzHgFBhvzyxC8opDy8lfRYhPymDcxv1ep8N5Pf19//VW+2i6RFcDqusp
+          ol89ySjelRxQkK50Fzg+GpzrvZR/etpOg0Z/E3+XSO+rLJLwQZ8pk1RHAP0zucQvVLWb5v17AbWQ
+          Euq+uVYXh81Sghlu7Ehup0s9S3PLwc1cI2rJ6f6rR1AMgTIF5JQP725IVpH35z/urnVVTNa+DnDp
+          2DGJniuqt4ry+Gy2LT+Mm+w8s4/BThHU3XiiWq/F6PGtd7zfKw45SK+oHr/zkw2zKp14w95Iue/8
+          CLma2FGnHzm9K6JP9fMnSfj1E+fV2JYowOWJWLuX3YlTcXjC8RasSXTNKjaFcs/9zUOOB+z6Ytyr
+          inTaWwUND1XaLYt6lmCKiztxYySnSxypTzgN0nEUv37HUj2mCl46UJpp67qb8fEgoyc534l23+86
+          sdzcY5y3TI8GZ6hRz9xkhK9fN4L/IvUUXo/9/8+JAuG/PlGQjJcXNbvulS7MMQLg8oVG8F5J9YSc
+          cwkntR6o31evbsSdsYK0niLqHZqMLbuEfuAune1xNqwFLY+SlhC+7IZqqhbV3DGWBKTAu6DOfh3r
+          wvt4tqCc4jUN+c+gL2RIGvyQ7w1xmeYgcacfBTDzUvqe5J8LmuO0QnXAhaTYojWa+aX6IPgcYqqe
+          t3I3hmvFgzDz6SirVPB7bvQFSHaCEG0u4cenzqM5wXWcT3QrS5XfO0bkQRNvOeLwB4pYOR80rN5Z
+          RsO0dBnvy+89nF+3JZIOsovG9lxEqN2TJ/WKl5Jyzw8XyR07bKLHYSt2bOWpT1xfJ42qSep14lk3
+          Wtgd8ooc2rgvWHYJRljcwqKR8axSdssGDSppLVFjq2M2Xa+JgXYhH5OiMwLG6ssugH2Sm3RnXpti
+          Ompejle+TqmTv1/Fhxu3MZ7Q50wCtGZsnmTsgWWoVxoaVYYm+15r+CCxGw0/4wFxO3u34Fk9HMeN
+          3rzqOSKKBEo2nYn9uo9sqcINgMxmm2q7UCjmWz1JePvKTtSQQO+4+ZF+sMf3T7J/e4Y+9OQuYa+a
+          ZkqSo9hN0f54QlF754mWJVBPxdqPwC0JRLOSkIIXVkeAd94caHkq7j73EroKN7dzRQpi3rvZl+8Z
+          fo37Z/RSdx1r6Lk7g6txBlEmwWACV91X+I7CiCgrR0NCdXzHECS0pJodyIxukrOBxlB50ORuPArh
+          eU322H0odnRKp20xXJ+TggenvdItxk5KFe6mIK7aqzTQshOatE+b/8XDur30VJz7zIGLxRSiuQ9a
+          LJ0f3/DjGKvfCTvzhwvHnnAFLR3h8soK4bmRLHxq2pjs28Ji3Iq9erhIZEut43VbL0YYA35pk0Kj
+          7VsqFpSOe+C6viH5Pl4QXz5ND+dattA02398HjNlj1dFdKEHdecjZgjUAsMvFbJ/CzsmFn0uILm5
+          bWgeqrdieAShBVIx9fT6cON02r4iGYH4dIkhfJ8x6d7ZApnkZTS4vLh0Vvd8jw9G41BTPUts0V86
+          9z1x8j364Ez6ckiTCl/S2Ka5dN7r/K2WJFCe80wjztA71pz2BtjF4RRB1K31qZ9jD09t/6JHfbf+
+          yy+8sSuV6u+z33GNHyd4vMIpWh8cDrHL49WDvV2JhISbqvjtF4ofkjs22/9J2plsKcuzXfiAHEgn
+          SYZ0In0QEHEmiAiI9AFy9N+innf4z/5hraolSO5m7+sOqeBMmewzR9C9aRj7Z1CH3YNYJYiBwmEL
+          aU069fppRpqdSfjF9q7KThLI4VPJXazamg3Yh+hB+L3VT/x4MuxAI90/INGQN6K9Osvmae7XUJmK
+          Fkt2TMMV9L6FcBRanmhDd58AjT10HJ7OdLmx4RyXEkRn0+u9NvYDsJnsV4dtk0SeWHChytqF3Yi5
+          VvXEYZv3sCxHr4RtZ6fE5iTF3r7lNQOSkasYUw6lJGaGCJxC5kkk7syncx2XBjRzF85H/tQCKjGF
+          AinyjBm9Bhlwab2J8HzlM2w/Ng+MrwvN0S+it3nl2HhgH2YRgUUk0GPiRQo54TE2QPfLZl5l2KSj
+          nK864i/PBl/ngKVL+vgo6KKM3Uwf5186Fy7WwOfd/7yDymuAP6lTC8rrvgNr45KK+vewR9dUkeeD
+          IwJ13XLbEGfitfObxK26HFVThwwiJ2w+TgAsTh0d4ME91sR9DR9ANPGhQyN8GMR4mdmwooO2wPrq
+          RCQArZwutIxKdL5VqscXWUsHwPYKtK+h6y1V3VRc+fj4UHu9bsS5v350ukuLhYZAj8lLd2zKnttc
+          h0I+f4i994etU9oYmryaYa0Hhr2yiSPAyuFccmFLK2xPC1rA/n2wHn4+w9qSUEEin8/ES9Zebbks
+          DqB1vJhzkcpfwD/mr47qxs6Jgg/VsPZBJ0HRgAHOwk4LedO4FEjrYIZfXdjSFUtB9rf+M1u6Ixid
+          R1KK1xc18N/99arv66iu3I5EgtukVO2PCzg2nkRsN+0rWuqfDKFOeRLdUi827fTMh4EgnbBk9fJA
+          vO/dEd+6kHrbIT8O9PUlI/x5p548zeZrL6fvGqMl4BB27q8L3SbjmiC2WgJy851rRTs98tHvrSfe
+          rxOrdAtk0YJyxpk40CwpZeLwrO2mVyXGRzgBOm/PBD5T7eANvyUfOCWNSnRxLo439UqjLpkl1tA5
+          ejV2uW9D1/O1jJCNlhDfIjAD9r2/w/pQqTrT5n5O+bkaINzrL36mPynly1FtkR45Llaj2FK5q3by
+          kfBp78RvupHOz8O0QehtHZE/5GYvqpw5sCdigP2pjCseOjcGHXHKYu2qvsDiv+UazeAjk3h7xMMm
+          TZYGuTO84UCcLgPLW2YMq5wtyTMJg3Cx/DZC10mNsW2vNN1O31BAdANnot30LmTn5KlB46nURFew
+          OjDpjZfg9g4Wcm7vj5Ttg05Bz1mi5K+eb6ZvidBVjCfJVFWwN2o4Dji4qPbEfGyq9dY9MthvjIH9
+          54bAdmyXGXGBP+PbkSPqFi1agYY8mYktmU7FcIXkI++7zNhrVMPevM8QwLfqMXu9OQ7bdCpFhCcm
+          xWqXDcOsaO0Cre9HIDmjVdX6Ia33l1/knEaV2l790EC/1c/+q9eaf8zBp2xe3lFLf4AcSLWhpTzb
+          xOzfhjpWX9tANU3uOFQ+tkrf99GA75T9ehybQzDx+j1G0Q1ecfKX33h/hysQlBOxxly0p+Mn2KAz
+          +oSY5zlSF4tTCrQdygNR0OYCzj1KBipfjyvRpPFUbbkhCah5tRf8PD6JTZVS/Pe8SO57XbpmnRZA
+          cuGfM119Ym8p/cRAuXo+8ZqDTNm66yDc+z85C4oN/q4HuSCYvU3jApXOzuKAw0czyA1cVkC/0uAh
+          +XAeyavxyoHjlPMBqkJhE+n9Afbo3uYC3k71lzziqEjpxyxquN6xSsyzUqbLW4MRuAtr4a0wXNN1
+          uT0P4HlLFBx/Y2nYOJ95glo9pjOvHEuw3EYewt9yh9jY9eXy4YQZQcN0sT15p3RJfPcAWbnB3omi
+          chh9rdrHoc/LzKQ+OywlXZt/3+dEkTLQx5vrYRu7DdFYo1CXi5AvMOuvJ6KtZ6bamtOiw/VaXAhG
+          d7XabvG1h00TnIjrntlqe11ySRQpvWBXUD8DPY0nDgDrWRFZz8Cwfm5WBn/nz4VcsRir69ZdGpid
+          I0K09zEAq2qIBlCT24voCzlW9JsaEB5S50WyH+PY9PZ0a2injT8zpfkOpyFPInEZE4/ojbKl9MIE
+          DJKMTJ2vbZuDSQABB/7iT7lXR7UXvMJAj8K3ZlphWk1T+vDg3+/1qnvR4SFGT8Rg/PI2Gwg2jQIj
+          EOfPccFS4SfDmH3mGNJ7lONgSs2BiQUSAO04JB6r6oQuoBQ9iJ9Cg5/F5Ub/9BFw83z7y+9wHMtu
+          PFFD98m5Mppq1q0HA5Nb8MPu4mzVKmq2BpNUVInSjmM6yqG8oAWDy3xIVktd93gEx+A+eJTVtmG/
+          Px9qOnsi7pXZABU+SgHGB64ILpf0T48+IexvPnG+uRpyV05ZIGDdhMgKL9lTYabG3/PHuvR1qllQ
+          BA5aTZaQ2zBcUn4Qkh7uep3ceeeTzjQWW0iRY2AXp4eQHG6eBDX/YWNPl68pw7fnGoJHv5KLXaoV
+          KzFBDe/NYpNrfPH2+r41yCkFhzyezK1iucMVwu9HpPMxP6qAW4/hAdYqSolUsl267fpcdISvMW9e
+          NFZbSo6RiKsqIRdHeaibdzmXUGBqh9yjszawl844gNS9yhg39JIu158u7gd4hORsam+wpdbL+Fs/
+          8lIv54HGWLZQ4R2MeeuSg/1Pb3b9d8Sv3e/wvfqy4K+UvH/xuOb3fITg0a4kveJZpZaxPSGD3Re2
+          9VCgi51MPYQFnEjOPcdw3PUMWkCbYa/fzuHimLwj4gtcSb7rozW/xzNUvOCJ8Z9eW/X1AMOokPAN
+          XK6AuApjweq9KTN75Ig9LorQwKW82NgVH31Kr8bXgQdbJsQFkaau+3qCOK8IxtX7pK5GeefEybul
+          2KmJbtMqSDZ4H18X4nWGF26JnyZQq+6SJ1w/bDWyJ2AA3pcTjIdSGxjCyx68Vi3Fe/yk3LnWRjC9
+          fh328NsGUzV2OgDHMiXOjfqUM3Ivh8TLOO/Uv1t1jdq5AUP+nInykE9017sasBXliL3L0Ia9IdQO
+          wqLFeKv9lABD774G5cHkvQW5nb2I06GHe78jNv1Nw/SyyRO4fBAQ0+HOYJQ+gge30mlIdLgHFeFq
+          swfnzHqTc/7o0u2vf41dcsVvA7zsZYBZg4IsD/70c7gkKQqg20oazo5haXe9vo5gC2CCjTlv7Nnq
+          ix5BKpTErWt72B5ckSFJZn44+YJzulz7RfzT83s+6wONpecMG4BYT+CCfYeKYynwdn+WRGcnDXCm
+          r4hIPAcDdlfOHBg/fTdQf5xM7ETmg9K2uFgQjE3jLQejpFOymS24/uLnnx+oNg/KCpS6qcFhIOBw
+          13sbiPMPmRHK/YHPm6BGuuwn+JEFfsiTiC0Rl4gWtm8HO5ykb23984vS+5OqNDATHfbujLFmndaQ
+          1mofAfemYyJ1ipxuDk0YMF4cD0vrt7Qn4Sb00Bte/QwkwaTbJ2sTUOV8SS5fu1HHId6e8HyIS2zL
+          dTpQzxsyAJX8ik14PqoENrSBo3OyPDYjWzqhg7bB56xQLItCqbb24cKh6RpesetbgC6sPx8gGOsG
+          y3u/4IXRO8DHK4qwyiQFXcOiC8D9yw9E3vRHyIjwqyFo2O68Knyhrsm+Q2LXc14Zbj97E7zCAm+T
+          87FSCwbl//rJeUgmbHIXoi4CZUQUareAyB/pDjhdJxxQlTedl+UHAVnMkw6/YlVjV72cK26vP2As
+          N3YWnlocjt8JxGDQ2Jd33P3Zpw8+Enp0/YQ9YHbD9g4FBxbL84llhS/s8fRdoz8+4H212qp4W7QF
+          +GqROrPBWqbroIoNXK/lxeM/93KgBVo9uOs7D1p2B1aYdiNUk/sLm2dFSZezpT2RQw6xR/muCkf3
+          1hTwerdHLMvTIR3/8mnUwxlLcdAONOLLHGWp98Byk2/hHAvMLN5OzRcr9+pt0+iljei1fll8+dyV
+          ivMxGuFDEAE5FxVL6VmJM8T14E10Nod0fSDhACJVyvb6ZIXc9v1oUJHV0gO3w5Buv0FuIV62q2cS
+          75zyaqQpaK9XXstPQrVUcTT/1TuPVwemIhD9IrjzhJlJbFJNr+eNgf1Fo0SZL2k6aOdbzQeup2E7
+          MXhAe/VlQNYVfzsPdO1lFqUIRoIRYaMjBqUFOjlwjPMMaz/Kpdv5YfhQsaMvMfgpGbbkbXJw0g4C
+          tm0ltrlvWCkwcB0Nvz4+SteMF2fotRWLle7X2Nsn4RJA73GOL0XWgnaLJ++PV3orqCmYhLFOEBLO
+          IrYM/gfWXrA96AcwIme7W4blruWtWBR4JMat/tisUb45WLxCzvvjRePf53FMvnpHHF1TGnJ8DXe+
+          +l89Pn1TAaSN9iF/+UCtyrfg02YWbJ+Y70A/RRaL42gnGDNMScdYID602LkhLk7zdIJpN8PvR6Ak
+          ij9COtJ7osG1NwSc1FcvpaLml/B9RoCYcVSE1POqDI1SvZEH8xLsvpXZHA6BFuM//b+JdifBWRpd
+          nDySehjPlpOI+Ck22BNoAXbeMkKlOTnk9dZ9MOzxB0Sv84gypV1Fg0Ll0F9+GTM+qBQ2oIa6jiLs
+          gnOmslfT6uELCALJI2UEkyqqrdi7I8ampr12vSYF6BXLHlaOWQ1IMIsBVIcwIFZsfsFS/N4JvPDc
+          B3v1Y7X/9Cb88wdmko9g2/kgDLlRJXIR6fYieMEMRSXmZjYjQbowxqeFxb0WiVaFH5vszxt99EuM
+          XZXyYX8DgPvHEy/j0RzWIU9i8McLNMFtwrE/pjN8Ru8S28+prbYg+QriWn0youo3b+CtOWugftzo
+          /ABfs9osOcrgU/5MWC8SEG5HLziAN5RCchGvv2r8y3f02Jx5637NPlHJRehIo4kzW1erSfKEJ1Se
+          +16RX5vSVWfcJ1xuy0asH/6F63GRPcjdjz3Wc29/Q66cfMgXjk7S28FOSbPC8bTzNvKOs8peI/4m
+          QWu0I3zpxCrkdr4HcHS1iBzTN9haGeWi6A0eNpVuArTymg1yBsmIMee6yuQn7fDH9/G7eFJ1XOjd
+          g76yNrufVO2Nrn4C+WP49dbfY7QHcz9z4PMZ0Hw8VYW6/qpqhHs/I/bODyjTvEQOAIbBF2re6Pbj
+          PA3uz8c7HXoMxuHbcvsGzc2jvYHCzQW5BZnv2cA2n64q9Y4QgoVVJeJVPR1IW2BL7O4dj12PjNX0
+          PHMRkGqpJ5LPzOqkFf4GUjeU8T6/qNY//Zr8QkgcVRXUvge8BR/W800Uei6q6d0ICqiUyiJmKG27
+          n4o38CLNez7y1jBsqDBKuPMN4qsloetfv1i1q0cMsNT2ZJ0OOfi2BcaKZkkhr51fDYyc5OjxylEB
+          25+f8fBJxY43vFWajj6HlAY4e38wKe/f0x7y0mph5eddBxbe+RmA7CEQxYucYbULtYG7P8Hm9Vup
+          00P0DnD33/OW3Vp7yU8O/Ndv/cIXqgGvegHT7Snuem+hQ+U1C0x06YFf81qFy3S6ZdCQXuH8s0Mw
+          TLu/++MB8ykENqXPoithF4kLMZLnO1wd3Z9hbdU8vk5NbZOxK50/Pj0zwcSAJYZXH9FJfmKzZVd1
+          +tysHFZE8kk42wbd3Bs4gL94+NMn665HYaENZ2yv15rOvTW1sCKKT87x90tXlPQNnPLCxA+HO9Nl
+          mq8SnD9omY9fwa2obKcMGNm5nwXqVin9NucEURi2Xref2rftPPSf30p/x6Zq3dtcAvlhldhwNiHs
+          ftUwg+DQ3vYdcKeqM8o3AxtwZD1u5+1b4Z/0f/qOX98H8MfzYUl9SLQfjcOxXsQMXic59pDxLcIN
+          X5bkX/+6SctFHSZZEtH59lGJLdegmi/B2/rnz900cFPmNh4hoD5RZt3btJTb8xEOt/0M4+WXge0q
+          4wBE4inFeokcte3o+ITUXR1s1dYnXfPGteCu/7xDUG/qep1rBuQfZZ2Xbg7t5TvRCNqKdCSJEVT2
+          1IrvDIgFH3jwenRDhtGOvTih5Y6xlv7oOuR+hNyl5zxRVuv0j0eJ/VNqyaMc6t2/Sj4IJ2Mlyu63
+          ZyPXcxjNP2Vf76u9FSkqxV+SQaw3K67W53Vqxb9+74KoVqed14qJEm8e9wlZe32jYQNX0cc7/3ik
+          W0jtHlSZls6sp3zV5QKN5a9/zLxV53/6qIS735vZTJUG5sGFEtz5yz+eTyNlhnBfX7L3T7CJ9kcC
+          Lz8OsDpsTLotnpXBTtyeRDIlq1rvNmzgw9N0nOx6ZntwbSZqOn/C6ouphyU6Jgl0HJZi+94qKReX
+          xgFWK9Q9kW6GvQRc4aA/fg5pFqqLn75ryIiHjWi3DajtuCYGVD8CmdduK4Z5iDMI2UIf/vhKtYS0
+          02G22DXxjOcZ8L/EZEDwdh9EviULXWH6GdFpfNuz/o2LauVJVcD6VY7EnI8oJX/+M24v8p9+qLaH
+          XB+gZ3n7DhMw0zE5myVY9KLGOb016twvmIEKXDisw8ar/vQeOqfukVjCzbSpFXLMn78kWvxJ0uXv
+          /qN3xGC1OvhgpTibIRpnGTttZVZrc4fJ3/wVGzmlwxBeXwksPoZDLh9DrDoqNT6SdW8isjzl6Xaf
+          BP9vPkP+5gWrJMkQZDa0PX7/+xkVUongrPBYSZt+WK/3kwa+INP+5kkVGxafAIEsFYhsvX/qNr3P
+          BjzX9Eoc/a6nSx9fCrjPL/BNzhL6bx7dQPFGzoodq388DQlBYuLHmqyAvLrrKIoFG2CbG0NApmbQ
+          4Wv9sTNRFJqSI0g1qHS8PC9DYu3zY92AYVPd5uZ7SaqdNzfwq8cO0UrzGE5DFy2wJRvvgUIshtW5
+          vgyARwDmfuPndGmVpwgdSx89cZDe4WIfLgw8l5KFDVF3beZpXWaYFSAmeL/+3v8MKOOfj236cysW
+          HZwFotBYCD5OTrhaURbAXF4qktorDcfb5bXB+jIoc7HXj50vlaBfTy3Rl8OlIgryPZTNhkeiQK5U
+          KnbbCP4fOwr4/3tHwaGrTO8wnUV16VY9gXbZrx5wDj3dBsbT4XIAA9E5b7HHOvzoKAju0cx6rEKX
+          UFZK1OedT6xCW2yaPWsdfoo2JtFTa9StOu8KP9ZXD7qNWfGn47LBOiV4Fo2zF26trElIcoNp3hbD
+          B5xhSwF0GWvG+mMKwFr0pQYJurNY8YPJpkab9PCkyz4x665Jaf2zGIhikBHn6riU/uyLBaOSPvGF
+          3xq7zWLJgJfDRyf5M5zC1b4yFhKGJPN4bHb2eq/LALXjNyXSgIyQo9Iowv6mR/iyfVd7uU/fHpzC
+          e09c4wxV+rngQPRiPyOaORgDDatSh3MkdkTyCgpota4Oso4vG6uf0EyZl/3YYDlnMwl7Xam45bp5
+          sJoPE9FJ+A1pHrkG2N6nDkcmbNKVAGOGYoBFYm3fA5hrU8wQCSaFZCqfgMVaIh+Vx/I7V2Z0qdht
+          Vjh0cguLBOZdSbnp2fbQdUyGGH5YDezMrQEqfr+vx5hDW3UXe9ygmb9dLKuwVxdXDjT0hcTFWuOc
+          ASfAMwPZb6diQznMwx4PTzTEvYUd66rafFHQDCWI0cnFOUsDa/4SEWyWbs3clHJ0HdKzBRM3upBr
+          zsUDlznXAk2tdSH4eKgB+15NBanCdcaXF2DBYvzYBDGcB4j5tmm4ScWXgbeBsbwDK4Qhq3/9A3pZ
+          kMX3/QzgdTZ5BhBD94jiMU97YjJUiqOWVzN9TN/9/kAO+9iwZ5HtlYE9/JwFug0XkXyP7/Evvmcs
+          v4inR3Y1ap9ZgqndNESf77O6Xq9jczqeMYOlA2FtqpWmD+vQsYhft4q6BYoVALQUZ3K9lF7KRmtr
+          IYt6HTZbUVN5D+AaeoRPiHkKPxX1v2uN8NcYsTwrR7u73qACX+/NxDFlJSB8+6MCv5VxJa8QvgAf
+          vVQLPV6BR6JspvYSQKDAmIMIh4/prDJxYASQ+zkn8m79j7pJh0+OqgH7RH92/X4qqaeJpSAP2LpU
+          wbAdw1CBz1T/EbczIroqz81AynjRPPKbv+GqX24RethGge85P6nbtB4aqC1+Sx4XflCZQ9XWsFPy
+          H7ZVkQ5r0Jk+FK6UwXLimJSzC31GayN8SWJkcsUE+TGBwqwrxGXU47CFVvcvHvBl7R6AXnwAocJN
+          gndE1ttmtDFLwGOOz+QyRzTc+gOTQUYQPvgKwyHdbOKUcM3vFEvyJFbjPP6eaNt2gvrer++BjUMf
+          N1/w2V6ONnUjyqBkXDB+P1Q0UCsAIuyDZJ65l/JIuRw4BpBLyfQWJ1EqLj05Hrztp8RY5/QTbn22
+          WLAPnvPMsUFCV1N8J1BZxAexJc626UXmLDi9vzesQH+p6L969lC7GRI2S7e7v5TIFE2B6NPIDJto
+          MA5cG/GLVfd7TnvPEsa/euTxe/zPuaVIaGzqfaLmsnQ6cU8LjjZYPWTKNKU1Z/TgEjAZzuCJViSG
+          xwPErWgSV3RswBVqpUO2SwqPCy5exSl5HIPgwXgzhxgmpJHnxjDdBVbWX3S6xF/vXz3CRltV6vBh
+          jBmOufwmZy1NACcVEwd6TXkRK5QEOj44rUSC9npipZZWe4rKPgLi7XjHl1OR0jV/0Q2ul/yF8Rq+
+          bN559x6UqS4R4/22wSbfjhmQ3+LTXS5fU6UcWhj41K8hOSuTai+DUVno7/5ccpVtmv+CJ0qh9sVy
+          HXUV9+LgKFq+rJGLoGpqa1pSiQ6ZQ/b4tFVmk2EPuR7G5HXymWGziVZA7iG+8VlLBTDaz2MBNz2e
+          sPsYvGopSKyIX8f1sDbaj3SNlbcHr6Z4Ic7FeadzFz0CiGmJ9vX6psw7XHp0O1k1MUTKqKTNswSG
+          z8jFN2VSVe5o1C2S7bOOwyk11UV9XCw4HH53oqTufaChyQogvhQsvqrSRlcjFxXolrJEsv7S0FZG
+          3QwauSjJexNdQI+fqYbNhB1vNB7M0AqRGIsof8bedPoEIb3PkgCz3AnxeR6aYeHm9wLP3XUkfrCu
+          w7ownw395dvzNUh0OxpjC9lLsBE9TsZwe/kFRCYTP+bvwRjB1rtiDDPXyHCSep7NDaoNYSXVr0k4
+          3xFY+0+toYI5HYm9BZ+UH/e3VhtL/hGViSntQWbmotx/PBwMzc1e5W9RoNwpBpy0wTflT2NVIq2E
+          nhemtTCs1/rEQOdTydiuzi+wmdJVQOZSQ2xdK1llxOOtQE33IsTon3H1vdYrh0SsS8R7KmRYL3AW
+          oOrHHxxSJlH5dDEYZDxMEe/10N7Omc5A/fUxyHXJunTVP3KB8rO84gg/DvZiP/lSYIrbQvxbmg+L
+          flJGdK5LDZvJ7ugsrYtgEGsrCdexobymnVvY37SIvB/HCLDrk62h5p4Vou2ft4XWx0LFYtzJwzA6
+          e/NPQwzrT52S29Xb9j3RzgZZMb97/HmIw9X8+QJSgsLAfrZKNrcQokMJLSN+n2+Pit5nQ0Dab+KI
+          FOiCugJCHVT8vl+iv5RTOOhOXULufLFn4BwssG3Pq46+z98ba91mqYuDrBombnwhvoU5u0lP+xkq
+          V8oQsyxOdHyKIUTZM+fIRUVSxS5H39iJbEl0aTvRPtCvM3pcHup8sL5dSONvL6GJzT74+esjdat+
+          JBJP4a2fv5s02c1Ru3lIw1uLlXjIbIo2UYTd44YIVgZn2FSpy+Gez0Q/yvvnxaMDX99fNwvwerHZ
+          +PqSYCyOT6yXhWNTu/drtNdnnCTWd9ieXifBUKbZ/HVP9sAvPTVgLgWmt/rNdxjBBWZwr/8ENxem
+          2uQbn8HgMhb7DpFbRVHrFDARtW5GHfXCpQ+vJaSarRO19SJ7JS9XR7nkmyT81j+bdafWh/015onG
+          cR+b2mmwoFHLql3P6jZfc0YLSmXRSO7Z7rByqq9B033EGPOZmS53OGTgaLc8fpPvLyWdwfdQzJ5P
+          rGY6k1K7MQ14vTsfHG+tDcZ8/f7LD/zenxfza0kJN44bsXlzJJthvB7CUtk0ct4k12bQdu+hmvDN
+          /NjXs/dWJ4fV037MazF/w+2iLwzqnydKzkMs2SSzmwA0Qf3A+KBzKXlJvACuXN8RXfi9w/V7dwPQ
+          Pe7o3/ddvifIgFEiOpH8TkpZMZsO4G4v1/35XFQ65Q4Hu6i+EWznX7DJNfJB+w5lgpF1VDetQQ5U
+          rdeNRN1m2exRfErQz81lfi3zw6YXHo9gKgaXqOZRttn24Trwzz+oJZYGyoS6COf59SVY/al0HZhT
+          BDPMKbMgDApdEkk8wJQPrzN6xq39vQ2yh25G//XQ57qBtfNyDqblG2EldflhmqpRgV/jteLoWloD
+          g83lAMNzcZprEzYhNY9zA5HxsXFg3stws7QuBn9+wnPCo03PxkuHS7Yib2Wmiz3+9MGBaXfgsYH1
+          Odz++nmFogPJf7peLeoDGxDB7UekZxtWy/n4acFh8hviii9zYNLDkEDo5Xg+GY8GLPrJGkV4uzcz
+          Y0rY3rqkmOGuj72yIqHKC4/chzxqVGIaDx1skwQz8Jij80yfp7e6KPLqwV0fk2t8ZIb1TlsdUvrz
+          iDrfi91PSCN68CflT3+EpHxm5ens9cKu58Jh+dO/lp5dsDkWRsiK12MDuLoKiWd/imEta9VCRPpF
+          xCiZIGSwCDSonNLcg1/O/C8fuFP8m99W6Q3b2ZQipOKg8zaJbeiyxytwo7kh3oyMarMPSgx0aT+z
+          7eIrgIvR4EFheGZYDmaOLpVteXA4fO8kQM4PbCCTM7TXR/JcXtNAb59nA9/pExCd1pW9nB7yghjO
+          Afi96w/m8FsUxKqf2ww+bbOfar34yP8tD5LV4krpN8sy6LFj7XEg/aqLf7RKqNyvCwnuPxqO7UPI
+          0QW9MHEIC9PRBFYP9avfED/QE7WVNGmGoScfvBLxn3R5RJID6tqXcXK8kWHk38UMg8tceMddf06B
+          7fYwEfWOSC8k2utxyJwT7ZUrsUYHg9kD2/7OmqJi/Zbd1M3VPz7skfzBen1K0yUl3wYkiNOJmqwu
+          XYzzqkElX3qMrXKuyIQvEvyso4uNWr4MjCMAQyC/MsaefOGGv/uD0e0lYczbQzXFxG9FXbAh0RET
+          hZTcJxFqaqlj7anp6pyPXnSCGVcQ57AqKnPtvhvIl8XAj2Sr0zH8xT46pIbgCYv5S5fFCGtYF0WN
+          vT6d7KWSQg7eTX2YGWZiwHSnrQZ6KZTIXt/VafOxB8EYfUjmlRJYEewKwBlcS1RT7Xa/32vgdXSu
+          OHfCtz2qD2whZREe2EpymzJOh2LwOzIn7OS+RpfZGnTxAfo3Vs66VrFCpYmI77iGYP3qA37juwym
+          dt3guMReOOXj2qBOyX57/JVg2fUXNEDYeoLrHcFa3s4xKOd8dk8ttgbOT38zlMaRJd7QljYFfaZA
+          i4M6fhJfCKn56gPAJnE3M3kjD0uyIg3OygaIwr94da93s5jtzTiVLwbdPvESia+jd/UoGER7UwSl
+          QKcnV2Pbr7HKaCDY/+ud32J3r5fz7yk4wBc4f2Z/5zKcHyCcAda3FEtuDSuqtoUDcyT1//T/lhqW
+          Bvz1HnuLV1D6r35B/Xibll0PTAXr59DYlooE2n4WrbDeM9HarA3rv0m2eZleGPDKlNFdqnSk85e5
+          9ZAm8IlNoTXUtaQShD76CR5f3ySVInnZUKyQD74cO0Ndzseuh+gVv+cyqizKdeJXgZMSQCwt71rd
+          dJJp4l98SbRyKq6Lrj4aS4bZ9flc/fkTcD2a817fDZv/i+9Zutyw6fRaxSYPIEDhJdk4btaOTmp1
+          VdAHQULM4ALTRVsqDtbvfCTnPOjspdTFDZTTk8fnJyqr7VWRDHxMEhIzPFpgPuTmE0L+FRFVfEvq
+          9mTgAmkTMBh7dEzp+fR2IDwcvBkp9yacXm1v7GceBVi5eCylHBIY+Ons1Ku0bd/i3A8NNHCNdl6U
+          2ev6ZBuouCP10N/PvNC1//RLWowz7dij0cBB5jisgGMw9GtstrCpxwuOt5M5LNt7Cf54GvaehwYs
+          ykZjtPebmaZ1Mqzs0ahhLMkAy4U6gK7+0BalWf+av3nQqdvlcYxg/rPfxNj7J6NZiYauXYSJ4wBf
+          3fWoiEr7as69/+voBq+pBlkxu889fkYDVdvWEX84FvHZbF3AthVt0KUwOCxfSi/k0+8zB4UtiPjl
+          /PhqO7L3CDpnJia6/ArD8XbYIDRv0MA6MkVAltMC/3gF1vRYsdkrKRx0FNIHOfdPA+z1NwPoJFv/
+          8mf9zKcCYhwWnpB6ntr/Jm6Ehzd0iH4cS0r+eM+uD7DWvNSBUy5SAt/694ftAo0DPYDrhqKH88L3
+          0qHDmHVeezr6uYP/eNScGpYOD/qtxipimmrMOr1F5sjnWH5/Fnt0Zm0DGqwsD0TsCOhX9HV4OZfK
+          Hm9ZyjEsl8PoOlXYXp4b6Fxp5uBamcZ8ouw80K+YaLBMvhWWsrWw//gWAKYFCb7NvLp2r6wH/vXk
+          EbNIjGGLUeVBjZHO+LzeaEX9GCz/1k9rXtUw6yTSYSU1r1nUlgvYfvt2uDCuF3zbrhZYgpx//vEV
+          olf2t1qwSDV0k+2BnAvpAHpzzRTwfhsJvpsSpLR8ZgWkfDxg2QSoIkabtOCQizd8FqmeMtfLJ0N7
+          PJBXfLbCPn7/GgDYU07Ob08KmWE+iHCrxwOOlLMWLt/7UqDT9514xy8XU273h/AB2jdJ5UtLWxKa
+          EFwCLvOEwyNJZ1O6iujP34rMbKcrD086bNEQEHc/04HsfBQmdwj2+HkC8hV97c8Peae601P+j+9Q
+          u5X2/uPao9pqB5AJ/MUThYgMO6/z4c4v53hcUNrvfOMfX1Y7y0435WI84cYxI3nWtwlsVvPtQXI/
+          AGJ/Tipdrf0d5J0HYcsNqnA9nRMJKoT5kLcvyQMLCPCgvswzueQNDrcuaUcorBh67PKAIZ3LLABN
+          Zt6IvTwDStVH/4R7vZ3X8LGEU3HPGNHXagbfE1dK2XrsFVB+ASJGMcoql0fLEx0+kkbsiB3pWGtD
+          APE1xX/9Vp3qZc3g2vcKURSU2Mt2k0VwvIAn+c8/vNoF+s1YYiNLq5CdosaB9irvelrq0pGE5gFy
+          7NfC9ztUUm7nV3D5xoRY8DuFGxOKFhSTFuHQtoFKDuCxQe8Xh17+05th254P7R/f2+sh+CbSZYSm
+          Jl5nKvzWaskEJ4PpNEf//PKUXd1M0I2JwbZkXys6XPMCMmb2wGbc8WCVgjSAy+E0zJnQ3igjMsQC
+          9cQgbOm/NNzMMnagduJroirde6B/+vN2MmoiHTRQjcxCfIC/1rivl5r281kVwYN1M2xH+zveE958
+          cDWFC0mG31At37tQiPKNiERWpYCu/uBvwk+4H4kTP5T9+j8D7HqMqKGC7JXvKg7yzcUkyjE6qUtD
+          XwKsQ8+aT49Dns7pSXPgKa/lGf/tGLq8Txvsl+WIzWVtbJrZsw/zB3/2jgc9Due/ePygA8H44hoq
+          uyGpR2f3KxPXRNbAf2JXF6E1buTc5ddhHb6ghCuiwKNG4qfbw7o2qC7Kej4xdyvdEt7xYWOpv396
+          kkpN64P2dGKJpIukmqlUC5C3khzjEytU832DHBCTHs21Rq4hSapn+cfnPOZSZipVan/v79OH6Ofj
+          DMbwl/vwoN9rfHZQFK7CZo3wlwDeQyzwbAqjow9ToMwet7UD3Rj2kENSLSfyrB5cOB6ifoGrLD+9
+          Ij+t6fi7Php4Gg8B1uqITydwWBUgCpczUbVtVBeB4/o/noH1QflVZNdTYNd3HvdSTml/94XiXzzs
+          9SjlL/LBgGwSdUS/2ny49Fo6w9OTqUmQMG216nFqgL95goPYnk5pSjlYA/OOnZ237fMqCWWulWHt
+          68/VuvN9yJj5w1uXzEw3VfpkyC/CN7HrgNrj4SdIEM9TS9QhgNVmsVUOv0s24Gswx5RWJX2iJzul
+          HpymQZ2LoBDRysOMaM6NHWjZBRGky6mdBT9Uh90Pe3CtbOOf/19e9mMBipxd5696eqnbdYYH6Pyg
+          huO9n/75P9QdsTEzqJhVmlRBIWoSVxD38fTBtgpnCN+e9iE36B/S6RzYIxiyd+dtQu+ELK2iHPLN
+          2cSKAHp1tYYiQhpe2r/nb2+QnxngTfZGZGHvlwXzdaD/2x5Eb9YOrI8wO4APqXNsyxcDbLZrOOKe
+          /x7DHIJq7V5RDxUpeXrLPZzsbXHLBcrHuMT64bzZVEpPIjBwg7w/Pk8i7xzDkJMEcg71olqwKcA/
+          /TOzgWWmm8ZSCU7rpM3lK67tteUlDj6Rc5s/I/dV93mJBKW1sEgsz6u91Oz3CXbeiZ2z5tn/8iPu
+          2mqfV/rqOsRoBuDp0D9eGS5v0wpgYjY9lgI9scftZopg5//e6Td5YM30t/7Xz+dp50nVX/1CNea8
+          VRI+1eoIWIA+q1HyzoxPWt27NYOB4LyJ5rm1Tf/mZ3PpqiTo3EZdws3pYYYZhWRu0w3rFM0OkGTJ
+          Jsr++533Q0i0eP2nT2ZPUmKAs584d5z3Aqtr3xzx5JYW1k9VXNGBPC04tcYFS6OnhyszgBiYT7L8
+          q69kKLIInuX4/U+vs/m4z6ey/Qy5smrAhptcg4fkFWCDeX53vxNJ8DPwMrZYIKXzY/nG4NoGCfZE
+          c1XXik8SKMAsxPs8bJ8nvSKIJEUnrlvNoGMXZUGDd0mIdikzm85QylCc1CO2dv2y83gJvsd8w9L2
+          Y8PFV58KVJABsJld+n/zzv8BAAD//6RdS5tzTNf9QQZxiqoaCiLiVIKIzJBEECGoQv3679L3M3xn
+          37Cv7nRQ+7DW2tvegJ6toy9bxbuW/vhqvf++qTUk0/YGyG+EzbvNKe4LoFPzV1qwettX7PkvU//T
+          UxBdjhE1v9Nb79X3xUCHh8vjKDmVYIsHNvzTh7WTfwVsKd0VnmolIj/0boZpcHEMNz12mznzBsur
+          0n24hMcL1lSh3fCsDOE3yFZq5xWft6bWpzBjB4EI2cHJ10aBLfjT7yy7CcGcYnmFTM6qrT7cR9S5
+          OfGfnvZPj//D+6gr4z2NTxV0Nj7V/eP7G77dtk7eVbjxYXo8mms+7OeVQ7fDd6EGz601g7VggiDc
+          HX0+li7OzH1nFR53lkT/8DfZ6uvIHelIUDjXbHktZ/X/M6NA/t8dBWotVdSUHLNmQfRpYW0KrT/z
+          +71O1E3hmRPuQK2ljfKfLF9EBPRsxMdzugPjSRNF9JqxSc3aVB2+vLkB/HjOSH1hImw2hSMP7ddR
+          9QWEbrUQfMtMkWv/QsYk2LEmJEuAbswwqLoiA/BmvczQruqAKJ+rHy1LLxSw299KfA7OL2fR36iF
+          J0s90biFWU7Ck8zDyxtcfXEHGmfyYjEBt2CnYg1OyjCzxUxAMT8Mai2vSWdGctPQOTk+sbVM+6En
+          WuyDyCtiqv4qRV++MgkhB5ueeqDLAYstSYTs1ApUH3O/ZofHtrd7Z9XYBEYZTUcwKVAZudZfDnEN
+          lotwj0G7+9g4qGnliOFpFqELow9VNbusF1PWVmRwxPa/9r2PlkqoDRjzLo+t6mJHS2dMBELJuVKn
+          fp8i9n18tndCxoJGZ+uYC0OmyNAy8cN/9+MjkhpJeMJng7etEZQbportK3jiCtdf+vGRS6ef8kOZ
+          OBLymR9wYNOT45XwVnT4EftuLer8Y5MIQons0F135rIlM3jt1hG7gfcZpJGzfPQ7yD8c0PsDrO3O
+          jNGrHF7YF9eXs9jD8IR7+tMIS5NOXznRisFpXXx6V0GRi3V8hnDZ6QM1X9XdEawLsNHhyLU+AOkh
+          5z11imGjTwvFl4c2TOfXaMBE7Vp/F1Yhk0TtYUJeUSOcKpaRi9HHDaA4KjYBJjvX671aZ/gMlifN
+          3cgAAvWjFkEvUvGxYSoQ7v57BT1dPvR6HY2c/YQagjKrGfUOas+WHS0D9LjdKqqiXqyJfJ4g3NfI
+          wDnjlXwpJ5OH6lrd6FnlE0eIvcCG8SO5UvzJKzA74u2JDkfY4miseMbyXadCtmgLNdjtoy/Bt0zR
+          UbZMGokaBYsXuSv0+DTBN+BOg/A1byGM2ZzTR6ZnOf8tVxu16TjSQn7KYNmpMFUc4xdjW3qWNWs1
+          PYHDOwlpImt2vdISqIA5rwcB0sTry+HKufBl4zu2B/3KmLiuFSLez6Bn7uzkzfpSQuX6RApZrvdf
+          Ps73U4Okn0+wa092xG/+gjSZf9Cj9nzXglGUMpL0OPDlu7ir5zw+GOh7bSVqqYzqS3/QWjQK75i+
+          vAHrUv9lKjotSkGk+9d1pPv3HYBj1778ng2tMwv7XwML2OT4tNZizoQ61eCxf12ovfm7RL+5AcGH
+          O+NAu5hAGvQCQoCmO850lTq01OoQWceNAe++ui6+nQ9Br/OiUqyb+rC8v6uFHpH6xviTa4Bv3KRC
+          XG9lWLVQOkjT+F3hn/+B5vLRV/A8tOjzomey352EaPSQbcM7oo4/PodfzdxnFKNHq92oN0M7l8ht
+          bhHFUU92y5GrmXG0t1cIiwq7QewAfj/cCQwaO6VeNV6GmXsqKbif6MtnD29gbNcnGewc26bHUtfz
+          Fe2HAt5EbvxnzyOVAw5KC9rhc2T4Ay9ejQqtj/CE4/SGHSoFvgZF3tOoa/LzIKX3OUE+mXTsj17E
+          BHIRCcRxeMNH+VAP0452Idy9bBNfrnc7X+vvqYWnXKH4+NlJ+aIEHK9s8ROrl9d52O6vQZoPLZpn
+          8uwQ5YNF+Cr7lw+PYjss30/XwUdbzFinDyNiWXwL4T2ER3z8yBdHEB6JDZP0Lvq7F3/ImZvJPqT4
+          0lMH4k++qEoc/PPvRoi2d3LJPUFxuX6woVc2IMzY+xDGX5maa4Uj6cUsAvcB8jd/OupUfeU+xMPH
+          p/FZM3XxXikzaNmLUf+4Uxjd9c9MiYKng4+IOjo75akFke1sd3J6M2opVws1sjhgF9y7/IekRUGd
+          euWofZrdWtL3TrX/wj3BRl4ODq8/Tua/+OpoCmZrvOgd3Bk3BZ+tVx3NGk4sRLq7g8+jVLDVtYIV
+          fuVmosciFZ21voIKdu31h19+v9dnT/0kqCp2F6pJ3i5nTvDbhneXF2qZXcG+Ma+2SrVyHOl10WHL
+          kK0yuE4/Sk0rToblL/4PH1fB2j1b8pkjhQzb0QL0PvdBvt5OgYLe1a/B5jPRBl5/YAPWHfyRhddh
+          tHrf64zM9laQXeKpjph7kgtyCSi+sF0vKcfyB6fivlL90DxzMVP9AM4oMwmUnykQC2VfwXflO9S3
+          RW5Yb6dUhrLxLnDGRk/nueeaomt+CrG2XIZ8Ua1jiCifNdRu3dCZTwGdgVy7F3zLPoneTZQfkf9L
+          NWwYd58JmiY3yCi+Jtl7WGUjTpQKyqqfE36zDzEBlwIdOq7AqZC2jD8dvACq7H7G9y2+UG8vuChU
+          tCM117x35uCqQvT0rj6Ofq068DyILWhGt/Yffpimdd/9xWec9OsCGCumGOz2e4/ar7KMRJEYGlCm
+          bEdxS14Df4177S/fUXXL78wbjRDtWTnQdIuP5OtKCfzkqUWf73elz5e3mqA4PM54wxcOay8fHrxf
+          bowLNw8dySldZU+FS4ITxHHDvPvtZ6SYhkXTkdJ6/Ds/kIiASL0OoqZ6lRX4XZM9Ps99EDHrLBF4
+          nG57Aq5K61DnLRZIqJWORlu8+R5y2UQrGz364qVQn9/8bUSvVPGpu8W3+TZrMhqdbusYEYN8vU5d
+          CQL/a/zzx7korQJt+ZnA8vsEc1HXDdBk8eEr2t4C45u/EdSeP08frtxjYH17fsJ0B32sXlM9X/J9
+          a8OP+Omod+4PuiD2jxJeT4OOPdABsBSnlwV/h+JCTenO9LUkmQv3zAjxlXSkns7Ub+Dgrzw2D50B
+          eGCyDjrc7uDzsqTmy+dhpFBufyU9nx9TtBT8vYCBN4b0OD+Keq3kUoPzLwuoLhhHJqShqwByvV+p
+          K1ZN3QHpbiIzurb/8q+IbKmB+msEhIVdGa0x387o9Rg6qubexRE+9z4Af/HM/lxJPt1d2VdOAV6w
+          SsPb8DvPFwjJVSLYinc56Iu2LBGnWwRr6k0F09/1XpP5gm/H6lNPjNfDrbOjwupqHhxxf50SSHkR
+          +UtQtQO7HpYWJFxe0SO7fABbURmif/iDnwe2GvMiK1G8uPjIWz1jmUli+Id3zTXeevRDzYfx570j
+          ILHCYTmY+agQLX9RFyfvet3iH7Rf7zs+KzcJLFoYmbDy8h/Z8VLoLOz4SoGzSj/q8MMzWvfXTwJf
+          76P8L38TZlYrPF/WB+EG3Nez7o0tUJ6Xhb5yQwfz5xgEwBhITDVlMSMpiKYGfnLxTUrmXHJebZ0A
+          6vXZIaB+n3KC6aqCB24vZPamxJl9EDTIOd/J399vXT5KCKsCXXxlsKEzqs2dwMGWXfKN+8pZ4aBU
+          8KHvI3x4aodhzG6jD5R7ecM3KYoAP9ZCDMbFj8hODbNoehynFm74HxdVY0TrGzoWZM9XRPU96IeP
+          LMQKEDz1TE/O8z7M1asr4dcnKdZD2tYEH3wLinYy0fOjV/T57346/HKor+ommzc8obhdiuhRG31H
+          nOdrCXRHjKie4jdYz/Mdwi0f0Q3fDTMZLw0UzeeRWu+ojIQnBinkh9uToExXIloVUQPv3LkhVO4/
+          A2sxDGF6r87UfOUhWNKIe4K7nj5p+o7UXBQqt/vDl/iGNbVm+7QKgRCzbMt/Vi7Ez9EFNYAldjb8
+          IX7dXSL/UmZR4zPXbLueFolvz/3Pf87Ub5UND1KDfpZhuQdRolye4Il9r34Dwp2KAm7xAfvp4eus
+          NulSwOY2pZbYPKJ5Ng8c2PIRjd+qmfOiPj9Ryk/91tGVRDOjxQpTJ5ro8VQ4wzTb8hOkc9rR/JMI
+          +UcL+gpu+JIa+d1g62/9qhD5B4bvxUMfpHnpFGh1yZuMm3+zUhsCuMTBh95bo2Gbvc5wXNwIR9v5
+          LYJrmxA9vhnV//y9nJ0VGQxwJNbzpl4RuGrQJq5PU263zWTqVR/Qd0oIyOTA6VD1KAGsJ9Pnheie
+          L6pcjTCQdZWMJhABS7yZg2YIPLL348WZ0kh8guUb5KTt96nDzoL1hKonCfi0PFudFd5dBdQuMdX3
+          vJ6L8+tQgGY/PcgUXEO2GBPkIDw4PZHivtLH6tVV8OaVM9aPyQWsmCoaOBH5g/X7tpfRzZP0jy+Q
+          X+gv7M9ewO8a7zf/eNdTvic2fNcG8r/f09dZ2m/oK+kdVgQ6A5+PHwtkwK++I8Y8J0eU9EaHMLAF
+          H72/7/yPD0LVywt/34ZLviRyUQLzpTVUdXZ+TkrEE0VK1tf2M8lHCngTSUls0rvRnmvq0m8Gxmz2
+          aJouOJpryCB8V12zTdXtnP4UjQQtx++MHan1GO/5SQDBL7KxtthHfXW8zxMu3zD30eBrtWT1BxNV
+          w8+khq8GjpjD2oBMuK64uBp9tPrV2YQbvvD3xaMe5se0T0F0iBHGYT4OVMXyuN/iN/VSSQfCW9mt
+          4CN+O3/pPzeHRb0yQ+Ohv3xJ9PVaHKYwgfvzLyFMbeN8Gts1gL1VEJrlAtOZ0kYuDIdbgE/NccmH
+          4TBkAKf2Dm/6CGAvfg0QFIMaa6tF9Zmb/Sfc2kboafv/ws87cwDcopjqQ2MPksQtMjQF9YJzvkn1
+          tbo/OyAEyopt9jHZ/E9POXcWNUPtqdN9lotw48NUTZIpWnOu6KBQF5gmDSDOMKyqiHprf8Ludt7s
+          8hohiMPT7Cs9d6wn6aNksDcaE7/K1Rm6eHvnVy2MC974Tb1yoro1BSU5dvpDNYybP8OdawVEOfVm
+          JLxyZQX6wqtYpXfENnxQwFMuU6qH1KyFcb+ToRk+eV859W30x8eg2wkJtb6Wr69q8QzgTrm12Pff
+          DhMBfXQw/MkSjn/fb81EdahgaN0yarrenLM6nQsEEL1T03OyiAUXXoWT8avwqRMIm6dn18GX7d2p
+          tiagpoZ+yaC5v4W+tIODMyTeDBF88Q+svaGW/8Mrvzt+42NrPoYVFqmmcFfi+FP1GaJZzx8l3OwX
+          F+pOA/y7dn5ALcwL9aPXFXRl264QiiPEuidXgMaCYkEaTnts6LlRi/UXt+DC7x2svpBfL9J3gHCs
+          6ic+v72DsyAeQrj5Gz6Yzqle7u7so0f7nP3vVUgHxnfjCA9jNhB+9nu9v/vveasO2VQH6p0tw/0c
+          /8enylBkG18L0K3JMfX6y0MfIy5aEcgdSORI/9YLzt8q2vAA/dN3GGffUwXdgzvOslkDa32bTBgF
+          hfMPj61NPT6BTsQzfjwEjy35z+mA7353hMVNXK/wp3Lw3MwRYdi45xseU+Ffvim+ueWIC2cWaNMb
+          KSZOmdOwOPAwaKwUh8vFieaJvXjofrkGq5mlMla5i4FW7QHoyb6f88nZFxkkvvGgtlOVjihUxk9p
+          x0+MN/t31v3tqUDlzs34fG6iuns5rqxseIaAA/wwFmZBiLL4vPGRsYjWSeosNMjDGXsp93PWcX8n
+          YLtfam/6A8t0xwXeM+uJoq3VIIQJ5pTydGq2/CXm7B9f3bY62Nk2UCrfE+sPL5HdexaGJQ1dWSlm
+          8MAqv7/royq9EoC/oYe9J5MYGw5DCtSVm7FJeqGe7VtWwO18qP4Ra2eczx8TPqKnQnFUFmB+nd4z
+          +nueyZVfhk1vymAmBleag/QdLcCpLLT3djcCq2bb8ngTOPgsVB2H9SGs+W63s2Cn3jisPrSdPr38
+          gwi5eT/4LMUHsPa1UcI/fpqUYQL+6cO7qk2I8Mev1bexogh4FxwdmmfEHvcgRalzmbAh2khfR+CE
+          8K2eThS/+ENE5eKY7n9fJdm6RCAY8l2pobwgGcZ7cmXTnx6Gyu+FWvNbd4S4UDq4nSd1zLGOluv4
+          sCBWAt1frx0Cy/mxVqiYHp4vlHY//Da8D34jvlEnW85A+OQZD0l/QX77RTWYbwt8/t0fVvP4ljPZ
+          rRTIkG3Rkx3qDmu3KfQ/EH/oQeQc1m38Hk7WrsbeeT5FM0gxAZu/UfwC32Ha60cFHntNpof11DnD
+          Y3xq4O9+/P2rihac99o/vVa0YrFmj9/4hL54AGS+vM41y8NpBJEU9/jW63m+at21UyTkNjSwkFwv
+          sZdaYP6lwT/+TvGR/8GBTyx85BMxIkIwN8rMhTLFxX5hNPFkDuqBc6V/8Wa2O1JCLaF3sufzmol9
+          wxmAOUCkhuVU0XQdHzaYc5rjf8/7ItwTQGIObvz+ORC+Fn3YfNLiH75anmqgwFmHDj2tReCw9D7H
+          UEsidZuB8HKWNBILpU/VK9aijx4tnzzkFS3MUn/EDXGodd4RuHSOh90jLfM2jdJAukxthP1iMQbW
+          5aAD3ieyqDWKAvuHJ7AqWthIqoPObG1fgEXlE5q88hWsk8UH0Bd1QDhx3TmjYicrvDr1gFV5GOv1
+          lOYuHJLjiF10Leoivyod8B4woIE8jAO9EVrCjZ+Q4iiawyQO2FBuB2mlqmHGTBhGH4Jufy3Jth0e
+          rKc08qG3d2bqfetrNIcHDkLrPIz0IErHYa6po8I6uwh0i9+6MD84E15Otk/PN7DmY5uM8V88xCFm
+          lNFSXQgaXycfu3XP14sNdxU8+4nnS1hTB2EyvBV6n4vlX0/XY77Gi9OB1JbijR+awzZJYNuqeSmp
+          0UIlos9OkeE0GF/89LtQZ/dvH/zjB8dkiWu26UvQJ1T397o4sEVuLF7R0FfyFXTU9PX+PvsQtGlF
+          lEQ9RvOQGNk/vuleHTp0OFFKZQGfAh+6ezl0x88Qw4pWBLubPrCC57mB4APP5GO8oU42PgH9nHrU
+          3mcxY7e8V5QtP1Cr8Z4Dmx5eIG/49F/8nM4PpQSh8Vw3+9UYz5FCgeLnyGPTScaINGGa7r9yO2H9
+          bH3yUek766/+QrXThQMkfvcWLPduh20Cab3p4T7E38Cjr0/V63NsRy7Uo0+Dj54+6dPPO0B4vLa9
+          X4m7cKvorQHa6mGERmXB5j89VyxXRr6HfgDsLjnz399T/039YeOfI5D6+YpNvNPrObMOI+yEp4zt
+          nvvUG/7O0ODHOT3qP1mn7xdH/vIx2f1CcxCOexhD0pg2tnO4zze86MOZC2T8EMYveG/6AhA0jHw6
+          +tk2ld9d4fubc9hu3dUZr+Zs/otX6uXV1yT3JB9eno+jvzuhUy5Vz6MF8e4lUl/ZttKJulxAz64O
+          +GxMlC3b///TwwgcQR2t9/fBRRJcHaIUDOXz+dyX8LnKFQ2Huh4WHn984B6qFzX4UwVmC/0MAEIY
+          0YN2j53hiNMVHi/k4Etbfp7NpRph0nbhX30rZ8/vOf7LH9idtH5Yzg+lgl7JnegBfryItKiYwaxz
+          DmE76VHPiXBr/vGd2eOL/E+PgrfS8LGbaSz697yWeOR8gbyt/BeeZh6iMcRUzWMpn7u56MBWD8NP
+          ixk5zxY/hkp2P9ND8szqhVw4AjN10vDzvJBo05MyOHyGCFtmBwHJufgH7yhY8Um+cA6THxwH4OHc
+          4wNeO7Aoz+8Mt/xDpA1fUm+PXOif+A/NQv/CGPL48i8++mVNNWe8JmGDeutJqHMzi3pxT8cMydHv
+          Rw/9MwBMdn8y2PyF8D8ysT8+hMbTevvT52veAnMFi8RK8KYv69P7q1jQv6XuptdyEVHeI4HVCrlN
+          n9T0rY+0BVs9FfvfBuYEE7WARvmi2E93iz6r0i1Gyixh+qfXbfmpgKK9mARoCgVzr/wKMOeR5svL
+          y3OEDU+CxEBfbFt3w1mFMjLgIPdnfJR300C7NCiRYgsVdWEwRqNTX3jk7c8zfQT1K2cjLmKw6T0Y
+          H89Ph9BLWsL7j16xdrjabNnvqlH58y/Lj37O+inPMVh5uyUoPrVbPUAliINtT3Vu9SKW4F+5jXqo
+          yGBRBOiuT1JofvY63vK5vgbesYLPoihoukz3Yfrj58t35PHtXn7Ad6tvgtMiF/SUjTCfwkOw/tNn
+          znP107t39LaQRGrq725hEC3seEsht62Hn7d66OIeI/Evv1ILGmk9JV+r+DsfwiKAnPXKwRbMq/P2
+          lX3Gg60elMGdcm19adO/13LW5z8+hgPYmNHS/oQE6dW4bvy4H+aODAbsicKw42aDPnMPlYdb/Yso
+          RDjn43x7alDS157qaWI5o7vYK7TxbvWLPkgc6fuYXBhJSb/V50tnbuyRAz3zBKqtWMjHl2PI/5+O
+          gv3/7iiIJFmjWate9AkgHEPhAkVfjPfUYXzyqqCFPz12zMc3msfolkC3/2n4+F2+0Xo1eRHJT8km
+          gjwdgDiKQATSu7bo6VKt9SKNjwxWrLtSIwhPkXirfxr0GNqm5EmGvii6p8JLqu4pvv6+zryoYgmC
+          2phw8A0MNul6kkLe91yc1tEwLL/yYqPKEGVqempRr9qwQFRfxiN+euV3WLj1VYIYGm/8QMoDMO/l
+          ZzDdprz65RFEdG8KFYq6RMZuu6PROO4tTqnFa0YPp0Bks6E2T6hro+fzj+kzLBXqZ6U3onxTeA76
+          XLN2RZgoHlmzGrD+fgM/GBTb1N3TFA6CcJ14+Ek/PU53A+8s6+MpgkhSNCKhuK+Xt1r+kNHAwhfC
+          fmGLZj584AZeh4/DQYnml1GE0Eu4I3VUuvUAJ4mFzJM70PMLUmeNwtlAMf9MfQnF51rM0tVX0Ck7
+          E2ZeV0AqrkyQyHMrPUxl5Ug/2LRojiKL6qXcstkrYwjZV99jp+FVnReXuECRxQiJ22E3TE5vlKhW
+          hzfGrMlz4fRMfQTTSMDa8nvX44qkDjWPaiKFpuCBd0+rBsOxwtQ7Hnf6Ys2ohInlH2i0AH8QX+Yl
+          hcd+q3gl5yyXVL7f9qqtDrbqlwkWo7ZTsFd8RM/H59npxugWw/nNCPWG9zHiNevxhH2/izf7O+X8
+          T1VUyMSvTo1aF/U1Xx4WnFfjQ8PoZehj4R5DBPv4jr3M4we2fi4N2M6DPjhz75DDr63gSJFLgzC2
+          6/n7s1Oo0ymlj/5bO3S3PHh4XpUHdjCTwKqctBSZUG5o0HGuw5+A4sJpakzq1d0KZuV4DBHR0Bn7
+          x51Ti7T3tj3J0oPafZfVq/ESRwiazKPOI2r0dcJPEfLPUsevav9z+F6fRdiaw52qY3sdpDQfbYgF
+          FdEwV0uHj9ttrjKuHHwojzqbJ1GzEbf/pfRmnOJ8reWKQJgZpr8jGq5ne7FTcNNDhA/nXMuZokoK
+          coPdldrXczOw72nrSYzUrYNCMSPqDIcOdZw94MN38nRRI/OK3PvNoe4jvTJxdxFHiF5aiHXXOoH5
+          +q1KJCfb3srlcqvnIP+tgJ+0nhrWMkTrTipjVBrVHp9t/j7wjbTX9uaJS7GdSc7ALtdKRaaoWthq
+          hNvAJ7qbQKo2OjWUo68L3HorgQXPO5zatTtIwq2CCHhc64vJbhym1kpsxMSP/heP8i4xdRsVqXTy
+          12fXD2t902P44tIzxQ3OwLqdNwri9oTV9yUGUtHfQhAOa0Vd991H8/frtNC1+Q+2GY71WVziJ3Tp
+          Q6Gne350+Pf7UKDHNX9QzTSOgNe7QwHH0I99cYs/Is+dU2h3TUI1p12GOb+eCawv5Ej6/jbV6+zI
+          Jsw47+CLVTxFq0bkGbqHSSMAgtfATGXoILfvUnq2+X1Nq4R7wrMHA6qr7OswRd0p8Hcunzg5DDmj
+          sfnloOZOBXVWvXbEYTeHSCrsoz/rNzXi89BW4ONDYnxQ6NGZiT/Z8Fz/bFzU2xhr72Vm8CI1Or7O
+          PM2X037h4K6FOTZWNx8khC4G6vfel97FdWD0ZloavBzsBz6cggTQIE4aeBGNHlvmk2N90vUmbOr+
+          QXZaEYF1yDMT9hH/IeynFfrcLCxD8+BaVHVbfVi4XFVRfO0Atrk5cKQ8gAn67dybT4Rt7Qmd7AZm
+          lOT0tPMdJko+y9A1mCT/N01dxIoxMWHVpBoN/PGgC0h+2GCIhIkG/OkyLHxeF0i5ZBfsHas6Wpvx
+          W0CRGFcatWvvTPd9pKJ+KAj2lJ8NFk54JrAu/cGPxnvF1tzd1pKbP3M77w8QftwywoGlElXF+y9f
+          pTlcQXO1JazJSlxvr1TwSJpZjp3d8xAJzugkcPWeN39/DD0wf39aBp/zocZBGP+Gca3fCeqEg0XN
+          7rS9E1AOPkDHr0y1QRpr9ubyApaLCH34SAW2WNdohZJ8b+hR+RqO4GkrRONhGw/KNTHg5/MKUabK
+          OxrzdZWzoQh+aEF7D0fjXQMi1FIX8sdTj093fGYsp7GM7t7TpXrFGbkU3B4ZqM/MwMeTdWXrh8kl
+          DPa4oAbiykHSTr4M7zk+EiHsL0BAIOLQll+pVlSuw4drFsN7DlecsJugM+sWyEhwRBM/oWI5wrF0
+          WggvY0BDmezzlYtRCT9SbmLPbXN9vk6dhdydbuBTxDuRsD4SHo0ad/DR0/w6JHndVsCVHoetkfyc
+          1bZfBvztOAMnQlXWEnrOJZo9K8UvotGaTlHsotn5cWS289RZjRdHwHg3ZqzvkhFM+7PWSebJH2hq
+          9b98TEzHQtWoOj50r74jSQ4l4PiuVno8R+9Iet9RADNBcjDe7Ft8jD8OWi+b+SAaf/k83d4tel2n
+          gt6ibtYXXX21UEEDpTbLcbTZ4xPNyjMnMlouYE4LZYTyubD8HZZChxmlECrepxjwPS3f+hRd9iWK
+          Ls+CKOH6YMz2iwLU9InoQS9WvXs/7Z8ycq8VH1YxHNY/vOV59c/f906Yr9JlddFL1CL6MnJDZ2qk
+          2eiQXER6/wrveh6zLkXnVX7QV4MVNgZpDZF9PrjY9govmqv9mMKRe6z40OO7swScNSJb+FZEHpVP
+          9J2CRwrP+XSmG37KR707PNHDWyJ/L1vFsPqvPAD9Hn8J096+vuTTbQXX3TuiSaPx9ZpmUwuehx3D
+          +HLnnKV/HUe0fR+1p31drwV9z4h9D3v8yt2QrW7jFVAoX4UvFEaQL1dNeipQc0zColQZ5rdWyugk
+          U4/6GfgwtiM3FTbDpd0mUK3RMl37GOYu/dLtedW8wLAIaELWbXddMvzZB7rd7yNNDtLNEefTxMMN
+          D/3dv85CXmuQK6WAXjf7l+Kt4qv9mItNqdj86/4l4JtbFLu8v9ajo39nKOJQxp5jYGfN+BzC88+j
+          Pv/6ntiM1bMKrJdSYr/e7wBBGbMRiKVtK04D8lkijx9sEy7A7o76+RiuWQJvvyLwrxu+5qGam3/f
+          T7XLva1XK8li5ccVP+zjrGdrLf9GMEH+7ovRrdZ7ElYV3PIpPR61NGJHtY/htU8bagyWzgRT+I7w
+          Z4cOVm9iki/+O6/gOngBYTd3l0+m3JbAGXWLnvvaYmxuTHvraDvSo16O+ehzqJLO0S0iQl1y0T97
+          pWYOqGP21bD+5cc8sTQa6tNOn6+uNUOvWBD23DXJV3ZQg3/5zaa7g06/HROB0Ekh4YpLUi/+Dhpw
+          rviQ2jdc6XOeqhUUk/MVnzd7ppl8TyCncy+qd7Hg/Pk3cKUM+OXH/uakGekTbHgR2+euG5ay5BuI
+          reJL9fvNBiwe7y384wvx8fxm62deCHTidPef/wJ0SoAV/Hyq21m55SethaG7O2K/Kt6gLdauVaax
+          rqi6zf9YVPppYNPcbZoBq62p474NeNi2iOF7GAyLd9yvkB9D+OdP+WjjQYVh/hr+8G6+RCxQID4R
+          hboCvNcCUw/kX/xVyl85rL6+lFAnmCNsi9/zLnZE5cFeHmG7xAXz1VVXeDh0b5o1JgGMDDsL0tO5
+          w6cNL1LfkE1o4vaNVRsXQEhl3ADzccvomZfPtdBVbgX+8NNLEppoHa/FE5bZ0Pmj0SVseatdB9c4
+          4Yn8GjRH2In9DP3uPOO/65+fwxLAk5Pb+LwYDWA3VUvh62Cc6MHTB7buEouDUOJ39KQt+iDItqhB
+          SRokcg+nUaeFJ1TonO0K6v6+Mlh6RAL4UuOMPt5krue31inwa4TR9vl6IGh3iVE7GTt8M++Tw8R4
+          /Id3qGrnqT4/7bsMBb0IaJGLZ7A9Lx66WkFxdKXvaP5Vlwz98ZMIBTSaH+bph5xq/8GHMz45a7GW
+          LbRu1Y8aNV6i2bo8M+g9G5+ejJfN5i0eQzTqT18QyVJPB3zv/vimDyDYDUsAzgp8dvnd51tJYZ0q
+          yAasSy4l3MZXiWfKCfwd7A77Q7vtuQywBbZ45ctJUDOa3eUVXqy1JSKtg2jZn7UfZG/3QC+HCrOJ
+          E5IESL3CYX9atEG473MNfvfQJeW6lGDVjZgDvl7l+OgCm0mXa6WBa1hZ1BcO13pZ2iiEW/6j9/Ad
+          MKYGTgL2+nKj6rconQUawRPKn+mL/dvprc8ngW+Q571/f/hYX4lgFVsjTYDtvdbrPTxcfsjPMoni
+          krnDunzvIrx1/W77eax/bnN8orBORxp1saAve/dcQG6uBAKM+qjTg2eU8LeDBr7w6XMojy9PQ7I1
+          5lhLlaoWR5Hx8I+vnE9NO2zP+wcESokvdSmNGHrOFYKSuKNYEpp8FpRUAZzur9g+gCWa29nSIHeZ
+          Unr8SEdHPDbblpYlrHz4uVg5G5Cswevh1VAtVbRB0PVnBlf3zrB3TDpnBs0S/MVL8j6QQZ83fvOH
+          76h6nVPGrIW6cPAIor520/SZ4rmFt27YYTO66c74tQYNyuLs4KxKvwP7w49/+U6D1dOZOr3m4Mqs
+          C1YBVdgKOE9WDL8S6KkcUme9K/dWeaFfgT0e1myKpjwEd2XFPqdQODD2iiCS9Efq7/b19g4m4Vso
+          RVpFT818ZAxzSgZeuBlpSF9bxxaHSrCdJzZljoB5RbtOfBDDxpveoFPrlirwfasEn6uKA5unuFrB
+          5VxF+C/+E78KTfQSwhYfKucQSaO0Zuhhnm0/Sd9yvQb8LoOT4AdYBfWuXsrdwwZiW7ikfN3WaOWB
+          1qD3yePoybxP+gSrRUPPaDUoNo1H1O/YDwKhE0KceCnT2bvgjX987g/fz1x2eKLxAB705LqVw/hH
+          2aD70Uho8HceF+mcKc7qyNvMhiOT7BGpf3iCHj8vv/7tPVWDq5szcg3qOKKCqJdogsZIWs+VWF9E
+          24yU6nTDVm+Ww/KuVBUYQJT9MdSvDrnszx2Uohlj67JTAJNCef2Hj4xOHPRptpcSnN2n6KNZEp2x
+          WUAKyqrp/VZH8VAZL5EABzQhvjjnPhpFKTTBLQMm9tVLM6x6hUrQWcWKcTURtq7PRwbHrLFxsrd/
+          zhxGywzlG0eom//IMCnbjDMMe5PM6+dYC3/6meqO8T89RYzdtUCh8cVU1Z2PToSkydDx/FnwZl/6
+          XM9fCE/z9UbVLX/OQV7N8PZ7Bn4DmOAwLl0VVOjFEftt0A7ML3AD0bg4//S6VZ4iCJPx1Pv6zd1F
+          TLeyWPnTZ/xtJDK7NV0LbxV/wgeaGbmwGExGz1mvqXVhu3qcu7xSbLRLsXuHp2FOudBAFtyP2Bif
+          /PCVh2j8d/02j0A+enW+QsxHAXW6rgG8c56yP3yNr/q46YFCkkDw262b/jDUz+979wPl/nX84+v1
+          UiyVhu5HM8FmFXv5hi9KtOUb/Mi8eKBge4Nj0zcI6r+6I6A734BNH6TRzX3lQmf+VvAt3+vGzxpn
+          vdyqEG75w/88uLCeVDo1f3ycgFvMwHwY2Q8Yx3tIdeFWsGW6vhPwx8/NdcfVUxGdC6Aq2o06l9XS
+          F11/pv/00w2fRXM0FgQsQn2j+G4K+RykA4SQm/KN71VgUZoxg1g1l00fS5xRP3cymo25IrvbbNQb
+          nyXwdfNF7B0e1//0iz/+lPwOxsDO69ptQyQ67BePC2OPqS9h6I0rTbd8PyPl8Z8//9k3MXO3gpmt
+          VvjC3fyBNlfdQKGLjtSIzVFfv2/p90/fSdIpAquvKgQcpf7t55u+NjXCvKINf/h8nY/15PRuCa+H
+          R0PPwdcfliLQOrTFW/99zrVIfLZTDKdhEf7prVRz7jYsf+6Kz/B+HMj+myQQW88v1STXy4V5TGV4
+          5Vjri7+DUS+icBchFUqM3d83ZWvXWwQez9+FnpOrUE/c+qrg/fdIaTon0TCL13UE69V8E74dXjX7
+          +74hkrYpzmRwtuv7QTptM6Fu7isaLf/WwvrL8/jxqYRodsfUB5+vAClu31G9XLcZi2+SX+hZj1A9
+          yTanAUysErtEfoPpc8lapdBIStUcGLq0IzcN6o6p45OrtqDLgl0BrFv5w2aZOQ71i1Pzx3f+8l3E
+          ELqbALfljT6Ox5ezqqHeQGEpVXzMqibiC/cYgOJ7m6gTjb+I6f1+BKwzHjh93/t8SarhCc/7dU8W
+          vrmDTV/nIAuAjR2j/jjDtc05UCnV4oOWaPlCwl8FFwsd/viGs+k3LahLd8DqdZYZwffHD1zfa05N
+          z70x8lYSE6xH3sdp3vI5GeBSwK8RRNsbIUP9x+fADi/bjL081Vcp3m99TJ+Tz+3U47Ba97sFd5/H
+          gd6Hh+OsUAtcOI3vinrrkuZ06YIMyVwa+/JrqHQm4z0HinvD+2A0mDPWjMwwmKWWaucRR8RX11Hp
+          HUciTe8f6mXTk6DIw5Vu+KPmG15xoT5rAd6uH6yyw7egPuT+tjfb1ufvdHb/8jvVrodAJ6erKcPb
+          Ek/YOlgkWnT11sK1kjJ6jDtN59e2sxTYtSP2nzWO1t8NizA2047MnHmohZHufgCAsPf3nlPmgvL+
+          tJCPr5AeK13NyZ+eTuxviA/weGVdfQ+e6E+v3vAHIKWeNfBAxwC/9rcgmj2DuYpIzCs9la+CzV+n
+          F8H52Gb+rskTfZ7PCoTfqGp9edNj/+oT//QvXz32OUH7iKD3Sx6opk8vfcnYI4X6xWXY+1TXfNMX
+          LdDw4wFrF2/IF6VpUlg4po8Nvq4iNn9dDroHqlG7fB7zRTvHKsybrqPBxjfnnTD4sDX7OzZOeuMs
+          qoMzGN03vlc+PxHhxEKBgQPu+BDeYUQqrkvAHz66L4bBno7/TmACbpCqokf16U//e8T9/FcviPju
+          2PjQ8iX4j/+uBe1n0P+mA7bTT8imoHjHEKYXAXtP86u33ieQ0RbfyG60vIgvpuYHpWjFPgNV57C1
+          Cle41UP86XDJ8nX5pR3sEpTSA/V7Z7BHQYVF8Lrg03x5s7XnGhEJjnn3l15HrFs+BYFhNnHUHbeZ
+          mHXsliD9aq0vvz/NwH5wbGHxvU54yy/RihVBg//iw+sZRey3GzJYuo6Gt3qUs9TByUZS71w2PvJj
+          hBNjBZ7s4eEL8G0O/Me7umBRvw71cXZmy1IXBYgo/tOHP/qmjySAHX7Sf/W0Ya90kJ6cjizO1ag/
+          Hz+fwZ8+fX8/gTNXuc7D5Ts/sS7cIGCPuYvh3lUUAvavSzR8XPP3d56+xPaOPk6XAQLdThJq9P67
+          XrI5aZQ/vC91buyQm2b/4N51PH/Pl0wfK3FqwF5xEeG2eh3Z9FtwEJuJ/uFHUj5hDO45t2IdBTha
+          YdFWcPV9w5c2/re8jhkP12L80dPr7Dr8y4gDqM0A4hM4c87SmdWMdjdiUd3QM2c83QoNDvEnw/Fj
+          +tRz0gzjn7348l/9LqnqAoZKZGHbUQ65NPGsgYWdvH3eHrxBmoJHBqpj+/HbylqddSd1MXxdaUG1
+          E2c55M/ehka5+OHiPXLmqR8edcIq0BPbD/pcpZ4Gw+NWA7Fz2fnL/3/5eNMb1HrpKqNE+7XhMSb+
+          j81PLxjBVo8kOwYXQPMIWDALzqEvu5eILeJ9NpGVBjesb/rYulj7EHiT+N7qG7w+Rb4ywiWwrtiI
+          TVcX69go0eY/RHGUd85C3m7g2fNr+qePr/rWRTce9g98mGGQiyp21/1WL/brzf7mPNRkZCYFh89x
+          HuWz1ooqCvS69p/nfA9G6/JMIXHiHQ3el3irX1AIGg+LfpWrpf4PL/7h2y2+OCPYtqpIVShQzbDr
+          gW58AobGB+PzEpS6UO6uNsys2qKHyRpqVkVSB+gzAdSuGo0tgwxNoPZTiY9iAnTyhz+IBK707Hdc
+          3n2dnld6+aVgo87dgeinfIWdakAfCMM49Pn0WoGkv1J/f0qHiDHEl6jB2gl736CISNceZBi6Xkoo
+          ltZ/fE0pdeZSrHVU3/TEURH0Z4CPDy4c5l2s8/Cb29RXntEezJb/amHjeSK2l/fk1PF4b8BWz8Hx
+          D1o1+bC5QtGlKOj9lDr5n54O70dexw9RbHW2vE8xaJ/NHatbvphFPgoR60SbYjVJonl3035IWCqV
+          en3+BcJnhAkoB+Ju+tM8LJqb8UruTl96SER1WDJ2Tf8/HQXK/+4ouHpGT/XDs6/ZqZ9LxKt6hu2f
+          Vw/sVAIRPOOLQF3rt+1Nl+YQAau2/an/LTV5gK+P8EXd04ILWzBMslACCbonWpz9C+PdVzhDbi0m
+          bFHer4UUX0bElV6FtQzhYR04OENe2rTVnf4GI/WnEaqLp/prPHjR7Owme5udfcLx6vDO+n1pBL72
+          okj2pVyD7vdRUni8Jhj7Y6Y7y/f2XOFc9RCrY/Ud2DicE7ggRafez28iphW9jY7R0cDR4+zngnna
+          P6Eep4uPtt+T6jVo8OsU83Z9dFgCpxZhTmWHKOn7m8/Unwh06UHCDo6cbW+4t8LfixjUSqwy5+ez
+          MkLI5gzrC+cA0WlfIXyn5YHelSbU6UkqfRT5aUIvuN9HswiqHzKXGePETIC+fG5hiGat4+mr7WLA
+          7I+toucdAvpijzJfE5In+6tCLtRXl0stfMvoic4hEahX8TAnqW9x4BC9In+V9dYRjRaYqB6Vihqd
+          +wCLpdscFMjYYMP1J32dT3cXKgAd/fdj3w3zedd1cMxggQsSZw7/OB2f6O6SDHvcizrzRbnw6Omc
+          B+y231hfZ+G2wt9tYfQ0JK+aPEvMg52IInqJl4uzpr4K/+yJHkpncMQpUAi6u2OG9fUigXmAdYCq
+          iojU8kkazVnQQhhz0Z2UF8LqGRlOC8fi7OLMvKSOeDtLIXykckCP489ls+B7GZjexPdltTAcyTrG
+          T9QtqkLg2P4GZrBMg+JrPtHsNPbDIl0FHuLj0FAvrSydl72yQc3XxPSCEy8nzVLKKGgSil0XfMC6
+          zOP87/O56fA5i3flDH2TN3zl5+nDnJ8sDvXjPcGPfOh1PrguAaovtU8N/iEO1C2jEgXH8kate/Oq
+          qf3RNMSrhww/Sf4cJJTPBJ6swaMWe1eD+Pc887NTUYwoD5j0UwlSj2JNjUs/DyP9hCbkivlJk/YT
+          DhOfjQk8niZK/bfa65MXhDHcNnuR7+ryjJTJW0RHUqQ0yM+SvnI/4QdyJ5vIW5BO+SidExlN/QJx
+          dsi+EftwRoKmtDFonhUvnW+R3qDoVpbUzRSid7p6UNFdEEZqqXTWiYLFFtUa5ai+5xBjr9qaQT7c
+          BYJK7DuCdsk4mBw+AS4Em0aL9FxLFDy7ktr67jpIkjRniEZ8Qe+mgZl0e2Uy3CtKhmNB7hxWjL8A
+          Km0g4YwafPR/pJ3JtrK8EoYviIF0kmRIL50EpRFngIqg9BIgV/8v9jc9szPeS2WHVNX7PmmKOH48
+          oAQeJv/YF11Nh7v2RL+rxPjC8SgB+lkCC00KH/p/+YTKkFYIjRPApzY5j0L/YAoggcO+IiIs4XqK
+          fwO8XR4uthDd93qtViPRaEB4j+dwucWPAkaUKYnDvlKX/TTZDO4z8/Nn8dRrdBXTGXDm6hLsGScg
+          TOKqIsswDayRqc/XV/FtoFv66bwoj44uhq36iJTZm6hTUro0SCQZFoLAzWL82Z3yQZdgnmwqPjGO
+          Mq59Hw3osSkVfn6FvWtE6DmwyfLzLHFSP7KabKswHYoTxkb1rNev/JXQ6FcFlh/5sZ5OCZOBu319
+          4aJ1G7BdjNJE7KcSsUpfiksV0apAeKtKIk/VaRSiQ7fArsIDMT/yJ+/+8lN98yN86tY1pzj/yVKa
+          ehCr17sDVvGi8lDdrCfObGbS6ME3KqnNrjIxIisPV9m6LGDWmPcMzbQYSXx+yH/xhK1rPuVr1YtP
+          FLLGD2MoJu6gjn0D30nAk6tqlRp3eygZagnj+4y5nOj2/jxViSaiRS65RsP1sngZXC6FTgKRlvky
+          fnkRsl77xEY0+e7UILeBByc1iVYJD7pyHtuhrhUX4pnGMm6vTfHR8Jp0opibVXN9/oHS3XlgHGhu
+          mnN9+WJA0G4Dkb0w0Nb2lixgvncEF7zIUPrlcAenN5fjVAlVunxZU4LxCJQ5MMUpX52l9CDqXwuW
+          rQZqnfNRrsiWOp9clJvhcvL6i6TOq3NimzoGZDTLDam3wxEbD9aouddt+8DbzbrMjB0n+eaktQ6v
+          3PVMTpw6uP/q9Qc7HLE4vOW/OKU+MgLuiE25c0dOzyILGUJTzSsnnEJhjNICWU/PxzEtFcDafM2A
+          +WEO2DmPskt1H0HwNx4m+8ldVraHErrD+0IcpBWAK9JbCkaLS3aHIwNhC2EHhdPU47D5rHSjxx+E
+          UJ0FbGvbOeeGM3dFWaBedj2h1at8eMjH/sVd8N0+r9o62mqAaF2ciZ3FUk2zeE3RYdgd+w+X4Xpo
+          8wwC44R9qnlHdw4Hw4G/2LHx69lVlMfY1JERCEdseIdOmybnVsJfkhyJ0qsATKH0jKD9UGYc5s4N
+          cCGWVFiqB5nEp88nnyKmk5HJ+Q4+exec86/zz4GkTN+kyFePsnGcQpgEaUnyxFZcYRqVBKrFfsbv
+          2oaA7bBSIQN3MiluZ0mbqselAq2xDsTqThL9sdknQaf28cWOffrl20VYEmhOuowDAJWcDwXIw6jN
+          chzu9fBf+8j37GGSbgjTVdRyC9qnuPQPTz3QqEGTCm7YANje5F++nH7DBqvjl/WXPX63HAwTXO30
+          S3IcjmBwTStDcnVfyA2nM5gt8+shejmr5JXufdnQ6csiwDvbzINs0aZYuzWg6f0H1uLCCrcztj+I
+          76QLUWMsu9vCvRaQfeeXvwXMMVyOm+3AaFfI5ulZjnQzPQbGwftHbHut6QDYlwP1dmqJuvSMti5W
+          D+HxEUpz6QkV6LyPsUB0Oj9m0Q4j+ptewIJztR2xqvFknI27m0I+MxRs7POV4zQBwmM6nGZxviOX
+          aP5awMm4+v5Scko9jXAMwJULzjj6SiBcgoPHQitHGnFtxa75dZk2yG9yTrLS/+WUU5InYgIvI3t8
+          upsnph2klili3aR9OA91KqLw5+1dqy4GYA3ZKOH5GjXz0d0O+ZKc+kDq9QfCbv75jGucUg9+1M+K
+          ZZHK4W9IjhA+XquPdfl9chd8aT+wYcQR3/f4WOAnZoD5uC7YsZnJXY24d6TXral9xJbi3tKHdeDx
+          ihTihWfL5SXMNDCMzC+x5jvStv6FfTghjxB3O13o3Ky/7Fg+k4gop77NSRXHjaRuzpPgff4sPa3M
+          v/Ehj+YdhOxzbFR41f2BuFnxctfh8x6kc3/+zAdR+eZUdcgmicyjIhb+ChrVFGVDsd3N2BNlvf4l
+          YrlBo6NX4ifOm1J4a1UQ9z1PnmWR5EvgwQ8Qgdb+qweLzdcQ2e/KmGklILAEHvuBzEkPZq4Zk5xy
+          qcNK4XGhBEf9VG+fy4GFIBRf8/BLPnSOvpkDeaZ5+Qco8u5UHO8+YFcdEVPiUf5LzLSCGlkMHMiN
+          4/JC9DXh5Q1vOGwtfVzsdWHQO60UotedCObv4WHC8XIosVMPUkg9mziQA9d+X01LRnrntQUeSjP2
+          P/jmu7Q7GR/JOBHi87a3hDTkJx28rzwg/nxYw5XTBAaKbfMhVjPdNPqNxAXIK5axcgEAbIun+1DN
+          sD7HBOfupqlZCdxLbhFj6zWXZ8RvBOYa9cTeBE9jmRcp4HCjdF4fcUxnS9k8+Dd+QDStcMLZFEC1
+          VyNybuGb7vW3BAZ3uxElOjP5EkpJBFRmqrATIjXc9a0Eu1WVsBV/fjX9LKkDf0JWYCOuPUrt4hzB
+          v/yrc0ofblyVXmEflynJvruvEfVRgs7BYbG2ZkeXCoKYQZYmFpHXu5ivtX/X4VAC5JeP/D5ugL1Z
+          wNSddkZZcBvJ+VrzABPW8gPl0YFFKN0rpBes+kwcaho95MiHofULSP64svkMU2ETdeMZEY9zcLis
+          c36FN189zQsymfBbEG2B9dDf8ItzA1fQREOF19v483f9XxNL+Ezo1bbjrq/V8Sc8pUoK2dMPn+H5
+          SZdfkYpgnx8k9BSd/o0Hkjudx7teAMvPmCp4LalA3OfvS5fD2vEg6cveF5/TVyNbcYwA/VFzFtrk
+          N26xvS1QbVabnB43If89RZ2B6xdtxCvjlc6v888Cmntt/VigB/A7PYcIOjUtZ7EbfzWlyp0Bad4J
+          +Ez6FmxcFVwh00c1uZrWfubsIBeSK5wSIq/I1YgwyBN03fnoH+21BqvEtRJ4yvw2U6dRXDZVpQX6
+          m//+84PaIv6EJ5T3HQDuzCz1ApP5+qf3/G3r1Jzsfh9+hsohHnFFMKe52MGBy0KMOVLR79/zar9K
+          wtpTPrtrkb5SKNPCxScbuflytuZMMptMIqYXX+rp9swcGKCnuL/5S73W/t6lp78NRC45ZaSfoNXh
+          7n9JvufPaRbg8hfff/ownB+A+EhUxIQYeaVp7KFeLSBV/H4Tum/R9RZ5M4iW8ovDnV8suC4ZZGvH
+          r7/Yjg/4dpJSuLKrirPQ1Gp2r8+odZ+LP9JKy4U9PiGZmC+Wzeg+Lgv1ryAWH+bM2nAY19yp/b98
+          5bPsIxkp7LsUnBwpxeb0sTVh90uwoL6KzdY29r6mcgH7DuXklpZmTc15LOHNNHgsd1xYT6cX0CW2
+          hRSfpuc3X507ZdEoTvY/vbCO24OHpS0we73q6/kXSBNw6Zjv+fngLqJei/Atvm2cyiTQFr8LGXi9
+          S+J83ORfOI1GKaOL+VWxw0l2PeXRMvz5IXK+MbG79cwSIElXP0RVLkq4PqCwz8/Dhk3AqEAQnFcH
+          m2JI5wXAd0g/ATEhO3Mr1h72nM9FqSVwaqu3zz32vq3XMLqi5BY+//mxJeTGRPq8c3tG+/ve3OYW
+          gO0Ydj7c5/96i/QJulbAEB9HU73y0l38xxOUX6jWqxfvdNWZruQ5Hy75Nkn2AvxvV+AXIhEVUC5O
+          8J1ceWwLc0HXrFEbuGmfK7mbYB7/+fU3SQ+zlGCddten9ASgjSus5qwfcgf1EME/vXxaY14jvNFZ
+          oLufynmt21TrTcmeoJ3sfUoVLI8c+dgR+IqWtOcLhrbH4lFJL4uG8zo4Ssh1Z3H5F08BLi7uP/34
+          3V4sUepW1MhKClUKcu+L9emhj/Q4GB7sWmnBKpts2lZZ/ABbXjjjnc/QhRqeBy8v4mIrvLrudhHE
+          CKb3fMInooUjfU82A0f2TLHxHD/59j17LNjHf5a6bXM7dKEB4ja3I07p/8KV8+AA9nxGCl580uUR
+          qBa6c8LkS7nJ1dsnY3eCyWX+O29N7e//hSLzqsheD0D3VK8FZPNsxhi/I7Bk93pAEqefsClyCuCN
+          KM2kaq1MnGxXsW6OzQClw0s5Yk1uB7qepquM/urHn59ceuXOQ+HZtn/1Nt+W08VDL1d9ElOq1/B3
+          uNcT1PhRx3Z1POQk+ZxFuPMyfOXOAV2+DohAp/oV0ZvnrFGlwQ5U69gnL5t+8318oOT0RYOfzH3M
+          V39k/H/5aden9Uo+dgL5R/W3W4av18V6M+jlF3dsCPRFqQVvE/qbb09/uYP1z38y2hBjY32jcS2G
+          dIMvaCpz+Z7KnDssqQkCVS/w4/Rzwj89hPjXdsLnksp0fRW/D4wMyyJ+WnzySd4YHQzwcsFaeSX1
+          JlTRAr2G2Xfotqz2++MFcl6tf/mfUoLnAgLDwDP6mgz49/2hRQL/nXlsTS6/SkS34Rvh255/OfWl
+          DuiTT3v+Vb9jF2UaRC3PnbETsTTcpvqRAfykP2zgTxLyJDh4kB6dHJ9bo9aW9j158Ls92H88mGiK
+          vfNZBMh9+vaUhvzHhM/JN2foAYNyvHSR0N4eHBueZYfkyZQW0gb9MI8nlWp/fgBksH0QzYsCuu28
+          CADfC7DZyUk47/oM3QT3jnUbOjXZxx/86Qn1gTd3GzTCwuvKX+e1I4SuD3/d73vSPJ8ZLHmkFnzN
+          cF4/NnZo12qTpkMV+NnlRHST2jnH178SeosZzktoGy4VnNcAMEzP2M6O57xzcq+BXgxmIjvKb1xK
+          NEr//JPBJP0/fQaPcxLPfCdw7qTarQo8Q2qIfXq98+mU8Blgn2xP4ukm03k8rSVc1IHF5nemdMvt
+          7Alv4Zzg00EpQyo79RXmJ7hh9V4x7qeIkAUzBSvY9qSvu9rtWII4VjWikUMcsnhNO+BOW0GMjAnC
+          tXgX5h/PwKfM5901h6CAfzzEaXU1pP2DLyB7d+t5vYAc0Kx+6+i4d3pFuz7ktkttwUucXsll/zux
+          /ItzNMKTjq2Sacf1fPxJ8NdTuPOEa06Gu1sA691fccz5l5EsDiiA2ssR8ZamGNefni+gfQKKT7/4
+          4653KgXQjqDsH6N21Do17zLAVVZJQpF9u9NlqhzIHNQV6xqK3H/xfRZldvfvljuVUZBAuTP5GYB9
+          R4Y2nz7wkrLnOdpIA0ZAzRSe2eb4jycvgWV/pF2v+PeDpYPNwm8JnmdpwNg015DGyX2SmIO8Yj//
+          GvmizK4Odj/kH/PF07Zs78KIWM6YqU2qcAZkckD3KTif2/kyrz6OIqiPe5ek9RICwbwJDtxO8pkU
+          r6f8V38+8ObLJxJ8D89w2+NJkm6PlrjRjGuB/UoV5EDQkzPpT5QuSJvhwf3w2AXYDhfmkJbgHC9v
+          7KeFHgoK0XhA3ZKSPz3Fa4Hjw9DPkr/xrrdLerqC8k1Tnw1BCQgUAhO9x4Pub3E17ScuQAqTbaBY
+          3gJB2+T2JsLbxL6Icc3GeufFEciPcowDNKzuT+KIBNTiif1x19trBw4JCJ5DSZRqOYb01IsVWqoR
+          kjPqlnDuGTGAxfwbsbLXSxKtRx0qJvCx3GYYzCZhA0jVfJtZn+YhDV/CDBsTOiSOrDz/pxeP7+1C
+          /OdsjuNWHBO489y55VTH3dL5xYLd75AzW4qU6N8hgVmp3DFWcFkvW19eAbl8L/gfzzMYOMFmS/YV
+          Mq8a6SVhyr94wK/IuFMqGtMABeifiLYyI128n1OAndfNADc3+jNXQ4XbTYuxv/P4rX+dPOTRyid2
+          dXyF23rsiz/9jp01ScCv6pcCtnXq7Dz/MK6H+zhLb2mJiPd6lmBpkNbACD8U/Pf5xT6UA/Lr+wVb
+          OY//6rv/r179+acVHo8fuJ7sbGaujx/YUsO7wm1LB/y8RUw9T/FvhuoNHYnfzGm9VoW8wfaopzj3
+          nAFsMD1ssIjbisgSC+jWTlsGh473sQLtHKzO0vlg5y1EzzsTzMDc7/y76T6+nPo23EL/HfzpJeK1
+          bkOXmPmUSODfCj73w2Xcv38BwSciJGWhmS85o3uA9b5Pgjd1BgsnxzP0ZFzOXHpXXbqmpBNrARfz
+          5k886NY5DIByDeqZXw/e+Pd8wCbB8R9vJXv+Bjv/2+vro9556gDMijkQp/MPIY0O3SYJJU///HVO
+          hLe3nygFwR9PHLc28VXp+Uh5/7DzLHrWoQWG8ohm7iSv43odLyk8e8MNm770AtusaTKSHtZvZrbp
+          G26XGQxgYvIXPsf3hi7K109B0le9/zHSG9jXH66wrTMHO30f0KlJSQVz2678QykZOfvnn5L1qc09
+          /5G1v3z9l89I/rRD7Y+PA29I9RmUVj+SlUQy2tcria6Z3L/6B+PrJhKVlXT6b31KN4oIn9Y4cUkZ
+          33WYE8n90497V2np+ecPfOlteO7mkRzCZ30y/G73S8tMlQ0Zr9GaP6MqgzGXLgwsPumGvfg11P/y
+          81/+3fVMvVJvsf7xPZ1T7HCplc1CQ21bZ4BYAWzySZGB+D4+sYprbqT4008wP6rx33oH/XU61KHs
+          /d448ikItyfPeihPFpXE7G37i98IfjwwzNyur2Yolixahua868c33a6DPEAFZBNW0d6ldBt6GQQp
+          yogdta67rw9e//znvDaPX8ia4PpEciOZROt/ocYDlLOwOrbsDMqDFa6CZVp/egtf3mpP+yAqK+ie
+          GZ3gUvqGXeZSiG4354JVZ9LBkkfiAKHOqP7h2alg66S5gcIvE8hf/pxFLOnQdx/CP/3/XCtOB89V
+          irF2ZB5g/uPdcVD/yGl6GiH/OQQfOEJH/+Njo+CfRRYoXhP59z9eu9yq7C9/+FzOYW1TI+SBnWf4
+          jPz81tNhLXlU8/JEHh3d6NSYcgSV67XGxgaJO7eJKcMG9uXffB8pnMMMle3zSf54487XVch8xQ5r
+          B1PL+bVCOpTVtCNuYr/dKvpeLVA95JZ49cqAfufx/8+OAvC/dxSwcavPC6xbQFKpWVC/CnA/KvHI
+          t5jXGrDBb48N0evo6ul3H37wQZoH/hfX2/VwqZBnegJJtqymvSTeGmilgkQMi2PrVfftDo6HnPVZ
+          OR7C9fvULOTx59Vnu6sN1kl9TFIUBjqW0Y/Nx6dnR4AVBBuf77eTJqziR4ZaoHXYCxfo0uZpfKDp
+          epSo32bfI9X2FqxfB9mn8l3J2R+mDrLO96eP+Kc+UrPPOiDrQkXw8YzqSS7IDP16Lnwea4y7nL5G
+          CQ8/acBudHtr6+E785D3bqEfU01x6Xldnojn7iqW3Yc8cmyjmsjrCohvbHce17CrZygs1PHp+upy
+          gW28EnybAeLTcXmPdBa7DF68lfFZU/3UVL/JG9rieSNuMkra8hsJDwcz9vCpNC7h+l3mKzSv8oPE
+          B43U++dTJPPQIMEWHsZtv9EHBtqozhSceY3evZpHXmPVWLGtYmz6yvJhr9s5Pk/exxWqSMzQD4ce
+          wWgZ8+FygTMcMrbwy54w4ZJmgYke1+FDTsrlni+BbbKwVn3PP+hTAfihik2kyKaFFQpITm/aTkji
+          6IRvyT0HZGseGwiskCWeMFnuypKwQ19b6UnG1ba2TJWSolhSQ+JVn+/IS/CuouwqST7KzSNYfwtO
+          oWgEGUlnvqsXKdUi2Df3Enso4zXqJ8EAvWw64Vhn3y6vzq0D49vmz7QyxfqnGnYCL5zwm8XxoIws
+          xd8KARdh4kyHE2C1ryLBwLqw5ARbq14dELDw4bISyc6NXbNyQSY4mecnian2dhfDZiCYlM+EM3qx
+          co6Z6xK1uEvI9a6CcWUEpEMmn48zkl6ithg2z6B0YyJ88swT5d5xocO4uMvEAjnKf8UiBpBRGwEb
+          d3RzSeku5r/3/WrLaORndE1hx6cNiUoxdrkxqhloW51OXnnZuNTRPxKCy90kp+ym5VyYyB/Yv1qe
+          WPfZ1Ca9MDz4+h0u/sN5Gxof12UBOXg1iH0ZI40qt6uKsDZYxDuMtbaR9/aB2ek9YayfknENE+uD
+          tqnKcZFGqF4vZ+uKOM27kctlelAhe3YsyoVk8avKwfVsfd4sKt3Ph2jCatFlSysfDebNI3LQWCF7
+          /ck8fI60JF5xNVy+CYonFEMF4+wVXsOtnJ4LvN85F3vHezUKb5Vm8JeRjqhs1oRLZj1nQKR3g5PA
+          vYazCKsA3T0zwXlf3rSNjLqDfuH2mFefSyh9LNuMcLE4xLdDZRRyTZGRf6m/GOel6QrL4/GEd09P
+          SB6/zyPnv9gBHI9zQM43owlJrVYODKcvxCrhP3Q93RIWSsLjgPPEpJSA/Uw0iqSEeFHouLz3+T5h
+          PNgcsdT9VmmRphNcp14jfmprI8/AaED92cczPDIdoJnWWkd4pXQ+PCENt+BXVBAP5YHYVPqO6/ZQ
+          FqT0nkPcrH+PWzklG0xCJ5lRVfrjgq9ahV6tavosXq75+m4PPjytSJuPif8G/OMuqiD+vjVs1QWl
+          9Hq+NWiPb3xz7ud8c655A731c8AX7cCCf79/v5UbNgIWj5MlqBV0rSrBetaw+VJOUYQ6Vj+T+2vv
+          f5bX3ytyHHQjqtLqQGCEoROb1zxg45c0YM0lwoDjMhbYCAeVLtz7kELiCQbRG7nX1rzTWdDiYb8l
+          8Vhr63lzRDh7vIutTNzG9Xs88LAcxw0b01qO2zDOA2SOiUfS1epD3s+dAUo9cfxWeMqUrYP6Cm+e
+          pxO3AGrItmPZILU5HIn7NHww0VQwYV6+TzhtwgysL+n8hBeO+xEff1Vttba0RNZFIvPxEP9oH72I
+          CosR+DM39l7I8hSmoARlgDPfxYB/MT8V/uU/8x4dwcS/71f4dr9PfKmot5+h/hXQJOsFp6+8DOf9
+          fSBZre0ZFZTNN8dWVchs7YpPdTlQctsJjGZjhvgM4wPB8HUTCoU/Em3rI7BSMUjA/rxY9R4mZZVv
+          6IAtu32IfTHPdNneYIH35wsRv7807kLei4UOaoHJfT6P+dyPlw1Zv5uD7U95cdmIVywkFrqGcWpM
+          9bLgXkQ2bjFWWiADzilvDvyOlxnr2/cBfq9nukF7FWKfz+VyFN5L6sGxPtxnSrYz4OKXLEHmqV3x
+          CbyiXKDKJwUH9Ylx2n06bamDMYC3ubCxHLUArEMj+RBymjhLt+OjplW076SPGR8HsG5p1wXWAh/C
+          WfWb5LHRFSiuDsklSohviy2distLhdBQGezcIiGf47p8opy5AJyGzwflL1ZtIfV+1rF2rouw8/hS
+          RmdTxyQSfisg3QNGQHULjB/y5eHyBLVXZFnbRNKoubqrsfUL8mreIQ7g7JEbjo8ELk6mk7yEMSXX
+          dWCBpzQGka32nPNn2+hgXEoXIvvRaeTJe2tQ7T4WEpXhTKncl1e0hel9puXtS7fm3qgwCa2EXH69
+          D7ayNny4KLcA2y1wxi3Jxg2ktz7DsVmHoNOEn4zeZyD80wfbi8sL+B8AAAD//6RdS5eyPLP9QQxE
+          QBKG3O8SBFScASICInJJgPz6s+jnHX6zM+zVvWhIUrt27apUccHefBGGv2Y+OJsr/Z0nQ63l/Lgk
+          txs0nt8KnefbMlKzUkzw52+vBbpTWn/zWlqRshAz9LZ8Mfl8gcj9xCR1PIeyErV7yT7UCn59K7ZZ
+          wtauxFK+GshYNlZbo7VK/+wFIXto8zU4Xyr4DBMPuQfuqW1lU5bQvawKCtCn9pZ7xvjgLU0KSdye
+          iYlyz1QoWQEhTsaCZnkdL7Kks1n4x+c0GvZ19WfvmKFtpuGvdd/A2ssXdH0Zo7dEEbLBdtFldNvP
+          4/RuXr24n2cSXTuc/3hyD09yZRfByZM6b/mF1+GkzjhB1vhm4mX/vj9/FFBbUpsjHw/dH38hyd21
+          vJ9Z5bUUiJVBEHO9jKzXNrJkz1cXBcdQ0aYNpb20n09yF69fsHzHvoVJc42Il1PsTdfZLyColxNC
+          /ngHq15NC9jxAVf5k2i/8Yt0eKq6AK83dqXYk40EPgHPIh/Y13yhknqDudMj9OBvG53Sw+0Gg9fZ
+          wxuHX/nanR4YnqD8C4RCNgGr8vvUpWv3JOdz6mscE/AcLFpGw9SdrvnuPxaYNPcImYZ4a7YTPLji
+          T/dyooyQzZd7ixa44zF5pptM+e/idtCcS4EY3vgFm4YdGRLlKeCTMZ28HgZ9DXZ/gE/nQW7Y5BBh
+          STTsCbkvIW02swUmeFqBTc46akeyddcN8N+fRTK9PI7Yn3kVqLKlEqPdPiMhT7mVyu3r4sf+/AVL
+          EwujVG2xjrzJW3n3svNDQSRWJOGRcm47SbGWTgFz5Ws6CWnRwaYtL8hzLQFsQ/3UwZGeL+QMVwHQ
+          aB042HvW9M//sxEXpSD+DCs6n67eSHPNUeG78qTdH9TatH8f9FyOoD9/wk8ns/7j18Qa7hc6tRbu
+          gKKm76BNTpm2wtg2gdZtGnIOYqL1ulPrQK6LI9G6nxlvLePWAL/TgRjG4xuTtZMF+Hd+iz6KR95/
+          gRK+TvkcnKIwy+lL0wdxxwNiTKs8smc5qACUUx/t/oByTxsJYPcnASSOmJPbiguogqeDLMnHFHdh
+          UcBfa1hEfhIEZrZTdagT8UDUL//78/eViLOvjJQb8sbt2Tgd3PcL2WzLNcuKmAWcavGL/Nxd6Had
+          NlnSkRQHv/WXNnRJPBH+8cHrehdjkl77Gn57/huwuHXzhXtfQpGzpjfS/uKvlXVlKHOMQc58v3rr
+          Y/32MOKZkGiZ8fF2vNjnZBcsspLB8QY9Fzf4unt9AARwiJctHQIoOY0dmEgrvTW+yR0A3gERpQxf
+          GnlqTxHmHuyJKcyCR4zA14HcrgGKs8HLuW/9KOGlKc0Az3uden48QukntPd9bmlK5/rE+n/8BHlK
+          wzfU0t8R9IRSJdGvbePJ401XGlEqIvPbvsGmfp4dTK7tk7jXTztunbRG//YHTb8invb4BlqH4Ye8
+          HixgzcTYhOdXcEG+2r09av+cFP75B4SvmoY9+XwDj6o/4lVuQT6TZ6zC93xsiPO6/+Jlxx/4HLcO
+          yZIdeGwxpRgyd9MialuEMf12qwxrkNXEfQnC+PkuTxnSr9js/vGer9Gtm/7x2SDFP/BjhY8P3RXP
+          GIyrrvXN9GLAtchl5O7xGevkNYaPo23sGUNPW49KiCV/oFfivuUJzFsXJ3CJKwu9Huwhp+U1EuDu
+          74N2ubRgva/RIsVFjYOjfUzGWaF+JJ4+SMbi4cKBmSRO+acXEKsGbbMAQzJhC0KWpCTHYPqFzx42
+          L0kmt7eQazv/n+DOj5ATvq1m9+cqE6Z29Bfv0yVt3sM//qE/n3lMA1CK8PfeMNIyw9A4xugicMxk
+          mfhRR8YVKi8IuGUJUS6V+9TDXyNCnEZf5C9HJf/jP+A4WAGer4/DuMTzFYNIDH+B8CSPka4eCuCd
+          b30kh4rqrXexTcBF8RX0qn5nulVWWEty/vOQt9xLjS7PawEPszCQ+34DZobFo4cFoxvIKd43jzDx
+          KwWvsRfQ5ahp49EDn0laSzEgNndsGvr1Djc43vyWpM6yaDg57HMv48NEDCfLAP0sOQu+pFVRGJzO
+          4I+f//GLgMNFkf/4eGjBo5YCpLhsE//D0+QuclhiN4nORyWcJCKgc8BAGzRTnIkdTEQbIHPcu7rn
+          pTQB5xxy5HpO63h7TXuPjUAdg41zTt46qjYLnXPEETk4nemq64IpwRYOmBNmYe855Rdik98ChK7B
+          p6Hhyprwcrkg5LJMTSdsMgVU1OwdYObRN4Q87fZPf0Du4CT5zCmfDtoXgZD7dpvA9rY+KfTe5Bvw
+          222iS57FE+zMrkfqWZ21LbicTPAMnJgkf3j0HatWYgc9Ic731sZ7fIbhM2Ek5O3x3DaafQre0cMm
+          V4M5j+un9Gwoq3uv50uhabRRBxuKHz1FfiUwYJ2TbBLuuHSQ2eKgWYPzowZaqPSYKs2Z8qNXcKIs
+          gRBZA0w9NmvSDO7xEXG0TIhnWFwGSfdu772nzb1Zdn4H04o+UTDZZ40qP8mFP1ImSGnSWKM0sSrI
+          8SRH1jlM8uX45jPxOOVfZJmS7B2bq8sCflld5Am3k7exiDCgRZKIYSWUlHpb2UJzhQ55SreDhg9D
+          kP3DC2U1oLZyn4WVXuwGMPOORu1PX4DhO7wSv+XMmCtDmvzFY8SMlU+zHQbzH74FgOlP42bPNJMA
+          EEGwuYuf09valVA7DzJ6Hr6T95OhDP/FS2rkvjXCCU4Fh9czIJ5pGjn3sPUImuj4QdlZ0XLWj4oF
+          +n0JiaMeN20RjcqXdDYNkTKclWZFwPPhZ4wxQq/3afyBGPsw418psTMhama3erni/Usw0V1N0Lav
+          9drEdyekKBnnJt8WNCXgdXd6pN9dS6PG/TnAKBDfJOgORrz96WGn0xQSNSyzfCWEs6U3J3XIDL0o
+          J3/+z0PPdyBNq5tv7Ds14bFIHsicZmc8ZmJugumpWciTyrPWbigd4OdZn9B5Bhxdy9PSQivxA+Lq
+          5XWk0fnVgrfJTPjvedO12URxfDqEBGXeesOzcVrw2YwU2flToFsZggQGWfcgbol9jR7wh4E3UTvv
+          PUP4cWAlf4Eonm/o8TBscKxPMICB/ajJecezmbq0krxb7BGD9lbM6R8/AX983cHbpeGCtvDhPLQm
+          uuleQ9fbzffhkNYa+YsPJ9/RfFhi40SefQAbSrrMBb7SGkRpq29MRFcNpAe+5cjUPY1SrGyy1Eb2
+          RrQ2PcVUNbQFnvskIclyaSnduucCKD5itPtDb3sxH1WyVy5H8ha/miWYxxvY+SD58//kclYnKV6t
+          M5a88dKwcLj6cHVTBmmdXnmLz/UqZOSoDBj3ibXZCMoKZPh+RYqk372jeVcgFD6+STymfzSzc1sw
+          3LQG4HFcW21qwjGCIIh7omdK4LFCmnRQ+sZSQH/VOT/q1BIg21xP+HN5XOmfXgn+9G0/Un75wgwa
+          A6+iHJNHW12093M927B/Rg2S59mg2+dxcaXcGRCyH7jTNvljY/h0qve+Xg1ds1Wp4B+fCI2ly5fc
+          WlrJOCx3ZJM8oGx+cW3IxtFIzuyZb4aWUWu4+NOTmGVxBuwzc0sIc+6LtHpkQO9HyQK+kcLgxRQj
+          MGc+E8IDL88kkEkVc+v7ncApmzBJfTUEi5SaDJCwmuJPge5g3uMVYcGwJrZ4WcBvUq8TIML5jC4U
+          kHgI2iSAxm24Bh3SSg2nkzTARmrjXQ8GYP++CeZMDDBkL7iZP16vgsbkGGJJ69Vbv89V/OOHgZDF
+          U06VeyYDbUyPf/zDW4L2sYmMDsK/nrzNklwkAb6iisOs5a8ecX5lCMFV1cl5Woec0gTVoicUKvJL
+          5Q5wYQEZtNL3jIzb+0eXk2kmQK1U/C/fQAXFEeHV+NyRZoW9th2NJZTCd3TFNDIH7/OX36i2VxEQ
+          Bw3Nzpdb6D7dgFhxnQD2GS4iFD9mSlSjF+LV+6ktWEshIMnewJhCR2CA9pwDvHhPueGW9CjCdbgN
+          mPY3GfRT7WTwNQ5CwFcD32yfvefCVdsr7DBTxfR9+i2ApJc35raD67FBLgdS7ZUcMcqqpX/x+D5l
+          zgpCud/iVYN+B3+kSIhxTtWcnSRGF71L/yD/vq+XKhXCRx0Q1yR1s/LkHoE9/4HOe36E3fUjUdv8
+          V7Au1RVsDPV0IAJBIdH8uVN67W1V8i7Dg+xzJeP6cfzpsJK4H7LYfm5wX+ME3D+PKwpaJ/I2jwMT
+          SD7jGMBxnjXyun43uPj4iZeab8DGDWEq7fEv0SH3GCkfp+4/PdXogw1sR6RwIN1g8t96hq8QwtA4
+          5MGRewUxvYsPAZ5v3hMvB+Yd/wp/VsU9f0W0Tt+neGZOCJkz1FHeg5CuAj65kA1ohnnwYuP15m89
+          3Pkdsvwx0BbbmTGMeBiiyBNTgKmpVFKk+OCffkl0iGQwoTBCNnoO+aYWvgyCQtjwH/5ucXOaANOy
+          kCjzD4NlMdMb/NNLbTVR6VGp3q4kWw5A1oWq3p8eAx3l9Atmg+HzJb68ZenPHsDn1YzbOogiDLWf
+          ih6c38WbFAwMVCwtJn7jjHTmuzMGjNryOx9/5dtqkgHe7tQmf/xmIW/BBfahUojmTtd46kIsgOfj
+          S4irrsD7pw+0XK+iAoZOc8TCxsLkLnDEDTxC6fn88SUybStROfyK8SQ1A7wXl4jEp7LZK3Tu4j/8
+          QLrFjf/WU+HgAaHqN9PFiS42HBvpQczJbuJNlo8lPA/9TGRpTnLanMJafJ7eeXDA4i/GlZXWMBFd
+          EKxqqGpHppCzP/2W2FGyxosVDSmELTMQG3P2uPGpa0OfQyv689eLlgs3iC6pFHBXwW/YH1uJ4Prl
+          C8yCfUr658Rz8NsGR4T6vgfbdRL/yyfq20ei65/e+eKNBGWdXdGV74wJRtlNwofWyL2lPiaDxFS+
+          i7T3e4sngIIBXjP9i6wNxB6fl8d9KsZpQ8auH/Oj6kVgjAWPXI2HFS8nM7iBPb5Bf3j6fS3vQtrz
+          a8Th/C5f2c41QdlyV6J+FK7ZJmnsYVzyFhaaIqZz2Fe69Pb2G3IXptB4ltxM+Hee7B0/J0AfAmwJ
+          lxLrGQwxZfDFlDrgmRhERpVPVg4YKHwCc+fXbLwxRhfC5cpWSA/jat8fqYBSItzIH76TQ6Gy0q14
+          iGg//95qg0H/05t3PVgej9PJ12FA14EoR01r5mJZQmmPt1D8pRvdRGPooRmpTxKA+t1sQa72f3pt
+          cBR8G1D/9hag5pwZgh7s4b/znYNDjZxb8Kb0dQL71HC5I+f2/POwjKMQOgrYp0ytucYuZniT2htt
+          MQu5R8PfVzuBs3Y6YAYzcvyH5/BQN0+E9vwa3vUGwST0QtDzLDfke0ps6SeKC9K+dAOrrQP8l//D
+          vCgG3lZYVIXgY0QksDrLm9Lm3f/x4UCITvfxn35aL9o1WPZ87NSOJxY6rYeJc6O9tlhOygLu+roF
+          sFR4sO16M4gYkAfLzxW8Zc9PwvNZyZAOL3G8XC5iCI2TAgNhzx+Me74ZPq7+g+S2NMVNcEt7aJyr
+          CF0cu2g28hZbCEtHRtbqy/Eq4NWVHtVw3OMdPd9Ob7eEezwSHC8M9GjcYxvwnwQTDbFOTLOy4iB7
+          rBeiTfckXzt1TP4/FQXwf1cUFEtxQMGC5mbZrK6GBzURiRE6v3h7v3MTslcok8v8IM3qXLoB9qP/
+          QjalXLNh/ldK+bpQUvona2RFhscQRtVCXrwq7F26LiHUmblB6h3XzdKsaw2rfFGQcwCnkXI6XaSv
+          zyDMHVMz535wwWLJFiMK3pGp0e36FqDEmgPSm3L0aPtYNskrcEn8RjuDpRQDFoYw2YI4e59H/mop
+          IuyUUxx88jMal8+9iMBdib4kOONam88fXZQ6KfOQUni3kd7exgTWjfWJ8v7OzeocgQoDzGQBAKxJ
+          KTMuESQM+yLOufdjUvOFCeNHHOwa+T0n75PGwX4lT7xuludxUDm08O4ZAcoGxs75V/xzYduMWgCM
+          eaWLezpysJYbRILnR4q3VcUMOJmlT2wiK5SbJ7WTBNqVJCkXNNLc+y1S8dVjUprKKV9nqVChYG0y
+          cs/MHK/fp9xJka4bJOESOs54OacCfLx+wQl5dc5f7oIqBX7Vk4hEZkOP85GBlm+B4BA5MKYfyraS
+          HUoVCQbxGE+tXFbQqZsMnfHzmnP8rDIS6KYIXW+vxJv4w8WWGq/1kF1OxbhC/pmCr7nwxMkkPSdm
+          /imh6fcBeS2fYDzy63mB8UdqyRVrqcdyVIwkzSU8OvfHzqOXroBw0C53ojhXrC33rypD79aY+AjS
+          RaMlSbAku52NvPxn5nyjAhnw5ntEymbtXSe9Ywr7snb2Oaj+yD4rTZWyKjgRs8Sct1w5OYWsHLTk
+          ieQlx277biWfKCdyB7dbM/CGUEFPXgcSHRiYU1/YWCDMRwV5xV31eFesWyk8+wN5hv0BUP9SZZC7
+          5w1BXZ0BonMuljyHXZD7SpmG6sESSmPqAuKI8RDjF8OlUmZyZ2Ir1S9fhz5QYfDwGZQ+th6wayK5
+          sA/g3/7FI4c1OZOYzBuJDssuXxS9yKAqphGmAfGa4Va7hfj7yhJJ3so7XhZxjeC9wQ9cXV9u87df
+          cFOzkJit0IO1JYYqfacqJrcnW8eTsKkVFO1CJWede9Jl/1lKb88FJWYZaqzKG1hqhw8kHvt7N6zw
+          eMqwUieA7LJqcxKtlJWyRv8SmSayxntvyYTnAxcR/ZrfG04hsi5tLasSmxq5d9TW2JUIifLg/f6e
+          R75LFVkaZ/GOZP789tizIkVwU9MQuUkixKtpvn2oe4z/9z7eRv2bDp3Pw0bpKxP2GVY9J2Vx3eC0
+          fnzilWU/kRSsnBOI2DkD7ur5ExTE9x2pMwlH3iUalrwDWIMRsB04XpLbBnknf5BAgbJGjbNcSwd+
+          yJCcnTttU25Z9+98GUvyo+SRVbY0vKSJ2K5lAO4mMao4I/mKnKpMKPucH714aSsFneuR5ot1c0X4
+          t39qbyk5d9/HWConT8LSU+LpmkcXU9rtk2jewuQ0usmh9CGvKth2fNwgzTrYlZ8ew3Z85ivCrCrt
+          +IRuaMXj0sq3GtrP8RjwR/zNuWSVOakb6mew/ARHO2LQRhLsvEPA9VqQ07bgODgjyqGnUP08urR9
+          BLv7QIKFuXzHLYhpKd1ZjQne4SMZWeWpCBIflRXRD8swkh1PJUrrADP+6TvO1uQJkJvtA9KFj5Fv
+          YK/97DL4QIX4vQBqL7SAkaSyRHtJ+ojZO8sAu2paLObxJ5/Q2lfw8NIQOndtEuPp9AgkTNcG2Sfp
+          l9POFTNIbe1K/DOMKXfEG4a4akNyKYte29wvGWDOQ4+4h3tH9/9XSlG45CSLszaf64VdwGGAOgrm
+          kzpSM7ix0J7EhFg2dOPjUtoQ2tW7RbKJQb61J+LCvSEbcViHp+Q4H6GYX6sHuvhVPPIlHG8wjbj3
+          P3xe7mzai89pTtGjI11MUfFhpethmlB5oSBeOpypMJ+Ufc7V8eDN5CVgeBsbF1nn4UPnoTdViY1z
+          nuiLAEcKPC+DaO478nyiIqeHN5lg/Co6FAdkHIdvcAygTTaVnNeiHpejIkOJxPcvMj5hky/v3k3h
+          8JkdUv7uh5yqYa/CdEI+MlbVyLk5fuoQDPc3XqB80+hU9MJ+5wYjaz8/23S6BJDE1y8yuCFpZtHM
+          OZD3ZCJGW1uAe9FXBbE+5kR3ftfx2MaXUsoa8xswZ6x6PPeTbUgv+x2dpPg26/YVdOnq5j6S7cPs
+          rXf3gqVuVXKi5M9sz9DvU5GKi4X2/fTWYxu20ssaGTwMoU/ZOqOLlJ4tnagfO9EWJrqpcLF/FlIa
+          txlX8lqmP/tHL6d+0SNqnjbkmemEbrCYwHo/3xfpz/8kumLmS8g8J+g4LkaG9HsDFq+uC2+OlZKX
+          ec9yGpTXVgpD/YGKRrDz4yDTFq5hmpFX/Xw3nLrcMvE1PzJiAnzxjgRfe3i/yxrJk+GUU2G4ldKO
+          TySvwkajK9Oa0gd6HOYIRNpiz/IgqRdq48qsNY+3GVf/22/0KDkj5/3KSMRIkllU3C8enW78bZKo
+          4Xzx2rhawysVnODCU4RpwQvNMuhMBitT/iC9FdWcszgsg9ZKf0g7OYK3nJEmSCJuNPL6vW1tiSox
+          k2YQEpKRqQGUre+69NO/HtGL+EM3H6cTvIfsSJ67PZJX/Hah8XZuyN14J1+SmbR//jsQktuFYt5Y
+          ajC+my8+xSFoplZKMykA/RJ0V/DO12IqBPD85ZjYZaXHdLjyHMQtehHf6mzKe3MO4cAsCBXt+Izn
+          /NRU0ptyJ6JzCW1WtpMTqU3MAzL6vAd0rxf4sx9ilOnVW5VvK8MtIkcsfg0rX/nVWOCYqz/k/fEr
+          Ghup+LJ+TMDID1dj19pbxJK5RMT9GlbMxZ83hB+QXcmZsYUc62PfSc7YVuR+xFZ+BBdxgj1jnvH9
+          9LXj9SrXrbQoOCCGYaiAvQ6/vdLkDvGbKiXY/p7nFGqBvHVI/uE5/ECHw0fIfvN5Q2r7zx9oMHk2
+          0/3ryn/rjcrT2DY/5rGZUsTxBTKNXAMcL7w5KZEwh5SlPXqY8bq9J1eCiMHvPaIkIXLhVStjoofv
+          hlKsagO0pM1H8qVzxnWQaQcaI6PEvnh+vphuK4u+x3a4tfxxXPTHkImGtQzktWaOxiYz6cDphJoA
+          8sWUzzY/bKIvMhi5h7sJNj55loBrbinRTBtpRzgyNQwwzILTR/ToFnKpCbQTf8A/qSR0WUiUwbf0
+          q8h9I6JH+7dgA/BrZaK17TtefoNYwe95XoLV23iwr38LRwbBnY8W+eLgkgGHq98HIPzIHmGsUYBu
+          NvXEsaJ3jnd/BzRL75GqS/E4e6wWQUcIOYQ4TvZ4BnM32KTRe8e/T07M4MbBP/+ZcpsGjtevJAOk
+          Opi4qu/FS9ApgxSok4JS1zLoyhznHmxrlZLcCr9gcNbFhw/xlOL4q4xak9xvHXg9Ng95sfqj28m4
+          61AaIpX44zDE+82AvQcBG5BzUlgNvZy/JWxP2Qt5bj7n9GZkrvgUhjMxt3MF6PdxT+BlDe7ESJcP
+          WH5aZcLr4xBiZrr5DZsTJZUSbG/IUTg3Xt6oEP7iGWRvp73LH21FkFzKJwkuNI/Jqe8w1BnSELTj
+          21JnewUaha/dPi70z7+A0yEDSAcq8dZU8mQIydnAcDXfzdrzfgqHLX8RGXdtQ6VErOH3F/wI0nyW
+          rnJhuP/8LxIVY+8xNZuQgd9nIP0aZeQVG/ngcannQAqiBKzcXWH++TeTM990PfR3FfK12pL9fQD1
+          5pyBib5Y+HRDPfjnD+GVzcgF9CBfS06X4TW0f+SinW6UbilhoBvI0W7/33zbHtME0+nso39/j2tm
+          gzs/DaiknrSewLcsNerbILnTRYDWsTXBy3oQAlH+XeONB0SEtTA9yQWNAsUKsc0/PomC2qqabYL3
+          CEKVmYi88+utMSUOyHq2IP/5vYEtYQNR3P1nUK4MAH98Wmzd8I4FbmvAGsZyIbrllSdqG1TxZDaX
+          +u/9iW1COV85psjAkVH5gGNOt4Y0g1HALK4aZKB8Hjc+uZbwqiMWr+EjaVZieQEkRTbj0/cnamt4
+          MyuROb+ugcgtbr6p71MJJX4+IZsawFtkQSpASmoukOS1bZZjwVbQuN8wPph3Md4YD0+icLEnor4l
+          PK44uXDw732mhK09cjwl3D98PeffieI60QsJV12I82R45Pv3lac7t3eh3+O5rY0KDP/4w36+tMlO
+          YhfSDGjEVAuu+WmujKVv/DyTzLgvYIb8NYWJnd+IeU50usejHeA630TpHm8suSup0om7Hom88xeu
+          024JrGwhRuqBKXL8FuUW1EeWoHjHpxX2SymRzXaJl5SK98/fVNf2h57mIwYv2/mpsHWjO/HTU57T
+          9X4ZQF2UWvBzfwdv6QXFBWYv9kT1AsPj6dteYG7GIvFD8PboJCQFtIJeR1kjKvSIXDJAbB2M4BPy
+          OqCdu6XA9IeA6K24Z+QZLgPZ8+oiB3BVvIlH9wadUnSRcTpmHl1dTZamG28FFCZSQ4OLyUhR9lED
+          7mY9mvXtZSYchJbHrBYPIz3rlwWyvnhCzvcneqRPfjUokeMiy9CcnH43nZXoc1uRR8GHzrMtMtCY
+          +ZFYVhV7W66fJ2ipcUIQx1XaUh0fMnh+thcKtonP6fphOjAwGwqgo8reSApUghJ57n/x2YEeBfDp
+          6y/SljpvVgG2g1RUIsCNNhraKs5dBI1nZSLliyDdRmAzsORvD6S0W6ytGLQhpN/Pg2ik/zYr9+kE
+          aUmRF2xm9Ws2h+UiSDRZJU/8vMbrIIMWut+UQVmqvLX1VFsp7J6shaIfhxp6F7wSptcJ7XxNGTlp
+          HVh4UXVE0kLCORkvqgrf3m9Asf45aQtqL4K04zlBlrjElOc/Oswq/4S7VrDBrsdA2A5fGByyRQBL
+          VG0p7D+atX8vGGf3ea7AQb2Jwemz96ixHsfpLx7C/L7+R0dzMKxN/CU6CWcPU/9mwqdTGwQdY9ej
+          f+t5XzKCfO/l0HUrHgFcYU+R7R7f8Xr4pBBePnGNXI0oOTWDkhNfpxtCvrsG+Xp9hKZ0dqsNnQ8j
+          zjeFyCYcnJIgZ/e/K665RRLH8UpkZ7bHNTNFH8gdV+/nHcQbNs1CUgpXDbbx8Wlo4iIGZu6r2Pn6
+          WWMF6Z2BAAwL5r7XaJzHg7aIzpXX0V88y7bU4uB+vpHyq/A4+1tcgygeHHRpN6otNvi0sLjzbFDv
+          +7Nwub+d5Gp0MXkCJl86HKnQmI8jsoSTD3j4sXxJ5FuXeOzxQReVP09//gHpWx6NP/OUVTA9hnGw
+          addW227koZ/+xeeP8h1v8bZ00FGlG/HvFw8cj+mK4Z1GT2Q5awtWP7BscefPAXcfUFN3hceBP/6P
+          Pv3Xm3Y+BA/j/EFWflpHnF2SRPpoyEI2w0vedrydAxjXt2uADzEF8707hyf2U+vkD//5nQ/A69bm
+          KOyv1FvVD4xENxSVgFcEjW5nu2jhZ+QuxCwuU76937EJbbesA1jMjbbqrV7BVZt05L0Plxw35pGF
+          ZfrmkMI3tbaR4zpJ+P52iFMrLqB/eHx832pyXnuRzirzXiT++KPotvNHdtKCAbKH20jUw9Zq5I/v
+          J1RAeHt9tYbf1msJT4TlUbIIRcM1psRCNf8SZK83n86wXwpY1wcl4PS5z3f+V8DpdrRI1FT6uFm9
+          Y54u6nHBmwerZlGeiijFB3lAl+oUaeslKRe46xmYkt7a5z7vPRcy2ydRfNynplq4BuZGQuTv+LIE
+          58SV/vQaxbkG3nZsYg7mNquRhDL2yF3CGEpHXFISZNmdbrPF+DBQsYKlXZ9cYegyULywGGlY9LUZ
+          19wG/v7/obvV4wJER4ABU5fofLw24/od5xqUw6Ah8z1r8axmzwx2T84iXtpk+TZXIwaWtPikpO8V
+          rCOdBKDEOkROSFe6+h0cYPu0DXKNOurRyH1n0Hh7ew+dM2kotfMIuGI4ITmqnjFGp5sLfRymWGJs
+          IaZZzpXQameTyMzl2yz2bPenrr0byCh+HaADnhMQrKxDHrGuxccvWhm4Wvc9oyjY8aoqRABnefKQ
+          t2eo1tu9SuBQH3OiDf7SzBn2B9hy1Rr0YOJHKi1Yh9vn2wQ0IGPD4/NQn6oxMUghbfd8gkbFQbe8
+          88iOsJpz7DIUQPUil5wvzyyn6JjfRFENB5Sr9RTPpxpl4q43YXaZ9Jw9FrAGyohNZEt2ANaztKXQ
+          aBUdLxl4NxsuAQRuoEbE0f1r3k+aOUCqpyaK7TLzaBM8BHDq5BBZ1t71nKNbCOs4lYg1DDI9Dn5W
+          wPv33mP4wEdKH1w+wbc3DiRw5G78Wy848hcN7fpqTByY3cSL4IXo/FOYmK71Up3kjq2RSa1kXDkm
+          SaEjWz0JWOnnrdwmVX/xdfD7/jLvnz6+lGlDgqp1G/590liY3b0NL0Gue1xXaCz8URLvPe6/2nbE
+          4gT38xRchmDwSL5BHwZRaCAt0TXK8eArwt1eiEE6rE0ZDGXIufP5X7y6yYdig4qQFcjZ3nbMXuph
+          16MKA3nLIo/8r5IzkdX0iKi8Z43LOg89+FnFCZ3Zr63947uMAAMUjH0IFmPpbTgm/AXJRbNqi2sk
+          C/zDV6Otv2Db9W7wt99q2t8a+hJpBxUhLZBhXIC3/PG1PX7C4pfW4I+vS6djlxAXvbV4MQ5nHQYo
+          vaH8miZgifYK9LqWlAA43QbmXZ+B+3kIAKzrhg7XAwebluHJzn/yaeAeKfhS7UL20tFxnqVElUIt
+          05CsNVew9NMQAVQxlz++pW3eIpfSHq8ib9dX5vV1D+FjNVdk7oPSCfd8R0ApbJVcSRd4m38dfagD
+          M0HB5atprINvDCiHXiNRmwkN7Q55AefrO0bnKnCahfKXDQ6j+Ea6ThQ65VKXSMazNvFaeNw4fd9e
+          Bl434Uns7VSDzfVYE3yPe8U2vb/AUjhGCHb9Dx9Rfh7JHs+dvj5EyNvtZeUdepN2fx+wxYvRxsP7
+          iyGXxZfglNkO4P7il0HoeHS+PMV8PcVDCYefTFHIn98aoeSD4Z4fCqDV9ZTu50H604se2kPOd3vL
+          /vIxxJmpT+drcXMhkBkOz/h5zFfpa08iCzXlL5+Xr+qHDeGAegtlGlHio3PBA+ivTw15As7oon+U
+          UDKPHUMs91XFuJ5WBt4i+0VK8SPEpMjMCez5MSz96ZHTkYcwLUY7OK7qJx7/4u2/980Xq4yn4TkK
+          4v68YBDSE1i839uGyXC0USBwg7etrMPA6+2mIG9e45xMRS9CON3veAsebsM6Yyj8849nxk5jKl+f
+          LFANc0OqeJu1+fu0W2BkQ4q8Gxgo1fXvJOz6K1GtTMrXPHroEEb1glCgax7N0iWVNt/nkQPslv4e
+          BGZwmtoG5ctSNZTXfBX+6bt/fOfob3klCkrbI5S11fiXTwH96zIgRf9u4/L4ZBEIu+8tGGP4bNbw
+          FlRi498nEjyqVludoGrFYRTexI27Yuw7HMnSPeRGLLWW2+zxVCkd2G3CwGKafJm+sQy1cyjj4wt4
+          3lrE31QKqwfzl89o2Js+Z0CbnR9y//iwc6Qq+OMLWuiplB1/6gZP1nkmlqBH+fQ9CSJkbnWBvBhK
+          zZ++AVMB1MjlVLzfgj6m0iskBrGZsQUzHJkKPoyvEJxeqjTi8hFP8J0VZ2T+ujZfvpBhxF2fDoSW
+          Q/mC5HMFG0xLdBYqR1uuF1aWdjwnMn9WvEVwF106nR8GSXSQaZt3zyB8tFkewMTsmu21XEoIO+eA
+          gjuK8vn6CHVwsKuIuPSjA3q1HBFu4wKwVAW/hvuosiqRG7riZvf3lHy7BMa3RxV8nAPXzK8QiYD9
+          VDoKd7ymWXjv4KmQHPJ8VXsF9q6nGMlTRonMyXQdD9oG9vxVMMuao5E6dPpT1mWXXW88ahO0sQD/
+          9Gbd71W6PghMgebAG7obmhMv0/EAwe8kjkROVqRR+pY3YMfH4J8eS/XnMxD9tILIQo/EO+7vB4F0
+          Rsi2W0IJ8LwUPieSon960MmnLtxu2ouYkVrlyyHsQmil8oKQ5id/epsOx+90QH4YMXT+iKMIzW0O
+          idEd1r1VghdAXT2ZSDM5dzzyDkjgrN5k4l5VSslf/BFEkUH0iDmP25dD/f+nokD63xUF8r3KiK7G
+          dUMrD08QFvYBHxLmR2dnaRl4fzYTQawXNsOZP5fwmWVFsJJfpy0OPHWSskQXYtmOB9bT04FgQxUh
+          bgNwswT5osN3okUBDxalYZnjI4NGsxQoMN3Zm4LeSSF3WmJ8TivP46rW70EZvT/4fkCyN2uTWEFJ
+          QRQFPEu8TcyrTgrc2SWqa1jjMXvTVupX7ke8rHqPRM21DjrNGwZsG6gNzZJzB69cmSPzfFA1bm8I
+          Kcl99UNGI3LjopTOAOWsJQFwzS3eNkdhYP/yWCy9Dr98Q92EAcMeA3I5n3VKl9S5gXEJFbzVySWn
+          rXHauxI5M2a+ugb4ohRYODx7Ft0OQQZ4wYtVab7fI6J6d5z/POvdQrOZWGIfoBRP9fshwOnlOwE8
+          xe1eYxRl8AUZmYTuiTQjGPlKGtEvIjmUhZzebHWBH/ezYfI8RJQniZ1J3wpNRJm9heLicq/F4FXq
+          COVTDfiEF0rpwXIesUpuoHi6VJwEXtHeNVKSm2Mjjy08hERFCD9qb5nSqYKFtYboJZ6ckT1kRgYl
+          iA1kJwjl1HGLvYZdcJASV+xIbUZlpPl0XYgtgiRe0CPA8KQUMUlE8B7ZufgUkB+XlZSkHXP2ybmm
+          BGOIUMIXmsf/pCiEy9qX5HbdB0OSRE4l/Sq4xINyGvOZxUaSb20GeuqXd3483V4DhC+U4I8lv+jy
+          4NIWcumm4418So8vg9SEr1ovSKmITo6/3UOGT7+syaVjHt5a2wDCrFkScnFiyZu5q61LtbsMJDe4
+          YJwnpi/gfLovSJ3k77gx3zGF0+I/yD0WZrC2ocTCmzovxI9PHzofKphJm5Ma6HmQrjF/A2sqSXjx
+          yPPW3LX1mx4mGLMnE2+3D/KWKDlEoGFSA916+PHYQ1diOD9fkJw1jgfUM30RBkx6JuecGONWeEIN
+          PqcGYPbCv+LNN20fwFVwSPzyWY/CUnal4xyvwdYY3IgP5j5F4KVckMMeDLDInqlKv3DbdvscAT6Y
+          QgpiFpgInWagrfVMBDjV4QuFiqXELE8NVWLve70MG1TNkbEfCzxzfYee+lfKsQsKAWo8DDHoUEOH
+          MQ5EyDxxQszf7Mb0hHIIc0URCGJS2zvqzwMH05f7Qxq4fsbl/CxtyKLHC1mmfwecP7Cu1BzokciJ
+          C5v5DesWIhCF6A4WZTxKYW0DjaIbUk5jmm/LJbpJG8gaLFTkRrehdjoo9k2LXNYVRnr3tUk6vVKf
+          BMOp3BW2ryrpnzMbHB6EePw4rRVo+28SiOVD12gJigxypy0O7ke1b7Y/+yC6M+93oB1KP1K/QN9x
+          jkS5MK7HD/jVii3XBOhcR4+YxUDM4Fe5NMjMXs7INuZZhOComuSmWVK+TuPEgom7akS+s3LMX/vj
+          JrmJPQSHM7Ka5ek2LBTnsCQKtfWGNbjnBiPHFJE3KVs8dXa/SW2Vtqh49b2Gk1Zw4T+8Wy43j7U+
+          kgoZHjTEGnSv4edlK6VQyQDy7epA17lJSuhUry/KBtvMl4lfS6kFoY3CweTyZdzn9oggSImex5eR
+          Pd0yCLvqkpCyug2AjieVk2SuaJC3qnm+vuRNh4EKAPLrmxLT7phmsI9Fgs5JaOXreoM9PEIxJNYF
+          ZB7Jv4r/50+IFfAtWO1UTaTnXoOqXdxaI90ZmHBR8RVpruTmPJfTTVID3yWpXZ013oomDH99pxBD
+          RdlI4vkcQKw4MobU9BvuN1wYib8OFim4z+jN93OfwgPCGzIv2j1eDO65ADtL2yBW566h6rcQ4I7f
+          RGXDMd9+DRn+nV+jGqG2nJfHBn59q6BrPkSAnXWXg3V7OOFj46zx5lpdBOaX/UYR3ruQfrevLIV9
+          w/6dv2arDml/4iGV9zu/E1i6wFpgtAoPFARZSOfb3qXoU90gcXr96G3PpS2B0j8VEh6PwkhOmtrD
+          e+CrqPjeh3hbmFsPj6OokIDZfG/xHk8fYsWTA+ZyQvHSHWcdegd4JM88KfJtYj6uxMTsAa9fwuQU
+          zfEEL0XAo8B+PjW6OmYvVf7OiFG4AHriThGclChESgq/2k8qaAmF54qI/UNMvhxOUQLPFn5iKOrO
+          yNuf6ia5KCf78ySPDue2hRWGJcr7OqY0s9hQMtTtHHCKfslpyr8ZyQF3D1kfrR4HP3pUsDvKAUqP
+          D4dijPVUstemxJyi73OO31oqPT+hQtLOczz8WEQG1ilugwyng0eH2gzhjmeoOHSJxknQ0CX0nHmk
+          ZTDMaRAEC6wSUyRmtFg5B+LQ//e9d82uR64wTB/KiLgkZr/TuMnSHUqugALkhLcJrNXK3+CEG7pP
+          leibjSi2CT/6y8bwe3dzfl0aCF+m3xNXP5sjT6dkn4spcsRZSO0R7wRqeFsiJziWqa8th+ycwazj
+          fkj/LbM2GBplJe1lr6j8Wz/VthjImVaITJc7jFv35XUpGiwHC9f1RFf+BGVYgruGaVDYlFQDGGCP
+          txgF3nQfsUu0Dm5rEyDbuUT5ZvMxhE+wW8rLZzV8vmWltNsXkcXKaLh7VU2SlRU+cR1y0dYPLPY7
+          9FlBFLGQGjxdeg7i2kHoD/+3p1bZsHLQlZydyI3HE4ohHMOxRnoU2c06c6EplesLIz3V794/+7g4
+          do23d+N7//DQqW0fr+7La47AYARoKEaP9vUdV3D6VPAIhZDk3xXH28wVC+QECyP9GRrNOuRZ+7ee
+          yEubGfz7/d6Lkdj7+sxbMIVwqXqEZHlKteUwO7KodVDZn283rLSGJWRR/sLUw6TZTosbAfZiGkRv
+          23Lc1CTpJD7FBTHva5hz9FH4sCsCgjyLqjFHPqwsOR//TcK3cc45cr6kEnvRDaKi7+Yt7KjZ0KtV
+          GLBfNaYLOPUsSLhaQ/LA5w2RX7EMd/99FjxiNb9WbysgqUKKLjvec/cOYRA/Fx8lMxpHWoOzCF9U
+          H5Fx5JZxmwevBB2wBXyKYnlkRWWNQDxbHNGppDc0/Hg3YNNAIxbqXe3P/oDlHmeSJf40zjvHgL6m
+          eMR5fk/jhrI4BYI7icFR84yYLZishE/mJRNvPldggUPnQnHwHeR+s7O3/qDRwcvXLIi3ES/+Z/+B
+          UnsoGJo6Xj/NWsAxSizi5ReT7u8XwggYKblVk5LzZtBsf89HuhSsDS7ZOYWDwX+IbukW2D7cukh/
+          329dgKjNdF4DuE+fxuCH6/0Ok9tDwD04hF6esl9We5kQMckH2dqrAavtrTK8FD4fiP7TzVe25Gqw
+          uMBAzlp9822FQgfZRQ7Qfdp0wIviFgDCbjkJzj+vWeZiLsCOV+jOaJeGOsethXZKDkT2uWZc9TTk
+          4NrNF+QErOoR+1Ml+90pP9jWetO2vrT70x8ftExJ0n5evAlwx3cMs+LRzFxw3WCtDzIK1BGPK4nA
+          BB0zSpGKWYHOkh6yUIqFLym3SB/570ZUcbdnvF4YV+NY4W2DDwgqJJs/D9CMPDqgLZ5CjMf80pYt
+          MVs4+p1EjNvPAVQaHiWoRtEhaqbPI01ayP7xS+Te869GkrSpAMtce5TLSeutrRDb8PbJ2kB4ju9m
+          ca9JCNqmOJFiu1oxx/swA9+vOhF7HqyGU8J4EG8qWTCnrkp+fMOhgx2Dzujc+Yp3dCATQpt52kgz
+          /Y+3fbuLCv/xrSm/5as3PSsQXLs93uL5cRO/qyhl7NKQP/422Vtwg6p+O5JbsT3GzY8uFTxm54LI
+          x6PQbGet5ODpWvHIuQh8vknNUEBBb0MUf3nFY5+/ZQGGJfTEJ1/sjdeiLf7wlsS8oVAqHJwe7vYb
+          AC74UepFZfAPf551dIo36X5x/+IlhGYmpeNwVETJcBmJKCgM6SKUzQ0sPz5ArkNWb2GUevuL15Ca
+          39icsofU/sPfgPUSb+SiM+3E/yPtbLqd87kw/oEMiiIx9F5FRUvRGapKq+olQT79s/Tcw//sGXad
+          dUoje+/r+iWyPV3/F3+js/XFpVk8klM91c2Sn86CrL6zEbkCbPS1eIUOLN/eDZn8hwOzYCn4V++J
+          026n5mLVWGHOYkB++W7x2UcmVSjZ+e/7pwWrAyLpT/8Y7KkbFmLVKTQNV8bD5qfW5sF5IMtlg5y4
+          Waar2ZEVaL7vkJ8fX057ZYVuhkZ/zYxTM7/vWgHSLvki9Mnf+tI85kpOfF9D9s4V3KUTvF4ajPSE
+          lP26rZihGkL+rd+IeljWfNP7DhSu7eOnB/XprOu9fGNZl0Syfx6Wk36xgcC/tncoTlhvX25ege+y
+          qMgpb4a7alHRyrdCGf0O+ZbOt/m0Ssr9a/z9vrVIVR6sVbVD5eNthrwwqrHcXq4Z2vjDQIXkFcFB
+          vCso7w0mnAzz6sPkHUFkGp4/cIszRGDiwxxpxveaL+YMCuldBIkvquKSrxculWD20G/kYG4rnGKc
+          dFB/OAvSLHdH5+oWaDJj7zLkta3pzorrK6DICoNcb9NOXzfeAGLvvZ35E7X6EoZXCX4s5Y7C7mC6
+          lER2KiO1bHCTikbIhQev/cufR9KE+fK7Xhl5DjHWgB3ozy/bUyH6UuMc8+6cVN2fPzqcHBZgH8sG
+          jMI9Qv7zMOtdc0n8X771l0bSAVUB84JZfSDECZ+PcI7z0JK/eBxQZCYJ4C0wpNAqdyJm5FZu1pOj
+          OnLjTBo6sbu0mZ2JF6CW5A7S5e7aLDsqaL/n7S8bH5kbyxTg5ucw3ed3SmS+X2GW7wzkz9eAkmsn
+          z/Da1U8s385zOFXL1oUAXGyk9yAO5yw3Wph/uZiYzXHJx96+BFJ5/UjItqtJp7tWU+AFHFJ0wFxA
+          F61bXvA6tDwWdm7q0p5TBfmn77d4ztep18sfX0LZYjUhvek0kmMal0T93OpwWrWMkTb+gFAD22G9
+          650NT9cuQebplAxL/YxfcKwvD+T4s+TOcxprsLUskzg3nOi/8QRBdTWQt0ily+miy0Opilnk0mpu
+          5libS9gFfOZLh+K5+XXbgVDQKHIUbLtkvDP8Tx+hA7imv77rLDg584ocyL2HpUdqAMU7+fgtlVQ6
+          ls8u+/EqZIKLRpctnkTHu2Y+dxWMhp4CwZYXr9JIxr5VsDDqp4B7votJWhYtpf5T294RTQfkx8fQ
+          HTceAV0MWz/ZoUqnVJhY8FHDxhc3f0KNPbEl+NFe6Hj/3Ib1tD8V0G82flbwz2b5KHdDurXSGT1Y
+          eKUryvIM6Nltj7Qre6RcrtovoHqfEzI23kOLfV6CMotZpF4Iyenmd+R9IJ6Ia31COv/0gwMZG526
+          x6JjtXjBf/NXHIRw/fGLa7h88HLWdiGWnocWNgE3EXeLb1IMjg1+zws9y3M+tz5aYQNJhXQDLeF6
+          O157SJk8wHMV2ToGrOvDqTA8ZDtMnZOXKXpyIt4qcnSHx0BvgddD6ZJJyDls75zZg9PD2PuYeF9f
+          j/k+Te8aNLX5hMq+0XLWOH95cCu0ceMDYkj2DzjCm6wAolLxCwhmQQZ2uANISW6JvljdYIGNr/pr
+          eriB73M+B7A+4Cuyn46i49u8QrljJQbZU/9pZk16+1C5D8Z2qvOOzjSbBXn0nlfk9aY+sOdmYWRz
+          GWOy8QtA8yoWfvkTuT+++QilDOybl4qCrX6wGs4FOE62ju7sTmiokIwR3GtyiH1MW/evvt6M052U
+          ixW7k5/DFUhNfsS7febqrB4eJLnI9yzS7/23oaNZeJC9PDjcbfqcGsythK1lmCTmCq+ZwsfHh09+
+          W3EstDEffZst5S3ekSU4JsV3JJYQWtqVHKvrCrb67P3py6uGpGYeJsuDkxHIxKlG1l3d7CnJU69Y
+          yEtSI59zbHsAF8qL+Pn1OazhpFRyU5WeT+/9saGfzEvBxjd8SW4dMPvdMQPK8GL+eAO9RH0LuOm8
+          kF99H+lH8+CJ71vifnEdrmWhrdBs+YWczjV2Z554Fth4GAlGFjWz3ur1zz//8eU+K84XeEwuHzzz
+          rUxnwSUYbnwQ6bHr6svCH1tYeLWD9D3/0fuL2gcA8U1MTMmKQh48nrWcqEFKNh6Uz4lUaX98zJEO
+          Xs4L4zEGthUpmMvE0zCW/rj+1Qv9eezc+bB1Edn8BPHW6yfcj3eGla4h/ZBTMWPapUzHwF3WJsg5
+          rLJLjdW2AVlGi2R8dwLrG0UljAp9h3xV+oabnp0B/jIlcSYPDEucJzPo7rNBXHGyw8l49Tbs2p3h
+          79v3SefGR1iJu5Amf/p40wcMeLztBV0NOQRUaLgOvB/xiE7k5OhLcqoyeEvOK+ZrMoVz+ig0sPFq
+          H+zQSrHjBy/YL+RGlA9t8xWtvAJtr+iRRt6lvvD+fQamr1+QJwhAXy9qfYEbz/NF99oMa+eeFKhk
+          LSHWPHzyzX9D+MV4IC5c5AFvPAGiPlDJAczPhisEwYbbegay5e9p2JcgSqFOTzE56HbdzNbxVoCN
+          15B8hy5gidaeBfUTvIhFEEuXqyY78BJIDt4L1n078stm4cp+mk1fCnTdUYhh9T4ct9/f63O3XDuw
+          y14J8a75PCzHI+phwVQdOu1kLiTLNWFgeX1LyCgeSs6f4kspb+sjmPOKSz6z5cLK58NDRpp4jkL2
+          cBtGMBjZCXPVtXLXNE1jaDKdgi5mljXLfuoxjPG78derbgxb/PiQDxiOHLr6qa+99PLlz9FYyaV9
+          cXrnHB0IN/+BzKkd6ZKuRQS/WX/46RsXy8nN/uN/rtsVdOK+3xVqdUBI8urEEF/qawCfyeGDQXkt
+          XLx/PzS4+XGf7t7ffJzTUoM7r78TtXMOdD3Lz4t8exxUHzNqC/Dmt2Cu6sLWpeYVfoaq8iBi4jfx
+          g+QFaJBKinToUgt5knFs9ls8Q4daFGeQMxt+uTcdfAaMjKXI84ZJMRsM70l8QL4cqMPcXBJPxu/T
+          i5gnpqALdw8K+PvsiPHHfR6Phw6i5q4SRxLOA63VORD1Xon9TguvOX5CcgH07Z3RuZEagL98c5E3
+          f4id6WPQP77MYwmTQ2a+h7WXRg9yGSp8Qa1nlxalwMMkeyTooELWHZ8x7sUkuyd47i0+XJJW1+CN
+          RwApfKAC7jWqvByrkYKUMmH1dbDkFyiy0kCeGBo5O1RQArR6JMRpfWYYZldQYH7a1UiNd1SfL/bE
+          7C8aDMjJU5pmcQo3+K1/YC6zu3xOQg1DMHk8+emDeVCOAtzduxGpR2yFhM/pDMNg27GzrWeswcXV
+          YNS9jght9WHpXDuGW3z7rHpgddzibyr9/BF8fZd81XcrD2I0ZL44E02n6bzrQfc4ssRdNBBSKrxZ
+          2L9LgyjqXQ5bnj2PcjJ7T4TkT03XvMo8uEAOINd+fSitLjcFbHyJWOIYhPRlLp6cV33isxJQh3U9
+          HiEMpgtG3vLSdM4NV0ku0OuB7sdLH85cva7QdSTNFzpRy1fBGBVgzvqCtK0+r272laChuoHfJsuc
+          0494i6VDS54b3xqan76Sktl/EuWYmPqPb0A7iRCxn1Xh4p33SH88nqBNL1HLH1a4xeumJ/cuSedd
+          B+qXLGJWibcupS43wiseXOQEk6/v1cTp//inTbyLS6/Fq4Ts2/niaA3Y5i/fbDz3X73Rb6svN3Cq
+          0DE1iU4LV/Z++h/99NJS2xTKH42xyVEVl5CM56mCY3ZHSO2cD6BnI43E9yMaEWLdeaDhZPrwt94k
+          Kbhz8V4ZjL/8WJasonN33prlzGE44lknRJfs2gdw8494KXJzoK+LU/30iw93LatP+HNNYaCmAF1H
+          7tDMYpwxoBWMBikfTXV5tDIK3DtrTQ6pmOj7cyEx8HDpZXLwbqjhuHtawL04Dhio5jecn9/6Apcv
+          bdHGd4Yn/ty39cvR3Hjy111vx3sH+QBy6OEvF3fp7RgCG6SUaLvdFNJW/a5QNuFEnKEP6JLfGk/e
+          nhf+bn7st/4KtnqMrkmGdOJOvQfNZi2Isi5N+Ke3dd+pcV1HS74+5zsjvR5phCy9tob9T58WqH0g
+          7wxXd+K+z1m+CLqOFHlnuGtQFjHEeyElwV79uHPKVFAOqsRADvtWKYttd4ZudhqJXqUaYLf5Kuvz
+          USXGcVX1njWkGuZVlyC/vMVgz7rfClrMK/VfM3b0ye8sCy7XSfvpgZw+Wmn8f3YUcOx/bynIVXDx
+          P6cp16ddtWBgJubJ399WP9zL1cOBHxieiR6TuZmvriCB6aN8kGl0LuVEOXVkiGuHnPzezAl4lCOA
+          uHLQYa+ydD5czRSeYl5HTpPKOo3ydyXVXGsh+53O7pxO50oOmXNDjjN4DvQRpDaULQ4ho71F7mSb
+          LAvj5nVGZwwFl/qJ4MDYvR1ICeo72PS0APk9pkR9izxY2itapc64XdEBjcNAHmfDlwF/BpiJ0bUh
+          5IYwDE98jg4jWHNKxI8kix6IsJivIKSDVUuysY4H4t7mL1h3T68Hw2qdSHR0pmZ90SyGLLNEyEov
+          KO+dVrnIQAp2eI58w12o2wpwXxkMujwjA/AvnY9lXr93xBluuKHRcM8kJizfRIXGSFfLCAO5YLZj
+          G6/zos/eoZ/BnD7PJLh4urt83p4iK9FTJVdRPQA232kX6LuPAHmX06GZy/nCyxiMErnt1Jc7oXRN
+          wU2ReBxpdj6wb7Zo4SGEPbl4R8Vdxr3qwEbdH5EjVhrdi/fag5d74yNkf+pw5ou0g5dbitClz/h8
+          YYdQk7PnltJ8p9Pn1ggEWTKjN1JH0IHlE1WWvNhsidR81sFbp24rvx25I95Ylu5eUUcB3kyVI45U
+          X4Z9o30dud0R1V+2/5/byF6BexQrcsxXkK/MrBQyZ4EUw91TDXnqmqs81q8cJTszHvjYGxz4BgFH
+          Ct+LdW5/f3Rw7l4J8ta7P/yuByaNrUmAC0fnb/6jg73cK6QsLp2Lb9idYfzIXOL1Nw/wv/HHhzsh
+          RRwkgNqc04NgBiIWD5/DsGcdUYHHFDrEg80SkkeQOpJcwwRZ50DOV1EeLJkkvIac8PQN+TtJX9Ai
+          1Z3EyaXcNqGjGcJ6cHwKlK8+CpOswViPHPRISDSsRKgkGfSKQILvzQb79Si0sOhfX1Iouh6uR7Mz
+          oB9+Ir+J8DOkz3Pawvr6VEjS8qq+fN9pJj+fIYfpcZBpz+aOBp1bmRIf1kQn7Yvj5ea4uljobvMw
+          fTgCpfTVTsgTqqe+Toc0hU2s+0gnwKJcUkAb7ARbINeGO4ecIO94SJC0QwrJj8OS3u0Kfg6ciQzV
+          GsASDc0LzkbiYX66R+48CWUE94UpEh1N5cB930Emm0RB6F6xXzAHt7CSERILpAnPftirC2Sgmesu
+          UZUWN3Op8CzkLJygewcAoGWwYyGUlhPyQ//q8m/WzyBJWM2no2Q1bDrdauh8vzfkFqw0zDjlX7LB
+          cxY61Bwa2FfIZrJRDypSBT3P97/vM3PVJeqpf4bT4VVVcGb3GTIc7x6uJIUrnEqhQRFTH92VHd4Q
+          aqfbDdOCC/O99cpicGC9BRlt3FBWuwSl/BC8HN2us9RM6d2uIfQvOsmWQnX3r0aN5Uz1dSxt+W8b
+          r1YG0mXnf9V3C9aOR4EEeEdBxuOsNlzH2Rdw0tgd8uvIpoTLv7GcdModFeJDadbD+1NBJcg0H/DK
+          DDjLUGL5o4ODz33EUN8/KtaSf9/nZqvYTHeyYCiXuwuKuvyqz6dOY2AKwYLsNtuBFQtFDUV7vpLL
+          LdBdlvZ0lLd4Iula1GBW1HBr0/Ih6OCP0oA5z/Lk+FMhVDjHN8CXnjHAHKUX9JjkDKzFI4qgq6kv
+          4gV+4q7MToigrIEzcW80cxeuSDz5ogYeiW5IoWvH5T58Pi0WL1v+4GOvceSwPVYkCUlO+WPsafAr
+          2hgn3706sMlkOrJ2TktiGAcv3AffqBINaTHIoxcdd66W/CJZpL4jV1SXcIWPbf4SaJJLcHpSKpZl
+          AJOHcENqs+fAItxfGqSOCvAUjFbI7e9JDyO/6ZHqdlXOn/aLBtn+jUhGMkX/noTTCswL4yNd/RQu
+          jgatgLYXsajIA5uyHJhTMbumNsnAc3UXrUEREO31Svw66uj45e8l3OeSQJz87LoLuV0FIF71lNjW
+          oWtoxlygPDr5GbnLbmvEZj4F+fFCCTG69AUWkj5Z+eU8Mn/f8k+X1uuLh4S1XXL53K6Ab+iJgUyS
+          Nbip95HLtwYzQ7auLjhVjU+4UnBnIWVCCy8jsMGCBbWVfbrP/d1bvOn0qVajZM+LSg5f6Uy/wpRL
+          0jZfyEF/2S5LxHiVh6Oob/m+AUs8jgrYpebqfyl80kF9nhX5/bFmX3xyJ7BfRHiBwxFr6BCRk75I
+          vl3DUJ0zlKbnySXuXq/k6lW/sMBVYfOXz283RSTFlm+3RnYx7DnMEPtof93f9WCtPiFKFqCH+/39
+          zsIpY97IlbUarN/4U4Pz+2iQe7l2dLg/aCc3rRGhm6y/Xb5EAgtPAzr+xSd34IxCfk5b18hXeh24
+          vUcx3OYzKcd7TlllGXiwnh6Dv75SbpglLJdi3LRnv9WMo8t++Wv50yckEqouJIyyy2B2ex6IqzWG
+          vl4Lq4dx2LhIvS5hTveTOcpcg7ZN8Q4c1ijPWqh+TxEq+9QZaLELU3inPkB6tCwNZeYwBaMVXIhx
+          OWG6+p3qQFa7uMh5JyewZ/OCAbrJ9+joTWL+SeWsggrEDom2+OBtrlllJm9qoiV+rVNyQyMkqa5v
+          881ufvkStGDRyfEroqZU1BsjzlURoRDoYrhUTy2Ap+F09Ad+N4cjEeMZwiw2kZ8FH/3LAp6BQbdw
+          KOXuTb5aJoqAbae8v7aJ5fIygRbsvz0k3vUshLRAu0yykneOheA05MveAyOEdXEkxVISd0m8wJM/
+          her47w9fA9q+OPY3PkR/ckbO8iN8QTdWOmSZ9o1OS1ZGsD/XIlLyC7vN+lqBj4PoI81xUDiK8nKR
+          H9C+IWNf7fXFFy42WLmeI2ZKHwP95Vd6cXfEGG7mdowd8IBP+ww51pFvVkZ5FZADdkNux689LMfr
+          YsjvQcO4eXKvnEbg3MmmP9uYMe21GS/OaMBIu3S+wD4jF3Oe7wEiGEfyq/fL2XEg4B/NF9lDojTs
+          IiU+HC7qAVmOC1ysaVkGQ8l++SGnps3CBHIJd2ieiKea1J1H27qA5Wu8ST5Bos8xtCHc4pH44tMf
+          8HRIM9DqqUiMS7DT5+569ACVnAPe+V/dXehg9dCk3hXZ4tPU9/Uyr7K1W1/ILsNp03OdBQ+e/vHb
+          9f526Vbv4AcaAblPwKP7Tb/C4xudkPZKr83Ke5cX3FUKR9Dhs+hLVipQPiWqQS5zdXN58NBXeEgP
+          X6QOwiFc2Zw3/urDIafyMA/tAuW3xX/xah6wSyP32ELTX22iSA10V+EUavBKLAYnX2mh9w8LBKmy
+          88IXEn8YpjsRMVy8ucBg2T1y+jh7HthViPcLy3y56yxGDPwkgoviIa3BcmOKFYATmInmeXNOlltV
+          wpHh7ijCd45O6uraEGq5TqwuqYdV8vMCtqDVkH89vekwJBYL11szYfAF92bu42MH952ikHswtiHN
+          SWLDsyHJxNzpdrO3WMcHDR9e0WmGp3xpGt2G151tEKvLLoByw9mHV0cQyXmrNytfaKUM6/JIvPs1
+          0zd/VEvZKTCQqb+dnH7M0gCcz2bkWLJTji/Sm4EWHRtSvtNbM7yjUwXiDwmIi4E/zDGUFYhC4Yki
+          8LSb6U1BAU8xq6NzJyU5R/N4hoEdT3iXkKhZLSMPAKy/DolB1YbLzTcwiCm9+KKw00P2WhT8Lz/5
+          nPUF7kp7UYLWfl8jlM+Wy3+vtJBrNd4Qz7el042JVljmqYG8w4ECUs4NA5SuNjHcPtPNfwC+qTCx
+          p5s60L7VehkfHoQkVidsh854ATywW2NW5n0LV7OhmtybtwUp31sHxs9rMWSXDwjKax7kdPmmMbxx
+          jkqsMX/la1yEiuzWRk+CV/wZFohk5afPffErkgHXmixBxS52JLyzK6XaqsZyCsUFOWteNatwZxj4
+          SJiIaDj/0LVabiVs7kyAWe7xAbSpWQjeJTP6v3pKv1ZcwfV0HzAzZu7A9/sDhvVUaphXzWOzzXcG
+          OuL4IMnmrxcWMBDS1tUxft0knTCKG8NKiR7IKqPZXS/SxICfnjZlI9UXUQ4c6J6EFwnRVDYz8J89
+          GMt37O/d7guosiiSnFYdQWegDvk6xPTPf+E7t8PN+qzvHRTG4USsJnrqaxmMEliP8YjZR3Sm/Xa/
+          YrubVJ+nhdEsP/2owNFBWeAdde4mdyOs5uWDLMvuwU9vQS4373gp6Dv8XsdHDzjgNMRdoeJyxUPN
+          ZDQqL3LspOCnpxkgLR0mD0bD7ro/WR3Y5rsfZnTNl84cL2A4s5a/e1A+x7sZ1n/1S5fUY8NVC/ZA
+          EhxypJ6mZZiT+z2AxF0VFMqmFOL7jq3l/voVkDnLlsv9/MnmJ5G+MxywbnpOvnG2+uMRLl2dsYOs
+          FrgouAaUTr/xzabKRXZ+OeVz8N2XcFFAQdSSjRpK7GGG74swoMOdK5sZElmC8fD1sbj9vuU6Jh18
+          XT4lQb3Yu9NTi1v40cUDMvjykP/lN+6h53/8YHmzn0o+UdvDlWM/w/WhcJrcnCWEfvVubiNlBZ2R
+          X302O8/NZJv2DKXzu/PFjOZg+XJsDEt0l9GpE9xhuTh7BVav6oXusjFTEtxePcy4fUAcsarpdB0f
+          nYQeTE2OvFyHyxB3DNz20f/yw8DzY5uBPeJ54g6Sry+nZGHAy4pypCxVRmfx9PWleBj8jQfY+mq+
+          PhY8yg4iVntVG/YkYAmcZcMit53BhvQseTwUWHr2JdW65oPD6xasjHuJ2/Np0leHm+0fP0HKcu8b
+          qof3Hr7atvYl7rGE9KFUF2njJ+g+gZEu55vEQj7PeqS84k9DxXvvw8vURv5S8deBf65mJW/Xw9Bz
+          WzA7Vh1LfvK8oIgUXDjvFKmDBiYacd4Zow9fsy6gpecxOibE1fsEAg9OjfX1d7flOoz2mxshbY86
+          MotFAdxduY+i0FU+QUpTNbN2CQo4DW7ri5s+4HMcOPLmz5H1vAbNstzqC+yUICZOGgIXN9Rk4Ltx
+          VqQP0jXErJs5EGaRSZCg34cxw2cGgJM4k9N5eIM+81+aJIzfky+k55NLE9lyoDlYZ5+RKqzjMPc9
+          EK0rjym0JECw8CnkxwGfiOuNiU4L5EbQNXc1Ocy7SKeHqC5kcYQZOsr6W8cdjy7wMF8oQpMs0flg
+          fEv4aZgTnuvIGtiy6iHop1ZCjvTYXmLOGeOXj9FhZ/INZcHRAL98kBCxDim5HTAo9YuKkP6q6HzN
+          q1k+nnXb1+2vr2NuWCMgU69ECl8NYH2H60uOPzXCvWpx+ct6cQEMW7dCh+tSALLp5d/8Qk5zHUHX
+          1IIPk5zJfXHzG+tzNWuIPxP/xw/Wc3aapZ+eNCXlkC9Pta9h01rRdkicQ3ncjT08BMpITOWThc/D
+          O2rlKzV05ND7HJKjSRWZ2/tvdKhYJZy9Q73CbfyROkl2zu7UFsPAyAhRZQsOqxEyKwzVNUPaM7nq
+          o4TVDCpBqpHToaP5KJXnSi71QCUF+0DD9x6cXtLJWVWfS+muwQNfKdBw7nu8YJjqa8YoCnDun+Rv
+          fD+Sj1PYpZ8MnXgG6+vGG2EyX2vMt+nyq6+zzGnzDRWWabi8x8+tLLBt8OfXVseKbPl1eZf4V7/5
+          WlN5GUr0hIzNf7PFjo8gj+IO/fzn6necB9Oi/P7FDx33RIFco1Tkkc+tS/ZQfkm/+Vpk52O4HAw2
+          k7U3fW367hMS37aDX/yQQ7PfhWs8MswfP/vp2d7jty2eTBv63smxcu6nb3HT7pFN8zqfpuSZwYe+
+          IqIujOPyJIUzpMzZIocHS9ymu2YR/LB+hJnn9ZgvkESrfJNvu83/aWDGQo/BZK0VBsbbdNdPBBRo
+          9BOD7DPKXTomoQciK/8g90zScLa5dwsjLehISstLs/GLWMqvUkxUDAUdz92iQJZ5hcSX60pff+P7
+          gM6NuMtul89SeavFeayv5JQv9o/XWWDjx+j+YNNmjSHmpfKQK/hxxe+QdlzrQOeexXhXc104K+dU
+          gxt/R8eZOYOFUc4VyBZ8Jv71ZFJeZB4OlHenG0KrPNBer6UW7KoTv+lvblgPb1L9eA5R4FvWp3B4
+          +HC5mk+ifAvRpXcyVJz4Nkp/dY+YcuqqO9vi/ER8krt0XrN9CR7FIUTo1DfDMh3EGlgUN0SvWR3M
+          +7uH/+LP67IiXx2rsEH2fZ02fS9u+uTNSFt9wnwTPd0VPPQZbn56qy9ayJuGL0mPykhJyj4jnXod
+          GsEs0BQZ5fWSj0vmptDKmxs6Gt+kWb2PMgLxecj85yQLzXb/NXzvfIo75BzpoKliAGVgZUTvhfeA
+          Q6cOYKbeDOJZ5kun+uX5kp6GvLVli0BI7g/Qw9j1LKIoJtkOlVBtuBMSkRwlNXS3+gfBxsewWVN1
+          2PdxtkpmLHmYnR8x3f/88ounjb/FG6AQtSx83twFM6H7oouugQrqtU1J5LnYnY2G5cHG05Dhub6L
+          n89vCy10iTGTnuNmnD5SCrZ6gU5nzOurcMq1v+fre86rWTjoxtKcxgEWNj+98YBM2v7+N77zgRUr
+          YDqsg9IzAu7Gh6Q/3q4HEwJzTdEMmtvOQ4ZmuiGX7+4+sAgpMXuwx4HSPl1/PAHdOmnWN945g62+
+          Yma5j2B+hUoED2OIyS8/k8u2pfB81CGxkkvpdr/1AYZMH5+7zVVInTb3IJ8UI+bXu6n/fT7t/JO/
+          S9fY/Zvfv/HJ9uWTUh6qFWwfhUSOwk7Pl/Fzu0BbGHVUsPWDruN2CPT9wrC+OAiV/ud/F5svfbrp
+          IcyB2v/lI/TjQ/Nv/kjvdiTGPdSGPdzxkjArue+LJTuFG39fwcaP8Bz6Fzpv+gR2r6Egtu/2+WSF
+          sw/ulrD6IymwPlVPJoJG+9r5zHL3KN8aDv7dP96f+p4uiZf68LX2L6LgondXblhjGK0zj4wkmN31
+          ub1qPLbVDV2yy4ViGVUXefpoH3859c98iD3GEumZYfDMPkE+XaRcgSb1r1ubHA9MgeT3sNy1R+Q7
+          djOs4OHO+wT69Y9vUQpJxsB+WgAytdYB3CtCMeSbGiPrcOjDnhwNARZx9sJyQPiBIiGIgOoBlxhT
+          bm3812HAPJIYmU9W0+nSGxncPvs0o2u4JLAK5C2f+fKVHAYK/HWVRe78IHnFje5W72ZoJZ8cQ1zc
+          9XWrp9Kn0hDycCG7y5R8M2DPLwWV4hM3+BuTCmz1F8P6OrjTuY9X4Lv34KefhkWn+gv86u/BHSR3
+          2t8fPTy/oE4s5MNwLueMBfhwwdvzrZs5zKMMNCfR+ekxSgDxXz//gBDHjCE9i8EK+2uxJ8cHHcHK
+          upktXgdX8WGTSIBq2hRB7fx+IIe9qw0vyo0lz+MUoxTWZsP5wsWR5Y8ZI1fQFjCNe84Xn1dtRSdG
+          n/PNT2twiwdfbAXizoARMUyDtMNNiAN9TjwxBruQoQTF81WfN54Oi3r44IGTq2E9JSn75/e9KZvC
+          LZ5rWKKHTPTbog38Yz4rMtR8DRnO8eRSbT1GsP28GHRZCjvkg9sNw43/oENBfbAOW9sf7V4+/PXH
+          t3/+4pOzR+KyjOp+Qunsy+WnsJAzlTs654/ShuvtOSH/eT2GLD8qlbz5SUz9AdPhEQS2fA0VGWlr
+          Kbg4Z1gDIvb6RcoFhfkSDokPHbm/I4t57V2SjLUn18+3hu5QU/W258NOvg5Hhaje8KQz+ropMC5v
+          gMwtXyxTd4qB5UYpOmxdT6llYE963o4L0k8jytfmsjoQvZzulw+HuW+nF7yoF49svLRZdJ20cMtn
+          BJVL2yyZfzN+/o1cb7MSsvMxtOBwHDWUPLZtvdZ26Nnb2XXE8r1xWDiXBrLBdyYKpAbq3YetGLh4
+          u7u//+6fzbI6uIT1rWQwQNM+XKOBr+TNf/viTn3pS3DUKjiP1RU93nxHl8jdFz8eiwkaVrp+Xr0g
+          fLqDh0zlI+V4019QrpkEqW8x3hoQmDX46UWD3q2/+QmTwxyjogj7nHZvx4dP0duTx/G7unT3fLxg
+          rwbr3/rxcmCDVI7iAWOJ3oNNf5xrGXVGS+x3Grjz80yr33qq/644bdjqvS13aXojllDz4a/eQ6sZ
+          7mTT8+FfPvh/thRw/72loDN0k6intGrmXKtLYH01h/j1ztT3uLz5UFWHC7HBKdVXRhgkCDG5+rSd
+          9ZCDSdvKrTdi4t1Dyf0CwYRgnerVt8NWpXwYDgJ4vZ0B6cqY6Nxr6QLIz8IRHYYO59N0d2cIkJpj
+          sXZGgGXpWUFV8o+Yu+tAn2wOVlDmGwu5dT8NVHK7Fd7IdrTWgczDckiyGWadphPzdojzxTAfGpis
+          29YQKiTDssq3FkqGsb0iNr/oatxUC76exowi6Trp+DvbrWxwleZP2bnVFxQZqyyFfIkOQU+H1bLo
+          CnF2O5IiKE86lXKhhOLuwPq0ejzcufQZCb4544Icm90NtFgzDaj36xfd4S5yOa43S7nGhkpu5aUJ
+          aVR9eHjfBQrS4HwJKe/QCu7t+YJOB0TozB/lTAyfHkNSFD4AvS+5Jzv6C5KQPjl3TZtaAYL72XZB
+          m2w47/q5lB/rhRCdPq/uS+gSGyTkLfurwJ4G1rzymuwkuUP0N3kN1LbTGv7uz/EsjXI1ljPp9/vp
+          9v3Yb/wIQP66IqVstJwTNsu7Hw8Wss8vI5x3LYLwOX0VzLpFOFDV3PmwDyyTHITuo9NuO+FLKOoD
+          8YykAHtZKF7Q9qoL0bolHfib5vJyohxSfzVMO6drP3Vwb68XDOMC5bx3BQYUeZnBnHc5hvyr0kpZ
+          2rsXZFiNm/M9+vqwxJVI0voa6+s6gBlePRb68y0sAFs8DR9G3tBtb82XdDnFTgohnq4kBzUZiHas
+          BHhvT4QcDqI6sHx2MiDxJZOU23yn3Hgb4evrtUiNzzbYu2ZhwBvo5en7NVcwBRoTgE/6yoh9eesu
+          b3Y3Ria+YKLE5dqQrycSQAVHd3KnR1vf53r1kgm2GX/3Ke/5b3ygwkEJJSVVc35gsh72oCLkES8W
+          2L+ZQgLT4AREe3EhJfX5a0FT9TKkPiyOzm7D13ARzw251b7tLqdYy2SoawxeRCjl+BB7PaRyccNc
+          oRXh6saLJFeOLZBA6dWhf5Zx+htfdGiHHmC+OEdyfqi+6BbacbM3Ds0FHpRMIkqfBA2Hx6L/jTc6
+          2V449OLR9OCURBq6MlbXfE71Xvtdj1yJsYZLnycjnJxHjTnJ0cF+8O4ltIvLG+l5GzczuIEK9oRP
+          MH7oVc7WD8BKy5VoxBUa3Mx6QjtAbfmOrnc9d5caRSw8hi5ARzVT831xWUq505cU+Qxvh3s+d334
+          MMUvsmyZ5vNUbm99nfM37rrScPdMIvuwe5DCl95a5e5jXXKkG51Z/7F2jvvixamHH5dIyLlC3FDE
+          La3cVmGAUDYNzRgYXgp3VjlhXh7TgR/u9AIHFMvIT8Sw2TdBfwH7ytXwXvy+wTrfvxiK+TMk5947
+          5NxYOi08x7GF9/ba62t2jrDc5cMTv49Jk6/Rk4OgktF2Kgz5hrPidBlgC7jzmW18ZtC7vlyL8w2F
+          1nbOLNM/bbioEutLKzsPLD7pmXwwSg75OZeGXHyWW5juc0yON+1Of/kc0iNyUF73p6HTDwMD59Pw
+          Qqhq9sO65T/Yjigi1pPjcnqqdwq0i+BNSk0ELhVuZiuz6qXzRQf5+Vj6jAATIasRKr9zuHjWUgFT
+          SAp0NOPWXfXj7MHzEYl4lZezOz5wUsB5dZ5Yzk4O4BRJjWSeZQwSOC+7weNkOPJtEHqi81WcL5Yw
+          M7JIrk+S6O2j4U8v05aOr1jHHHDrYY3ZuZZPpcQS1/vy4Ryacylco8kiRZZG+tLJEENYNy+EktX8
+          m89Sn0Cf6HMd6Xc3tTVw6t+eP/DfT07JScMyaQOO3FNkN2xTyhrAvVsgr328cm5Acy0/+v5CMn7m
+          wpF6Cw9f1yxFRhdIwyTkT1aGDALIvLpLM59rbEP9Jugkua/1sJ45zYJi3oT+7kuef/VHHgDWiaPA
+          2KWxvjpgu19iVudG/8uvpxV9/YurYH0J2ewFI86v/u1av1HHg5ZxfhLvdbfAbO+dDuxfnkyudagM
+          e3K/ZDJiVg2zbRzmXHzmXpAWdMByFOqUQ5E3Q6V8OcTzX2+w7lJv6yROGAyk9+ji15zZ8BsyI25N
+          ZnUpB58prI/X1O+yuzfsS+XlyJe3+UaHLX9Rl/Mi6IQJ2nrORflXjskMm+vJ8Of/AQAA//+kXcm6
+          qrCyfiAGtJJiSCfSBwERZ6DYgIp0AfL092Ptc2dndh5gsYRU1d9UkpqSUynwt8hVuPn5nJgYZmOR
+          wAmBVd4E584l32512/tKmQQaDt4mXy+8fLIVV1Z6klzaNNnyWVYs6TUTp0x0j1+CZ6rs2D2Hy88n
+          ToTdb2ggfeYxTk4D7leHfhvY76MTudNfi7b1HRS/+L3xmS/uHpclxgDIQm+y1etylg7irLycfY+j
+          w+/Vz7f5ZCLuZArkpDtlwtFA6oDyn2iaw0dQirFhSkjKrz6xCsbthdZjXyDkh4YYTK9R8tIuBdQD
+          /5hYdDsb42m9MnC9fywcr8yzH2d5GkBOmgV7A3mUC1NcKuDVI4fjdGUQ7UbaKkKaZJMwI4uK/NVV
+          wUtedShclJ6uUx++4KPO3TQlSoi4gZUsOKXEwiYTZP2cHLctdKmrkjT7qAnXTHtZkZriOAmEMbz+
+          udrX/6yXfVtr6jfQ7v7w6nD6nr1ZzfAH9oFd4QsYAV3ncckVbq6f4XjzfW8+0doE/uhjfDlZF++n
+          Xn4M6KeWw6ctP+md3D5w9rUrPpjCvuTI0DOQRzo7CTxHjOWoiVcIprIg+zjy+mVUehu2+kfMtNGR
+          eGjwdg/2dwxX/fBGq7l7XJX9fE8mTj6NHv1VrxmU1yBhTxM+dCz00ob9SW+3LV60HOKFVVHINBMO
+          NemaLFb7jJWOdh+s19Anf3+P/uLL50OpH/NGcpXgoPbk0pqcNxe1I+zmyxlCXndQsl6eQw6xsLWM
+          A0f05uvT9JXQPazTyqwErdlejRTpkWXYSzmNzj7FOVL1zsLhuKb9eDGOAMZtOBJsTQtaw0S5IvbN
+          bhbeG2rqnJ+Fkr6rAGumL5SrwYkMXOjzQMKt/k+xmeigvVKVhI5lllTbiyFw7e9MUvX9M2jDJgN8
+          uXgl1peK/RCco8euL40TwepM6vFWiRsfyW7T6eQd+w2fJ6g+FwNfPtZSL/FpFpTKZXK859HNWyrD
+          LuCA3iIxqvcZzcnx3IGI+AspmyalS6Pg6l/9N+iTN0ar/UU7/OwFEujZoZ4v5sNUvr6sk4jJ3JI/
+          q8RE1iDk2ImcUz2+1cX/49dkX5g7RJ3DcVCU9eZhdf/de1zUbHfVHuQLUcWyKecZcRwga/eePg8u
+          q//4DBKvKcVOnmrGmq77DqZl2YW7s1mW1K8VAbR3QImRrG25XF7yC0RzOOOiV/ZIWHu6wk5gGWI5
+          ckaX8OR9QLpfZFwePxdK2yga4HttTyS+ae9+8YwDBweyu4e7Mnl565utVeVX3+JQVkEw1jK6SfLG
+          d4i3DDVdDOvGIb8rpQ3vXYP/DZ8QZdxFxR4e1n5mPweAi/jViRknTEnKZI5Q6v9aUshD4S2PY1FB
+          /D68yZbP9Rzr1op4bp1DLT41xgw5N8FxJ38J/la35LfhE3S1lJDUqvtk/d4sHRF9SonpVEs5uqbV
+          ws7ot8nbdVg/vtEl/4ePB7kze+HAv3047qTvtBnayXo8/R5I8VKbZLo+JssphgKJancl2KAe5b5k
+          x4CDmPsktVXjrXpWhfDlohUHl+hQCrdEnRX+GGIcfHWRDg89lyEs1QCb40P1lmv0DoELwoHo5r4t
+          l9HnTFhctQuFz5qj+WAfQ8UykycpnSkqKf60A5ya+4X42ZWU4w7phXI//QD7z5eK1o49XhW//6i4
+          EOZTyWvtKgB9zjpxR3H0VtNecuVnq19yvtSIfuqjlEMmBSHG+5Eaa32cC2XVLZtkLuLrJUr2jELk
+          vYGDEWsGn6ePCn71PSbWIrnlYLRehiz49FO7nIdyW48BGon4WF8ZrRcDag1Q4u4+8btALTnd7lbU
+          sZ8uZGTZqNfJJS4E8vlOTDY8lUNwsDl4+88CuxWfoWH2ghXObYexZrl5vw5X7grGr8vIYXydaop6
+          J4MgWxmiuk1bU3E3MDB+rRPe+FC5ZNoowzeajtglO7en2qdq4KlcErL/eJ0x/4IryOU5AqJ7amjM
+          ulSqcrmOX6LeH29KxnukwjcXDiS4u2FC71jR0WrZBo53/JCsoTHOwHZGiA/xsHhzO1EXvSZLm1ju
+          tZTfGWQAIcIlMS/PTzkHB1VQ/vRfukvDZJoZzfrTG+Sil6rHPeK2gZMqSZNCHXvbkl9PYCqFR1yb
+          u/cLh0pGKY2pJK4rKAb10wZAE+wQR/dG81b7FHBISI8ZLi4STTr/o6/Km7fiadn4N/cesQlDB4Bv
+          2qlFU57EGRyVgQ+XFPflfDFbUy7PMWBVy6tkzVdZ/qtv//jrdKUP91+9MtuoqEmv+C8kHZ4L1m/M
+          z5jt88tSbnH5DNffei5XTt5z8AzzAym+6qtckBQw6P2MzUnc+Mb6m9UPqIpo/On38h/fMzXNCr+m
+          rNYC80t0RQu+GVbR2fGWU8wVYO4PWvjNWp6Ozj64IjtKn3g/7I9lL58xyB16kakZ7s9kcARPkJKZ
+          PW7+RdIP52Ws/uVDWry2Df+nfkbu17uRoGfLZOm3LR8bf8H+93pA3B9f/9PDGnE6uljtL4bxnOnY
+          GRQo+z//ggm9F/6Lp+m2/jJgq0yYhEomyQp91KJEm3LiLts94aPPWQgqRceeuQz0B+LBhi1ewtfK
+          aLXonepZYeT5hKuTKBj0HYwCiE2ohO0jyes/PQR/f3+MEqvmywdIQHMyb3juJSLvyakUtzs6zc/K
+          LoU/vrLrj5+QZS8lGrb4/KtvONbcV7I878MDPNaX//SAN5r7swrpYr9DpuWPxnJWv+Y/P0KK37U3
+          X7g8BS7fmdgKhLJfeyZu4R4sK7HbyjQE3r4DFPjmTbv4cEaC6LkC6KeOI7rPRPXQsroA+l5+Tkqs
+          38t3h0QLRXrW4A3fk1nyHzGoxQuHw5rEdSt06fzH74hzSNvyp4umhJ7PCbDxXXRv80s6eC1tjW3t
+          SZJhrewU8pI1Jh4f3XLZ6r0yHIpyWu7WiS7RK+PAw6CROxkbY/7M86wUXhmFN5bhjQVE7ML5MVvE
+          0CeuXmQnfcE1LJmwuZdBye0GVUAne9FwHpyUfvG/uQTqIrNYDyFGy+ceg2wOtxlfNjxe0to0oc/Y
+          Agdqup26DcsZPv40hQKWdsbocMqAqFJd/vhYubAS9ZV9zgkhkR9JPeNTIUGWMyesGTiuZ/fN2X/4
+          EVbWoTY2vjDBfZoCvPlZ9apb4Uve9Om0u+MP/T1Vg9mhs6ASj/s66B//m4rSwerpnBu86hQx7M63
+          nASBc/Yozo66Mq2MFi7acU5W43TOlb/8Cq7eWA6LazEAGe9jt2VMJAoufUFj3hSi9YHlCXJeWYg7
+          WQLeVyrTjxs/Qpvft/kTbULv972kNPt2T8p11PtF+Fkt9GJnT02cVMmSm/kqGyfBnn7pXk7mJvg8
+          wF1FHnu6uK/HRptVZX02Ew62+jiW2837WSoX0z++0wVppBi7/IcvE3tBHDfEhaLKOk8OL/Lxxqt6
+          e0C8MPtQ+SkKXfK0rYDTohafn9MhEdJ4NGF3IHvss+dPsjxVD9DeSSOs87tvP3/JAsDt05oUm7/5
+          S436gT6zUZND+16N+XWngoLY6xXn35BJyO0ot9CMOjfxTXCgot6psbLxeaLbWmDMj9Wx0euKduFW
+          X9AMP7MAoqAjDiqH1EvFnAX4q0/qls/id55NZYz3LAn87d7rOnpFcIi8YGK9ni/X5fqSFRrxX+zX
+          D8P4DZX+QdV2pGjTB2jdH+QGqu7KhPYzPZdzvLA62vRQOEam4g1WKvqAdzVLwoK79PPnGPkgn7uB
+          aN7hasxJZtsQz4uy4adr/OMvx6LQiNWpKuJ5ZzJRUTo1CZdpVw4DaXMwUzHF3u8pJaR6mBOo++cu
+          7GyOrZfuYYcwClwbco/8+OdfrogbSIyD7PQsW/tQFGirbyQcV66m2HIyyM/IJuarCGsar7wt5T3J
+          cNhLY7m+714LkXstSColuN/8l1B5KZ8v+avX466TBzSS0Av5JCAGd11KCal6a5H0VUz1uOXr7vuo
+          Gnx4TodSrGedQ86sl8Rcb3U9z/3oAlnYAGuixpWDdBBX+Z2WDXHfrNELjfqT5U/0UMi+Cb5o2RW9
+          Cq7xgYkr7229xRODWOs6bv5KW09bvURPvnaxsdRb/WDlbe7XvOC//Cb3UPCR6bgQPpILX1NX8FwY
+          mMUlbss0lOgXqYHvHIrEPDSdR+XBtuH+elxJ8NXPdG5ZnQP+FLtYz82XQb9wj9HmN2F7ulrl0rOP
+          GYyxaggWmgnN52W8ovqixcR2C7Gfj8ktQg7+XibqfiqD+6v/uqLVYa0OojezSLWhrROGuAa7JvNc
+          nq5/ep4cZWXnrYKvtWC18y3kSKUndJdbK6gve8BY2+bWlIvf/Pl32Dy9tVr843N8UKh//YOalsbj
+          A6ii09TlItevsm0VqHqd7XB3Ozb9Oo+7HCUOF+IssTRPPCfbqfBdVJLjDopy4IYiB0X+ajjc+PzM
+          l4YMUtQYIal+czmOSu9Ccmme+GbFQ7K8tjkGxraZdc9YbT+zRbuiLX+JrT7mmmapJigsTz5k/0oe
+          9fvyWh9g7IofNpicReNb3flwVvc51h+Km3ygVEEJ3f06bXwFrfZpL/zpN3J0nB2dxd8NdtbN3JND
+          QIJ+ZXbdBPwpcvFRUE/1eKEbrmz59se3Nr/B+sNnbM9ZUC++tXuByjPbrUW5bGx+mqmwo3jd+g8H
+          T0j1W456w/ND/nE0DCGvOx05RtkSbTmM26l3cYJmFg3i8H5tTJufgm7V+R3W17ooF/YSt8rmn4Z0
+          83NEw35ZcHypVTh7z8F7F3rp/uM3gbhNzKYQr4jyTbTpfa6fdperhDb/AFsJp9C2Dt+ZsvsGfciK
+          76D+8/Mkw9yN5PDufG9cmChX/t7Hn84Bmu8vKYbNH8T2O7p6C2p5U37qfjbxoiLX1Dz0MUjXxwHf
+          z+u+F4v7vQJ05lRcVlNLVxEtkvgIWGtiy6kr11eAJXQTbZecP31Hf6rsZAif2ArbQsTVa/10YhAI
+          O+J//ZIdcnMIPOszMZp0Lan7WWPA8nQgFzf59V9sOSnoz/M2J9zM//BrRjc2VkPu9NZ6DnVG+Oc/
+          42Dx9rQdsaOiFisOxqe59eiD+VRo49ckvC83b13LwULJfk3Cde3v/bIwUaFwbX+e+Ges0b/+FbLp
+          70RMMevpou3ch/y53rRNT8f91E8ugzZ8w1qtF/Uqop0MtfMYps1vRUTJQYVZI4+Qjt/QE/iWTWHR
+          JA775hnohiex4iQOwqXYWYZgOFIIYpQT7Bh5XY8D82TgHq8O9pVL4604ZjrY9Ag2ih/1VlfKLfBD
+          Ngjfm6Jfvgx9gWL4McnW3c9oDzs7Au5Bw2n86iIi7ynwIeL7fJrVTquFLPpWsIfyTPx3te+F0xJy
+          MJThL1x4IzaGKL2acCueZ5w26YmK4PTbkeBcwPhHNETv90BG6bOIp+F7/dJpJ/EVsLvlGdKNz423
+          io3/9B+2+9v7n/6Dzb/Aziu0DW6rX4oE+ki0YF/Vs9y7Khx5bphqbeTROg5Bijb+hp1Nr82dEoRQ
+          76sF+9alSIhrxZJiKrmHj7MV1sOpTnKFLo8TufhBROePZj+AeVN2Eo1ORlMxKzFiq4Pzz08h31my
+          0Oa34M2fqmmrOir6UH7584v65VLvUjC5k0z2u0BNxH/9RvQim3+9p5xPDzls/Tdyi0zFmHNVyNE3
+          5w44Y6fMGOHLRPLvWwdE5a51uT5Da1A2vfmv/o3Ofl/BWAR8CDTKjLW8TysYS0z/38/58L6y6aGQ
+          D1OL0oiTZ4T6NsTJ/OIMuvW3dlTbpwTHvG4Ie/HFwb79OtjxgwjRgFoTCjOI8QEvnscJ8nBFz/cq
+          huzneUE0WZgU6iCTsX5VP5ufcakgm097ctoFj3Lhd1WKCu8STStLp3pO30dVYeT1RPDQfdH6HW8T
+          fMFFxLNoV/7xMWBC54X94TSX1JbUWYnCx2lak12W/PVXQPLeA9n0Hm3fSrXK/8uWAuG/bymo0kIm
+          h2NiUe71OVzl0z0ysZUgveayazuBgj8v4pNBq8fXI3DBkN6IGNadTZY0OsagGufPBKKV0rm15ByW
+          S2hNjf0kBg09L4WoHAusPX3BWBjkNwhO15Cop9hHczHUjdJe24Q4yeHlrYVtN2Ag/0e8YTU98mOe
+          NqAUcnzT433PrRfI5HD8nIk/K2YpdMPehPHWYHImyY3Ox1GOIVKfHXZwV5Q/4adncIsKn9gV+dV0
+          TJaXspdlJlyqCff0+T1WypAxFNtn9VH/smPEKLOPbazigSsX/JFidD5ZKrm0Dz1ZHxfDRjdjZ4bE
+          WDyPvFdLRhN1o0lAUenNeV4AyG1wwlV0uvdi+iwzMHf7hRRnvqgX0wzVndaizyQPmmjQ7cALTNd1
+          wt73qZWLkvQ6lM21I6dfoJU8r7Wzcm45So61sS/5QXkCJF7fTZxwfiVzqBofxfL9E7nWyehN2qkr
+          4CnDhxR4KZHIv9EKZqN9cdBkqB8RJ1Zw4fUIu2EjGTSTdznc7bINaRI4dYcOZYMGw1dxEBeGJ6zG
+          VYDqdzOn93YKYm1K/aG0NWKxwR/ZnjrXVIYdm5+x9hS0mta1MSnHtVnJ8cvGHs+LAQdKMSdY0z6f
+          XnBUz1J+o+uEf8+b2WJUoay9ZNphNvJmJ+uav9+DHSqfEpHJnh9l8S5fXHKCjgS5aF5Q7FSGFP3e
+          NdbonUXg8EGFdTGrPY7xngJcxRKmxeneJVlOs6T8umdJ8iw7eYPPGhxg/4Em5nmw63VHdUlZottK
+          zjjPayoLQgryvY1xOiyux4unJVc+/ApTD2+PLsfbyULWobDI/tSCtxrErhQpbD843+KLc41LCyXb
+          FuQkBz7lZOF4VRCVAGPxnhvz7koLCH1tj4/HqfEER95awOSKyfnA7EsBSyREf/mVHQKpnKpY0GHw
+          qL7tklcTTmMLAUQy6ORkDHU/XTLfgukwNNg6vCaDqLf6A/U8v6Za6Tjjtw5RoYS+sZ/Q4fhJKP9G
+          M7hqiacvHqeEvJyxBe+g9vh01PaGoGrHVWnobUfO90dmcEYuMeDwuJoWMpXeei4ugsy3zYydGXvl
+          ejQ1QRFzzSUVvsT1qp1eOTz37ZMcFePaL8wXdf/e9/as7wnRXouq1IxDt3gs/+Uz4hw7I1HMWYlo
+          rkhG4/Nk4EKcWUSb/V1CkjkHONvya2nR6wPkazM4KBzfEKl/mpSt/oRLi+7e2ldao4i54WKtRay3
+          SlYiK1v+TkpRPRF3IrsWMGMGxAv7S718U01Xiri/4UNqGcbCXiQXhnyjuKo5GvMUkwf4n/s+jFj7
+          4fEnxY7QPXlfsP0MZmO5OooNH7USp9HVpp7uRvWjpHnbkio6sf1ypqMFtphmWOsPpiG+QmoBzY4j
+          dh6HEfVJnNuSxMjPcDxrskEPb8kCSF0Rm4f3y6Pe/H4pXOU/cG6UZTIfRFFH7lXKSCQUJ8TZxGbA
+          /9z207fdBlXIxsNV7FfsE+ewtQg+x28G5uNFcKW+RLo0JVtAtxQUu2de7udDdLqCOls/ovLB6m3v
+          bwO+CTHBKX/q51LlYuVxKj7YF8uup3xwy8H9rPPf7zNoezjHsLuoAlZf39DjX4+9jdKgEom3W3FN
+          jtKqynHMjcSQLsd+2J3fACC/cmJSPJfPU/OUlVurhHhzNnq+YeZO2Z5H4qh9UkEU81YMEt8mVvq6
+          oEUPfiFkDbcn2S99UfGR2AJ07pKRe9ku/eJTzkL+fN5GRxOfclt9RJkRUaz2UtTTuLR8+LxUB5en
+          gEFLjC8fSHxFI/p+sSjfQRfDyswWjvzs4fG3NNL/4ovgBN+89XN1Q1j6/oUP6lOhk6odZ8VtzQM+
+          Wa5r/OrDZ0WRYeek8K0jEmUwZjBS3sbXj/VLlm1CDozwoOT4F58Gsa9g/cQXcdyLiFZ+8R7Q/N53
+          EtBASOj96afQzuWBuBZPjLWGCuC1NAbZP51X3X7UMQLWPByI3vPQ0+ieMMrnVGs4uEJbrpfMNME+
+          WyXRqonU801DAM2d7MMds8zlgq+3ApLdZIYcxVHJB7qYwn73/eIwEC8eZ/BsC+4TnkT7vd1eNC4n
+          QbleVjIhmaVls+E9cj/zjAP8tRBN4shWeDvlMf64GRIF9igo02FqsHe/vfrhstd1lNPAwPHZso05
+          svJB6QehmjJpTanwbK1Uieo8IUcyIWNU/edL8fZjOrGB5fdic7J9WEL99Fcf6GzrR1Cca/HAbhnr
+          JW865RXw6cCQoN5LaOVeaqqERmXgzD11Jc98aQc26W8kxb8HWl1sRUpZOwk+v1rbE+NH0skOtCVJ
+          oTlSjiyJCoLN9VgPvL+Lb6GD6X1etvXAHre9D1zw5G58ZkCLM34YtH1PgidJ80iwjCHa7SMDZ6xs
+          1fN7SBtlq+9YlU6Nt7TCzZYY433EaSob/axc3qnyF99quncM3hkngHzZDcQtWoGuuUYrYN/tBVtm
+          cS25YFEzpZd0jPFUe+Xa8o9IsTs+JPePK1DK+NakfJ9hSwzxZtbztb1xUJ2GIwncrEmWp/U14WIf
+          gykaHVpS59ll6O990/vVrvnrLTQBVcKCncfV6Dt4LC34w6fHqnbfG1Rj2QnO13OPtcv6QrRiZld5
+          U5aEqzCI/XIXewF+q59iVTSf5bSUyAdS9MLEcHJUzouaRRAKXjHxG99ZPxnMQGKLw1hvvuU60W30
+          saUIIf9RlPIvvuCqH8RJOWcETZ5AOPgUpygUhJ/pTfQ8Vujg5dy//B/umf8BTiOXUE7dqF6X7L1C
+          1IYvgneFlUzGD+nKl+wiom34JkbNM0d/8Ree169BI4tvlDHlEblx1cVbTcZrwWyML7ETKewp9W8D
+          CFagTAz8LmWH/J8Kh3mSsWaJjkE/kVzBa46G6evrTdntRv4K1e9u4lD2VrrkgReCll1dHL3Jw5vX
+          tZhA7dWKbL8f9YVYZ2gfzR0+BMLdo+XdZIAxpZZkTScmy6A8GVndQ0oKm2HpMBhNqlx2wp2YT3k7
+          lfNIWjgI3AMHaU3REnJ0hutdOuGcsx3KDZkSw1SZe5J/+6wcIiuagN6mjjhXFqH521oCSMrbJRcp
+          6YzlI79kZd/Ut1AI7E89zvb9upOdaAnZvR6g6UMSG6bydv7LX29leD0CdcnV6YMi5JG6vdggcNGZ
+          lD/2gLjly3LQBYUZSlHQ9LPwY8M/fCTO3Dy9Da9yeKD1MPHerCcz/9AqMPiDgf3TL/OIbZsx6Orn
+          Hu52fN3Pu/ObgbszsThUXyKaibOrgKl5HxvzKUZbfcvQ9GXuJLRue7r82mMLY7uq+A9PFus2mAgO
+          PSH7fA49+ghoAR4yLpOShpU3dwfZRNYn7bBePVCypOMdgAnu3XYLzZCMKK5TRWweMbFQhIzlwg0r
+          2vQLCb+y2a995TRwbKd8erm5gtb78nvsXpZ7Jti48uXCuliAm5dpxDznDuXxZ47+rY8jjEs9/j1v
+          WrgdyZvTy/uHv4OSpiSviFPPRciuSM8YdeLnb9pTzS1aqPepgO+VJXmEZvPnH7/2Ew/3yztAEhz1
+          TMRWfUlrqgoOoPfzkZDo9Z0MGgznFbJflxLn9jskgi4KD7ThbTiFT7MX+3jHgSmXdshdf0O/VvK9
+          gTUIuXAVtTGpTInxkew2NvEtbxvtTNlY7iUVkyy+/LxFfkOMHC2cQ+XCZcls5DOjNPS+I+fbeV/3
+          yk9+wLQIOxKMQNGyDx+ysum/cE6Vs0dGp2oBvrsOF9L+hLhxTlJkTOOT/OHfgs7zrNwamyUbn/N+
+          x9frAay053EYndh6jcjiKmJoz+SIsFYv59Ka/p4/scbSGyMv7oW/+kYOoxuUS2HzFdrp/Qnvj9re
+          E35m/0DM265xta3v0p3sSXZkY8Re3n36ZdPL6JPibuJ+P8kb3Pe4og2/yN5yO28c317zTz8ZWr3W
+          fY4WCx0PjkU2PdUvjioy6LgcS+LEt6yeFfHoQidgfYvfN930WyHr5Kljs398y/lhOxa8/coO2XHH
+          ollJ7xKSNZ4j+z2fJsPzbEbwbcgjbI15MZb9rOiwTFE0SUaJyi8fd5xc7l8W8aftVrq/eP/jj0G9
+          z1F3Sv0BLlM7TbQ3Z28V8tEG8bhriC8NY7nsBfcK7aFxiHcRu5IeafNB6lKoxHVar5zxqup/+pEY
+          7NMo+TyOWjTF0onsH9yAZm7yGdnyw1MI6QHqgWo3RgGuWENmPaSGuBvtD9zbX43V+x08YgfHCRjx
+          y2L9tDMTIR1qS3kpPyUU4ovjieLzo4J1jB747GDbG7rDav1Hvzpx2Q/X9sSBZ2c6dqljo4V7Kb5s
+          1SNgSwMejc3TF2B/v8UhY3ROP3R60yJmJG/sj/yarPtZdRXR5y8YD/v9Pz6HQvG6kEoch2TUg6cP
+          y1kmWLfxWBL2cc7QHz74dRJ4fB4YvrLV/5Bufsp6MCNBab+COS3iekFroOsVAJev2FC61Jjx6zwB
+          fzCq8OFrXTJ8CyOF0X7tifUwHskazxIHg3vrpzYEjS6vx94FD9I5lC8lNWjCNRHkv/RHNj1R/jJL
+          sJRy9+Kw/g3fPQ1ANZXdtZ6IE69DstAbWdHGn0K5zjRjqG+XHHZAfxvf+6D18rvp8GtKHh9eyZpM
+          9c/S0c0mySRufhKPw6MMcPiRbb2uxjhQbAJJco+E92dk8DyU2y1Ii4v3nz1O/vQ1bP8/ZCT9Va5s
+          +Ngsw1TFmQYnKhJcz/CnP8v77bWdUnJlBAE9kXDjE0L9s1Sl6uaS/OlT+uPCGArV1/Ax9bt60NhY
+          QEv2zif0RD9vc3VcNCJ8naQpKPtVHJQZvSJ1mQRW/vTLeGdsYNnHOrHeI/WWRS4lsCyTTLy4l+l8
+          3n1zYOYsxfb2PTe/ZkWidNwTf27DmuR1MUEiHrJpzc8dokY9T7DVA+yXOEwWRo0tQAdnh037Sbz2
+          g+4z+uqxGdL246Bu4Y4f2OIf259q25ImRYKSDpcSb/5bOTJf1MoOdGUoWaHYT4Z/8BWYjGiab1lf
+          rzvvl6KNL5GQxx9j3p1HgLfQ3Ig6Vj9jvhVxpvx9P/tD7WSR7l8XUQ3XZM8fVU98i2kK4ZCo2PCL
+          oOe3eIFnMfT4vK338qd3mAus+K9+8q7kcPAIyDhRs2/qOdKM6B+em9yDemR48uofXv1Hf4lKDMpa
+          PTwSlyPrEX0BX77bl3a6UJlP5h17eEBWv6ZQOWeYDj0pWzDgpRJ8cr1yrqaOgZY9kPD5fPaU7jXW
+          RX/6339fla1d634gkKIYH3/vrp5Hvl0h7JQfVjWXlOR5ZT5gPh4EG0ah0T+8VJCzfidycvtyYp67
+          DOb40/2rj/QhdwXIHWcQvD8bvWhUfQyc42YYP3ZpQlhJmVCQhDYJhqIol3gZh7/8Jc7p3fbL7yas
+          //SBn1/jhJ7m6oFkbriRw0UndLlHegQyKwfYtFBW030j5XCqVWlalHxNVuZcVXJ4Qndi/44TGlzj
+          2CrHJSmxhrDWc41rV2B7Ukdy23mj4VDMqdJnlUKwOUoGZY2oBSdla7z3Sr8W85zl4OaqHbaYUadc
+          HQsPUBtJwtaSa724+S8AbnQj913xSQZ0oyZckoohAd8DnZPmbCnsxTv++Tk1vbiwnXJmLOLJh30i
+          2GH4gehB38T+29p4e5gd+r2JP73y4dHPB5HVofl979jamwe62PUSgmxeXyRy50PPOyoLUErBbWp/
+          xxBNuuIBykqRn55KvpazaJwyZfN/w93jWvcr97JTOK9jQHRPPpf0Dc9ZGdx7Px27NUzW1UgF9Ej3
+          GHtq5SZ/frR8uy03cq/muiZKM1oovo5WuO44p1/rKbmi567HE3qYOzpufiw0zrmZePml9+KPebow
+          gKpMrw1/aOs3A8xx0+HDeK8NWqw3Gab3aQkb7kEN+pjxC8j+PWATmoUOocxY6PFozrik74f3x1eR
+          ZdKcaL8opeKtEEO4ze2BBGW/7zvhJn/gzw8LNr9pdS8/+x+fOpFsQSPXuh0IxPdJJJpaubiNJCN/
+          lHSi/4bB4EW9vaIPPwM2qB3S4dzX8nYr0hXHrCJ7C3wQwPFbamQ/g5ssB9F1waoJ4G39vDVDUwxi
+          4iOifdnVoxwVYxDHwzZIop+82WrR5veuCQne6uHP7zKRWtXZVm8c40/vIDZqMM6q24F2u1GpZPbd
+          XULOQlk/cwXo6BXpy9Sj9GcspwLrQMs9TOL2Pev27ZkoOn05YldFglavbVL48+e1Ft2NkRtzEwbR
+          n/D5fD0nbVm1JjoMYP3zc4kuah1owrAnt+s3rsd1yHM433p9kmT0RnMyLQAKUkOin5qwFhQj4MBV
+          7OPUmWcbCUkgftCfn6IbySNZpIZroAty84+v9qu4Gv7f+oWX+Cb0NarODNjuO8Hhxh8W8lA4eOS7
+          CuPLXTOoOsMLxoc2YC19qglX344F2vonxE1C458fBpixAmJv+EQDYe3Q3pIifA2xjiiq7gxo/Ssl
+          B/V5Q90IBwBQCxMbH9tIhPnw+vzpL4KreQOf+PuStvjARROHxprDG/7FE2Y0pWySQ1RA3Q9n4m1+
+          8YAH+QMkzFSiNWnrDcXQN6AKPsbXEk/Jmi7SA3YvFIVTNz4Q4eMXp5ztRiJ6xR+9WV58HWbhdCM3
+          4dQmNK+LAVpTKbHujkn/16+RH+XDx1ERXuo/Px3yAz1PtIC1nte7tsLrmNnYvDKvZHDKWYWrUr/w
+          AZST8c/fqU1t88tevsd1+tACOni7cCreRiJsfAp+79EnfnVWEy6fdylKM03+W99+/PN7N/0XoqDi
+          0dwdVlM5CMIjnPZUTuaT4Xag+ApHwq1fMFJUNn/+Ezmcr2JJ02eS/fUTiNN6fj1pk73++fO4DGjf
+          rz8205Ut/0kYne71XFzi7q++hzMFUpI/P/J4e3wn5VnfS+qcDxUwb7fGIfx25WLw+grG5Xcmxoan
+          g8QfV4ijgP3HV4WfIzHAf5MLNqMoN5bLNSxQd44rfGgWCf2qOvHBX1qBpNnElOPTIhZqC+ONdYgj
+          Q7xtt3IS2y9wam13nA5f/ap0HOonbtWHfvWErwCJ9+tIwogimlYGp0jn5YbYS2eUA+yvg0xoZ4X8
+          DF25JkmgI/sV+fgmqGu9DTatkK2rR+weB9cTy6bI//TaJA+ZT6cc3gxs+umv3tb0zy/0g+7w54fS
+          pfkUEZKkGf2L71ks6YCK820gnlp1JQkuFofU6plNu3LaGUtIpnS31NaD+A/+g+aPp9ryX3xb9eSW
+          nCxcroAethPKtdYna9vvC+UePkN85muNimcHOnn7PqFbKH6/WHnV7gQLKzh83G7ectmtoeJb2MV6
+          1D7Rsvkff/7C9Nz6X/w2tQD++F+xvd/Gp2x0pLIz7fQuREt3UgelyBeDGHNn11xZPSxlX/QGcZtv
+          jaii5wB//SiN//0SOlxJA5sex3/PE9JFeoHtfhPsE13wFnkbTLr1V0LlOJ/rf/E6C+fbxGz+1ZLu
+          mQL6Ds/4IGl2ssXXFVhzf8ChZUeG8GWqFN6OWBA7y3iP/OWnUTDTNL/3XL98vxcAV3GP2K2zp0cz
+          eSkgCaP8X//ln19NW6nEyVZvqBTeTIij7Yik90iN2QtwDrV3vU5sUOrlxhdntPGHf3x5ul6PIWx+
+          K042f3JVum+OIo3XcMge2H5VOpKDe5UzYs1TZnDXsY/ByOw7+etnzMa+VeGL9hopZTZJaOO+U7jx
+          VYYtJh+NVXCf3e5BYvZfvKyqIHDKdbaPuDqmDzQHi53BWr288GnbHp31QZOVSImGaXkKz/rh74Rc
+          2bHFGbsX65asUyuqUOTUwHtvfiV99m2vaHcT9yTQFEJX1ikA7MF+h/JlbulPCU8NCCLFociHvic8
+          qryR5bq0sGszd0RDz0hh0j5MyB2ioZyQ/9QVLPwiks2T4FGbjBxseniSz7xcz+xdvoK5cBZJVbvo
+          Nz/UVSAWHwRn+q6cuo/YIVL8BGz78xWtdhg2wLKvdfrluYnm28mU4Qh3j2BGuyXzjz6L3f+ypUD8
+          71sK3v19R/yTyyU0RVIh125ApstaOcmqBVUIJHnaZH++jMZ8C/ax0nQsH0LX8TUlDBaAc0ydhKNZ
+          lYtnFS38UL4Q53C1Eq5LLRsJ+Orimyufa7Exdlck7YWEWEd8LCkOI1NJzv6A8ffrlcuXhSuI+wyH
+          K+PU3qpe7Aes72OGbzEjeQtXKhnwd/NAcq3RPUHgTVVxrtusbbHH5UBEaYWX30b48kIMGsgrX+Hr
+          yO/paOjHZDlWVFd6XxWwVXyXuuWjYwjDGjMTXd+fej3LGqPAS/hMnTJP5VrCK4JTzPAkHYlbk8Eu
+          dGCbK8W2n77QpE5qBfeYv0+rHvFoMYTvDAf7EuOMfya9+N0vk3IxlT05j/mjHjbzFN62dMPq74Hp
+          GiSsCtzTuGOzOXXeaJtyBbDCl9woW/fzVdzbyuwcPVJ5V4TmUhhNdDvyZ2yeK7Zev3ptKZTxVXLj
+          z403DUVfoOc1DqbJMsuefympqzwC9jVBXHHeGtyfGXht5WHNLI6UA+bSgHPXrhiP2uCRfA1TJAl+
+          hg9HcioFnLIucNPHwHt8Gmpa/KJZCX7XBEfuJ0CCelFfihUcw+ny/N6NaVnrSUnXgZL47eeeWLIR
+          p6Svz7pd1Hf3RHgjXdk9s2LaLfavXPrzdYbXOT/+xUvJp4/Eh209sbp8olpw7Z8LXJhq+GKuVs/x
+          5iTBzhvO5Kwe02TdWzYHpzzkiD8NJyRwd4+BbGcirB/1DE2Xj2tD9XB1krwnKCdztieZ47ZZiG00
+          GnOffnxF1porudFITtas+QE8DLydwttZpVCrnYCi1kPEWASlJFYS2PA6xBEJuu2WgqviZkqNt1nS
+          6XqvOVhJBK/rfCQ32Js1NwhRo3gj54Wr0F2MBWWqC3Z1OeOqKb+lmB2lFxher5FARWrPaUlfgCo/
+          JJJ/SnmbNSbPEJ2aOw5H/ktXFg/TX7yRizpbf7+XgaXQjtjYyTe6+q3AAMcIe+x19lLS53iMlUP+
+          TCac8129BomoAv/Z1dP2e7zVLNVZcc+chW9EMRIu48xV4dnUIXHX5AmX3wcZ6u624IO8l7zJ/Qoc
+          oLRGoeKeNG/peq6Dy33ck+gvniVv8AGU4UuCkuqeyPnFQ+aHpN7Wc67pZ4d9sJTfcfqC/0RcLb8/
+          EPRoIj7s1GS51EkDJvwcXBp22HOjMLfQtI8C38Zkywf/HULYDNvFaeiX0Pdl94L+I0nEJpFJZ6uv
+          ByVownT67iqt5PjLR1Dk8MmHjeh4ntgtO07GfHYjlnufa7qwmAGbWQrsh0VSL4hTVpDUKsHBm9vV
+          tKzCBzwv+EtM+2CXvDHIFgRCT7AB2qEWh3UHoCThD5uxxZarrwgyHBOPIbhRvF501aAC3gwtbLMs
+          1AvzzFTgasYkoRjYBmnqgw0nhROmQbIFOjTLrCtSW894zxKC5nX8dMo+djxc2d6+nA3WuKJ7cKuI
+          Lavvnt85twnM4+9APPI41hzxpRX8nRWFVOWNRDx/pQY+l9cLJ6ZY183l4TNwujUcTi3yRTSJohQu
+          LHsgdpTV3jaQbVCevuqRiItnNHsX01JK7xRgV9QDtLBiZoJ1u2B8EnuSrPjxq2RXsmysE8h7QQ+z
+          DP2tl7tPXUSupxyAzv2VhKfvx6P1dMwVEVkxCQv2U67pIFaKPdV5yD/4yZuHh/SBL9u25NwNj0Tw
+          704Md1RYJHgGOqLlbxbkTzg/STlwx1qwrrYkmcvOIYmIrJ5KocwBrSsVa6uV1kIudzrsvz8/5B1l
+          SGiQjQ0Q5umGqK+WctmdVF1J9ZCEqLOPCbd9f2CJtWJdKS5IPEF+hSIKXHKFnVoOn5sUQnxMW3zZ
+          n28GTT9Rrohx3WP8QF6yJsFzBS5CFTGU06Hk8U2WUf5jdayLpV4v4WEVFMW+tiR62KSnT/0Yw6Ns
+          bXLn90M9mmNhgz9lOOSQ1ye0Ni45SEE9EoM/WmiRudsAnfdaQ6JFd9oPN1mFZ3j/ELvsZY+Sexwq
+          RlYboWJJ55Lnr7cc7p5uTLyX+AnHyRcT3p9AJv42BYFeOatBj+RUE3Ufjv0Mv0v7F58Y++SH1hy5
+          JnTqvJID+BoSM5hfyrGsamznRd2v7+0E4G65lvjMjdspATpHyqQjAwfLyUB8tm2pq+Muw5rgnoyF
+          fQomxJB2+DQdJ7qUu7JVHt7rMPH0G9J/9f3+aC1ym5uiXF5fcOEV60fiOybt6fk7N8ru87nggqpB
+          zcNg+8ByWzwHtzah8+8ggSz6FUn9ifUW80k4ePldhG/M0vTCBXU2gJswpHpcvt7CGvZV0U65jLf1
+          qPu8zVukYUkiMdDaEFchzuE9syicZSZC4nBbdXCujUY0NV/6deebMfBpdiAO5X5lr398AXgPPcjB
+          Pp+M9XhiY3BT9MTqlk/rJeps6PqVCxmdmkh8vQ4qCOKzxRevsnpu1j4PxVfLlrgvaahn/9LYyiFJ
+          0i1/LKNnvmUIr5DtcHBeNMR5k6/DJPUWdp5qW1Lt3cyKX38wicc4p7SpD67y9/0M41knNEFRqmz5
+          NDGjySRr6GlXRT/JTPg6fHhK1Mhc5aJTXHyVRNWgB2bDO0VPps6ArqSjLZsIF1WJzUxs6Xgwi+I/
+          eHt2HuUskEmG8dVr4exEARKvXPiB3teFiY0PnTGIaBeBqdVWiD7W4n3MfTvDM1FaEvT+Ba0nnW3A
+          ufRX7BDjThs8m5Hi5ZmL3Yfv9xteNMBVVx0H3bs25t89z6HoWHcSx1zt17PBTqhWXuewjlizn3jX
+          15G8wB0HNqsZ3Oc2h3B93lNiXt5OsqLALxTcfa7h9RmdPO5bRRH8PH0K+XY31csrTm1FTjWRXAMr
+          QTO3XbT+9HVv4s23781EnGdlq0dYlw8vSoIh+MhGvzMwXs++0X+LpdulcTqHtF7e5YaXLdSHOf0X
+          /7OsHCuFtpyK71p0RzTaRZaiBEIbCuFzRkPrRBmg9VpPTH769Et4KwQY10UnUaRO9WrO9gB36fYk
+          +8wYjYGh382CtgSy1eOaEkcOkaX0x4kx5xLNdLdTob5FJxxfa80Qw/O+g/u+M4iR/JqkCx9aoZjD
+          0yYBo5be4pVrB3/86N4lfb3I6sNVpFnIsDqEUTnv4uMVSV6YhKz6cwyaRHkKx9uOIdvzvb/1R+fd
+          hw0pES2PNBI2ITkKA3HU2SqFRNIGReD9kHhlOnsb/n9k+ZpdJ0bkKrQ4xfoA+TjdwnMdZOVwl8YM
+          jE9vYuf5Zb0B1m/0Vx+nfIG3N8PtmyE/F0JsuLhGNGoiFxw2qbAuLToSjqLWwJMLOmJnyY6S4Ys7
+          RELngAsun4z5ss1afUdkGxxhseXI5nYBL0VNwjW6PcpFtG/NH5/EbiX4CYEImeALXk8cJXITrv9N
+          NkS99SBhhPdUrPOlgDDbLcSb3KWfRH4wZYkmyzRGqVQuCV/aoEctS66zKZTz673/gAyfioRpYiXi
+          SAsVws7Kp2SoFGN8p9cCSTOXkbPyePaLOxcv2IXaQvTkoBpr1jzhjy9Nu7VySm5XEhPYChb8p+eo
+          Ge91ID/NwSfmgtHcVlIKn93cbPyVrSmgREVngwNyBImvV/WVuXBdvqeJ87L/I+1MupXlmTX8gxhI
+          JwlDemkTBEWcASKKDdIkQH79WeznHX6zM95ruTVJVd33VVDh2uWn8T2UejHGB+4gl4twuFVgqHwJ
+          R74iDQt2CxmeknrBrsbkgUjN6MJ3PO2R8JJQuij8mUBbDSLqRNESsF+5X//qF5qXTzwMt37m1C1f
+          0LqCajnZeRLCup91WtO5TudXFRtqkJ+2u2l9vV3Vdmyge/Xyf/lTWnwlA42svqlxmaVgur1LHjJ/
+          n9PreNmBCdxVHqocGnCklq92BhVFcDn9jjRXU9gy8/jiVN4kBtYO8pQurw4mMAyCjgZWg835kA0z
+          eJ9VmXDlC4I5OVgNvBCew8WcTWzcebRRpPfnuPlDA/ARDX14zkMeSZ3RDN2JhT10zd/lL16DZb6q
+          FmDspaNpfinp5JAkgWV5Q9QrkgPgp0DuFPDMnjTo1mM6N6+ohpJsjtSJgpjNPN778Cqc3T+9OjAm
+          KjV8/9w79gqJKxlhq6OcXYdt8fZJF0XrfNCIfUA11BFzy1/rf3odXR5/9VMBYY0QPbxPWrAaVPvA
+          ffpg+IBNIejcceBhENUukpwdP8y8+ZmhEqISMV9UhvUzmRrcf15X6kXlbZjv8juDn1fSbr/ny75b
+          PlTV9tTTq+M45aLx3+df/aOHps0Gxj/WRtn4AbYbgZiUMMVRn2ucIqXyw1RK3nEMoW3ukPp8PIfZ
+          vZgiCJrgg30+R6Zw/6Sc6txKTID7xKV0b5wZ/s4zxenz9m5J8vwR2B7WE45ad1+uRSHJ0DwGJjb3
+          6cEcT8zq/vwZ9WNrGNbRbms4yYFB3f5elku1B6Ny+ooLRjt5G+yNPQ4O2SxRE+wOwTz4cgNonbrU
+          cuiXjf19roBnBSVqmX8LFnBXxf2KhxRrGE5gjstXrgaG6KF5O28sLZABpek4oGanLSYlKRb3p6Ra
+          yG69jMG46QPIBq/A4TXkh3E8OzEIh2uP7cKOSvZ54Rg+0l2HtcrB6W+11QRK/gdRZN/f5prKHgFy
+          eTrRtMqqYMG3VVERz19xVv8WMM86aeBeHAuqdWFszn98pIL4hG3uStnwK5cZ7n4iQwI5IrAYDSUQ
+          J56IPftyC1bDaro//U/T1oAme1yWBHZ9NxCJkKadf/c4V4vz26dIaXYm+wFcwcdYIOxVuGYzssME
+          fizjQZgvFsM49SQGp5s145p6h1ZC0aypsV3VeNN3Jqu1xYHb76fRbd4H8we+eXXemYgwF7zM1Yfi
+          Cpv60OKIfM10JlI8gj20CmrnyzSQv79v/pAau3sb0ESZY5XfVzY1bu5l+OcX4mRQ0O5w/gWLKp4h
+          RJLHaCDDtaVXLh7Vz3Lb4S3fA94egh6aj9Siwa5L2nU7b2BPjgw7R7yk5Br3PjjLlYjxSx3aZbxA
+          EaDT54JRyMN0YI+kgu1FPuPoLhjlapcN/8dLMM48P+1pFXzAH6+x4qADCyWCBls6zdS47fVgyaO0
+          UWuhMahFzb4kxBleMFrKmbpu8myZHr949UnxiH4u77BFHxZOhfdlRzX5dDQZvrwJ/JdvlDgphT+e
+          1e52K7YKfwcYC5QK0LR1MTodPyaT0SpCfyUuPuFiZ7LJXS3YPucL1pKepCylNoQKn+4J57swoKTD
+          M1j7X4bt2ouAkPDMUrd8Sv/ib4VtWMH+4QrYE3yvXXq9GuGktgY2UpKxeZ9cK0jy0qL2iU4pCTuO
+          g1u+Rs32/xbMEgXq50L502Ngma2bAx35ePzHN1gG5SfM0npPCnVG6RIsRQ/DZuixtfG9rUkN1Tjy
+          ZOxHeVC+VUPL/j4fR+ZbKyVJXkdory2ifu6GgaQeUx9+79sj97OVleuh6jgQR4GM4KcT0lclRS7c
+          zgdFb00F7NrJ45/+JazgyrZPt6ld6305bXp5YuyI8xBuPA9HcAQpPdsfF/YPjadRVupM+Jz4CoLd
+          Q6U+06aB9YejrF6umYqDexaX/KYXVf9tPrf9XIJfMK8vdTy4Cg14eTXZ8SUjSLqHjUviH1s+F0L5
+          H4+8brxW0IcFwvpd6Ijd7HfJLkn8Uo83wFED3AwwL9Eswz8+dOzIKf3jTerGc3C1azXzTw9AIInt
+          Vn8v5uox3VE3Xkpvkf4oZ8n+auAvXxrpoQnWZR0IfPC4xxiXHqNCuH7gJ6CAaimHAl4Ex1DlHbMi
+          6u9opottJjF43dYr4c+tF/BW337+4o/0sCItsUp3BdH+c6KBiNZgNk9to85oCNAsdDH4x4c33oqW
+          C0oYo94aQu4XEnoeWpJObQLRf/xk4zs0Nm0ebP6I0Mu9T1ffffjKX7wFkYKHeQrkHoZpjSiaH53Z
+          m0VD4NkQeWxZdmeyrjxuUyv5nEAOa6k03hQDbt+PSJ9eC75/+/GnT61EcAd2Tef1Hw+NUYdM3h7M
+          Hgqv2sXR79imszsOIpQi/YcdLM2Avq39C9axoVPzlJQtGe2hhp70smhSBs2wUFPQYNf3AzkJbDB5
+          mn0q2J2TgZrwU5vzZ519eNOnDCcpEcFMtE8Ou9mfyN6nt5T+1aNCEEKM3OBdTo6pZPD7mr9kj8Jv
+          wLb6/I832kEalnwZRi5s1eZCEZ25dHlzfAaEu3Oghmg/W/adjx1IRNjj+k2/ATlquwzofr9H6z1q
+          2Kxy+x7IqU6pv+mpzZ/X8BjzL+zs25TRRkQ82PwLvbk7PVgPBw1CK0wdiqt9YLLRLTTwx882/svW
+          uusrEPGnM9ZZxQ8jr1wdsOkHGv3ph6XNZWgdhwN1DiwuFztXYviLC7r5gyvjaUYqyC9eg92fbA1i
+          ft2tQM0/PyJx2yt0Twlxyt9+nU08AIbaWIPz/oeI2krPdomZAsGmn6jr7q/lzHIywkABH2ycWZMu
+          dp4gtXFiF5vZ2WqljcfAeyLdiZi9j+W6wy8Cf6BYEFvEW8qkaCUAhOMdxyOay1USRgcqC3enls7V
+          5WrSroO1tLthZN/tze/mPnA980INPIrtnz+F+1LMMRoO33Smz3iGVlUekDy+Xmyuo+mkxG/SUDfq
+          n2x8H2Zfje26pm7ptMPyp6eTY9Zt+z2m89edY/X2veRIvWhyOq2FGsKjzT3J/tpyAy1OuQhxEojU
+          RXifzvNz5qEmuDKRq6oFi+7THvSfgseGbFxbprxWBQYqF1IjvjXpQkjFwdAlEQEf5xgs9qQg8Knl
+          J06Dd8vYcLtCiIuqpL6rBYFo2c0MNr+ND59eM8Uc+A58fDfA50t2ygvK4sIcHzQynzM0jG9or/uN
+          h+AgfOFhPawaVL3dscaH0dKY5HJv9MejCZ9jL2WJEocwsI8vij/HoSR22YiQB+yOgNVQk2BRr4Al
+          pR7We+NkLu/TqVD3LluJYnz8dL49IwSL9xRQgyPXYeEuzQnen2qz6bep3OoDgdKF9Th4WjMgKnp9
+          oMqFA914HHvW3bP+d36im2eU//pLYoECetj01WJkqwHhrXpjPW6SYXmfqhxsep1ag+4Hq0a0Cr4m
+          ucAn+/pla2nvOPDHj4wh1VueYzSHWccpaOw0AazJ80GgJOvjX3+kXEk5NFAZq9u/+BzvyM7hGDYK
+          9SxTLf/2C/zlg81fAxJ4OvrjoYibH12wxK4CYaDCECkbH16JM3zg15PfZEeHAXzCu57AseoBNfJH
+          w9b44T3BdFh3NJi368tXt+X3bAgKxC7Dq126M3ztx4OvoHvm9Sm7JPnnr39DVguKAa12swwvKadi
+          S2BDwIbbEaqFfffIpl8ZLx+XQo2W64yvj+/OXPOrtMLnlQ7o7/y+VJcWUCsS8Y9fBrOfjwaMfbHG
+          kXHUBnEqzQomo9Nhy0zq4Pm3vhi/rjQH2btlwrN4AeyfCupu/crlQNaTuvUHCTBd1PK9H9eq8dF2
+          ZG1FCZAz+snwNd8ldHtj0/zj8396DsdRgkzhGrGP+hyVBAemS4blLJYxVNlnpCaVPuZ8R1EBt/4A
+          kn5Zk478a3H+6h92v6GV/ukLsPUzsOc4Tio08KSpTdAcaHUYU3M9VA1UMYUV1eeVBwtaFR7eRcPB
+          3jNzS96bAlnZeA/Wef5sjn5uwr96gi35HpRTJrKTuuU7sszrCczepZXhy9Su1NIv34Ekg3QCGx/Z
+          1oszu5v0DqFqWh2ug6pk1PpOGhCf5hc75jIHi3MtTrA6nlMaPj9nUzruZR+iK7zQfHxZTDrbxAV1
+          4xp081Ns5nb7FRyHzwv7wELmPz77kHxMo61/9Et9Fqrm52fheHrN6QjU/RO2q5KT9x+PG5r4qYr3
+          T0BNYy3MdYtnVXsbEtl4HNv2X4MyeGU4cZLMnE9u2IE/Po+KnZOKkaCfYAWj08an7sECm7OsDHoh
+          YVfR7IE/19ao3LtLg+32HoHFLYoKLqfhiC2rr9t1Xs1ebfGN0OM3Y4AG56mCSTZ2Gw/apkhBaMFI
+          AyNhiVKafXE0kLrpc4qKWA3opB7IXz+XatccAaHTTi/41OWYeqegLlmUvT/wyVNCzVB9pctqC8n/
+          6+ID+X8/UuDR+0y9+TOYzH41UJUvc4wtdk/T9YJTbnvsbU8eSqIxwfTYU80nvMOH0z1gbKg+heqi
+          /kDt8+dRTgJ+QbVVxQlbb89I+Sj3Gvgpj+1mvStT2DNWwWM+11RXbLeUOgYgaBTriENp7YKlrDgZ
+          eCRbyZoUr4A4kVqAS30ScewjOsz6+dbD7Jgj6ifpFyzydA/hWTUPFO9uU7rm5xmpu+uNYpN0j2Ep
+          G9tQ7UoM8AH4eSt2J7dXv2aI6HFuECCgO2ZqlXEfHP0iKX2aI8+pYqKoCGZJC/hBnC0onnRA0+wt
+          pO+1d3v4K7kHdsDMsfW+nn2oo9JA8QDGkulTF0P2UFqs5+9XwN9nzoGzt2i0ehtOsJa6vMLPvHQI
+          PBBIe0/8zkCA7wwfRv7QznI0KNAht4yeW7iWzH51UN13yY3Wkq8MC7xeZjAt2kQjKdSZSIp3puqv
+          8UfPYHUCMW3TAsbNMyeDu78PogVMpH4Bm6mr0aQcX5+HoR7MYI+1fHymaxL/Yni/ggZHUn0sheZi
+          ExjNHcRRLIOBHbvnB3bSu9wMqheQElQVlB1IMDrnHzB7o8apLHhBiu6GlLIWggSW3kOiZ/mISul1
+          YJWqyp2DL8DWB77/7T7w3B9t7EhHEcxnYIyQr2yPHiNBB2t2eECVtfoDu/n7Y0pE4Ff4I2OL4yWN
+          BqlAewj0sHrSe4fadFbD7AVlJwgRLE1QroukizCtV0rUiW/YK0nyE3TxYtOk5zOwDNflBSneC1Rb
+          0qhd98nVUqvL6tKCy85s+va9AlQt1/DRel4Af8jcEcr2uUei40plUz+/PPxw0oEo/ALL9U0VRx3k
+          t4D12/HVSqfjTYO/wN3R2/TtUmZlxQhrAXEYbVfHzON2F5lpXU/4Jh1zICF/PkHhckio/XuE5grY
+          sVJJ1iw097ogZes7qAFX+z419vdnMF3tToFYaDAttv/Pdjs7g4qAnW0QYDeM1rL64C9+0/vwAazd
+          Tz58JD+eCC1MUsbR3wns7A8lr3B/a0kXr40KVvWIt89L2d1zV1W42Ak9kUebih/jlcB8ON6x5tHv
+          MF34sAaH73NHJHwyA+Glz09VFOsH9Xr6AKxjgANng39QzdelgI2t3yn2oyG4vC+tyb4d49UCtYS8
+          +kNSitd3heCwLy/0sFsMc30vBxFU4Ofi8/mjl0IrzifYmrsWa7fbnpEQNTXc65xE1otjtWJspIYK
+          Vo9iq3nKYGqKlFdvR0gRh0/vYHyerV49zc6durc2C0Q0JgR0sSvRUP0dS9qBN6f+aGPjWj8r2+Cf
+          2ofIzDl8Oydhy5ikviBt7wU9xElaCvM580Fdg8t2Ph2Tt5bVha+PkmEEbxYQf3dt3XqRD2q/OhbM
+          i3pawfCtJKrnPTDZCSEF+pYqYHst3sGalI/i77xSKztyLVkNw1D5URmwn6wYrPvk6KhHR6vwyZhf
+          jLRZO8LmOwzUG8pnyX/2qa+OvGPRv/zGQ7g+VR8qd7TnIDRn5aPE8P1Q7vie7g8lf+ryD/yLx7Ss
+          UcmoSzh4eWKJGvvlOAjjY3jBX5Wq1L9Yu5LF71OtDt9awsiZyuHv/Cr98Atx8ab9sM7y/QNboERY
+          R4elZFMPe3h9cyGNAuQDFu1LGY4hBwgnXrfz4vUNNNwkovo8fcvl+A2eML2DCaNXIQ/rxxgTaGlo
+          Rw0uvraSEHb+v++P0/23XHYaX8Pnyr/poSuyls+brwZ4N/RoyedoILHScBBZYYMapc5MCR2eSD3C
+          zMAaKl1zgdf7DIuomMl+/gzB+uG0TC2edookfC1MfjQdAy5OHuHL9VYCcVhWee8ctJUWiYPLObqD
+          BrCQKtiS37a5HNegg9k7Rdg+uGh4/W4PAg8/3yQwREb5Vz/Uj9s+sDucXXPukt92t+6G7O3nYhKz
+          OzpQTGSVZjqo2b/fd9MMjJghmaYgZl4IE8WJqBkVI2CXaWngk/gGtVkAhzHwQAxuUsWoZrtcMA/s
+          wan0fVUQLDirFNvrnoenuQRIMbPIFAWyDaizdgKSnxUO1nnUCZzr+xO1vPbc7uaqV2W8jyqapFAH
+          gvqJY3WrVzTIhmVY4Y5m0MyMHMHixcrfCTWh2ujGD1vN7cHGqJpyWDeSTLqP2AS8rdcIpqE24dOm
+          J+b07Vcg5wQTH9FhG8y3HE6q43oz2sWOZorm24fwdDDELR7qYbk6Yg9HZO8JM71kWCvvuf7lQ3wO
+          zgeTLy31BcF0D7HuKTKYivhhwJCXjhS9XC/gB/aA6hfUJ1zS6A2E7LUNsi65B80csymFPj6E6pKf
+          JnzdeadhnkK5gC23G2gYPPVWEA4rVG/JfMSO/sMl6z0+gXcu+SC1COxBHBZFhpJv2GSfjOd28KI0
+          By+8RDSavm7KdsqjVrOPM+NwOw8ScncQhk9lwTasHbC8j6fn/pasR3wzzkfAb/GlHiB+UEvOx3bl
+          juas1nTcUxw7WvDzH/ADV6M94ZA532G9dv0JcKzD+DLpWrA4PZPVuLYXerwdrWGpVslXTzSG1Gmz
+          0JzhfszhI/IoWhs1Ast8rl2oprcaPT9uli4HmCdA3oU8vh3tk9k942ZWT7OsE3EqIjBmXM7DmpI9
+          NmizvSX/WkZgcnG95etHufjZNkbjMjDs9107rK/XnoOP++dMzUNrthOKBF8RX8KN6p6nl0x7dj44
+          lVNO/b98ErhhCK/adjer13kpia7NSxVf0o2AykyAkOp0hEGKtoft55qx5lx3sCy5G/ZdQkp2mfZP
+          aCU7fdMnDpvpc+7BL/B31Drgkf3VNwit15keve5X0p/z8VVjmo9oVgY0/Kv3Nn1FWM8DZi4Kx17q
+          Oz6W9O6z7a1V79lAc5BkIi27OWDcmIbwd7Pv1HLdj7lGBXcCgf41sH2+un/11wWvl3/CPj++yyV1
+          zY0A9Sk232epXTY9DSQ+HHHKWB8wuS8KaPPxjN3h2QwzFLSnWlechMBx0gdJfYvcv/xX9h+YLtp7
+          hErxlAWaxeHVXDzsVxBl8EkLegXBYre6Au+O0VA/94+B6MudCy3deGAELmy7y7owIL98EXU7vQOz
+          rWm1Cn+JtK0vB6j9KESoPG8WjdbUZTP/aLO/+o0Pk94EHRTcBu7PTUnvJ+nSkv6zzip7RjuamL0O
+          BHz++vBwmlW09wzbXCa68GqeP2400/1DKcyeFkNu5SfqJw4t5zYvVrDpdRo5T79kta108PpVxs0/
+          QHOInyiDP1Y6GNnQCuZ9/tJg+EgnAg0gB0sNHw3k3faNjWRkA1vq7zZVxqFYU0JnEJV3gEARu4xu
+          er/8HqTUgi+ZF/G1UjGbu3R3ghUY3K3lqg3Cnx6MDt8Uka1+jZp/a+A5K2zCrG0O9aZfwJJnEw0i
+          XTcXeZ5O8KZpmMz53WOSQx/W3nQuPBoqcwXkoyghrMPgTNb+W5frFN4tQJNTRY+X6Alm91l/4OJM
+          P4px5LBtfRLgTFmONdv8mkTo+QK+H5ZKszYbzddHtxEkOo/pGc5iuqaNnYB5FX9U00Q0SPrUxOo1
+          DX0ce/Tb0mn99NBzRYKjMzeZYxh5OTS8a4DROSDprNn7DoDqhLb9lNrV7K7OfuTLhgaZ7Jt8ZsAV
+          NiDCuG4rtxQS1VyBVxUJDh1RDwQh0C31R0iL8a5/g3+f34nVDVd/+9OKcwaH1F6oZbRjycLlocB8
+          tHXq3mbSMmSdHPVPP+RjM4PulgY8fEQBpfarLFrWH0kCmzMNyBju9mw6DjSHH0444EvnHgLh+5xl
+          GJzOFnV3t6CUfhy0oLOLZRp/BIVRrU9clfwShvXhK7T/4k1FM6DZ9W6a/Ht5+OqgGSlaLpHBKDeW
+          CG716Z+fWC5jUW1TGd5I0OdoWC+pxAMB1DmONj3PLustB9v5Qb9BX8rFS+KPKl/HmTpnOwkmbomL
+          f35O85t7wCZizOrm19Cine/pSjJv09fNjjD2WYb1kvQOwPJ3or4pnM2ltsoPnGfZowfx+jBXcSWF
+          cpoVHYfb+m76zIcN1Bbs3SoJrB5RZ2AH0YEaVf1mq2QZPOzw5USEV5eaTBeKBCjH4UkNutul46Yf
+          FEu/qNR9VthkqVPJUHneLaxTnpjsGh9G+P3KMz5o4jzMkrKMai2EHFFMQQhW5W2G6l98HcBlCf78
+          ALQFJ6H2K2sH4guJoR7ztaZ68jBKyeFDqLQj/hGK71m5elNnwPBxnHAUlG3a/9qxV7lGjohCP6wl
+          ++TqgDffPZDCL1X5T4/er/sGVyzgWhY380dlEdJQy999wBd7IYcGHRd88ad9O37UnQEr+/zExw/m
+          B5a9Di68assLu1rcl6tzDBq4rScSJgYGagbJ50//0sNNGtmKacdD7fSQt/qQmkI3LYZKY5ojvrhq
+          wVQ2kQbLXaPTqnixdE5vcQX+8rl1F+VgmIi/AvI9pdh4VAKYfuZyUsFh56K+HDrGwlXq4YX6OfXs
+          QxWwTzG9oFFqGEdFP7XEmxoDdvh8QurmT37H4ZsrbvzpCIvGKp2H3Y6Henxu8WHTs2xaSQ/a/PRD
+          Y5O9h0GxuJeCzpeaOl/hFKz3aUnU7TzhU3bkhiXVrhASa5uVvYZ9Ozd8iGDecyY2j3uesc/D4+GR
+          +3IYn96uyTqaO/BeTAZSXILSztYzBG3BSujtV1BAuvz1hG1WeBvfcc3+67kNVN/OE+3tAzRXJxJy
+          9c3vVOzTQ9COq434/e0tUaJw1qdctI45oL08Dvj8EoKWPx7yEAKH93E9pUvLpGdXKOVSPanPebdy
+          7c5Zrv7l43DdY3OW7b0CM9mMicLYaC4hyyGc3c6gGlLioEsf/EmmaJWIuq3vu49xCLbzTQ+K/AvW
+          P/2MhSfGdr14gOlr6IIL06/U/QDJXDoFyfCG0YjNKDqV/PEQI5WiWaLuqd8eAQvl/p/f026SDwh3
+          /vZ7eaxXbDzwLV26tVIUbhWnP71T0i2+FEu+2VR/GlWwkhuoYFGjhEhfIoBl4s8EZFJEsH2WZzDx
+          cJ0hgvWHzM82KOcj9kOV33UT/adn3Wf92ltJcKeosL7poo1eCIW+E7HjmE26ZI8ig388xT3lU9td
+          Tkcf7j/DlRqW5wf/9Ph9V256rJzBKlm+uH+HeYHz3W0ou2I1C5jo4gEfP9VkMliHHGR3aODD0mxT
+          TrofD9NQxOhlzBZYjVsOYa++Rxz0H1gucavm8Dg8GxqwMWITrCYZqPvFwtaLLun8+oozPBVAxOZ7
+          Gwwe7WYfLoNypFEwrYwiV+LgocseSHmY/n/+QctciPPb8TOsyu2IVIyuDq7x6W2uYZbU6seuTLzV
+          91ZQv8SAWXoi+I8f/vEhte8bRE2ZW9lyLkQHVhZ6oZVOFCwHGCfgWdKKhuFuD5ZwNvh//t+9WigY
+          6OGKIMd2A9bSgzswQ0sqlZ/rmEjLKg2r5n5qeFU6Ql0L9mz+46MlIffNDx9SwcNGBWHWCNiPRi4l
+          H047qbLjhVgbKiuV8HmuYJCGDnYt6P/xR1f981ParU9KgSNGou6ud0r9Sz2Xy5EIjXrsCYd+qHBM
+          8U1XB3TtWOKLOMStROBRU//8mfcx9VTgNRdCj7on6kyRUDIr+3vFgE9psX0f6vpRomiZDwmU1ZjN
+          J/5Tq9VldvH5Hi9gkW4s/Je/Nv/WMiESFGjha0UPrXFOl6tRrFD94AYtO1Mul5f3fELhIvMbb2Yp
+          +dOjMSUCkjvdBXzGxbz6q2aTXpZJa1c0iy/FORgrUs/3d/C3nupDsAryyPvSXEPUVfDIxU8aqv6v
+          nF9tIqtfsMxInKJzSu/RqwdaJ3OE4A/fdl3yeIGYv+a0fHtWeud4e9u/DqOnGc9svM+cBcubXVJn
+          TT+MRbDXgGmHIg0Yp6a9JakOWPVpe6v0egBMXq0cPG8DRAONbLYIeOTgpqfRapXXlvmVlIOPXZtY
+          E3bInPMqcBUsJ0cckIf9H2+veUejeJyGdMYCW+GjvWdIYSw0fz3a19DFzEbAxAgwFDoWXM9pRjce
+          D4hHhBX61wxQv1eYOe+3Fkiehzq9ctLblAi8bhcTiBL1LuvBnCa6iLDaLr6xg/MhEI0H5KHP8Tsa
+          bvViDiOvAF5zTLGxT7x2lWC+gh2nMhLVkhPM3jzyyt/5+dNzi33fLqL9uQcibnpKcsaEg/gYeFTf
+          9BS7BgOv/NXjU3PTwbq7FwY4m72Jg1JUzLF+UhHWwZlRR9x90zXMihr+3EJF66M6g3nunx0k2XPZ
+          9AkNaHlzQ6h9V42IRay047xWHbSOvInvXbOCUU8ODUQ/8sL4pnrBeO2PBfzTU6gWaraell8HOdZj
+          6t2qC1hacAqBVK0HajDFTQUmaLOaJY8QcRv/kvZy1/975BLHTmNOdqvL4LLIJU0e+FbOgzg7UKrJ
+          FzF5bFMyz9MHdiLbE5XkdzAv8izDuGlyfDvf39sUBtGC7HHQEXD395ZNn3iGf/livgZ9OfdTRCAL
+          85qAbDi2m7/IFHp8Kvgghu42aH67qO7nm9Tj+rHtze5qQSYsLWL3wWGiryUIBJLSIf7X9eXAeFqB
+          uIJs698MYLThGivB6WIR9a3VYJljJfkX/xmoacrG1ujVtL0lZHkaMCCS5fPwWeYaroLno13s16VT
+          z4b4QPKm35fPZZ7hxpfpoYs4c9ntjQKY/ajQ/PM7tBI3liEE5XPE6PtrANvvxXyPo8RFYG9Rc4uP
+          AtaBdqFHC3Zs1VxSwzab3hRJ27M8G7+GM+sCXDsTaJn2tTTouTwhL6MNy3/9giBtn0jyy95cFPPe
+          KM/p1JDdbY7bJdWOHBTxJadWQHjALkLbKOLJBHir58Es24usBrh/ELXt1HaVDzH3V3+oMxUTW1d0
+          kRXr+2sR6Xhv4APn7e83vUEgKQ7Bz9V4BN93t8I3ku8AMWfTUfU7elHvYz5SZqJmBJdwWjd9QoPO
+          /Cja3hUVhNSNl84M7RH8TMTDzlN6BfQ4oRo03xOmhiv0Ldv85N/vw+4p36YGvZpa3XH95U//pEzw
+          CgNyQe7iwsyiYH3T1YKn4lPQg/bSzblLpZPy2XU/fCInn839b/cCt3SPyXOLbzH9PBoV6VWN//jy
+          tr89LCdF/devo+MdNv/iZVs/NtbXxoKKEDk0/D1QMBvdYYVbvsabvjLZEMfPP79H6MZP1hCLIeRd
+          5KF5LE7tGEZ6AW/vJiUbf2nnP/6aoS6gumZS9qfXgB5fWor90wTIrlVlUDdNhc3EiVKaPfe5Yjpn
+          Hp03fzgbd82Af37a2fzxPE/zrFrvyUQgCX8Dg9eHpRpeGaCn5fmmdJPG7M/f4Tv0kCm9L/dayTDz
+          0A5rKGVjmfdw878YdefHMGZcLKpBkr9w9RKmYX245wq0gzNgg+7u5e/7lBXoX/KSHs2lMVfIHV9w
+          0/vURSwFY5LxoVoHF0b1rZ8qJaAicPYBh+gUsnTppr0Gf9VqIpatE+hQq/ewUuMQVyzVTJFvjUJN
+          tMjEuXjZg9Xsjpb6XolOw/fp3E7Urw2YHyhAIry47crnfgJKr5Vw+COfdlE48IJDsp7x4SaFTFJ/
+          tqVqSmPQfOuPzOZwk2Hk1Ge09rwIlncod9Cj/gl7Vvbappg9NRiVroQNXs4HssUDFMyWYBc9rFQi
+          e0NWSyrdtngt2sX55Npf/xipWAnKf7xMU54GPsR2waahPlcgOrxTnD0lK1i7k9apF68/Eb6JvVI6
+          Hc+aenfOkJr7IEgFeL2vMJP5lf71a5Zq3bmgQyijxuOHUpFFhv+vnuLaf7TLnoEKVGoS/vVjTKqY
+          Da/S9lbg9LjnwVTkqQH30Lv9p4dbGTkwmg6Xf/z5vfZap677viIiqHG5NMmD/8eL/voX004lMfjL
+          V/mW/2bpgDuYX087JG35eQlZzEGc2j+i0MMwLM8pgvBsMR2H6kUH6+3qzH96j9rtzU/Xe9/66unA
+          51u/+Vz+/vjYidwL1CcrZavNXU+wuJ9jGjLnMEzV55GoW3wTccsX6+rxBpSHISZrcgfBrJ/PHYyr
+          4EjmfLvacgfzp/K338e5IYCWHwz/Xxcf7P/3IwWPek2o9ePeKVvLc67+lt2Mw2kY2HxptBXSNYcY
+          mWBtRyjHDaR1dCcfMdNM4YDYR80t36a+3kwpKbkj2l/vvkQAfiSmuEO7F0y+8hkf5d2RCXloQ7WS
+          VQtbfvAYJJC1GjycLhCt8B4GyzeZHLjj0xHNu8OjZaXW1ZDb7mu9vwM/EKxr2wGX2+3JKEnfgH3F
+          4wpGOAe0bNhQLvSXPyEwiI4kvt0N87gsBQzXxsRGB/yU9zhzVeUE3qn7fHID+/6qBvq/t4T1Q9YO
+          c+03vfrWJ4O6iiKUazm6Phz3T4mii28Hi5JSDh7r8I5rxfy0cwCbSn1OxhUJ38fW0udbDVLnqmCH
+          xc9Smi+jAusNjveRX5Wrcute4PW0JVI3ksIWnP9m2D+Qgzd2yFrddSvgbteJl/R0Bqt8N2W1lCye
+          5v5vGpixiDkw37uaPLX9q12s8FOoQVhn1PE+STu8e0mElzvXYbMJn4NQX+41ZHPxpkHDgpJ158ZX
+          F5zZ2BXfjrl0Ekaw+0CVRrCxAX+YagNe47uAAxw+wfrFTb21u2xC1VZjwhKtI0wqocT6OSnLFa83
+          ESLp3aALET02a2qfgcFwES18eXurNH6doDi7B+yEFz2Qbt5rVd/3q0BE7m0E/J0aMwwTWaanocrZ
+          eqg1Te0j44fNXXA0RROpH2jjpMZ26uGS50l/gn2k/egVf0HKxpvnw8urUTC+HS+lOFz0GlYQ5TRw
+          1IZNJTQ0mBzLCS2HrG0JAs/XNhSzpXZAjFS02GVURRJRerAkH/BpzGWQYemIIx1PA2vJPKurOFJ8
+          WOB3WHdIev3tF7W3eYCL/xJ89XkOeVz+xGcqRUND4McNWmqnHk3Ha7sk8Hp3JeyWh+Mwu4f4Bbfz
+          hc9nxwXCORMNGCu6SY9YOYGF61C4334vvam/3lybt/aBYpR+Eef6J7AG+TDLrU8jivIuB4u+5BB6
+          MJWxZopeKk7FnlN2UXXGhlbUbJRNrVd3aEkI93sb7eo95ie4frgbPlhSz8aBVSPUFa3F9TxoqTB+
+          zUz9ywcXN7kwqSXy/O884zszB5b+IKcQ9XjBumDvAtIm1IHZTmvo0a/0dI0ubg3L9dlSf+DBQOsO
+          OBDvbzO+gYueSiSFOZSODsbO4XcN+GR/LeDh8o3Jv3h/DG4Cp2f8xpXRuqXgPL8caCrFwPe1nVOG
+          bq4LeuXU0UNUQpOoyfxU4U6VqfdzxWHVPk9ZjfU0w2a02IHw2waHCUdZo0dhfwqE2LjPkL+2ByKe
+          Drd07SXggNw3U4zu1685dJFIoOS/P/gSh3278q/GVf27f6HO6XArBVX8rFBM13hbz2f7irPkpd60
+          uce2eddKvtqvIjx2v91/+/mGQw1fVdhhM9O+pvSxlAI6p9zEByXw2mUokAVxuUIaLB+unZ7hN1ce
+          wveH5hk4pZDveVH9XZ4xEub9OBD+8pTV8nL70YNUlIGkfuNVpb/kSu0oFE0mCoca6mh7JKfVX+nK
+          ++8eft/WhLf4AoLZhQiCkm9xoTdRylu2Wah5/X2hPRzzkl24MoQzmlfqn+Ii6ITGW1W5WC3EtSQr
+          122/wXeUE1yh6FKutH7W6i9wr/gSKTqQjvK5Uae7HtMkRlY5789XAtbpFZLxXXlseV7HDjaVbOAw
+          9mfWB/mwwq+vM3zFSRiI36+dwYePdJq2U9VKoOT4/WT8Cup8gdGKw69oYFOHO6ylTZ5K5jFcoXu9
+          N1Sf+iRYLXYn8DG4Hg4HkqQ8aXNR9WerofruqJQL+hon9WafAPZiRy8l0uw7sM8dk5rnMDH/DwAA
+          //9cfUnPgsy27nz/ipNvanakk6o6MzoBaapAEDG5uQFFFFSkK6CS899P8N25gzsWIsKqp1tLah5u
+          Qg1p3NvkFuyuHSd+TRPNnjkTawkFNvhaiVE8SgdyM9OZ9QWqYsihvhrlCDjdUimFBqtR+mJe0CYw
+          mafZRpXQE1K4/Ctve7uR4dAaDs1x6gMRNtAEGPMLVcJbCP7qO0DBeeXbY8WfT1kPDw/hQo2nsXSz
+          Jws9RGwpqTWxKhrMJ4VQOXsuDcujC/g3K0botf2N3J2H04nK6Evw/N27xNnqWOeL8/mK4q7liWbH
+          czeRWUqBNOSbsf6uG6+coafA7RNSjNDXcfm9tNvIp9NQUpV93FwYvUsGslC3iZHQI+stOQpgImJA
+          VHXcVVP++S6gvaQlzSHY5CwwNw7ki6YjedAmlcDtL8LOwteBJOfFB8vH2iqQ1cpIAhpd9YUW7RXe
+          0jeh9rmvqzHY5iY8yBAQa+XzZZtzAjx8hg+JufukT6bNZBgtwCYGdw/c4WHlNXRZNFCzul4BF/p7
+          DsFbz8j1FPGu8KquAhytZUPP29gHvNmGKVqeTjHy96Oh07P+MdGqH+hNrrycuYOnwEZ424QAkHS8
+          YAIHlm/lRQ/zOe3Y+PavcF96H1pIshaJx2NgIFvtAN4eBr6axPCLof+wtwTDE8nnnVpIwJte/bjD
+          hw+YBZPZKL0FH6I73SdfuLqxET4lOTW5i8mEUDBjZD0bh/hqfQWTerKOqIOeRPSIpJ1Ym6cRkt19
+          Ik4VtPlUnxYbJVutpPfwY+fz+ZSNaNhLIz2I8Qzm1JRtKLiwJYdd2EVvm68ltOlLn+heozD+4KcB
+          0K8cIMWskIqPRdrALxnblY9ndz5zDw3qX28h6sssGZPbewa05xSS2NpXgH0/OwzbS1YS7bO8wNzM
+          IAPGLcN/emvUUi1GzTHtKX7fa7b0oHrDwSIdnu4bmC/qxpogZ0DhTx9OxFBL9GWSRZVyCbtlxb8f
+          ftBr4ARsmXciB00+cYl7PCqd2G4TE3ZY3tE/fXMcqxjWqkvH9iQZepcf5yda9SyN6ySPGFVsDbk8
+          NojPhRXgP34YIKMYXGrG70pnbiZD2HDd/Ic/E7WAApuuwpSEUOymUo0k+EK9TvSiTvP54AdHVC7Z
+          9fd82fKejqY8iWVJNTsOq2V+Zhuw4jWxOBxEi8Md3sDJdzq5ebUf8f3umsGC5K+xMkbsMuZBAR4e
+          3AWjXnsAJj0iU17xl0auklT8O98JoBF3e2JHfrHWBwngfn7UVLsbhbu8qpiDS3XLcf6ws46XD3YC
+          vPu3InqgD2yyL98WYiGoqCkkisvTax4AdbubqX6UD/m04j98Km01Ljh0mFCKYYuy3lDJWdyVjD0K
+          XYGnnQypXj5CMDzjlAMrftCiFDO2pIdDAhO1f+Jk1cOza7gl/KaiT017QR3jd0qKKkn3cLm1Hh2j
+          iqIhWb93I/fxBbA8klJB60v6iGZqNliGpRDg59tuiHV/tO507K/2T39g8BW0fK5vNweYfOySxF5u
+          3YRr5iB2bkU8Z/I+Z6JmLMBLbI+elMNWH2MT2iAEICKeb0T6vLXiGir5dKVRM0VgSfHNkS8RPRA8
+          jiaYz59qkRtT47G04RqdbXZfAajah41PQov8p18hSxFHsJzybAZbUIIKBCXBj0TpRMf+SsAE/oUe
+          uJuQzxZ8FlA8HL/U03uvE4Q8SnerfiC+cXIjXj4oMXSC+wnP1eETMXtpAuhLsKaXx3DopvZ9vsJO
+          c/A4H1BeTaVJe7i9CN4oxGCJKPKTALJMckli3MZo3LSnTE43mBLXzuNKTG+PGl0p9yBXTRIrto/5
+          IwD11iTGFYzRIh3NBoDGy2gUpEifTxU7on3XnIjaPHwwf5NbDTYe0Ih67PRcOElRiYKnX2AZyQfQ
+          v61LDWfBvY7c8ky6JRRwDGii70cU7GBHoy/3p++oKXc07/t5l8FZHgZ6geeD3q/1B7yY8NQTyE1n
+          VisI6DRUe0KOU18xus85sPF2GklI1QH2lZ0A8qGsEMJ/scvZBX9F4jH3qAnudd5DKS3hu2hdit/K
+          Vp95rtVgFIWI7k1TiqaxCoQf3+LJzmTQf2brCaR+r47z6QjyhVZC8OePNBw6QPjEXLCOeEXE+uI7
+          G6LSWaB3vh+ImUldN90uqgefgg1H2Xi7+nS/7kb4QqNO3UKYATseUxMe7ciheEfCatgEjwRGmHbU
+          PIR6Li7J4Q1BgzNqIDGJfnoRrHqM2t2Lrxb4uJhw9ZPkIGekYhE0ld0cSWcSFbWU0yoWAhRcY0Lv
+          wq3SmY/mCfpu8yJBUdpMdD8ThljbW3j6SFRnJ6tUZGrmMiUG0dwZ7/IU+PFsEzdkMmgaVzNAjZwJ
+          bzCo3HnL61d0uB0tYmtUBss37Sc4tKaDhddFyOey8TEcNk1OD3JGK3Y4OwW0RC+nhbbgfEJUnyC/
+          Lw+0ODsvfQkn+QlfckHx5pGUP31sorV+6EGZ1WqGniRAv6+H3/2txt/xP38Sy3Na8S9VjOWbX52I
+          udbDcuO9Kyys1c2WTQxYuYENaNUakPDYVTnjet2BhGrByld7wLztyYRNi0Oib85XvYcEFfCOjiI9
+          LFut6kKm9rDF5zdZ/U01HI61iWRZzv/0x6hcXg3QrlFDDc4pGWuL+gpuu+eZHHTUR1O6hJzM37cN
+          8fS+7xioygA9HVceuc76VlOe1U+EiqM5ltvrBMbW8TAI23eCuczZstn3LQEqCq0wqqHhzmBoC+At
+          Tx33+QtVk+3VGVAG6zJKaviJBje0F2CqNqVaejOjxRh1CZpD+SJe5agus41dhn7r5xj2ZzDt67BH
+          41T29Cy734rVqiojeIz2P7/Szdc8SNHTv2CyF7eOu1y28Ama7oHpYeW7nvenERrv5zxKuRWuO3Pf
+          3vJQfZ5jl63/Cv3xJRfGA4YRr0bzC3M1ynivJmk/Zjr9+RenKzENWkcEU7QZMGS68SZXhuJouchN
+          D9ffh6l6MCJxULtE/uF5Yfp59Hf9mfB40PX5ua3nuBgeA3MimrKn1RJWKADzyY+pk6t2NN2KkwbK
+          vXnD6KaFEdumk4d+/EdQpQBOWDIOnvikw7td6OYzfVADqkbpUNc8dIxp8yaVV/4h2PTznF/9HjBM
+          8KaHe16y2ZM3I+Rsz8GCVnJsfsapIKUvwydB09fRsOIpnMezQBztG1ZLlQUlCjPkUt+rh5wKPCl2
+          FYkKosaHs8vZXp9CW6+PRB1OGuAk3W7hmqdhfgf1ShShJKGTVyO68l81tFNgIjPdxFSZwDvqBpp5
+          gJRnTDxfsvUlLB4FEEZCqffTU5ehK8BPL966KgfLqSsCWCN7IieIvWixQmf503vp0kU5fZUBRm02
+          3MhFR17E7YHwBlTY+nhbn0/uPPvyCO+3rh93X+h2q35UgDDfDWIb55c78b7UQzt1KvrDFwr5XQwf
+          3LHHjGctW35+ZBl3B6K+8kmf+ppzgN4+6LiUvZ2v+ccGjs1gjpl+e1fi/T7GEPslTwKs1Yx+anv8
+          5VEEJ909mjvtgdFWfCjju7pCNt264xEq1VsjGkisrgVWUqN82ITjbNwbd6hvJwftlKAgRrBcOm6t
+          RxCnlwc1LvzUzarQLzBk/Y6omfyKhi3vFlA8NeXKP6Pe1jOfwl8+ZGTfV/e3flv1DYgS2lrF/Gob
+          gLvu+TSueVXnwEQNGBhlS25hqkWzcPE8wNW1ivlTLulztTEXCLv8M655VkeF27NGpziriFWkeyCO
+          5dyiFd/oOc37jt3v7xgFzWhSKzyNbDl/Fwga52USfS8eo788zXHeJlHZp4uWdb0ATT+YeONHO3fN
+          YwIg51FAsZyewPoK+iP4vMwBB0E8VNPSzgLq4eISVb20eVvEGQfEkWnU8G5tNOvWsQZrPkp1l+u6
+          xgliGcrboieq1R3ZUjwCE9W1hokepMhtpt0xQ/XTEsddWe9zkTqfBr6enTMu6e0dsd6RTOBzaUS8
+          kQtcobzdHciuvk09OyjcAem9Ab3DciL+8fHuGL4pNvKjwiX+8r1Us/acUoSTLh2bDWfr7Ml/JFlw
+          Ny01hsuSz0gYJ1AbuU4I2W70odpLAdQDI6BqVSQV63WzgbuYo/QSkn23POCrgHAPMNUe6rlb6MSe
+          AHVQXvMey10AUTRkjHFCFBiYuvjjHys+wVGhiwrYK8o2sn54fYiamVO+0Ak8oeCHH7yoOGJ/x//8
+          s/d+YbcXxLGFsmaPxNMksWMXTV/grz4tTGa992I/gys/U72ZGJvDx9uULXXUqKWP70icdamEl8lS
+          6H4eHl2fGi0Emu6amGbbsWP96d7DuCTWX148Gc4Zy+t6wMKVf+nLcsH4p28xb5ppPi7qN/nz4472
+          nbvlp4f4D8MjJ1+/0Zzohxas+pLaKx81LrfV5Kd43xHNPKf6MnO1tm60scVS5K8t/sveg0/PDPE8
+          5WU1nUt7gqR6P0deii8Vu760Gup20BLVNbdgXsLO2xVZcSdml5CueSwsQLud4pI8+RD241v42d+f
+          RJd2TTcfQfn+80crfrKJU8AVvh6JT63Aa7tpQZoMsvvoEOWrsmreBI8YPXdZMG4mL+uEFJ9W/6+9
+          RsnCSsWDob2CKIoQNWxs50trSCm4BDeeag/e66g2CxlYhJES63y0XT4RjASqslZhwaEJ435+cs2D
+          Rza172qyrbT+yxtWf54vVzo18PNtNvRA95R13YSeYM1z8Q7gupuOuzCDz10a0GuUgIqanP4GYYHv
+          VAXCEC0e8mVoHm/9uheGzsbk/JXhhJeFaOHZ7uZgMhu4P5vGyARI9cdt+F7hL+8pOpTmbKqfDpLN
+          fku083d2Kdb9Bvzy/yO8ezqjTOtRcimtcdV70exrpQda0L2ojbyN+/X4Vgbr5CjVmPXqhtRc7L/8
+          zeepmA/Qkzh4jL/qWK/6lml9+4Z9Dd80XvFjuiRQAavfpKr2xNVc1HEP9cAMfudHf3539eu4/dWf
+          chla+NNbZNO/dGaL0IBxJAZj/gw37O9+VdqTUVwE2qpX6wWGtTWSA90ZYG6rI4e+PDHGJcB1NHEm
+          OcKqKgBmwrpZCOTnGK6731JLdr8d491LCW8bs8G7kLw6tnGXjZxfMmXkt9lZpzG+OxDYH0L9uFD0
+          nx4Ca56MoSha+nKwNgokcH2nkCVLf3gA9fDik/2wvmJM8hYB2teRo0SiHRgWY4CgvFeQOPtMcYU1
+          T4DpwtRxkzYpY6aqSDK1k5BoK3789ZdyKZ+pn012NW7h+wpeW1Uhd68wdOGRNJq86nfMhrphS+9z
+          Pfq6zoVq3GsXNQXhCkTDgqOKsh3Y/MvTG1/IqJrl94q+2q2wW9fruPo9Nq/9PXj+Wi7R1n1f5lHc
+          mfDaf3zM6VvTFbuJL8GKlyMrailiD3e3gZ+gvBKCqnVkm/tqUEkre+23vfSFlMYEbKIo9NCGGViy
+          uyNDeBsZ+eUzc1tlnPyVrinNrhuViVr/fMOfnhOTg86W/fFtwPthnxOt7Jt8idUy/j0vYnzv24rF
+          TdgiUagbkjiZVQnqfj7CIrvesXSsyqhv2iBDqWHv13z7wabn+VSgX352z+MbWK9XgfxnxmT/Th9r
+          /+uZwLDptlSn2jUa815x4JoPYATLF+uTR22g2z4B1PTOD3fuOfsNnZf6pv4StxHj3bBEWcMN1Le8
+          N2DSe5TgnBYNWflNF+WX+4bc1z6SSLu+crr2C2HjfEyq3SWmLz8+XOuHGlbG9BFozvKXp7m8+9Rn
+          i748WJGwWPtDG8Zer14BBKQWzdvtu2MpXRq0EdBArNk+dpOpNwb85YG615SAGvxbhsHmFROFU2uX
+          +RMs4Xo+CfYx0mdWlFeEjkJKFcVN2Q8P0EaSDKJGnNWt/bAA5nJij7sVz5gCXAfu1FdO0uezqLpz
+          0Me/fGN8FazIpxOJFESq+kkDTq31Xz8arfxCPeyL0YKO0lOGmwbQjO8sJmaPzITjJzuQAzx/dQYb
+          aMBObD/UGeVdPhbvSkLPR92RLAJtN5dBmcB7079XvwD1aXa7J9zKU0zvA83z4RQ+PaScIg5PcuR1
+          NN/DAO52mkv8y6fORyt0pl+eM0aCnekzvUYBDNs6ofZNKt0xVpsE3A9Wjl9N63S8GyrTT68Qd7tc
+          2JJvLh5E95dOfedYr/3SPoHbMbtRk1QumzP0iuExSgyy5v/VX94FrOBIbdTpjAURyGCVJek4vakD
+          +NA5YPD1bgnRLuAC2LVNMuh8PyJuVn5l9ywr4cewXzTp75zb+0qowP+MFPzrv/7r/6wDAv+8m1vx
+          WgcDhmIe/v3/RgX+Lf67f2ev12+w4J+xz8rin//+zwjCP9+ueX+H/zs0dfHp11kDJEv837jBP0Mz
+          ZK//76N/rV/4P//6XwAAAP//AwDHbHt3ugUCAA==
       headers:
         CF-RAY:
-          - 96665c82da4167ee-SJC
+          - 96a9b549bc1bed3c-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3001,14 +3002,14 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:21 GMT
+          - Tue, 05 Aug 2025 22:25:03 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=97cvrx_C7IwRXYGUDUWu9qhIwwq_PzDWMNEzfgPEnAo-1753726521-1.0.1.1-VjiApjkT7Bpj.tZtfqtyQZr1ovBXKtPZ60M6lIqNPzEHeFJnaFhs3mhk8AOQIKC3NyTAA5pCgMnbKYV_m93yfn7aX2hkL1FwGKuGhk3pi.0;
-            path=/; expires=Mon, 28-Jul-25 18:45:21 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=kKF8kk0rBVZy6VdmdGaPLEkzFFBomP0I7VhkNh_z8Do-1754432703-1.0.1.1-uHVOgw5jxO07snk8H5HFE5U59lQblKS4.3S69sL1Dt71v8_WITDXeDSw3B78H50MxzXFT5zW95uL61hE27uS.6z5BiKXe.fgpCH7tpWWBdE;
+            path=/; expires=Tue, 05-Aug-25 22:55:03 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=sokXYWby45WRqU6BuuNesctQL5oo7pn2NmnNeJpaW.I-1753726521487-0.0.1.1-604800000;
+          - _cfuvid=5MUMkUaoplmI.56MY5DgJYEYRXQXlODVezdCzA2.XU0-1754432703468-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Transfer-Encoding:
           - chunked
@@ -3027,7 +3028,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "518"
+          - "344"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -3035,23 +3036,23 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         via:
-          - envoy-router-f55cd75d4-rxpxr
+          - envoy-router-6bd4b5cc9f-pwnfs
         x-envoy-upstream-service-time:
-          - "611"
+          - "385"
         x-ratelimit-limit-requests:
           - "200000"
         x-ratelimit-limit-tokens:
           - "200000000"
         x-ratelimit-remaining-requests:
-          - "199998"
+          - "199999"
         x-ratelimit-remaining-tokens:
-          - "199979818"
+          - "199979820"
         x-ratelimit-reset-requests:
           - 0s
         x-ratelimit-reset-tokens:
           - 6ms
         x-request-id:
-          - req_c3c80c85f5e74fd16f99405ef30b8643
+          - req_8519daf19e4d1e2a20424ae8a9d69213
       status:
         code: 200
         message: OK
@@ -3075,7 +3076,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3085,7 +3086,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3212,7 +3213,7 @@ interactions:
           98//e/xf61/917/+FwAAAP//AwBe8oQ14CAAAA==
       headers:
         CF-RAY:
-          - 96665c87afb567ee-SJC
+          - 96a9b54e1b52ed3c-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3220,7 +3221,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:21 GMT
+          - Tue, 05 Aug 2025 22:25:03 GMT
         Server:
           - cloudflare
         Transfer-Encoding:
@@ -3240,7 +3241,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "80"
+          - "56"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -3248,15 +3249,15 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         via:
-          - envoy-router-f55cd75d4-56n7x
+          - envoy-router-5586665d5-v7v7c
         x-envoy-upstream-service-time:
-          - "97"
+          - "61"
         x-ratelimit-limit-requests:
           - "200000"
         x-ratelimit-limit-tokens:
           - "200000000"
         x-ratelimit-remaining-requests:
-          - "199998"
+          - "199999"
         x-ratelimit-remaining-tokens:
           - "199999930"
         x-ratelimit-reset-requests:
@@ -3264,7 +3265,7 @@ interactions:
         x-ratelimit-reset-tokens:
           - 0s
         x-request-id:
-          - req_650e2032020a48d793e8a257070a763e
+          - req_752844b89e5f4bbabb8c050514a20ba9
       status:
         code: 200
         message: OK
@@ -3284,7 +3285,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3294,7 +3295,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3421,7 +3422,7 @@ interactions:
           /+fyv74/9J//+j8AAAD//wMAofreDN4gAAA=
       headers:
         CF-RAY:
-          - 96665c89ac996897-SJC
+          - 96a9b5519e24f9d4-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3429,14 +3430,14 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:21 GMT
+          - Tue, 05 Aug 2025 22:25:04 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=mdE.4gn3lVyZ89IwkpfYXiWsc7ixQCjMFvlDQLD56J0-1753726521-1.0.1.1-Ef2nu5nAFQgL2xrAd.8p0.m0mbiORQm5boNtUlMC64ntU5zrS3HSluYj8beB3VcEGcg.UjyCul0vjjrMD4SN7j9U9gpjvF1kk..5zDLdpGc;
-            path=/; expires=Mon, 28-Jul-25 18:45:21 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=I6TfPxAzxjTmjcGYt5nnw91i7Ek2Ag9DMnFAxTzk0Wg-1754432704-1.0.1.1-eZj8f1Hluwh4OlI.jDwCn3nKHGs12IOV3OPNzM4V2MObG5oNgZbR7DGmQhli01FJMpqDLWUJQG1KrzvOZ9NiVZ8k2U2oyejBjMVuGyktS9c;
+            path=/; expires=Tue, 05-Aug-25 22:55:04 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=c0kRS6QhSdwcRKm0BrJF5K63TEWkBW40HO8BTtQknQM-1753726521977-0.0.1.1-604800000;
+          - _cfuvid=U4uxa3lN0XpyceKIYEA7LAmn4RW9j5lxkKz55e2Ybc8-1754432704502-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Transfer-Encoding:
           - chunked
@@ -3455,7 +3456,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "40"
+          - "174"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -3463,15 +3464,15 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         via:
-          - envoy-router-dbb4fbdcc-2z226
+          - envoy-router-6b444774fc-lcl5n
         x-envoy-upstream-service-time:
-          - "44"
+          - "185"
         x-ratelimit-limit-requests:
           - "200000"
         x-ratelimit-limit-tokens:
           - "200000000"
         x-ratelimit-remaining-requests:
-          - "199998"
+          - "199999"
         x-ratelimit-remaining-tokens:
           - "199999996"
         x-ratelimit-reset-requests:
@@ -3479,7 +3480,7 @@ interactions:
         x-ratelimit-reset-tokens:
           - 0s
         x-request-id:
-          - req_d16b051056b5a4d174b42b58f0dbbfbb
+          - req_db88265159f4fd69a0ac22af15518259
       status:
         code: 200
         message: OK
@@ -3487,11 +3488,14 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 25-28: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\n2021,
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 25-28: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\n2021,
         25, 1315\u20131360.\\n\\n\\n (9) Wellawatte, G. P.; Seshadri, A.; White, A.
         D. Model agnostic generation of counter-\\n\\n     factual explanations for
         molecules. Chemical Science 2022, 13, 3697\u20133705.\\n\\n\\n(10) Gandhi, H.
@@ -3558,7 +3562,7 @@ interactions:
         Ribeiro, M. T.; Singh, S.; Guestrin, C. \u201D Why should i trust you?\u201D
         Explaining the\\n\\n     predictions of any classifier. Proceedings of the 22nd
         ACM SIGKDD international\\n\\n\\n                                      27     conference
-        on knowledge discovery and data \\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        on knowledge discovery and data \\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -3567,13 +3571,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5833"
+          - "6109"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3583,7 +3587,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3599,25 +3603,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//dFTfb9tGDH73X0HoYWkB2bDdpkbz5mHpYAxB1+5XgLkwmBMlMTndaSTl
-          2Avyvw8nObHTtS+CcN/xu4/kRz6MADIusgvIXI3mmtaPf9xflXg7v/3l/urjb/sPC/nzr9u6uPz0
-          2Re78yxPEfHmlpw9RU1cbFpPxjEMsBNCo8Q6W5y/Wczfnc/nPdDEgnwKq1obv43j+XT+djybjefT
-          Q2Ad2ZFmF/D3CADgof8miaGgXXYB0/zppCFVrCi7eL4EkEn06SRDVVbDYFl+BF0MRqFX/bAOAOtM
-          u6ZB2a+zC1hnl7vWIwe88QRLMS7ZMXpYBSPvuaLgCF5dL1evQagkUbAIDVkdCwUMBRi5OvA/HSlw
-          ADwy8CmD1WjQ4F36IyjIsXIM4wbvOFTQSnSkSgqxhOUKdK9GjYIJBm1RKFje80krZElq3r/dhYIk
-          JVz08i1C3TUYdAK/1wS0cyStDbqTCgWlLQl6uI9yp4Mo2rU+CsH1cgUuBketaXrM+a5I2tgUCio5
-          cGq05k/J54Bt69nh4TjpcTV6T6EincCHKMAhiXOUw1KEyRDIAP0EXs2ns/evU95bLggQkpWEagrK
-          W4K4Jdky3adyXC9Xp3IMdzHEhmmofmzbKNYlcaRQRgEhbWNQ7vu5msBHq0mGhHPQztWAKfOoBDd7
-          uOqkiK7+Sldi/rkLIT34AyxrPAA5FKyuU33ZDGjQ1RwIPKH0QYngp+XnX5dn2he2lVgJNnmvjpzx
-          lvy+75ISqHVFUk9NW6Pyv4NHuEmZpeqlKph0ajmUyBJID+VG52IXDG/Ys+2T/Y7eyQE9V72Ye7Ya
-          hKrOo0XZQynY0OAAzwdHXv5xpnAmXNWWbJRMgaFv7eRsneXD3Ah52iZFG3VRKM3PbLoOj6fTJlR2
-          imnYQ+f9CYAhRBvMkub8ywF5fJ5sH6tW4o1+FZol62m9EUKNIU2xWmyzHn0cAXzpN0j3YilkrcSm
-          tY3FO+qfm52/XwyE2XFpncCLtwfUoqE/BRaz/BuUm4IM2evJGsocupqK0zffvHtOAruC4xGbjk5y
-          /7+kb9EP+XOoTli+S38EXBpoKjatUMHuZdrHa0JpsX/v2nOte8GZpuF0tDEmSf0oqMTODzs3G+y3
-          KTlUaUC4X7ypn6PH0X8AAAD//wMAvUSTLHYGAAA=
+          H4sIAAAAAAAAAwAAAP//dFTBbiM3DL37KwidWsAO4sROsrkF2x7cLXrbNkW9sBkNZ4YbDaUVOU6C
+          IP9eaMZx3G32MgPoiY9PfCSfJwCOK3cNzrdovkth9nH+6Q+/uvt2swh/3f7y9+cl/pni5dVvD58/
+          /v7opiUi3n0lb69RJz52KZBxlBH2mdCosM4vl4vF+dnl6WIAulhRKGFNstkizs5Ozxaz+Xx2droP
+          bCN7UncN/0wAAJ6Hb5EoFT26azidvp50pIoNuevDJQCXYygnDlVZDcXc9A30UYxkUL3dbr9qlLU8
+          rwVg7bTvOsxPa3cNa/frYwrIgneB4CYb1+wZA6zEKARuSDzBT7c3q58hU01ZwSJ0ZG2sFFAqMPKt
+          8LeeFFjgZgXWokGH9wTWElTkWTnKrMN7lgZSjp5USSHW0KFvWQgCYZaCDhVTsIyiCTOJDTlYjHLK
+          ZINKi9D2HYqewMoAuRs0pRx3XBGwKDetFTXlYnwomvRJjToFzJl3BGhFG2dImSr2xUqFmA9idQok
+          2uciCb2PvRjecWB7moLlXm06yKqRs5DqCdzerIAVKlbfq1JRDDvMHHuFwYhH0ymw+NBXhZRNAVMK
+          7NHGfL7FEEga0pE7phSz9cLGw5HCA4VQ/iW2OF9yZNIURXkwbwUV7SjE1JHYCXyip9GyYqHukxM8
+          xHyvEGWs9QwbiWrsgUofyKsc7XOODRrtLdk/OGNHI0Ed8xhSWmcozd7+falP1m46dlumQDsUTxv1
+          MdPYdR8OcK9UbbjDhrRANQaltbysZbvdHvdzprpXLOMkfQhHAIpEG3WXSfqyR14OsxNik3K80+9C
+          Xc3C2m4yoUYpc6IWkxvQlwnAl2FG+/+MnUs5dsk2Fu9pSDe/WC5HQve2Fo7g8/M9atEwHAGXV1fT
+          dyg3FRly0KNBdx59S9VR7PL84vAI7CuOb9jp5Ojt/5f0Hv34fpbmiOWH9G+A95SMqs3bCL13LVNZ
+          nT+6dqj1INgp5R172hhTLn5UVGMfxq3mxr7a1CxN2QU8rrY6bc7mF3SxXPhF5SYvk38BAAD//wMA
+          tKsSSuMFAAA=
       headers:
         CF-RAY:
-          - 96665c8ace4d4705-SJC
+          - 96a9b5553da215e1-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3625,202 +3628,17 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:24 GMT
+          - Tue, 05 Aug 2025 22:25:06 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=SBwPH6wNVhgwcqUjORsl6rhewNnrY3T4HbKxvSVw0SU-1753726524-1.0.1.1-0jaFk7Erch_bdzGebSJIlLZqthCtyBJOfSt6Dj_q_.8jVDNn4fRAlj_nhsYsVlIKEB1ygPlEn61VANv4weIhnEDdjPcAQKFQLHSQ436YwQI;
-            path=/; expires=Mon, 28-Jul-25 18:45:24 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=9QxdLCR8FKvU.8b0IWL3nViURfpVykQEHawdLyA2wu0-1754432706-1.0.1.1-9B_lchMaWl2I_HeGcc296_z8TL.nACjKbVVr0sie5yzTaC65UXm2PTA8addwy__NQF7z_PPsFMb12RmIrhRnsqc2T3qlimqOS48zF.j9kjQ;
+            path=/; expires=Tue, 05-Aug-25 22:55:06 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=0ZpgjodlTFvuaj8a0mlqcisBTY.shk8fzLyQOCj_Yhs-1753726524682-0.0.1.1-604800000;
+          - _cfuvid=k_ewve5RirBY0orrQ0EM6fOUGowmrUOBthRtBwnZkIk-1754432706410-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
-        Transfer-Encoding:
-          - chunked
-        X-Content-Type-Options:
-          - nosniff
-        access-control-expose-headers:
-          - X-Request-ID
-        alt-svc:
-          - h3=":443"; ma=86400
-        cf-cache-status:
-          - DYNAMIC
-        openai-organization:
-          - future-house-xr4tdh
-        openai-processing-ms:
-          - "2544"
-        openai-project:
-          - proj_RpeV6PrPclPHBb5GlExPXSBj
-        openai-version:
-          - "2020-10-01"
-        strict-transport-security:
+        Strict-Transport-Security:
           - max-age=31536000; includeSubDomains; preload
-        x-envoy-upstream-service-time:
-          - "2549"
-        x-ratelimit-limit-requests:
-          - "10000"
-        x-ratelimit-limit-tokens:
-          - "30000000"
-        x-ratelimit-remaining-requests:
-          - "9999"
-        x-ratelimit-remaining-tokens:
-          - "29998614"
-        x-ratelimit-reset-requests:
-          - 6ms
-        x-ratelimit-reset-tokens:
-          - 2ms
-        x-request-id:
-          - req_4d87a34beb214789412137cee164731b
-      status:
-        code: 200
-        message: OK
-  - request:
-      body:
-        "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
-        the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 3-5: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\n
-        a passive characteristic of a model, whereas explainability\\n\\nis an active
-        characteristic which is used to clarify the internal decision-making process.\\n\\nNamely,
-        an explanation is extra information that gives the context and a cause for one
-        or\\n\\nmore predictions.29 We adopt the same nomenclature in this perspective.\\n\\n
-        \  Accuracy and interpretability are two attractive characteristics of DL models.
-        However,\\n\\nDL models are often highly accurate and less interpretable.28,30
-        XAI provides a way to avoid\\n\\nthat trade-off in chemical property prediction.
-        XAI can be viewed as a two-step process.\\n\\nFirst, we develop an accurate
-        but uninterpretable DL model. Next, we add explanations to\\n\\npredictions.
-        Ideally, if the DL model has correctly learned the input-output relations, then\\n\\nthe
-        explanations should give insight into the underlying mechanism.\\n\\n   In the
-        remainder of this article, we review recent approaches for XAI of chemical property\\n\\nprediction
-        while drawing specific examples from our recent XAI work.9,10,31 We show how\\n\\nin
-        various systems these methods yield explanations that are consistent with known
-        and\\n\\nmechanisms in structure-property relationships.\\n\\n\\n\\n\\n\\n                                       3Theory\\n\\n\\nIn
-        this work, we aim to assemble a common taxonomy for the landscape of XAI while\\n\\nproviding
-        our perspectives. We utilized the vocabulary proposed by Das and Rad 32 to classify\\n\\nXAI.
-        According to their classification, interpretations can be categorized as global
-        or local\\n\\ninterpretations on the basis of \u201Cwhat is being explained?\u201D.
-        For example, counterfactuals are\\n\\nlocal interpretations, as these can explain
-        only a given instance. The second classification is\\n\\nbased on the relation
-        between the model and the interpretation \u2013 is interpretability post-hoc\\n\\n(extrinsic)
-        or intrinsic to the model?.32,33 An intrinsic XAI method is part of the model\\n\\nand
-        is self-explanatory32 These are also referred to as white-box models to contrast
-        them\\n\\nwith non-interpretable black box models.28 An extrinsic method is
-        one that can be applied\\n\\npost-training to any model.33 Post-hoc methods
-        found in the literature focus on interpreting\\n\\nmodels through 1) training
-        data34 and feature attribution,35 2) surrogate models10 and, 3)\\n\\ncounterfactual9
-        or contrastive explanations.36\\n\\n   Often, what is a \u201Cgood\u201D explanation
-        and what are the required components of an ex-\\n\\nplanation are debated.32,37,38
-        Palacio et al. 29 state that the lack of a standard framework\\n\\nhas caused
-        the inability to evaluate the interpretability of a model.  In physical sciences,\\n\\nwe
-        may instead consider if the explanations somehow reflect and expand our understanding\\n\\nof
-        physical phenomena.  For example, Oviedo et al. 39 propose that a model explanation\\n\\ncan
-        be evaluated by considering its agreement with physical observations, which
-        they term\\n\\n\u201Ccorrectness.\u201D For example, if an explanation suggests
-        that polarity affects solubility of a\\n\\nmolecule, and the experimental evidence
-        strengthen the hypothesis, then the explanation\\n\\nis assumed \u201Ccorrect\u201D.
-        In instances where such mechanistic knowledge is sparse, expert bi-\\n\\nases
-        and subjectivity can be used to measure the correctness.40 Other similar metrics
-        of\\n\\ncorrectness such as \u201Cexplanation satisfaction scale\u201D can be
-        found in the literature.41,42 In a\\n\\nrecent study, Humer et al. 43 introduced
-        CIME an interactive web-based tool that allows the\\n\\nusers to inspect model
-        explanations. The aim of this study is to bridge the gap between\\n\\nanalysis
-        of XAI methods. Based on the above discussion, we identify that an agreed upon\\n\\n\\n
-        \                                      4evaluation metric is necessary in XAI.
-        We suggest the following attributes can be used to\\n\\nevaluate explanations.
-        However, the relative importance of each attribute may depend on\\n\\nthe application
-        - actionability may not be as important as faithfulness when evaluating the\\n\\ninterpretability
-        of a static physics based model. Therefore, one can select relative importance\\n\\nof
-        each attribute based on the application.\\n\\n\\n   \u2022 Actionable. Is it
-        clear how we could change the input features to modify the output?\\n\\n\\n
-        \  \u2022 Complete. Does the explanation completely account for the prediction?
-        Did features\\n\\n     not included in the explanation really contribute zero
-        effect to the prediction?44\\n\\n\\n   \u2022 Correct. Does the explanation
-        agree with hypothesized or known underlying physical\\n\\n     mechanism?39\\n\\n\\n
-        \  \u2022 Domain Applicable. Does the explanation use language and concepts
-        of domain ex-\\n\\n      perts?\\n\\n\\n   \u2022 Fidelity/Faithful. Does the
-        explanation agree with the black box model?\\n\\n\\n   \u2022 Robust. Does the
-        explanation change significantly with small changes to the model or\\n\\n      instance
-        being explained?\\n\\n\\n   \u2022 Sparse/Succinct. Is the explanation succinct?\\n\\n\\n
-        \ We present an example evaluation of the SHAP explanation method based on the
-        above\\n\\nattributes.44 Shapley values were proposed as a local explanation
-        method based on feature\\n\\nattribution, as they offer a complete explanation
-        - each feature i\\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
-      headers:
-        accept:
-          - application/json
-        accept-encoding:
-          - gzip, deflate
-        connection:
-          - keep-alive
-        content-length:
-          - "5792"
-        content-type:
-          - application/json
-        host:
-          - api.openai.com
-        user-agent:
-          - AsyncOpenAI/Python 1.97.1
-        x-stainless-arch:
-          - arm64
-        x-stainless-async:
-          - async:asyncio
-        x-stainless-lang:
-          - python
-        x-stainless-os:
-          - MacOS
-        x-stainless-package-version:
-          - 1.97.1
-        x-stainless-raw-response:
-          - "true"
-        x-stainless-read-timeout:
-          - "60.0"
-        x-stainless-retry-count:
-          - "0"
-        x-stainless-runtime:
-          - CPython
-        x-stainless-runtime-version:
-          - 3.13.5
-      method: POST
-      uri: https://api.openai.com/v1/chat/completions
-    response:
-      body:
-        string: !!binary |
-          H4sIAAAAAAAAA3RUTYsbORC9+1cUOmXANrYzMw6+eUgChuS2kLDrYKqlkrvWakmRqs00w/z3Rd3+
-          mmxyMW69eq++62UEoNioFShdo+gmuslT99XibP6N7tdP1eenz/4hr79ufD7Y+NffalwYofqXtJxZ
-          Ux2a6Eg4+AHWiVCoqM6XD++Xi8eHxaIHmmDIFdo+yuQ+TBazxf1kPp8sZidiHVhTViv4ZwQA8NL/
-          lhC9oWe1gtn4/NJQzrgntboYAagUXHlRmDNnQS9qfAV18EK+j/pl6wG2KrdNg6nbqhVs1ff1Zgwh
-          wafn6JA9Vo5gnYQta0YHGy/kHO/JaxpDIkspgwRoSOpgMqA3EFPQlDNlkBoFGjwQSE1gSHPm4CcN
-          HtjvIVhYb6AvRh5DxCSsW4fJdWCIIjjC5Ivhu49f7i527IVSTCR9aMVf6w2lkqcpT1P4VrMj+Pjl
-          RAFMBMEKeah5X7sOUOs2odC4hNX1uHSRNTrXgaOc3zqZwvf1BtCYdM6KM0hCQ5NgLVQdWE5ZwNCR
-          XIgl4rMHqFqBEPFnS5dovClufREsplQK7bGMTV9J7TCx7SAmMqz75+nQjbNRTOHIhqDv5LP0ihrb
-          EpoN6ZY4hmAtpeKGfeZ9LX1qoW9HXzbXFbAhXaPn3OQh13M3NXqoqFBS4Wt4V7Xs5KrR53RX5oWe
-          LzYxZJnUQd9N4dMRXYtSfBRd9sfgjpQBRRJXrVAGxwcC7OPFih1LN4bTJpGnnMtXSqRl+DChQfaA
-          MTrWF4JlQ8O/FKo2n2xLYXKrNfuBPd2q8TDyiRwd0WvaZR0SDaM/n23V1r/eLksi22Ysu+pb524A
-          9D7I0I+ypj9OyOtlMV3YxxSq/AtVWfac610izMGXJcwSourR1xHAj/4AtG92WsUUmig7CQfq3c3n
-          i4dBUF1vzg38cLoPSoKguwEWyzPvjeTOkCC7fHNFlEZdk7nhzhb3lySwNRyu2Gx0k/v/Q/qd/JA/
-          +/2Nyh/lr4DWFIXM7jrjvzNLVO7yn8wute4DVpnSkTXthCmVfhiy2LrhZKrcZaFmZ9nvyz3g4W7a
-          uNOPy+XjB/v4YaZGr6P/AAAA//8DAGBrHiJABgAA
-      headers:
-        CF-RAY:
-          - 96665c8adc875c18-SJC
-        Connection:
-          - keep-alive
-        Content-Encoding:
-          - gzip
-        Content-Type:
-          - application/json
-        Date:
-          - Mon, 28 Jul 2025 18:15:25 GMT
-        Server:
-          - cloudflare
-        Set-Cookie:
-          - __cf_bm=2Ra8VgCBGKwIGuC_frPBbWj7C8CcaBrMSkw.nFuQ0n0-1753726525-1.0.1.1-PpyAr4EPYhgbqj.8OeehfQgHQ_0VEEGygMh7fYrvwk8Z2l0eA2M6cHmRz7RVXPhikHypC9_NH4vSeZvMi.7GaBqhwUUpOUtSZuo2Pjgs2K8;
-            path=/; expires=Mon, 28-Jul-25 18:45:25 GMT; domain=.api.openai.com; HttpOnly;
-            Secure; SameSite=None
-          - _cfuvid=XKOmb7iV_hkoED7v_kPk1YRbbH1vITaB81RgNewigiE-1753726525138-0.0.1.1-604800000;
-            path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -3834,15 +3652,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "2988"
+          - "1499"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "2994"
+          - "1506"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -3850,13 +3666,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9997"
         x-ratelimit-remaining-tokens:
-          - "29997881"
+          - "29997137"
         x-ratelimit-reset-requests:
           - 13ms
         x-ratelimit-reset-tokens:
-          - 4ms
+          - 5ms
         x-request-id:
-          - req_11be38ecfd9daa05cab103989b8617c4
+          - req_4ef95248f92648f79b59ea4659ae4cf7
       status:
         code: 200
         message: OK
@@ -3864,11 +3680,14 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 1-3: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\n
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 1-3: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\n
         A Perspective on Explanations of Molecular\\n\\n              Prediction Models\\n\\n\\nGeemi
         P. Wellawatte,\u2020   Heta A. Gandhi,\u2021   Aditi Seshadri,\u2021 and  Andrew\\n\\n
         \                          D. White\u2217,\u2021\\n\\n\\n     \u2020Department
@@ -3935,7 +3754,7 @@ interactions:
         a passive characteristic of a model, whereas explainability\\n\\nis an active
         characteristic which is used to clarify the internal decision-making process.\\n\\nNamely,
         an explanation is extra information that gives the context and a cause for one
-        or\\n\\nmore \\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        or\\n\\nmore \\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -3944,13 +3763,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5817"
+          - "6093"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3960,7 +3779,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3976,24 +3795,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA3RU224jNwx991cQetkEsAPbzW39lsUWRdAt0L6lrRc2LXE8jDWiKmqSDIL8e6EZ
-          J3b28mLAc3gonsPL8wjAsDMLMLbGbJvoJ5+6Pyo8v5653T/46Ub+emoefkP8+/qu/jOJGReGbO7J
-          5lfWmZUmesosYYBtIsxUss6uLn65ml9ezOc90IgjX2jbmCfnMplP5+eT2Wwyn+6JtbAlNQv4dwQA
-          8Nz/lhKDoyezgOn49UtDqrgls3gLAjBJfPliUJU1Y8hmfACthEyhr3q9Xt+rhGV4XgaApdG2aTB1
-          S7OApfn1KXrkgBtPcJMyV2wZPdyGTN7zloIlOLm7uT0FVkComLyDSmyr5EACxCQP7DhsgUOmFBNl
-          LN4oYHBAJXvYf6gkgSOK4AlTKJSTz19OobcJYiLHtg8cAzqXSLWE5Jrgw8aj3U028vQBAuY2EUhV
-          EKWBrWdwd3MLyI1CFmhwR/D5yx6DRhJBGxyl4pLrpZbicmo1P0rKdQebDqSqKA1ClLd11qJIyjNc
-          6rasLGHS4K7ExCSWVEnP4HfqwEqwFHtKXwkH61tHR55s2HPu4KTocbRN1Guo2wbDu9qGMA45lTJs
-          kYODkNMx3Lfad2hv6Ml/LYbMxfAHgoZyYquQa8yD8Rzgse5e+aC1tN7Bhgbl5E7H3zfpxJHaxHH4
-          J1WfQSPZ8u5xl+CREkGDjk4H91khYspsW4/Jd8BNlFTGspjSz42C5x2Bralhzakbw2NN73ozWHt4
-          w2KAbcuOoO6i9B3fD1bQMgb7DmMiCJIPg6WxTSytgpWUyA/azpZmPGxAIk8PGCyt1Eqisgmz6TK8
-          LMN6vT7eokRVq1iWOLTeHwEYguwHvezv1z3y8raxXrYxyUa/oZqKA2u9SoQqoWynZommR19GAF/7
-          y9C+W3YTkzQxr7LsqH9uNjufDwnN4RgdwRf7w2GyZPRHwPzjK+9dypWjjOz16LwYi7Ymd8Sdzs/f
-          RGDrWA7YdHSk/fuSfpR+0M9he5Tlp+kPgC1bRm51mJEfhSUqB/tnYW9e9wUbpfTAllaZKZV+OKqw
-          9cMtNdpppmZVcdiWPebhoFZxZS+vri6vq8vrqRm9jP4HAAD//wMART5t51kGAAA=
+          H4sIAAAAAAAAA4xUUWsbORB+968Y9FIb7GC7bgN+C23o5RJ6D+WgcC5mLM3uTqOVdBopiQn574d2
+          N/b22oN7WVh9M9I338w3zxMAxUZtQekGk26DXXxY3X7Wf/jfNf957b+Y2+b65vPd5acv8pt1Xs1L
+          hj98J51esy60b4OlxN71sI6Eicqtq8t3m83b9eVy0wGtN2RLWh3SYuMX6+V6s1itFuvlkNh41iRq
+          C39NAACeu2+h6Aw9qS0s568nLYlgTWp7CgJQ0dtyolCEJaFLan4GtXeJXMf6eecAdkpy22I87tQW
+          dur6KVhkhwdLcBUTV6wZLdy4RNZyTU4TTL9e3cyABRAqJmug8joLGfAOQvQPbNjVwC5RDJESFkkE
+          0BmgcrsbDiofwRAFsITRlZTpx7sZdOpAiGRYd4FzQGMiiZSQ1BC8OVjU94uDf3oDDlOOBL4qiFCf
+          LRfw9eoGkFuB5IFcg4V3illSxyMLHthyOsLhCL6qKPaMhesmSaHu4bE5Ag5sWrwnAQmkiyBjchdw
+          S0fQ3mkKXWb3Mjtts6GRBsNz08LfUB2p49zkFh1kZyiWRpnXMF+9Pj2bw/csXR8G2aZ/Z3SJi6wP
+          BC2lyFogNZjgAS0bTIMKfb2PPqaGHYnM5j/3YGpIdOTQ//lqqPpcIDyiQIuGZr2mLBAwJtbZYrRH
+          iGTpAV0qleuGWpYUj3N4bCjSqLAi78e7sXCg0UGd2RA0x+C77g1D4qS0tO8kYCRwPp2HREKO7LOA
+          9jGS7Qu52Kl5P80DIU170T5SP9Wr5Qkvg7rnFmuSglVohXbuZWyRSFUWLA512doRgM75YZyLOb8N
+          yMvJjtbXIfqD/CtVVexYmn0kFO+K9ST5oDr0ZQLwrbN9/sHJKkTfhrRP/p6651br5eB7dd40I3hz
+          OaDJJ7Qj4O0J+eHKvaGEbGW0O5RG3ZA5554XDWbDfgRMRoX/zOdXd/fFs6v/z/VnQBdjkdmfR+dX
+          YZHKKv6vsJPQHWElFB9Y0z4xxdIMQxVm229JJUdJ1O4rdnWxLversgr79eo9vX+30RujJi+TfwAA
+          AP//AwDPvyp+MwYAAA==
       headers:
         CF-RAY:
-          - 96665c8adff467e2-SJC
+          - 96a9b5553e4dd039-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4001,15 +3820,17 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:26 GMT
+          - Tue, 05 Aug 2025 22:25:06 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=4xbqp0qWQq_NT0B1QSbzakUqOSQZbm82t80QDt6cCkg-1753726526-1.0.1.1-fEApRDGmR5i6Vf5IH2KT89wm8cMcifZNddRjFl9axBPtLEutkRV4ag3nOwUyijybBU3Qzx4QbbgXYQA5_i2Y2x2JuMrNoYL7GA5BPv.pk1c;
-            path=/; expires=Mon, 28-Jul-25 18:45:26 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=.2Hn6nHMGiZeCdZR_TcULy9RsKyalPRdE_sEP18J9a0-1754432706-1.0.1.1-dE2Hjx2OKvqAAPD9FNxD2Re6MkH_9dAQSZywt9rcmXWXuyKHCwY5KjYxapaezqukLAhC6V6G8j1KjHNH2_BU9plyMwgGHcFpyurXc9.NCko;
+            path=/; expires=Tue, 05-Aug-25 22:55:06 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=V8vDJcJ1ezAw_snRLxvXL5v_cMV__JegPt67i60DbyA-1753726526587-0.0.1.1-604800000;
+          - _cfuvid=str9wXmj0mmnhlbPOXYXkywVpLM0rSXAT7R8d4aMc4Y-1754432706424-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
+        Strict-Transport-Security:
+          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -4023,29 +3844,27 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "4302"
+          - "1509"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "4437"
+          - "1513"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
           - "30000000"
         x-ratelimit-remaining-requests:
-          - "9999"
+          - "9998"
         x-ratelimit-remaining-tokens:
-          - "29998613"
+          - "29998544"
         x-ratelimit-reset-requests:
-          - 6ms
+          - 7ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - req_cb56755a22f46dd21f810f69277e274a
+          - req_499d492c8fcc4b05bcdf4aeb69cfaa62
       status:
         code: 200
         message: OK
@@ -4053,11 +3872,14 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 20-22: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nnal
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 20-22: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nnal
         molecule.  The counterfactual indicates\\nstructural changes to ethyl benzoate
         that would result in the model predicting the molecule\\nto not contain the
         \u2018fruity\u2019 scent. The Tanimoto96 similarity between the counterfactual
@@ -4122,7 +3944,7 @@ interactions:
         the input to models informs the XAI method.\\n\\n\\nConclusion and outlook\\n\\n\\nWe
         should seek to explain molecular property prediction models because users are
         more\\n\\nlikely to trust explained predictions, and explanations can help assess
-        if the model is learning\\n\\nt\\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        if the model is learning\\n\\nt\\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4131,13 +3953,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5811"
+          - "6087"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4147,7 +3969,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4163,23 +3985,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA3xUTW/jRgy9+1cQc5YN20l2W9/Solgki2K3H4cU9cKgR5TEzWhGHXIcG0H+ezGS
-          P7TbpBcBmjd8fHwc8nkCYLg0KzC2QbVt56Y/HX6t8JP/7enhjw8fP9+7+9+bv7bN1cenz/v7B1Pk
-          iLD9SlZPUTMb2s6RcvADbCOhUmZdvL+5er98d7Nc9kAbSnI5rO50eh2my/nyerpYTJfzY2AT2JKY
-          Ffw9AQB47r9Zoi9pb1YwL04nLYlgTWZ1vgRgYnD5xKAIi6JXU1xAG7yS71U/rz3A2khqW4yHtVnB
-          2jzc3hUQIvyy7xyyx60juI3KFVtGB3deyTmuyVsqIFJFUUADtKRNKAXQl6BkG8//JBJIQmUP4yOB
-          NgRdpJJt9kggVHB7B70ZAuyVYhdJ+4yZJvmSYpZf9kcaoEktepnBne+5+kr2mnna4MgmhxG6GDqK
-          ehhlKuAh5zkqdPxIo/s2pJy5QqsJHVAu2+MgMKsoSWzkTkP8Dot0ro4Gr2Dr0D5Ot2F/LGoGP/8P
-          O/tdcDuClj236EA0JqspogPboK+pNxZPWnsH0ClFYJVTeVSeKmaSAp4advSm5iQEkmIMNSqdjM+s
-          qpG3Sb9tjwaQjmzu/MivilBTJJnBnw0JnW0l36C3BBqTaAFoLYnwlh3roRj6qf3Pse3soQwtsj92
-          xDbUsmg8wDb3Luy4ZF8D9mLOb2IImUZytEOvwF64blRma1MMj/kIWdqIDZHyo/5x7V/GExCpSoJ5
-          AH1ybgSg90EHr/LsfTkiL+dpc6HuYtjKd6GmYs/SbCKhBJ8nSzR0pkdfJgBf+qlO3wyq6WJoO91o
-          eKQ+3WI5vx4IzWWRjOEfjqgGRTcCrq6WxSuUm5IU2cloNRiLtqFyFLu4WZ6LwFRyuGDzyaj2/0p6
-          jX6on309YnmT/gJYS51Subk8vteuRcrL9q1rZ697wUYo7tjSRpli7kdJFSY37EEjB1FqNxX7Oq8c
-          7pdh7ufkZfIvAAAA//8DACvbb6UKBgAA
+          H4sIAAAAAAAAA3xUwY4bNwy9+ysInceG7Xid1LdNUKBuihzaBkhaBwat4cywq5GmIuWssdh/LzTj
+          tSfJbi866JFP71EkHyYAhkuzAWMbVNt2bvpu8f6DXR8/vrcff6Xl29/efH4rH375/a/Pdv3H0hQ5
+          Ixz+IatPWTMb2s6RcvADbCOhUmZdvL5ZrV4tX89XPdCGklxOqzudrsJ0OV+upovFdDk/JzaBLYnZ
+          wN8TAICH/swSfUn3ZgPz4ummJRGsyWwuQQAmBpdvDIqwKHo1xRW0wSv5XvXDzgPsjKS2xXjamQ3s
+          zKfbbQEhws/3nUP2eHAEt1G5YsvoYOuVnOOavKUCIlUUBTRAS9qEUgB9CUq28fxvIoEkVPYw3hFo
+          Q9BFKtnmGgmECm630BdDgL1S7CJp/2KmSb6kmOWX/ZUGaFKLXmaw9T1X7+ReM08bHNnkMEIXQ0dR
+          T6OXCviU3zkrdHxHo3gbUn65QqsJHVC27XEQmFWUJDZypyF+h0W6uKOhVnBwaO+mh3B/NjWDd//D
+          zv4Y3JGgJk8RlX09kiUak9UUSeArawMte27RgW3Q1ySgDSpEkuQU2PeliCg9ybkCTFLA14Ydvegh
+          CYGkGEONSk8foQFQNfIhKYFtqGWLbkSaA6QjmztiJLgi7OXO4M+GhC7lJt+gtwQak/RSb7fjJigA
+          uczX19/OHi72p5f/jOQG2Q13Oc2XwG0Xw5EArSURPrBjPeV2KIk6cITRZ7Jcn+xDNJ5mO1MMXR/J
+          0TFL24sNkYbuX8wveP7cPbdYk2SsQie084/jUYpUJcE8yT45NwLQ+6CD2jzEX87I42VsXai7GA7y
+          Xaqp2LM0+0gowecRFQ2d6dHHCcCXfj2kbybedDG0ne413FH/3GK5Xg6E5rqRRvCr9RnVoOjGwE9v
+          imco9yUpspPRjjEWbUPlKHdxs7yYwFRyuGLzycj7j5Keox/8s69HLC/SXwFrqVMq99f+ei4sUt7a
+          L4Vdat0LNkLxyJb2yhTzf5RUYXLDQjVyEqV2X7Gv8+7iYatW3X65WNP6ZmVXpZk8Tv4DAAD//wMA
+          L4xVAF4GAAA=
       headers:
         CF-RAY:
-          - 96665c8ad96117ec-SJC
+          - 96a9b5553bb61566-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4187,14 +4010,14 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:26 GMT
+          - Tue, 05 Aug 2025 22:25:06 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=lLHrFEkRvvydEzNYVFn.KIVzajWf8GfA.fMkuCqUQEo-1753726526-1.0.1.1-f4kahxHEawTUn36QmstUqVNZtpb7_5vL_blKcEmvAwb.tHewD8rBfHyUtHSuHKdmZXDkODbXgpVxhMpcao49wjX3pklksPR8VGxZFsp.gcw;
-            path=/; expires=Mon, 28-Jul-25 18:45:26 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=77vFpgkgSziQaq3pRj7CO5BiaRSpYEPU5e6dETJ086o-1754432706-1.0.1.1-f5f6qpvm81tew5vO9C82wVGdpDhB5LqipbAmAIe4RlQAQa8s5_xtKOYyeNIG.x2s7hOCWqdFWlPbYxvUBQZdOsdRLwrO.BLgRejKpVtdQR4;
+            path=/; expires=Tue, 05-Aug-25 22:55:06 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=WPB2eW2btm_pxjZCr_8KhA6XH0HXhzZAwc534XC8lpY-1753726526704-0.0.1.1-604800000;
+          - _cfuvid=Vg.iltW988Xzxj6S10XqEU4wDLsiw6ae9EZLkB8XQR4-1754432706550-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Transfer-Encoding:
           - chunked
@@ -4209,7 +4032,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "4574"
+          - "1625"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -4217,7 +4040,7 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "4579"
+          - "1631"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4225,13 +4048,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998611"
+          - "29998544"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - req_48ec3071551302fabe02936850da29ef
+          - req_8d6aad85fef4c388133d25f1369f7f92
       status:
         code: 200
         message: OK
@@ -4239,109 +4062,110 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 22-25: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nut
-        to models informs the XAI method.\\n\\n\\nConclusion and outlook\\n\\n\\nWe
-        should seek to explain molecular property prediction models because users are
-        more\\n\\nlikely to trust explained predictions, and explanations can help assess
-        if the model is learning\\n\\nthe correct underlying chemical principles. We
-        also showed that black-box modeling first,\\n\\nfollowed by XAI, is a path to
-        structure-property relationships without needing to trade\\n\\nbetween accuracy
-        and interpretability. However, XAI in chemistry has some major open\\n\\nquestions,
-        that are also related to the black-box nature of the deep learning. Some are\\n\\n\\n\\n
-        \                                      22highlighted below:\\n\\n\\n   \u2022
-        Explanation representation: How is an explanation presented \u2013 text, a molecule,
-        attri-\\n\\n      butions, a concept, etc?\\n\\n\\n   \u2022 Molecular distance:
-        \ in XAI approaches such as counterfactual generation, the \u201Cdis-\\n\\n
-        \    tance\u201D between two molecules is minimized. Molecular distance is subjective.
-        Possibil-\\n\\n       ities are distance based on molecular properties, synthesis
-        routes, and direct structure\\n\\n     comparisons.\\n\\n\\n   \u2022 Regulations:
-        As black-box models move from research to industry, healthcare, and\\n\\n     environmental
-        settings, we expect XAI to become more important to explain decisions\\n\\n
-        \     to chemists or non-experts and possibly be legally required. Explanations
-        may need\\n\\n      to be tuned for be for doctors instead of chemists or to
-        satisfy a legal requirement.\\n\\n\\n   \u2022 Chemical space: Chemical space
-        is the set of molecules that are realizable; \u201Crealiz-\\n\\n     able\u201D
-        can be defined from purchasable to synthesizable to satisfied valences. What
-        is\\n\\n     most useful? Can an explanation consider nearby impossible molecules?
-        How can we\\n\\n     generate local chemical spaces centered around a specific
-        molecule for finding counter-\\n\\n      factuals or other instance explanations?
-        \ Similarly, can \u201Cactivity cliffs\u201D be connected\\n\\n      to explanations
-        and the local chemical space.149\\n\\n\\n   \u2022 Evaluating XAI : there is
-        a lack of a systematic framework (quantitative or qualitative)\\n\\n      to
-        evaluate correctness and applicability of an explanation. Can there be a universal\\n\\n
-        \    framework, or should explanations be chosen and evaluated based on the
-        audience and\\n\\n     domain? For example, work by Rasmussen et al. 58 attempts
-        to focus on comparing\\n\\n      feature attribution XAI methods via Crippen\u2019s
-        logP scores.\\n\\n\\n\\n\\n\\n                                       23Acknowledgements\\n\\n\\nResearch
-        reported in this work was supported by the National Institute of General Medical\\n\\nSciences
-        of the National Institutes of Health under award number R35GM137966. This work\\n\\nwas
-        supported by the NSF under awards 1751471 and 1764415. We thank the Center for\\n\\nIntegrated
-        Research Computing at the University of Rochester for providing computational\\n\\nresources.\\n\\n\\nReferences\\n\\n\\n
-        \ (1) Choudhary, K.; DeCost, B.; Chen, C.; Jain, A.; Tavazza, F.; Cohn, R.;
-        Park, C. W.;\\n\\n     Choudhary, A.; Agrawal, A.; Billinge, S. J.; Holm, E.;
-        Ong, S. P.; Wolverton, C.\\n\\n      Recent advances and applications of deep
-        learning methods in materials science. npj\\n\\n      Computational Materials
-        2022, 8.\\n\\n\\n  (2) Keith, J. A.; Vassilev-Galindo, V.; Cheng, B.; Chmiela,
-        S.; Gastegger, M.; M\xA8uller, K.-\\n\\n      R.; Tkatchenko, A. Combining Machine
-        Learning and Computational Chemistry for\\n\\n      Predictive Insights Into
-        Chemical Systems. Chemical Reviews 2021, 121, 9816\u20139872,\\n\\n     PMID:
-        34232033.\\n\\n\\n  (3) Goh, G. B.; Hodas, N. O.; Vishnu, A. Deep learning for
-        computational chemistry.\\n\\n      Journal of Computational Chemistry 2017,
-        38, 1291\u20131307.\\n\\n\\n  (4) Deringer, V. L.; Caro, M. A.; Cs\xB4anyi,
-        G. Machine Learning Interatomic Potentials as\\n\\n     Emerging Tools for Materials
-        Science. Advanced Materials 2019, 31, 1902765.\\n\\n\\n  (5) Faber, F. A.; Hutchison,
-        L.; Huang, B.; Gilmer, J.; Schoenholz, S. S.; Dahl, G. E.;\\n\\n      Vinyals,
-        O.; Kearnes, S.; Riley, P. F.; von Lilienfeld, O. A. Prediction Errors of Molec-\\n\\n
-        \     ular Machine Learning Models Lower than Hybrid DFT Error. Journal of Chemical\\n\\n
-        \    Theory and Computation 2017, 13, 5255\u20135264, PMID: 28926232.\\n\\n\\n\\n
-        \                                      24 (6) Duch, W.; Swaminathan, K.; Meller,
-        J. Artificial Intelligence Approaches for Rational\\n\\n    Drug Design and
-        Discovery. Current Pharmaceutical Design 2007, 13, 1497\u20131508.\\n\\n\\n
-        (7) Dara, S.; Dhamercherla, S.; Jadav, S. S.; Babu, C. M.; Ahsan, M. J.; darasuresh,
-        S. D.;\\n\\n     Dara, S. Machine Learning in Drug Discovery: A Review. Artificial
-        Intelligence Review\\n\\n     123, 55, 1947\u20131999.\\n\\n\\n (8) Gupta, R.;
-        Srivastava, D.; Sahu, M.; Tiwari, S.; Ambasta, R. K.; Kumar, P. Artifi-\\n\\n
-        \     cial intelligence to deep learning: machine intelligence approach for
-        drug discovery.\\n\\n     Molecular diversity 2021, 25, 1315\u20131360.\\n\\n\\n
-        (9) Wellawatte, G. P.; Seshadri, A.; White, A. D. Model agnostic generation
-        of counter-\\n\\n     factual explanations for molecules. Chemical Science 2022,
-        13, 3697\u20133705.\\n\\n\\n(10) Gandhi, H. A.; White, A. D. Explaining structure-ac\\n\\n----\\n\\nQuestion:
-        What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
-      headers:
-        accept:
-          - application/json
-        accept-encoding:
-          - gzip, deflate
-        connection:
-          - keep-alive
-        content-length:
-          - "5820"
-        content-type:
-          - application/json
-        host:
-          - api.openai.com
-        user-agent:
-          - AsyncOpenAI/Python 1.97.1
-        x-stainless-arch:
-          - arm64
-        x-stainless-async:
-          - async:asyncio
-        x-stainless-lang:
-          - python
-        x-stainless-os:
-          - MacOS
-        x-stainless-package-version:
-          - 1.97.1
-        x-stainless-raw-response:
-          - "true"
-        x-stainless-read-timeout:
-          - "60.0"
-        x-stainless-retry-count:
-          - "0"
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 3-5: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\n
+        a passive characteristic of a model, whereas explainability\\n\\nis an active
+        characteristic which is used to clarify the internal decision-making process.\\n\\nNamely,
+        an explanation is extra information that gives the context and a cause for one
+        or\\n\\nmore predictions.29 We adopt the same nomenclature in this perspective.\\n\\n
+        \  Accuracy and interpretability are two attractive characteristics of DL models.
+        However,\\n\\nDL models are often highly accurate and less interpretable.28,30
+        XAI provides a way to avoid\\n\\nthat trade-off in chemical property prediction.
+        XAI can be viewed as a two-step process.\\n\\nFirst, we develop an accurate
+        but uninterpretable DL model. Next, we add explanations to\\n\\npredictions.
+        Ideally, if the DL model has correctly learned the input-output relations, then\\n\\nthe
+        explanations should give insight into the underlying mechanism.\\n\\n   In the
+        remainder of this article, we review recent approaches for XAI of chemical property\\n\\nprediction
+        while drawing specific examples from our recent XAI work.9,10,31 We show how\\n\\nin
+        various systems these methods yield explanations that are consistent with known
+        and\\n\\nmechanisms in structure-property relationships.\\n\\n\\n\\n\\n\\n                                       3Theory\\n\\n\\nIn
+        this work, we aim to assemble a common taxonomy for the landscape of XAI while\\n\\nproviding
+        our perspectives. We utilized the vocabulary proposed by Das and Rad 32 to classify\\n\\nXAI.
+        According to their classification, interpretations can be categorized as global
+        or local\\n\\ninterpretations on the basis of \u201Cwhat is being explained?\u201D.
+        For example, counterfactuals are\\n\\nlocal interpretations, as these can explain
+        only a given instance. The second classification is\\n\\nbased on the relation
+        between the model and the interpretation \u2013 is interpretability post-hoc\\n\\n(extrinsic)
+        or intrinsic to the model?.32,33 An intrinsic XAI method is part of the model\\n\\nand
+        is self-explanatory32 These are also referred to as white-box models to contrast
+        them\\n\\nwith non-interpretable black box models.28 An extrinsic method is
+        one that can be applied\\n\\npost-training to any model.33 Post-hoc methods
+        found in the literature focus on interpreting\\n\\nmodels through 1) training
+        data34 and feature attribution,35 2) surrogate models10 and, 3)\\n\\ncounterfactual9
+        or contrastive explanations.36\\n\\n   Often, what is a \u201Cgood\u201D explanation
+        and what are the required components of an ex-\\n\\nplanation are debated.32,37,38
+        Palacio et al. 29 state that the lack of a standard framework\\n\\nhas caused
+        the inability to evaluate the interpretability of a model.  In physical sciences,\\n\\nwe
+        may instead consider if the explanations somehow reflect and expand our understanding\\n\\nof
+        physical phenomena.  For example, Oviedo et al. 39 propose that a model explanation\\n\\ncan
+        be evaluated by considering its agreement with physical observations, which
+        they term\\n\\n\u201Ccorrectness.\u201D For example, if an explanation suggests
+        that polarity affects solubility of a\\n\\nmolecule, and the experimental evidence
+        strengthen the hypothesis, then the explanation\\n\\nis assumed \u201Ccorrect\u201D.
+        In instances where such mechanistic knowledge is sparse, expert bi-\\n\\nases
+        and subjectivity can be used to measure the correctness.40 Other similar metrics
+        of\\n\\ncorrectness such as \u201Cexplanation satisfaction scale\u201D can be
+        found in the literature.41,42 In a\\n\\nrecent study, Humer et al. 43 introduced
+        CIME an interactive web-based tool that allows the\\n\\nusers to inspect model
+        explanations. The aim of this study is to bridge the gap between\\n\\nanalysis
+        of XAI methods. Based on the above discussion, we identify that an agreed upon\\n\\n\\n
+        \                                      4evaluation metric is necessary in XAI.
+        We suggest the following attributes can be used to\\n\\nevaluate explanations.
+        However, the relative importance of each attribute may depend on\\n\\nthe application
+        - actionability may not be as important as faithfulness when evaluating the\\n\\ninterpretability
+        of a static physics based model. Therefore, one can select relative importance\\n\\nof
+        each attribute based on the application.\\n\\n\\n   \u2022 Actionable. Is it
+        clear how we could change the input features to modify the output?\\n\\n\\n
+        \  \u2022 Complete. Does the explanation completely account for the prediction?
+        Did features\\n\\n     not included in the explanation really contribute zero
+        effect to the prediction?44\\n\\n\\n   \u2022 Correct. Does the explanation
+        agree with hypothesized or known underlying physical\\n\\n     mechanism?39\\n\\n\\n
+        \  \u2022 Domain Applicable. Does the explanation use language and concepts
+        of domain ex-\\n\\n      perts?\\n\\n\\n   \u2022 Fidelity/Faithful. Does the
+        explanation agree with the black box model?\\n\\n\\n   \u2022 Robust. Does the
+        explanation change significantly with small changes to the model or\\n\\n      instance
+        being explained?\\n\\n\\n   \u2022 Sparse/Succinct. Is the explanation succinct?\\n\\n\\n
+        \ We present an example evaluation of the SHAP explanation method based on the
+        above\\n\\nattributes.44 Shapley values were proposed as a local explanation
+        method based on feature\\n\\nattribution, as they offer a complete explanation
+        - each feature i\\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+      headers:
+        accept:
+          - application/json
+        accept-encoding:
+          - gzip, deflate
+        connection:
+          - keep-alive
+        content-length:
+          - "6068"
+        content-type:
+          - application/json
+        host:
+          - api.openai.com
+        user-agent:
+          - AsyncOpenAI/Python 1.99.0
+        x-stainless-arch:
+          - arm64
+        x-stainless-async:
+          - async:asyncio
+        x-stainless-lang:
+          - python
+        x-stainless-os:
+          - MacOS
+        x-stainless-package-version:
+          - 1.99.0
+        x-stainless-raw-response:
+          - "true"
+        x-stainless-read-timeout:
+          - "60.0"
+        x-stainless-retry-count:
+          - "0"
         x-stainless-runtime:
           - CPython
         x-stainless-runtime-version:
@@ -4351,23 +4175,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA3RUTW8bOQy9+1cQuuxlbMSum7a+pdgWcIHuqUALrAtDljgjNRpJITluvEH++0Ia
-          J55+XQx4Hj8eH/X4MANQ3qoNKOO0mD6H+dvTx9Zcf3r3/mO39J+d6b/f4T+f315/4L9fv1dNyUiH
-          b2jkKWthUp8Dik9xhA2hFixVl69evni1un65WlegTxZDSeuyzNdpvrparefL5Xx1dU50yRtktYF/
-          ZwAAD/W3UIwW79UGrpqnLz0y6w7V5jkIQFEK5YvSzJ5FR1HNBTQpCsbK+mEXAXaKh77XdNqpDezU
-          l5ttA4ng3X0O2kd9CAg3JL71xusA2ygYgu8wGmyAsEVikAQ9ikuWQUcLgsZFfzcgw8BoK6xvEcQh
-          ZELrTdFojLVoPNd/qYWbLVRpGIZokQp1WwlIAjf0OvICtrHWqVPcS8nqU0AzBE2QKWUkOU26NPDl
-          Zgva95VlpnT0FgHLcFGPNMRpAYchF7ZlGhpYapNK5i/+hbRmRmb47lAc0iUUPENATdHHDkwiQiNg
-          HPbe6ACZfDQ+B+QFfHKFhEHKAtazGUpFME6HgLFDBh8r8TbRWICFTg3wYBxoBsJMyBiljlBEmE7U
-          gMXWVxIXbWx9CePSuiFoSXQCwrvBE/YYhZs6Gh51GLSU3FHmOkQs4/7UZQFbAeyz0+z/Q67hvs+J
-          apsSXZVnOARtbueHdP+0XCEd2VfmdSfaSBVI5xy8OW/FR2g9BssQ/C2CQx3EGU040oxHTykW4joA
-          G1/e42KnmvFFEwY8Fhp7NomwvOw3u/g4tQFhO7AuLoxDCBNAx5hGXasBv56Rx2fLhdRlSgf+KVUV
-          ydntCTWnWOzFkrKq6OMM4Gu19vCDW1Wm1GfZS7rF2m754no9FlSXazKFz85XkkSHCbB+85T3Q8m9
-          RdE+8OQ+KKONQzvJXb2+3BM9WJ8u2NVsMvuvlH5Xfpzfx25S5Y/lL4AxmAXt/uK134URlov7p7Bn
-          rSthxUhHb3AvHqnsw2KrhzAeQ8UnFuz3rY8dUrGmnPc5e5z9DwAA//8DAOmdSqoPBgAA
+          H4sIAAAAAAAAAwAAAP//jFRNbyM3DL37VxA6bYCxESdOsptb0M0h6KIo2kPT1guDI3E8bDSSVuQY
+          cYP890Iz/so2BXoxPHp85OPnywTAsDO3YGyLarvkpz/Mf/zJSvON+vrv3y//kF+fHn+ub37pmk/X
+          cm+qwoj1X2R1z5rZ2CVPyjGMsM2ESsXr/OZqsbi8uDlfDEAXHflCWyedLuL04vxiMZ3PpxfnO2Ib
+          2ZKYW/hzAgDwMvwWicHRs7mF82r/0pEIrsncHowATI6+vBgUYVEMaqojaGNQCoPql2UAWBrpuw7z
+          dmluYWke7x4qiBnun5NHDlh7grus3LBl9PAQlLznNQVLFWRqKAtohI60jU4AgwMl2wb+1pOAtqjQ
+          4ROBtgSOLAvHMO3wicMaUo6WREggNnD3AENdpIKEWdn2HrPfgiNK4AlzKJQPn7+cHew4KOWUSQeV
+          JXQfHOWSsitPM/itZU/w+cuOApgJYqMUoOV167eA1vYZlaoicDvguk1s0fsteBJ5G2QGj3cPgM7l
+          Ube2LKAZHU1j00C9hYazKDjakI+pKN5HOEgoFWopFC8Fp1LogGVshkpqS5whZXJsh8fZ2Iu9Scpx
+          w45g6OOzDv4s9kVNE98QK4hNQ7kE4SC8bnXIZggxVspvC9iRbTGwdDKmt++lxQA1FUoufAsf6p69
+          Hn0MGZ2VaaHng02KotM22rMZ3G/Q96glRvHLYRP9hgRQNXPdKwl4fiLAQS/W7Fm3Fez2iAKJlK+c
+          yer44WKHHABT8mwPhIYdjf9yrHvZ2ZbCSG8th5E9W5pqHPhMnjYYLK3Exkzj4M/PD3gv5Fbc4Zqk
+          YA16oWV4Pd2iTE0vWJY49N6fABhC1LFVZX+/7pDXw8b6uE451vId1TQcWNpVJpQYynaKxmQG9HUC
+          8HW4DP2bZTcpxy7pSuMTDeHm84+Xo0NzPEYn8NX1DtWo6E+Ay8tP1TsuV44U2cvJeTEWbUvuyD3e
+          IuwdxxNgcpL4v/W853tMnsP6/7g/AtZSUnKr4+y/Z5apXOv/MjsUehBshPKGLa2UKZdmOGqw9+Mh
+          NbIVpW7VcFiX08DjNW3S6mJ+TddXC7twZvI6+QcAAP//AwDLWsJbVgYAAA==
       headers:
         CF-RAY:
-          - 96665c9b7e5d4705-SJC
+          - 96a9b5552a7a67bf-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4375,9 +4200,15 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:27 GMT
+          - Tue, 05 Aug 2025 22:25:06 GMT
         Server:
           - cloudflare
+        Set-Cookie:
+          - __cf_bm=cgICt4bcPOwUmu_gNGjPbvZVvmblhU8y21u9TEOwIXo-1754432706-1.0.1.1-kqK9HWWhH.X3PQ8ZOXwcCaCie_K1ZcnoTOjOE2GUcvOvMRsWsPxGKxzOl4DNvE3JT1flgDIUkkL9LWt5NFGlqbjmIJEn9k_0klFZlTtGDzM;
+            path=/; expires=Tue, 05-Aug-25 22:55:06 GMT; domain=.api.openai.com; HttpOnly;
+            Secure; SameSite=None
+          - _cfuvid=hL5zB5JFc1zq786CVUOAPsKQdYUD4eUNM3C7j3klGfA-1754432706699-0.0.1.1-604800000;
+            path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Strict-Transport-Security:
           - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
@@ -4393,13 +4224,197 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "2397"
+          - "1795"
+        openai-project:
+          - proj_RpeV6PrPclPHBb5GlExPXSBj
+        openai-version:
+          - "2020-10-01"
+        x-envoy-upstream-service-time:
+          - "1800"
+        x-ratelimit-limit-requests:
+          - "10000"
+        x-ratelimit-limit-tokens:
+          - "30000000"
+        x-ratelimit-remaining-requests:
+          - "9999"
+        x-ratelimit-remaining-tokens:
+          - "29998551"
+        x-ratelimit-reset-requests:
+          - 6ms
+        x-ratelimit-reset-tokens:
+          - 2ms
+        x-request-id:
+          - req_90bc6493b13641b8b6e9fc8978a443fa
+      status:
+        code: 200
+        message: OK
+  - request:
+      body:
+        "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
+        the relevant information that could help answer the question based on the excerpt.
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 12-14: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nnterfactual
+        approach, contrastive approach employ a dual\\n\\noptimization method, which
+        works by generating a similar and a dissimilar (counterfactuals)\\n\\nexample.
+        Contrastive explanations can interpret the model by identifying contribution
+        of\\n\\npresence and absence of subsets of features towards a certain prediction.36,99\\n\\n
+        \ A counterfactual x\u2032 of an instance x is one with a dissimilar prediction
+        \u02C6f(x) in classi-\\n\\nfication tasks. As shown in equation 5, counterfactual
+        generation can be thought of as a\\n\\nconstrained optimization problem which
+        minimizes the vector distance d(x, x\u2032) between the\\n\\nfeatures.9,100\\n\\n\\n
+        \                             minimize  d(x, x\u2032)\\n                                                                                           (5)\\n
+        \                              such that   \u02C6f(x) \u0338= \u02C6f(x\u2032)\\n\\n
+        \  For regression tasks, equation 6 adapted from equation 5 can be used. Here,
+        a counter-\\n\\nfactual is one with a defined increase or decrease in the prediction.\\n\\n\\n
+        \                          minimize  d(x, x\u2032)\\n                                                                                           (6)\\n
+        \                           such that    \u02C6f(x) \u2212\u02C6f(x\u2032) \u2265\u2206\\n\\n
+        \  Counterfactuals explanations have become a useful tool for XAI in chemistry,
+        as they\\n\\nprovide intuitive understanding of predictions and are able to
+        uncover spurious relationships\\n\\nin training data.101 Counterfactuals create
+        local (instance-level), actionable explanations.\\n\\nActionability of an explanation
+        suggest which features can be altered to change the outcome.\\n\\nFor example,
+        changing a hydrophobic functional group in a molecule to a hydrophilic group\\n\\nto
+        increase solubility.\\n\\n   Counterfactual generation is a demanding task as
+        it requires gradient optimization over\\n\\ndiscrete features that represents
+        a molecule. Recent work by Fu et al. 102 and Shen et al. 103\\n\\npresent two
+        techniques which allow continuous gradient-based optimization. Although, these\\n\\nmethodologies
+        are shown to circumvent the issue of discrete molecular optimization, counter-\\n\\nfactual
+        explanation based model interpretation still remains unexplored compared to
+        other\\n\\n\\n\\n                                       12post-hoc methods.\\n\\n
+        \  CF-GNNExplainer104 is a counterfactual explanation generating method based
+        on GN-\\n\\nNExplainer69 for graph data. This method generate counterfactuals
+        by perturbing the input\\n\\ndata (removing edges in the graph), and keeping
+        account of perturbations which lead to\\n\\nchanges in the output.  However,
+        this method is only applicable to graph-based models\\n\\nand can generate infeasible
+        molecular structures. Another related work by Numeroso and\\n\\nBacciu 105 focus
+        on generating counterfactual explanations for deep graph networks. Their\\n\\nmethod
+        MEG (Molecular counterfactual Explanation Generator) uses a reinforcement learn-\\n\\ning
+        based generator to create molecular counterfactuals (molecular graphs).  While
+        this\\n\\nmethod is able to generate counterfactuals through a multi-objective
+        reinforcement learner,\\n\\nthis is not a universal approach and requires training
+        the generator for each task.\\n\\n   Work by Wellawatte et al. 9 present a model
+        agnostic counterfactual generator MMACE\\n\\n(Molecular Model Agnostic Counterfactual
+        Explanations) which does not require training\\n\\nor computing gradients. This
+        method firstly populates a local chemical space through ran-\\n\\ndom string
+        mutations of SELFIES106 molecular representations using the STONED algo-\\n\\nrithm.107
+        Next, the labels (predictions) of the molecules in the local space are generated\\n\\nusing
+        the model that needs to be explained. Finally, the counterfactuals are identified
+        and\\n\\nsorted by their similarities \u2013 Tanimoto distance96 between ECFP4
+        fingerprints.97 Unlike the\\n\\nCF-GNNExplainer104 and MEG105 methods, the MMACE
+        algorithm ensures that generated\\n\\nmolecules are valid, owing to the surjective
+        property of SELFIES. Additionally, the MMACE\\n\\nmethod can be applied to both
+        regression and classification models. However, like most XAI\\n\\nmethods for
+        molecular prediction, MMACE does not account for the chemical stability of\\n\\npredicted
+        counterfactuals. To circumvent this drawback, Wellawatte et al. 9 propose an-\\n\\nother
+        approach, which identift counterfactuals through a similarity search on the
+        PubChem\\n\\ndatabase.108\\n\\n\\n\\n\\n\\n                                       13Similarity
+        to adjacent fields\\n\\n\\nTangential examples to counterfactual explanations
+        are adversarial training and matched\\n\\nmolecular pairs.  Adversarial perturbations
+        are used during training to deceive the model\\n\\nto expose the vulnerabilities
+        of a model109,110 whereas counterfactuals are applied post-hoc.\\n\\nTherefore,
+        the main difference between adversarial and counterfactual examples are in the\\n\\napplication,
+        although both are derived from the same optimization problem.100 Grabocka\\n\\net
+        al. 111 have developed a method named Adversarial Training on EXplanations (ATEX)\\n\\nwhich
+        improves model robustness via exposure to adversarial examples.  While there
+        are\\n\\nconceptual disparities, we note that\\n\\n------------\\n\\nQuestion:
+        What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+      headers:
+        accept:
+          - application/json
+        accept-encoding:
+          - gzip, deflate
+        connection:
+          - keep-alive
+        content-length:
+          - "6053"
+        content-type:
+          - application/json
+        host:
+          - api.openai.com
+        user-agent:
+          - AsyncOpenAI/Python 1.99.0
+        x-stainless-arch:
+          - arm64
+        x-stainless-async:
+          - async:asyncio
+        x-stainless-lang:
+          - python
+        x-stainless-os:
+          - MacOS
+        x-stainless-package-version:
+          - 1.99.0
+        x-stainless-raw-response:
+          - "true"
+        x-stainless-read-timeout:
+          - "60.0"
+        x-stainless-retry-count:
+          - "0"
+        x-stainless-runtime:
+          - CPython
+        x-stainless-runtime-version:
+          - 3.13.5
+      method: POST
+      uri: https://api.openai.com/v1/chat/completions
+    response:
+      body:
+        string: !!binary |
+          H4sIAAAAAAAAA3RU224bNxB911cM+JIWkAxJke3Yb4YQB0YRow8pmqIKhBE5uzsVl9xyZmULhv+9
+          IHV1k7wssDxzOWduLwMAw87cgrENqm07P5pPfnukbrO9bn//Mp93fzyO/9Tpp5u/2un9TTDD7BFX
+          /5DVg9eFjW3nSTnuYZsIlXLUyfXlbPZ+ej2+KkAbHfnsVnc6msXRdDydjSaT0XS8d2wiWxJzC38P
+          AABeyjdTDI6ezS2Mh4eXlkSwJnN7NAIwKfr8YlCERTGoGZ5AG4NSKKxfFgFgYaRvW0zbhbmFhfn4
+          3HnkgCtPcJeUK7aMHh6CkvdcU7AEv3y9e/gVElWUBDRCS9pEJ4DBgZJtAv/bk4A2qNDimkAbgi6R
+          Y5urszN0ZFk4hlGLaw41dClaEiGBWMHdA5QiCXBQSl0iLYyyYx8cpSzLlSeN0PQtBrmAeeyzdYVW
+          e/RAWUrAfcpEgLCmLWDXpYi2AQ7w9e5hmDNv2GUOWPjlsEPgkHNYGnnakM+/XDcqsNoCOwrK1Ta7
+          2AZDTZkncOh6hYpQ+3SQ/xR77wC9UipVKKreyVk1LuA+JqBnzMOT00IbPdneYwLbUMuiaTsE+0ab
+          gMUA0tc1ieaguU97pbkhxwiiqbd7PhHQNkwbAkfCiVxW3lFSJrmAL6fGeV4TfP58N/9YCj6/H316
+          fNwPBqVSyl7IQRUT1BQooZZSvKU4hCfWZh9nRdmiqB9hHaIo2xIcu86zPXRyFbWBRHUiybNRLKzP
+          c3zQB4qylovcOUAvEfI8JxSVXTp0G0qCKQ+tJuTAoR7CU8O2gSravgxYyLMR5UgJNr3PMlbsuVRj
+          YYa73UjkaZPHYCk2JtrtyM0RzmVYcos1SYYq9EKL8Hq+b4mqXjCve+i9PwMwhKi7nuVN/7ZHXo+7
+          7WPdpbiS/7maigNLs0yEEkPeY9HYmYK+DgC+lRvSvzkLpkux7XSpcU0l3WTy4cMuoDmdrTP4an9i
+          jEZFfwa8nx383oRcOlJkL2eHyFi0DbnznJfTowjsHccTNh6caf+e0o/C7/RzqM+i/DT8CbCWOiW3
+          PG3gj8wS5dP+M7NjrQthI5Q2bGmpTCn3w1GFvd9dXSNbUWqXFYc6HzLend6qW04nV3R1ObMzZwav
+          g/8AAAD//wMAq+WXOoMGAAA=
+      headers:
+        CF-RAY:
+          - 96a9b5603c4a1566-SJC
+        Connection:
+          - keep-alive
+        Content-Encoding:
+          - gzip
+        Content-Type:
+          - application/json
+        Date:
+          - Tue, 05 Aug 2025 22:25:08 GMT
+        Server:
+          - cloudflare
+        Transfer-Encoding:
+          - chunked
+        X-Content-Type-Options:
+          - nosniff
+        access-control-expose-headers:
+          - X-Request-ID
+        alt-svc:
+          - h3=":443"; ma=86400
+        cf-cache-status:
+          - DYNAMIC
+        openai-organization:
+          - future-house-xr4tdh
+        openai-processing-ms:
+          - "1609"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
+        strict-transport-security:
+          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "2399"
+          - "1612"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4407,13 +4422,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998617"
+          - "29998553"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - 5189f953-a0bb-4b6b-9eb6-fbc1a9908793
+          - req_a9036b85a491258e23a03134e0370ca9
       status:
         code: 200
         message: OK
@@ -4421,11 +4436,14 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 14-16: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nsame
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 14-16: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nsame
         optimization problem.100 Grabocka\\n\\net al. 111 have developed a method named
         Adversarial Training on EXplanations (ATEX)\\n\\nwhich improves model robustness
         via exposure to adversarial examples.  While there are\\n\\nconceptual disparities,
@@ -4489,7 +4507,7 @@ interactions:
         solubility prediction is a classic cheminformatics regression challenge and
         is\\n\\nimportant for chemical process design, drug design and crystallization.133\u2013136
         In our previous\\n\\nworks,9,10 we implemented and trained an RNN model in Keras
-        to predict solubilities (log\\n\\nmolarity) of small molecules.127 The AqS\\n\\n----\\n\\nQuestion:
+        to predict solubilities (log\\n\\nmolarity) of small molecules.127 The AqS\\n\\n------------\\n\\nQuestion:
         What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
@@ -4499,13 +4517,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5773"
+          - "6049"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4515,7 +4533,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4531,24 +4549,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA3xUTW8bNxC961cMeGqAlSEptuXo5sIpYDTpra2BKhBmydndqbkkPSRVCYb/e8Fd
-          WZIbpxce+Obrvfl4ngAoNmoFSneYdB/s9Of918Yk19o//2g/P8zvHmn5a/z9t63bPT3NVFU8fP03
-          6fTqdaF9Hywl9m6EtRAmKlHny6uPy8X11eJqAHpvyBa3NqTppZ8uZovL6Xw+XRzi6s6zpqhW8NcE
-          AOB5eEuJztBOrWBWvf70FCO2pFZHIwAl3pYfhTFyTOiSqk6g9i6RG6p+XjuAtYq571H2a7WCtXq4
-          va/AC3zeBYvssLYEt5K4Yc1o4d4lspZbcpoqEGpIIiQPPaXOmwjoDCTSneOnTBFyJFNgdokkCKXB
-          IDtDUgozkDoCQ5ojexehR0NQ76FH3bEjsITi2LXw09cvHwZXQxTOvu++fIBBzXgB926INvDbJfAN
-          6I56jkn2FTzc3gNHwBAsjyVlp/2WBGKSrFMWmgbxgSTtQchiaWPsOMQKYtYdYCFTcmqfC5kGdcpo
-          gYpObjQ/VBi1cEhepjUW+q/SvJGhtqgfp7XfHcv/xQvQDssIVf+bRKODmNuWYgLU5W9oUu9NadLB
-          qLTEW9LZ0pi5D+K3BAeOTBEsPxLU1nszrQXZQY0iTAKBpCes2XLaX8Ddkc/bMko8NgRPGV3ihIm3
-          BOwit12KhemxAhRoCIvEsQJkM6g4tqYYno1DQQY9IAgZHsjFi6F5tbBpC5eOoMUANaV/iByMS/cq
-          49CBLvfoTlofmFTQ42NJwAm05GGaGy9gfI/sDnIYyS0YjsNo7C/WqhpXRMjSFp2mTdReaFyVT2u1
-          di/nuyXU5IhltV229gxA53wahStb/e2AvBz32Po2iK/jf1xVw45jtxHC6F3Z2Zh8UAP6MgH4NtyL
-          /OYEqCC+D2mT/CMN6ebz5XwMqE4n6gy+XB7Q5BPaM+Dj/KZ6J+TGUEK28ezoKI26I3Oe82pxJIHZ
-          sD9hs8kZ9+9Lei/8yJ9dexblh+FPgNYUEpnNaZreMxMqZ/xHZketh4JVJNmypk1iktIPQw1mO15Y
-          FfcxUb9p2LVl9ng8s03Y6Ovl8vqmub6ZqcnL5F8AAAD//wMAE/IEr28GAAA=
+          H4sIAAAAAAAAAwAAAP//dFTBbuM2EL37KwY8JYBs2F4n2fXNSFPAaJP20EOKemFQ5EiamiJZDmXE
+          DfLvBSmvrGyzFx30Zt68eTPD1wmAIC3WIFQjo2q9md4vfnlC/qweavX7093meNT/PiHbP38rn91W
+          FCnDlX+jit+yZsq13mAkZ3tYBZQRE+vi7ma1+rS8m99moHUaTUqrfZyu3HQ5X66mi8V0OT8nNo4U
+          sljDXxMAgNf8TRKtxhexhnnx7U+LzLJGsR6CAERwJv0Rkpk4ShtFcQGVsxFtVv26swA7wV3bynDa
+          iTXsxMOLN5KsLA3CJkSqSJE0sLURjaEarUK4et5sryFghYEhOmgxNk4zSKshomos/dMhQ8eoMywP
+          CLFB0KiIydlpKw9ka/DBKWRGBldBK1VDFsGgDDah2SUuwMsQSXVGBnMCjegvIVc//Xo9xJGNGHzA
+          mLUnLZ3VGJIBOv2awdZmGdmBl5iKqgZb4hhOBTxvtkAM0ntDve6BEEoj1WFaupdzscyunLWoYqKk
+          AJh8szJNP3vCMXQqdgGnPjiPIZ4goOnxhjzP4GeXsmTamWKw0NAB4dEZzA3DYyoHm9o6jqTg3nVJ
+          UyVV7KSBh3HNq8fHzf3DdZamkVUgH913umTAYSrYDxraodhZKCEXwJ1qQDKUxjk9LUOKLGUIhAE8
+          hhYzYy7GznQlGYon8AE1qYTM4I8GGd+X98EdSSPIHJLHRJapbuKoYut0WrqLkxeBg6dc5MqpnfdT
+          L0+gXZvE4kvq5Wzoecw824mi3/mABo/SKtyzcgH73f8ywMmkPbWyRk5QJQ3jzr6N7yhg1bFMZ2w7
+          Y0aAtNbFXn664K9n5G24WeNqH1zJ36WKiixxsw8o2dl0nxydFxl9mwB8zW9D9+7chQ+u9XEf3QFz
+          ucVy+aUnFJfnaATfLM9odFGaEfDp86L4gHKvMUoyPHpghJKqQT3KHVhT950md8Hmk1Hv/5f0EX3f
+          P9l6xPJD+gugFPqIen/ZwY/CAqYn+0dhg9dZsGAMR1K4j4QhzUNjJTvTv6aCTxyx3Vdk67SA1D+p
+          ld8vF7d4e7NSKy0mb5P/AAAA//8DAGMPC5hbBgAA
       headers:
         CF-RAY:
-          - 96665c9eeab75c18-SJC
+          - 96a9b55f6d4bd039-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4556,7 +4574,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:28 GMT
+          - Tue, 05 Aug 2025 22:25:08 GMT
         Server:
           - cloudflare
         Transfer-Encoding:
@@ -4572,7 +4590,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "2722"
+          - "1850"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -4580,7 +4598,7 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "2728"
+          - "1897"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4588,13 +4606,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998624"
+          - "29998556"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - req_6ea7b5adbedcbd910186af8f429215cc
+          - req_a0ec9056157f8e2b8cbd6c3d4bae4bcd
       status:
         code: 200
         message: OK
@@ -4602,77 +4620,81 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 16-20: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nssion
-        challenge and is\\n\\nimportant for chemical process design, drug design and
-        crystallization.133\u2013136 In our previous\\n\\nworks,9,10 we implemented
-        and trained an RNN model in Keras to predict solubilities (log\\n\\nmolarity)
-        of small molecules.127 The AqSolDB curated database137 was used to train the\\n\\nRNN
-        model.\\n\\n   In this task, counterfactuals are based on equation 6. Figure
-        3 illustrates the generated\\n\\nlocal chemical space and the top four counterfactuals.
-        Based on the counterfactuals, we ob-\\n\\nserve that the modifications to the
-        ester group and other heteroatoms play an important role\\n\\nin solubility.
-        These findings align with known experimental and basic chemical intuition.134\\n\\nFigure
-        4 shows a quantitative measurement of how substructures are contributing to
-        the pre-\\n\\n\\n\\n                                       16Figure 2: Descriptor
-        explanations along with natural language explanation obtained for BBB\\npermeability
-        of Alprozolam molecule. The green and red bars show descriptors that influ-\\nence
-        predictions positively and negatively, respectively. Dotted yellow lines show
-        significance\\nthreshold (\u03B1 = 0.05) for the t-statistic. Molecular descriptors
-        show molecule-level proper-\\nties that are important for the prediction. ECFP
-        and MACCS descriptors indicate which\\nsubstructures influence model predictions.
-        MACCS explanations lead to text explanations\\nas shown. Republished from Ref.10
-        with permission from authors. SMARTS annotations for\\nMACCS descriptors were
-        created using SMARTSviewer (smartsview.zbh.uni-hamburg.de,\\nCopyright: ZBH,
-        Center for Bioinformatics Hamburg) developed by Schomburg et al. 132.\\n\\n\\n\\n\\n\\n
-        \                                      17diction. For example, we see that adding
-        acidic and basic groups as well as hydrogen bond\\n\\nacceptors, increases solubility.
-        Substructure importance from ECFP97 and MACCS138 de-\\n\\nscriptors indicate
-        that adding heteroatoms increases solubility, while adding rings structures\\n\\nmakes
-        the molecule less soluble. Although these are established hypotheses, it is
-        interesting\\n\\nto see they can be derived purely from the data via DL and
-        XAI.\\n\\n\\n\\n\\n\\nFigure 3: Generated chemical space for solubility prediction
-        using the RNN model. The\\nchemical space is a 2D projection of the pairwise
-        Tanimoto similarities of the local coun-\\nterfactuals. Each data point is colored
-        by solubility. Top 4 counterfactuals are shown here.\\nRepublished from Ref.9
-        with permission from the Royal Society of Chemistry.\\n\\n\\n\\nGeneralizing
-        XAI \u2013 interpreting scent-structure relationships\\n\\n\\nIn this example,
-        we show how non-local structure-property relationships can be learned with\\n\\nXAI
-        across multiple molecules. Molecular scent prediction is a multi-label classification
-        task\\n\\nbecause a molecule can be described by more than one scent. For example,
-        the molecule\\n\\njasmone can be described as having \u2018jasmine,\u2019 \u2018woody,\u2019
-        \u2018floral,\u2019 and \u2019herbal\u2019 scents.139 The\\n\\nscent-structure
-        relationship is not very well understood,140 although some relationships are\\n\\nknown.
-        \ For example, molecules with an ester functional group are often associated
-        with\\n\\n\\n                                       18Figure 4: Descriptor explanations
-        for solubility prediction model. The green and red bars\\nshow descriptors that
-        influence predictions positively and negatively, respectively. Dotted\\nyellow
-        lines show significance threshold (\u03B1 = 0.05) for the t-statistic. The MACCS
-        and\\nECFP descriptors indicate which substructures influence model predictions.
-        MACCS sub-\\nstructures may either be present in the molecule as is or may represent
-        a modification. ECFP\\nfingerprints are substructures in the molecule that affect
-        the prediction. MACCS descriptor\\nare used to obtain text explanations as shown.
-        Republished from Ref.10 with permission from\\nauthors. SMARTS annotations for
-        MACCS descriptors were created using SMARTSviewer\\n(smartsview.zbh.uni-hamburg.de,
-        Copyright: ZBH, Center for Bioinformatics Hamburg) de-\\nveloped by Schomburg
-        et al. 132.\\n\\n\\n\\n\\n\\n                                       19the \u2018fruity\u2019
-        scent. There are some exceptions though, like tert-amyl acetate which has a\\n\\n\u2018camphoraceous\u2019
-        rather than \u2018fruity\u2019 scent.140,141\\n\\n   In Seshadri et al. 31,
-        we trained a GNN model to predict the scent of molecules and utilized\\n\\ncounterfactuals9
-        and descriptor explanations10 to quantify scent-structure relationships. The\\n\\nMMACE
-        method was modified to account for the multi-label aspect of scent prediction.
-        This\\n\\nmodification defines molecules that differed from the instance molecule
-        by only the selected\\n\\nscent as counterfactuals. For instance, counterfactuals
-        of the jasmone molecule would be false\\n\\nfor the \u2018jasmine\u2019 scent
-        but would still be positive for \u2018woody,\u2019 \u2018floral\u2019 and \u2018herbal\u2019
-        scents.\\n\\n\\n\\n\\n\\nFigure 5:  Counterfactual for the 2,4 decadienal molecule.
-        \ The counterfactual indicates\\nstructural changes to ethyl benzoate that would
-        result in the model predicting the molecule\\nto not contain the \u2018fruity\u2019
-        scent. The Tanimoto96 similarity between the counterfactual and\\n2,4 decadienal
-        is also\\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 22-25: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nut
+        to models informs the XAI method.\\n\\n\\nConclusion and outlook\\n\\n\\nWe
+        should seek to explain molecular property prediction models because users are
+        more\\n\\nlikely to trust explained predictions, and explanations can help assess
+        if the model is learning\\n\\nthe correct underlying chemical principles. We
+        also showed that black-box modeling first,\\n\\nfollowed by XAI, is a path to
+        structure-property relationships without needing to trade\\n\\nbetween accuracy
+        and interpretability. However, XAI in chemistry has some major open\\n\\nquestions,
+        that are also related to the black-box nature of the deep learning. Some are\\n\\n\\n\\n
+        \                                      22highlighted below:\\n\\n\\n   \u2022
+        Explanation representation: How is an explanation presented \u2013 text, a molecule,
+        attri-\\n\\n      butions, a concept, etc?\\n\\n\\n   \u2022 Molecular distance:
+        \ in XAI approaches such as counterfactual generation, the \u201Cdis-\\n\\n
+        \    tance\u201D between two molecules is minimized. Molecular distance is subjective.
+        Possibil-\\n\\n       ities are distance based on molecular properties, synthesis
+        routes, and direct structure\\n\\n     comparisons.\\n\\n\\n   \u2022 Regulations:
+        As black-box models move from research to industry, healthcare, and\\n\\n     environmental
+        settings, we expect XAI to become more important to explain decisions\\n\\n
+        \     to chemists or non-experts and possibly be legally required. Explanations
+        may need\\n\\n      to be tuned for be for doctors instead of chemists or to
+        satisfy a legal requirement.\\n\\n\\n   \u2022 Chemical space: Chemical space
+        is the set of molecules that are realizable; \u201Crealiz-\\n\\n     able\u201D
+        can be defined from purchasable to synthesizable to satisfied valences. What
+        is\\n\\n     most useful? Can an explanation consider nearby impossible molecules?
+        How can we\\n\\n     generate local chemical spaces centered around a specific
+        molecule for finding counter-\\n\\n      factuals or other instance explanations?
+        \ Similarly, can \u201Cactivity cliffs\u201D be connected\\n\\n      to explanations
+        and the local chemical space.149\\n\\n\\n   \u2022 Evaluating XAI : there is
+        a lack of a systematic framework (quantitative or qualitative)\\n\\n      to
+        evaluate correctness and applicability of an explanation. Can there be a universal\\n\\n
+        \    framework, or should explanations be chosen and evaluated based on the
+        audience and\\n\\n     domain? For example, work by Rasmussen et al. 58 attempts
+        to focus on comparing\\n\\n      feature attribution XAI methods via Crippen\u2019s
+        logP scores.\\n\\n\\n\\n\\n\\n                                       23Acknowledgements\\n\\n\\nResearch
+        reported in this work was supported by the National Institute of General Medical\\n\\nSciences
+        of the National Institutes of Health under award number R35GM137966. This work\\n\\nwas
+        supported by the NSF under awards 1751471 and 1764415. We thank the Center for\\n\\nIntegrated
+        Research Computing at the University of Rochester for providing computational\\n\\nresources.\\n\\n\\nReferences\\n\\n\\n
+        \ (1) Choudhary, K.; DeCost, B.; Chen, C.; Jain, A.; Tavazza, F.; Cohn, R.;
+        Park, C. W.;\\n\\n     Choudhary, A.; Agrawal, A.; Billinge, S. J.; Holm, E.;
+        Ong, S. P.; Wolverton, C.\\n\\n      Recent advances and applications of deep
+        learning methods in materials science. npj\\n\\n      Computational Materials
+        2022, 8.\\n\\n\\n  (2) Keith, J. A.; Vassilev-Galindo, V.; Cheng, B.; Chmiela,
+        S.; Gastegger, M.; M\xA8uller, K.-\\n\\n      R.; Tkatchenko, A. Combining Machine
+        Learning and Computational Chemistry for\\n\\n      Predictive Insights Into
+        Chemical Systems. Chemical Reviews 2021, 121, 9816\u20139872,\\n\\n     PMID:
+        34232033.\\n\\n\\n  (3) Goh, G. B.; Hodas, N. O.; Vishnu, A. Deep learning for
+        computational chemistry.\\n\\n      Journal of Computational Chemistry 2017,
+        38, 1291\u20131307.\\n\\n\\n  (4) Deringer, V. L.; Caro, M. A.; Cs\xB4anyi,
+        G. Machine Learning Interatomic Potentials as\\n\\n     Emerging Tools for Materials
+        Science. Advanced Materials 2019, 31, 1902765.\\n\\n\\n  (5) Faber, F. A.; Hutchison,
+        L.; Huang, B.; Gilmer, J.; Schoenholz, S. S.; Dahl, G. E.;\\n\\n      Vinyals,
+        O.; Kearnes, S.; Riley, P. F.; von Lilienfeld, O. A. Prediction Errors of Molec-\\n\\n
+        \     ular Machine Learning Models Lower than Hybrid DFT Error. Journal of Chemical\\n\\n
+        \    Theory and Computation 2017, 13, 5255\u20135264, PMID: 28926232.\\n\\n\\n\\n
+        \                                      24 (6) Duch, W.; Swaminathan, K.; Meller,
+        J. Artificial Intelligence Approaches for Rational\\n\\n    Drug Design and
+        Discovery. Current Pharmaceutical Design 2007, 13, 1497\u20131508.\\n\\n\\n
+        (7) Dara, S.; Dhamercherla, S.; Jadav, S. S.; Babu, C. M.; Ahsan, M. J.; darasuresh,
+        S. D.;\\n\\n     Dara, S. Machine Learning in Drug Discovery: A Review. Artificial
+        Intelligence Review\\n\\n     123, 55, 1947\u20131999.\\n\\n\\n (8) Gupta, R.;
+        Srivastava, D.; Sahu, M.; Tiwari, S.; Ambasta, R. K.; Kumar, P. Artifi-\\n\\n
+        \     cial intelligence to deep learning: machine intelligence approach for
+        drug discovery.\\n\\n     Molecular diversity 2021, 25, 1315\u20131360.\\n\\n\\n
+        (9) Wellawatte, G. P.; Seshadri, A.; White, A. D. Model agnostic generation
+        of counter-\\n\\n     factual explanations for molecules. Chemical Science 2022,
+        13, 3697\u20133705.\\n\\n\\n(10) Gandhi, H. A.; White, A. D. Explaining structure-ac\\n\\n------------\\n\\nQuestion:
+        What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4681,13 +4703,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5791"
+          - "6096"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4697,7 +4719,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4713,24 +4735,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFRNbxs5DL37VxA6tYBtJG7rJL6lQbKbQ4Mmu4cF1oWhSJwZ1hpJFal8
-          IMh/L6SJ48luF9jLACPyPT0+inyaACiyagXKdFpMH93s8+OXpv18vv1+dm1//3r948ZetCc3D/aE
-          bq4bNS2IcPsdjexQcxP66FAo+CFsEmrBwnp49OnD0WL5aXFcA32w6AqsjTL7GGaLg8XH2eHhbHHw
-          AuwCGWS1gr8nAABP9VskeosPagUH091Jj8y6RbV6TQJQKbhyojQzsWgvaroPmuAFfVX9tPYAa8W5
-          73V6XKsVrNVfp5fw7vwhOk1e3zqE0yTUkCHt4NILOkcteoPvIWGDiUEC9ChdsAzaWxA0nacfGRmk
-          0wK93iJIhxATWjLFnSHRoiGm4Ge93pJvIaZgkBkZQgOnl1BNYiAvmGJCqWIKMHuLqZRl65EE6HKv
-          Pc/h0tebaoUPUnjKLz4YTFGmUEojBh2jI7QFaDrsyWhXefvg0GSn01jqFDibDjQDB5dvyZE81mw2
-          6GXGkrKRnBASOl0RHUWew597Gxxti6ZcCmm0kawdT8Eim0RRQgIsbnv9ct9bKQ35ttRPXhje4byd
-          T+H87OJrTftyenb2x4iJ34NOCJmH6siiF2oeoQv3wBFN6eOIm/Ptq/5idONy6ey4/DlcVIG6vOzB
-          wQ5d5EEz+UqtrS0N1IYsGWhTyLHQledf2jkyLqQKQBZM0GRfL9Fuh6ke1o4aTFLoq8vFzg4ZgTxT
-          2wmDdtR6uCfpYOvDvd83sjhlKDocHlmxw2KiO7TQpNCD1aJnthz43Qur/bm5uhoQv11d8XytpsNo
-          JHR4p73BDZuQcBiRk7Va++fxTCVsMusy0j47Nwpo74MMnS3T/O0l8vw6vy60MYVb/gdUNeSJu02x
-          MPgyqywhqhp9ngB8q3sivxl9FVPoo2wkbLFed7g4OB4I1X41jcLLlzWiJIh2o8CH5Q73hnJjUTQ5
-          Hi0bZbTp0O6x+82ks6UwCkxGhf9bz6+4h+LJt/+Hfh8wBqOg3eyf8a/SEpbd/V9pr0ZXwYox3ZHB
-          jRCm0gyLjc5uWKuKH1mw34wmtaQ0cWOWR0fL42Z5fKAmz5OfAAAA//8DAFZjL/VkBgAA
+          H4sIAAAAAAAAAwAAAP//dFTBbiM3DL37KwhdtgUmhu06ya5vQVEUxgIFivYQtF4YjMSZ0UZDzYoc
+          J9kg/76QxrFn2+xlgNEjn94jRT7PAIx3ZgPGtqi268PFr8uPf1Brf//ry5+rlT3cPrmvW0L+evjn
+          /eVnU+WMePeZrL5mzW3s+kDqI4+wTYRKmXV5fble/7K6XlwVoIuOQk5rer1Yx4vVYrW+WC4vVotj
+          Yhu9JTEb+HcGAPBcvlkiO3o0G1hUrycdiWBDZnMKAjAphnxiUMSLIqupzqCNrMRF9fOOAXZGhq7D
+          9LQzG9iZ25ttBTHBb499QM94FwhukvraW48BtqwUgm+ILVWQqKYkoBE60jY6AWQHSrZl/2UggUHI
+          FRjvCbQl6BM5b3ONxlhH1kv5izXcbKGURmBgRylLd0WARmiHDlnmsOXCU1w8as7qYiA7BEzQp9hT
+          0qfJLRXc3mwBfVdU0mgK2vjwelMgTAy2pc5bDNAnz9b3gaQC4hbZem5A0yBa9BLLkMpRSyPFuyNH
+          PsTgGxZ48NqCWE9c6jZx47mZw0d6AttiCMQNCXguGj3bMDiCRH0iIVbM+rPBoprLr1TgqPblsrNv
+          V7qcG4IOe83gNAfqmGPqmhKxAg7O5/YJ/ETzZl6N5kUzd7Qak/xclfxYjJ5KIz1aKlw2DqyUarQ6
+          YJDq2MkDhdjnFHkSpQ7VW6gTdvQQ0/1Y/wOGAZW+kzeHv1sSAqrrmFQAE4FNQ3luKJNHoQlZfKmK
+          RugTWi3CsO+Dt0evnsGzG0STJ4Hg7wlawqCtzbxjCw8+Re5yicPYJkvznanGaUgU6JCruRcbE41T
+          sVyc8Pyk977DhiRjNQahHb9MRyxRPQjmCechhAmAzHHsaxnuT0fk5TTOITZ9infyn1STWy7tPhFK
+          5Dy6orE3BX2ZAXwqa2P4bhOYPsWu173GeyrXLder1UhozptqCl8fUY2KYQJcXn2o3qDcO1L0QSa7
+          x1i0Lbkp6eL9yUR+ePGMLWYT7/+X9Bb96N9zM2H5If0ZsJZ6Jbc/r4W3whLlbf6jsFOti2AjlA7e
+          0l49pdwPRzUOYVy0Znz/+9pzQylvlLJt636/Wl7R1eXarp2Zvcy+AQAA//8DAAR7lfx2BgAA
       headers:
         CF-RAY:
-          - 96665cb0de315c18-SJC
+          - 96a9b55f5f1c15e1-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4738,7 +4760,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:29 GMT
+          - Tue, 05 Aug 2025 22:25:08 GMT
         Server:
           - cloudflare
         Transfer-Encoding:
@@ -4754,7 +4776,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "1563"
+          - "2086"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -4762,7 +4784,7 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "1567"
+          - "2089"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4770,13 +4792,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998614"
+          - "29998549"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - req_3d629873f8c6017885e969be24845e29
+          - req_fe0d0a398caab7d1e387c0b266460a33
       status:
         code: 200
         message: OK
@@ -4784,75 +4806,267 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 12-14: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nnterfactual
-        approach, contrastive approach employ a dual\\n\\noptimization method, which
-        works by generating a similar and a dissimilar (counterfactuals)\\n\\nexample.
-        Contrastive explanations can interpret the model by identifying contribution
-        of\\n\\npresence and absence of subsets of features towards a certain prediction.36,99\\n\\n
-        \ A counterfactual x\u2032 of an instance x is one with a dissimilar prediction
-        \u02C6f(x) in classi-\\n\\nfication tasks. As shown in equation 5, counterfactual
-        generation can be thought of as a\\n\\nconstrained optimization problem which
-        minimizes the vector distance d(x, x\u2032) between the\\n\\nfeatures.9,100\\n\\n\\n
-        \                             minimize  d(x, x\u2032)\\n                                                                                           (5)\\n
-        \                              such that   \u02C6f(x) \u0338= \u02C6f(x\u2032)\\n\\n
-        \  For regression tasks, equation 6 adapted from equation 5 can be used. Here,
-        a counter-\\n\\nfactual is one with a defined increase or decrease in the prediction.\\n\\n\\n
-        \                          minimize  d(x, x\u2032)\\n                                                                                           (6)\\n
-        \                           such that    \u02C6f(x) \u2212\u02C6f(x\u2032) \u2265\u2206\\n\\n
-        \  Counterfactuals explanations have become a useful tool for XAI in chemistry,
-        as they\\n\\nprovide intuitive understanding of predictions and are able to
-        uncover spurious relationships\\n\\nin training data.101 Counterfactuals create
-        local (instance-level), actionable explanations.\\n\\nActionability of an explanation
-        suggest which features can be altered to change the outcome.\\n\\nFor example,
-        changing a hydrophobic functional group in a molecule to a hydrophilic group\\n\\nto
-        increase solubility.\\n\\n   Counterfactual generation is a demanding task as
-        it requires gradient optimization over\\n\\ndiscrete features that represents
-        a molecule. Recent work by Fu et al. 102 and Shen et al. 103\\n\\npresent two
-        techniques which allow continuous gradient-based optimization. Although, these\\n\\nmethodologies
-        are shown to circumvent the issue of discrete molecular optimization, counter-\\n\\nfactual
-        explanation based model interpretation still remains unexplored compared to
-        other\\n\\n\\n\\n                                       12post-hoc methods.\\n\\n
-        \  CF-GNNExplainer104 is a counterfactual explanation generating method based
-        on GN-\\n\\nNExplainer69 for graph data. This method generate counterfactuals
-        by perturbing the input\\n\\ndata (removing edges in the graph), and keeping
-        account of perturbations which lead to\\n\\nchanges in the output.  However,
-        this method is only applicable to graph-based models\\n\\nand can generate infeasible
-        molecular structures. Another related work by Numeroso and\\n\\nBacciu 105 focus
-        on generating counterfactual explanations for deep graph networks. Their\\n\\nmethod
-        MEG (Molecular counterfactual Explanation Generator) uses a reinforcement learn-\\n\\ning
-        based generator to create molecular counterfactuals (molecular graphs).  While
-        this\\n\\nmethod is able to generate counterfactuals through a multi-objective
-        reinforcement learner,\\n\\nthis is not a universal approach and requires training
-        the generator for each task.\\n\\n   Work by Wellawatte et al. 9 present a model
-        agnostic counterfactual generator MMACE\\n\\n(Molecular Model Agnostic Counterfactual
-        Explanations) which does not require training\\n\\nor computing gradients. This
-        method firstly populates a local chemical space through ran-\\n\\ndom string
-        mutations of SELFIES106 molecular representations using the STONED algo-\\n\\nrithm.107
-        Next, the labels (predictions) of the molecules in the local space are generated\\n\\nusing
-        the model that needs to be explained. Finally, the counterfactuals are identified
-        and\\n\\nsorted by their similarities \u2013 Tanimoto distance96 between ECFP4
-        fingerprints.97 Unlike the\\n\\nCF-GNNExplainer104 and MEG105 methods, the MMACE
-        algorithm ensures that generated\\n\\nmolecules are valid, owing to the surjective
-        property of SELFIES. Additionally, the MMACE\\n\\nmethod can be applied to both
-        regression and classification models. However, like most XAI\\n\\nmethods for
-        molecular prediction, MMACE does not account for the chemical stability of\\n\\npredicted
-        counterfactuals. To circumvent this drawback, Wellawatte et al. 9 propose an-\\n\\nother
-        approach, which identift counterfactuals through a similarity search on the
-        PubChem\\n\\ndatabase.108\\n\\n\\n\\n\\n\\n                                       13Similarity
-        to adjacent fields\\n\\n\\nTangential examples to counterfactual explanations
-        are adversarial training and matched\\n\\nmolecular pairs.  Adversarial perturbations
-        are used during training to deceive the model\\n\\nto expose the vulnerabilities
-        of a model109,110 whereas counterfactuals are applied post-hoc.\\n\\nTherefore,
-        the main difference between adversarial and counterfactual examples are in the\\n\\napplication,
-        although both are derived from the same optimization problem.100 Grabocka\\n\\net
-        al. 111 have developed a method named Adversarial Training on EXplanations (ATEX)\\n\\nwhich
-        improves model robustness via exposure to adversarial examples.  While there
-        are\\n\\nconceptual disparities, we note that\\n\\n----\\n\\nQuestion: What
-        is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 8-9: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nrepresented
+        with equation  2.\\n\\n                                          \u2206\u02C6f(\u20D7x)
+        \u2248\u2202\u02C6f(\u20D7x)                                    (2)\\n                               \u2206xi
+        \     \u2202xi\\n\\n\\n\\n                                       7                                                                     \u2206\u02C6f(\u20D7x)
+        \  where \u02C6f(x) is the black-box model and       are used as our attributions.
+        The left-                                                 \u2206xi\\n\\nhand
+        side of equation 2 says that we attribute each input feature xi by how much
+        one unit\\n\\nchange in it would affect the output of \u02C6f(x).  If \u02C6f(x)
+        is a linear surrogate model, then this\\n\\nmethod reconciles with LIME.35 In
+        DL models, \u2207xf(x), suffers from the shattered gradients\\n\\nproblem.62
+        This means directly computing the quantity leads to numeric problems. The\\n\\ndifferent
+        gradient based approaches are mostly distinguishable based on how the gradient
+        is\\n\\napproximated.\\n\\n   Gradient based explanations have been widely used
+        to interpret chemistry predictions.60,66\u201370\\n\\nMcCloskey et al. 60 used
+        graph convolutional networks (GCNs) to predict protein-ligand\\n\\nbinding and
+        explained the binding logic for these predictions using integrated gradients.\\n\\nPope
+        et al. 66 and Jim\xB4enez-Luna et al. 67 show application of gradCAM and integrated
+        gradi-\\n\\nents to explain molecular property predictions from trained graph
+        neural networks (GNNs).\\n\\nSanchez-Lengeling et al. 68 present comprehensive,
+        open-source XAI benchmarks to explain\\n\\nGNNs and other graph based models.
+        They compare the performance of class activation\\n\\nmaps (CAM),63 gradCAM,64
+        smoothGrad,,65 integrated gradients62 and attention mecha-\\n\\nnisms for explaining
+        outcomes of classification as well as regression tasks. They concluded\\n\\nthat
+        CAM and integrated gradients perform well for graph based models. Another attempt\\n\\nat
+        creating XAI benchmarks for graph models was made by Rao et al. 70. They compared\\n\\nthese
+        gradient based methods to find subgraph importance when predicting activity
+        cliffs\\n\\nand concluded that gradCAM and integrated gradients provided the
+        most interpretability\\n\\nfor GNNs.  The GNNExplainer69  is an approach for
+        generating explanations (local and\\n\\nglobal) for graph based models. This
+        method focuses on identifying which sub-graphs con-\\n\\ntribute most to the
+        prediction by maximizing mutual information between the prediction\\n\\nand
+        distribution of all possible sub-graphs. Ying et al. 69 show that GNNExplainer
+        can be\\n\\nused to obtain model-agnostic explanations. SubgraphX is a similar
+        method that explains\\n\\nGNN predictions by identifying important subgraphs.71\\n\\n
+        \  Another set of approaches like DeepLIFT72 and Layerwise Relevance backPropagation73\\n\\n\\n\\n
+        \                                      8(LRP) are based on backpropagation of
+        the prediction scores through each layer of the neu-\\n\\nral network. The specific
+        backpropagation logic across various activation functions differs\\n\\nin these
+        approaches, which means each layer must have its own implementation. Baldas-\\n\\nsarre
+        and Azizpour 74 showed application of LRP to explain aqueous solubility prediction
+        for\\n\\nmolecules.\\n\\n  SHAP is a model-agnostic feature attribution method
+        that is inspired from the game\\n\\ntheory concept of Shapley values.44,46 SHAP
+        has been popularly used in explaining molecular\\n\\nprediction models.75\u201378
+        It\u2019s an additive feature contribution approach, which assumes that\\n\\nan
+        explanation model is a linear combination of binary variables z.  If the Shapley
+        value\\nfor the ith feature is \u03D5i, then the explanation is \u02C6f(\u20D7x)
+        = Pi \u03D5i(\u20D7x)zi(\u20D7x). Shapley values for\\n\\nfeatures are computed
+        using Equation 3.79,80\\n\\n\\n\\n                           M\\n                                  1\\n
+        \                                   \u03D5i(\u20D7x) = X \u02C6f (\u20D7z+i)
+        \u2212\u02C6f (\u20D7z\u2212i)                            (3)\\n                 M\\n\\n
+        \  Here \u20D7z is a fabricated example created from the original \u20D7x and
+        a random perturbation \u20D7x\u2032.\\n\\n\u20D7z+i has the feature i from \u20D7x
+        and \u20D7z\u2212i has the ith feature from \u20D7x\u2032. Some care should
+        be taken\\n\\nin constructing \u20D7z when working with molecular descriptors
+        to ensure that an impossible \u20D7z is\\n\\nnot sampled (e.g., high count of
+        acid groups but no hydrogen bond donors). M is the sample\\n\\nsize of perturbations
+        around \u20D7x. Shapley value computation is expensive, hence M is chosen\\n\\naccordingly.
+        Equation 3 is an approximation and gives contributions with an expectation\\nterm
+        as \u03D50 + Pi=1 \u03D5i(\u20D7x) = \u02C6f(\u20D7x).\\n\\n   Visualization
+        based feature attribution has also been used for molecular data. In com-\\n\\nputer
+        science, saliency maps are a way to measure spatial feature contribution.81
+        Simply put,\\n\\nsaliency maps draw a connection between the model\u2019s neural
+        fingerprint components (trained\\n\\nweights) and input features. Weber et al.
+        82 used saliency maps to build an explainable GCN\\n\\narchitecture that gives
+        subgraph importance for small molecule activity prediction. On the\\n\\nother
+        hand, similarity maps compare model predictions for two or more molecules based
+        on\\n\\ntheir chemical fingerprints.83 Similarity maps provide atomic weights
+        or predicte\\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+      headers:
+        accept:
+          - application/json
+        accept-encoding:
+          - gzip, deflate
+        connection:
+          - keep-alive
+        content-length:
+          - "6108"
+        content-type:
+          - application/json
+        host:
+          - api.openai.com
+        user-agent:
+          - AsyncOpenAI/Python 1.99.0
+        x-stainless-arch:
+          - arm64
+        x-stainless-async:
+          - async:asyncio
+        x-stainless-lang:
+          - python
+        x-stainless-os:
+          - MacOS
+        x-stainless-package-version:
+          - 1.99.0
+        x-stainless-raw-response:
+          - "true"
+        x-stainless-read-timeout:
+          - "60.0"
+        x-stainless-retry-count:
+          - "0"
+        x-stainless-runtime:
+          - CPython
+        x-stainless-runtime-version:
+          - 3.13.5
+      method: POST
+      uri: https://api.openai.com/v1/chat/completions
+    response:
+      body:
+        string: !!binary |
+          H4sIAAAAAAAAAwAAAP//jFTbbhs3EH3XVwz4ZAMrQVIUufGbmqSt0cZJkV4CVIEw4s7uTsUlWQ6p
+          WDX87wW5tqS2LtCXxYJnbufM5X4EoLhW16B0h1H33oxfz76/pW7/1fv9+7fTrp7/+ubVN1/36cd3
+          X7pfflZV9nDb30nHJ6+Jdr03FNnZAdaBMFKOOrt6uVi8mF9NlwXoXU0mu7U+jhduPJ/OF+PZbDyf
+          Pjp2jjWJuobfRgAA9+WbS7Q13alrmFZPLz2JYEvq+mgEoIIz+UWhCEtEG1V1ArWzkWyp+n5tAdZK
+          Ut9jOKzVNazV2ztvkC1uDcEqRG5YMxq4sZGM4ZasJrj4tLq5hEANBYHooKfYuVoAbQ2RdGf5j0QC
+          SaguMO4IYkdQk2ZhZ8c97ti24IPTJEICroHVDRRdpAKPIbJOBoM5wKDqHThLAoZ3OQx5MITB5iAX
+          b364LJnbgL4DSymgAUvxiws7gYtvb2/lsgK2kYIPFAuzbJ9sTSHLU5en6KBLPVqZwKfVDaD3waHu
+          SICtNqmmnKBmsnG8xczsifUFTdrJkKANueFHQ6nK7+vVu8tqIDfG1jqJrI/ehdHH71Yf4GII6yx8
+          7NAbOsAeTaJcfC53z5LQ8J+YB+xcZkm6AxQQNExWH4q1cM8GA8cD9OhlAj91JHTqFPeZMMYYeJsi
+          DdWBD1SzzglKX9n6FKEhjCmQVMA12cjNAbj3LuTBAknbontWCbLUFbgANAwR9M5Q6WNutacQD+cp
+          BqFZQIdUhqzJrlZSyH2NAa14DJlSBTEkiYMQSXDLJjNjC0OnDOsii1RA4ikHMwVumMyTyLqjniWG
+          QaBTaYU623ayVtWwEIEM7dFq2oh2gYbFmE2PeB7sDffYkmSsQSO0tg/nWxaoSYJ5yW0y5gxAa10c
+          is37/fkReThutHGtD24r/3BVDVuWbhMIxdm8vRKdVwV9GAF8Lpcj/e0YKB9c7+Mmuh2VdLMX0/kQ
+          UJ2O1Rm8vHpEo4tozoDF8lX1TMhNTRHZyNn5UTovTX3yPd0qTDW7M2B0Rvzf9TwXeyDPtv0/4U+A
+          1uQj1ZvT7D1nFihf8/8yOwpdClZCYc+aNpEp5GbU1GAyw6FVcpBI/aZh2+abw8O1bfxmPlvS8uVC
+          L2o1ehj9BQAA//8DANG9V1B2BgAA
+      headers:
+        CF-RAY:
+          - 96a9b56118ea67bf-SJC
+        Connection:
+          - keep-alive
+        Content-Encoding:
+          - gzip
+        Content-Type:
+          - application/json
+        Date:
+          - Tue, 05 Aug 2025 22:25:08 GMT
+        Server:
+          - cloudflare
+        Transfer-Encoding:
+          - chunked
+        X-Content-Type-Options:
+          - nosniff
+        access-control-expose-headers:
+          - X-Request-ID
+        alt-svc:
+          - h3=":443"; ma=86400
+        cf-cache-status:
+          - DYNAMIC
+        openai-organization:
+          - future-house-xr4tdh
+        openai-processing-ms:
+          - "1993"
+        openai-project:
+          - proj_RpeV6PrPclPHBb5GlExPXSBj
+        openai-version:
+          - "2020-10-01"
+        strict-transport-security:
+          - max-age=31536000; includeSubDomains; preload
+        x-envoy-upstream-service-time:
+          - "1997"
+        x-ratelimit-limit-requests:
+          - "10000"
+        x-ratelimit-limit-tokens:
+          - "30000000"
+        x-ratelimit-remaining-requests:
+          - "9999"
+        x-ratelimit-remaining-tokens:
+          - "29998537"
+        x-ratelimit-reset-requests:
+          - 6ms
+        x-ratelimit-reset-tokens:
+          - 2ms
+        x-request-id:
+          - req_45bbafbd02a343540aa5f4508e50084c
+      status:
+        code: 200
+        message: OK
+  - request:
+      body:
+        "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
+        the relevant information that could help answer the question based on the excerpt.
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 16-20: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\nssion
+        challenge and is\\n\\nimportant for chemical process design, drug design and
+        crystallization.133\u2013136 In our previous\\n\\nworks,9,10 we implemented
+        and trained an RNN model in Keras to predict solubilities (log\\n\\nmolarity)
+        of small molecules.127 The AqSolDB curated database137 was used to train the\\n\\nRNN
+        model.\\n\\n   In this task, counterfactuals are based on equation 6. Figure
+        3 illustrates the generated\\n\\nlocal chemical space and the top four counterfactuals.
+        Based on the counterfactuals, we ob-\\n\\nserve that the modifications to the
+        ester group and other heteroatoms play an important role\\n\\nin solubility.
+        These findings align with known experimental and basic chemical intuition.134\\n\\nFigure
+        4 shows a quantitative measurement of how substructures are contributing to
+        the pre-\\n\\n\\n\\n                                       16Figure 2: Descriptor
+        explanations along with natural language explanation obtained for BBB\\npermeability
+        of Alprozolam molecule. The green and red bars show descriptors that influ-\\nence
+        predictions positively and negatively, respectively. Dotted yellow lines show
+        significance\\nthreshold (\u03B1 = 0.05) for the t-statistic. Molecular descriptors
+        show molecule-level proper-\\nties that are important for the prediction. ECFP
+        and MACCS descriptors indicate which\\nsubstructures influence model predictions.
+        MACCS explanations lead to text explanations\\nas shown. Republished from Ref.10
+        with permission from authors. SMARTS annotations for\\nMACCS descriptors were
+        created using SMARTSviewer (smartsview.zbh.uni-hamburg.de,\\nCopyright: ZBH,
+        Center for Bioinformatics Hamburg) developed by Schomburg et al. 132.\\n\\n\\n\\n\\n\\n
+        \                                      17diction. For example, we see that adding
+        acidic and basic groups as well as hydrogen bond\\n\\nacceptors, increases solubility.
+        Substructure importance from ECFP97 and MACCS138 de-\\n\\nscriptors indicate
+        that adding heteroatoms increases solubility, while adding rings structures\\n\\nmakes
+        the molecule less soluble. Although these are established hypotheses, it is
+        interesting\\n\\nto see they can be derived purely from the data via DL and
+        XAI.\\n\\n\\n\\n\\n\\nFigure 3: Generated chemical space for solubility prediction
+        using the RNN model. The\\nchemical space is a 2D projection of the pairwise
+        Tanimoto similarities of the local coun-\\nterfactuals. Each data point is colored
+        by solubility. Top 4 counterfactuals are shown here.\\nRepublished from Ref.9
+        with permission from the Royal Society of Chemistry.\\n\\n\\n\\nGeneralizing
+        XAI \u2013 interpreting scent-structure relationships\\n\\n\\nIn this example,
+        we show how non-local structure-property relationships can be learned with\\n\\nXAI
+        across multiple molecules. Molecular scent prediction is a multi-label classification
+        task\\n\\nbecause a molecule can be described by more than one scent. For example,
+        the molecule\\n\\njasmone can be described as having \u2018jasmine,\u2019 \u2018woody,\u2019
+        \u2018floral,\u2019 and \u2019herbal\u2019 scents.139 The\\n\\nscent-structure
+        relationship is not very well understood,140 although some relationships are\\n\\nknown.
+        \ For example, molecules with an ester functional group are often associated
+        with\\n\\n\\n                                       18Figure 4: Descriptor explanations
+        for solubility prediction model. The green and red bars\\nshow descriptors that
+        influence predictions positively and negatively, respectively. Dotted\\nyellow
+        lines show significance threshold (\u03B1 = 0.05) for the t-statistic. The MACCS
+        and\\nECFP descriptors indicate which substructures influence model predictions.
+        MACCS sub-\\nstructures may either be present in the molecule as is or may represent
+        a modification. ECFP\\nfingerprints are substructures in the molecule that affect
+        the prediction. MACCS descriptor\\nare used to obtain text explanations as shown.
+        Republished from Ref.10 with permission from\\nauthors. SMARTS annotations for
+        MACCS descriptors were created using SMARTSviewer\\n(smartsview.zbh.uni-hamburg.de,
+        Copyright: ZBH, Center for Bioinformatics Hamburg) de-\\nveloped by Schomburg
+        et al. 132.\\n\\n\\n\\n\\n\\n                                       19the \u2018fruity\u2019
+        scent. There are some exceptions though, like tert-amyl acetate which has a\\n\\n\u2018camphoraceous\u2019
+        rather than \u2018fruity\u2019 scent.140,141\\n\\n   In Seshadri et al. 31,
+        we trained a GNN model to predict the scent of molecules and utilized\\n\\ncounterfactuals9
+        and descriptor explanations10 to quantify scent-structure relationships. The\\n\\nMMACE
+        method was modified to account for the multi-label aspect of scent prediction.
+        This\\n\\nmodification defines molecules that differed from the instance molecule
+        by only the selected\\n\\nscent as counterfactuals. For instance, counterfactuals
+        of the jasmone molecule would be false\\n\\nfor the \u2018jasmine\u2019 scent
+        but would still be positive for \u2018woody,\u2019 \u2018floral\u2019 and \u2018herbal\u2019
+        scents.\\n\\n\\n\\n\\n\\nFigure 5:  Counterfactual for the 2,4 decadienal molecule.
+        \ The counterfactual indicates\\nstructural changes to ethyl benzoate that would
+        result in the model predicting the molecule\\nto not contain the \u2018fruity\u2019
+        scent. The Tanimoto96 similarity between the counterfactual and\\n2,4 decadienal
+        is also\\n\\n------------\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4861,13 +5075,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5777"
+          - "6067"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4877,7 +5091,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4893,24 +5107,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA3RUTW8bNxC961cMeMllJUiKv+qbY9iFD056KIoCVSCMyNndibnkljOUrRr+7wFX
-          n06Tyx72zTy+N1+vIwDDzlyDsS2q7Xo//rR5rOnlr/9eAt88XH7B7o71yx8pfmrsvZiqZMTVN7K6
-          z5rY2PWelGPYwjYRKhXW2eX5x8v5xfn8YgC66MiXtKbX8Vkcz6fzs/FsNp5Pd4ltZEtiruGfEQDA
-          6/AtEoOjF3MN02r/pyMRbMhcH4IATIq+/DEowqIY1FRH0MagFAbVr4sAsDCSuw7TZmGuYWH+vnmo
-          ICa4e+k9csCVJ7hJyjVbRg8PQcl7bihYqiBRTUlAI3SkbXQCGBwo2Tbwv5kEtEWFDp8ItCXoEzm2
-          pUACsYabBxgqIcBBKfWJdHiucOTgKBXtbvilEdrcYZAJ3MZcomu0mtEDFZ0Bt6SYCBCeaAMaowcO
-          MNjpU1yz49AAh8JpaexpTb4CHNQMT3AQbloVWG2AHQXlelNSbIuhoaIROPRZoSbUnPbmnmP2DtAr
-          pcHj4OiDnHidwH1MQC9YpqMqPLaljkXTpgL7zo2AxQCSm4ZEC1Up+85bKXL0ZLPHBKIp252KCGhb
-          pjWBI+FErvjtKSmTTODPYzM8PxE8Pt7c3g0lvr0f//75867PlIbiZSFXGBsKlFDpR30VPLO2O5IV
-          lfoMhsfYhCjKdmDGvvds941bRW0hUZNIhGMYIqwvs7k3B4ryJJPSLEAvEcqMJhSV7XPo1pQEUxlA
-          TciBQ1PBc8u2hTraLCQQQxmFKAdJsM6+mFix51IKcDkVcE8wWZhqO/+JPK3LVCzFxkTbPfhtYRbh
-          7XRxEtVZsOxtyN6fABhC1G2Xysp+3SFvhyX1selTXMkPqabmwNIuE6HEUBZSNPZmQN9GAF+HY5Df
-          7bfpU+x6XWp8ouG52ezj7hqY4/05gc9nO1Sjoj8B5ld75B3l0pEiezm5KMaibcmd5E7nZwcTmB3H
-          IzYdnXj/v6Sf0W/9c2hOWH5JfwSspV7JLY+b9rOwROVG/yrsUOtBsBFKa7a0VKZU+uGoxuy359PI
-          RpS6Zc2hKceKtze07pf24vLy4qq+uJqa0dvoOwAAAP//AwDGgzdTTAYAAA==
+          H4sIAAAAAAAAA4xUUW/bRgx+968g7mkDZCN2nbT1m5G1mTFsWLs8FKgLg76jJM6nO+1IOQmC/Pfh
+          pNRWtgzYiwDdR3738eORjxMAw86swNga1Tatn17Pf/mt+nS8edN8aj4+3BzKn/DnzwFv3jd+cWuK
+          nBH3f5LV71kzG5vWk3IMA2wToVJmnb+9XC7fLN5evOuBJjryOa1qdbqM08XFYjmdz6eLi+fEOrIl
+          MSv4OgEAeOy/WWJwdG9WcFF8P2lIBCsyq1MQgEnR5xODIiyKQU1xBm0MSqFX/bgNAFsjXdNgetia
+          FWzNl/WmgJjgw33rkQPuPcE6KZdsGT1sgpL3XFGwVECikpKARmhI6+gEMDhQsnXgvzoS0BoVGjwQ
+          aE3QJnJss0FDoCPLwjFMGzxwqKBN0ZIICcQS1hvofRLgoJTaRNqLyYldcJRyZa4/0gh112CQGWxC
+          f1Nf5L1mnvxL95ZSqwV8WW+ABbBtPZPLibamhi36nreJnmznMY2lFiCdrQEFJPpuz571oY8WS0Gn
+          oqmz2iWCRB77jJpbmcHt2QbPh6ypy4WUaLVDLwU4Epu41ZiAstsBn+97KaXkUOX6OajADzSrZgV8
+          uP74ex/26/r6+o8Rk/wImAg6GapjR0G5HPTS0NLeEQ6l73ITs0Xnu6Tbn+oRiGHswwxuaxJ6oRXQ
+          cxXgjrWGQ4h34exnFmy59dRX2sQgmlBzm+t41zfCYuaPR3ZZj3BVa9/sCMMk3b90FBwlPpKDMsUG
+          HCrOtqYYnnAiT0cMlnZiY6LhKb8/wdmOHTdYkWSoRC+0DU/jsUhUdoJ5KkPn/QjAEKIOKvJAfntG
+          nk4j6GPVpriXf6SakgNLvUuEEkMeN9HYmh59mgB860e9ezG9pk2xaXWn8UD9dfPF1dVAaM7bZQQv
+          L59RjYp+DMznxSuUO0eK7GW0L4xFW5M7556XC3aO4wiYjAr/t57XuIfiOVT/h/4MWEutktudX99r
+          YYny+v2vsJPRvWAjlI5saadMKTfDUYmdHzajkQdRanajScshZbtbzK/o6nJpl85MniZ/AwAA//8D
+          ABNH4DknBgAA
       headers:
         CF-RAY:
-          - 96665ca7dfd867e2-SJC
+          - 96a9b56bfe61d039-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4918,11 +5132,9 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:30 GMT
+          - Tue, 05 Aug 2025 22:25:10 GMT
         Server:
           - cloudflare
-        Strict-Transport-Security:
-          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -4936,13 +5148,15 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "3958"
+          - "1562"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
+        strict-transport-security:
+          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "3963"
+          - "1565"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4950,13 +5164,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998621"
+          - "29998546"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - b813a259-3ca0-4aef-9e3d-335414aa8a2b
+          - req_f291c30000d38657e4046d5d4855044b
       status:
         code: 200
         message: OK
@@ -4964,11 +5178,14 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 28-30: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\n
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from Wellawatte2023 pages 28-30: Wellawatte et al, XAI Review, 2023\\n\\n------------\\n\\n
         M. T.; Singh, S.; Guestrin, C. \u201D Why should i trust you?\u201D Explaining
         the\\n\\n     predictions of any classifier. Proceedings of the 22nd ACM SIGKDD
         international\\n\\n\\n                                      27     conference
@@ -5035,7 +5252,7 @@ interactions:
         2018, 16, 31\u201357.\\n\\n\\n(56) Harren, T.; Matter, H.; Hessler, G.; Rarey,
         M.; Grebner, C. Interpretation of structure\u2013\\n\\n      activity relationships
         in real-world drug design data sets using explainable artificial\\n\\n      intelligence.
-        Journal of Chemical Information and Modeling 2022, 62,\\n\\n----\\n\\nQuestion:
+        Journal of Chemical Information and Modeling 2022, 62,\\n\\n------------\\n\\nQuestion:
         What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
@@ -5045,196 +5262,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5809"
-        content-type:
-          - application/json
-        host:
-          - api.openai.com
-        user-agent:
-          - AsyncOpenAI/Python 1.97.1
-        x-stainless-arch:
-          - arm64
-        x-stainless-async:
-          - async:asyncio
-        x-stainless-lang:
-          - python
-        x-stainless-os:
-          - MacOS
-        x-stainless-package-version:
-          - 1.97.1
-        x-stainless-raw-response:
-          - "true"
-        x-stainless-read-timeout:
-          - "60.0"
-        x-stainless-retry-count:
-          - "0"
-        x-stainless-runtime:
-          - CPython
-        x-stainless-runtime-version:
-          - 3.13.5
-      method: POST
-      uri: https://api.openai.com/v1/chat/completions
-    response:
-      body:
-        string: !!binary |
-          H4sIAAAAAAAAAwAAAP//dFRNj9tGDL37VxBzthde74cR37ZAA2yCXooWCFoHNj2iJK5HHHVIOess
-          9r8XI/kzbS4SMG/4+Pg45NsIwHHhFuB8jeabNkx+2f9WlvH7Xxt+3H68f3oJH2YaPlH75yf/+4sb
-          54i4eSFvx6gbH5s2kHGUAfaJ0Ciz3s4f7uazx4fZvAeaWFDIYVVrk/s4mU1n95Pb28lsegisI3tS
-          t4C/RwAAb/03S5SCXt0CpuPjSUOqWJFbnC4BuBRDPnGoymoo5sZn0Ecxkl71er1+0ShLeVsKwNJp
-          1zSY9ku3gKX78vQ8hpjg19c2IAtuAsFTMi7ZMwZ4FqMQuCLxNIZEJSUFi9CQ1bFQQCnAyNfC/3Sk
-          0CkVPYxbAqsJCvKsHGXS4JalgjZFT6qkEEt4eobeIwVLKNpiIrGeksUotYms12MR6q5B0Rv4oyag
-          V0+ptUFNFqaww8SxU/gW01bBajSg1zbEdFSxoxDbhsTGQDsMHeb+jftc2LaBfX+QRX3Jqo7V+RRV
-          oYgNsiho52tAhYYK9hiAG6xYqjE0aJQYg4J6HqzKzEXqKihIuZIb+Ex7yE1JvOlyLgUWH7qChlNU
-          4x31qlF6MTqGktC6RFdxgIJhr6xDjt5ASFR1ARN/H8ooY7q0kAPbfvDOs1FxsAmDRihYfacKvsYQ
-          SCrKuk4eSdX752NK5E1Ih477wNI7cPBuyHB0D0MVE1vdZIkK3yiE/M9ENavF1IfSLoZjQQWoodEk
-          lhOraYLJsojrR9Cgr1kIAmESlupm6cbDe04UaIfiaaU+Jsrv+sNS3peyXq8vRyJR2SnmiZQuhAsA
-          RaINnudh/HpA3k/jF2LVprjRH0JdycJarxKhRsmjphZb16PvI4Cv/Zh3V5Pr2hSb1lYWt9Snu314
-          vBsI3XmzXMB38wNq0TBcAPPpYT9cU64KMuSgF7vCefQ1FZc57x5PRWBXcDxj09FF7f+V9H/0Q/0s
-          1QXLT+nPgPfUGhWrNuV5ui77fC1R3r4/u3byuhfslNKOPa2MKeV+FFRiF4bF6HSvRs2qZKnyu+J+
-          O+Z+jt5H/wIAAP//AwBBLGdLGwYAAA==
-      headers:
-        CF-RAY:
-          - 96665cab39b74705-SJC
-        Connection:
-          - keep-alive
-        Content-Encoding:
-          - gzip
-        Content-Type:
-          - application/json
-        Date:
-          - Mon, 28 Jul 2025 18:15:30 GMT
-        Server:
-          - cloudflare
-        Transfer-Encoding:
-          - chunked
-        X-Content-Type-Options:
-          - nosniff
-        access-control-expose-headers:
-          - X-Request-ID
-        alt-svc:
-          - h3=":443"; ma=86400
-        cf-cache-status:
-          - DYNAMIC
-        openai-organization:
-          - future-house-xr4tdh
-        openai-processing-ms:
-          - "3521"
-        openai-project:
-          - proj_RpeV6PrPclPHBb5GlExPXSBj
-        openai-version:
-          - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
-        x-envoy-upstream-service-time:
-          - "3532"
-        x-ratelimit-limit-requests:
-          - "10000"
-        x-ratelimit-limit-tokens:
-          - "30000000"
-        x-ratelimit-remaining-requests:
-          - "9999"
-        x-ratelimit-remaining-tokens:
-          - "29998617"
-        x-ratelimit-reset-requests:
-          - 6ms
-        x-ratelimit-reset-tokens:
-          - 2ms
-        x-request-id:
-          - req_bff0c1c063b5175d3ceca3d7df1a8d71
-      status:
-        code: 200
-        message: OK
-  - request:
-      body:
-        "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
-        the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from Wellawatte2023 pages 8-9: Wellawatte et al, XAI Review, 2023\\n\\n----\\n\\nrepresented
-        with equation  2.\\n\\n                                          \u2206\u02C6f(\u20D7x)
-        \u2248\u2202\u02C6f(\u20D7x)                                    (2)\\n                               \u2206xi
-        \     \u2202xi\\n\\n\\n\\n                                       7                                                                     \u2206\u02C6f(\u20D7x)
-        \  where \u02C6f(x) is the black-box model and       are used as our attributions.
-        The left-                                                 \u2206xi\\n\\nhand
-        side of equation 2 says that we attribute each input feature xi by how much
-        one unit\\n\\nchange in it would affect the output of \u02C6f(x).  If \u02C6f(x)
-        is a linear surrogate model, then this\\n\\nmethod reconciles with LIME.35 In
-        DL models, \u2207xf(x), suffers from the shattered gradients\\n\\nproblem.62
-        This means directly computing the quantity leads to numeric problems. The\\n\\ndifferent
-        gradient based approaches are mostly distinguishable based on how the gradient
-        is\\n\\napproximated.\\n\\n   Gradient based explanations have been widely used
-        to interpret chemistry predictions.60,66\u201370\\n\\nMcCloskey et al. 60 used
-        graph convolutional networks (GCNs) to predict protein-ligand\\n\\nbinding and
-        explained the binding logic for these predictions using integrated gradients.\\n\\nPope
-        et al. 66 and Jim\xB4enez-Luna et al. 67 show application of gradCAM and integrated
-        gradi-\\n\\nents to explain molecular property predictions from trained graph
-        neural networks (GNNs).\\n\\nSanchez-Lengeling et al. 68 present comprehensive,
-        open-source XAI benchmarks to explain\\n\\nGNNs and other graph based models.
-        They compare the performance of class activation\\n\\nmaps (CAM),63 gradCAM,64
-        smoothGrad,,65 integrated gradients62 and attention mecha-\\n\\nnisms for explaining
-        outcomes of classification as well as regression tasks. They concluded\\n\\nthat
-        CAM and integrated gradients perform well for graph based models. Another attempt\\n\\nat
-        creating XAI benchmarks for graph models was made by Rao et al. 70. They compared\\n\\nthese
-        gradient based methods to find subgraph importance when predicting activity
-        cliffs\\n\\nand concluded that gradCAM and integrated gradients provided the
-        most interpretability\\n\\nfor GNNs.  The GNNExplainer69  is an approach for
-        generating explanations (local and\\n\\nglobal) for graph based models. This
-        method focuses on identifying which sub-graphs con-\\n\\ntribute most to the
-        prediction by maximizing mutual information between the prediction\\n\\nand
-        distribution of all possible sub-graphs. Ying et al. 69 show that GNNExplainer
-        can be\\n\\nused to obtain model-agnostic explanations. SubgraphX is a similar
-        method that explains\\n\\nGNN predictions by identifying important subgraphs.71\\n\\n
-        \  Another set of approaches like DeepLIFT72 and Layerwise Relevance backPropagation73\\n\\n\\n\\n
-        \                                      8(LRP) are based on backpropagation of
-        the prediction scores through each layer of the neu-\\n\\nral network. The specific
-        backpropagation logic across various activation functions differs\\n\\nin these
-        approaches, which means each layer must have its own implementation. Baldas-\\n\\nsarre
-        and Azizpour 74 showed application of LRP to explain aqueous solubility prediction
-        for\\n\\nmolecules.\\n\\n  SHAP is a model-agnostic feature attribution method
-        that is inspired from the game\\n\\ntheory concept of Shapley values.44,46 SHAP
-        has been popularly used in explaining molecular\\n\\nprediction models.75\u201378
-        It\u2019s an additive feature contribution approach, which assumes that\\n\\nan
-        explanation model is a linear combination of binary variables z.  If the Shapley
-        value\\nfor the ith feature is \u03D5i, then the explanation is \u02C6f(\u20D7x)
-        = Pi \u03D5i(\u20D7x)zi(\u20D7x). Shapley values for\\n\\nfeatures are computed
-        using Equation 3.79,80\\n\\n\\n\\n                           M\\n                                  1\\n
-        \                                   \u03D5i(\u20D7x) = X \u02C6f (\u20D7z+i)
-        \u2212\u02C6f (\u20D7z\u2212i)                            (3)\\n                 M\\n\\n
-        \  Here \u20D7z is a fabricated example created from the original \u20D7x and
-        a random perturbation \u20D7x\u2032.\\n\\n\u20D7z+i has the feature i from \u20D7x
-        and \u20D7z\u2212i has the ith feature from \u20D7x\u2032. Some care should
-        be taken\\n\\nin constructing \u20D7z when working with molecular descriptors
-        to ensure that an impossible \u20D7z is\\n\\nnot sampled (e.g., high count of
-        acid groups but no hydrogen bond donors). M is the sample\\n\\nsize of perturbations
-        around \u20D7x. Shapley value computation is expensive, hence M is chosen\\n\\naccordingly.
-        Equation 3 is an approximation and gives contributions with an expectation\\nterm
-        as \u03D50 + Pi=1 \u03D5i(\u20D7x) = \u02C6f(\u20D7x).\\n\\n   Visualization
-        based feature attribution has also been used for molecular data. In com-\\n\\nputer
-        science, saliency maps are a way to measure spatial feature contribution.81
-        Simply put,\\n\\nsaliency maps draw a connection between the model\u2019s neural
-        fingerprint components (trained\\n\\nweights) and input features. Weber et al.
-        82 used saliency maps to build an explainable GCN\\n\\narchitecture that gives
-        subgraph importance for small molecule activity prediction. On the\\n\\nother
-        hand, similarity maps compare model predictions for two or more molecules based
-        on\\n\\ntheir chemical fingerprints.83 Similarity maps provide atomic weights
-        or predicte\\n\\n----\\n\\nQuestion: What is XAI?\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
-      headers:
-        accept:
-          - application/json
-        accept-encoding:
-          - gzip, deflate
-        connection:
-          - keep-alive
-        content-length:
-          - "5832"
+          - "6085"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -5244,7 +5278,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -5260,24 +5294,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFRRb9s2EH73rzjwKQFkw/ESp/WbN2xrgCUY0AJbOxfGmTxJ11Aky6Nc
-          e0H++0DKsd2uA/YiSPruPt59/O6eRgCKjVqA0i0m3QU7/nF/X9Mfn5xtfpuaD/cSHm7f1NP3/bsP
-          +vO9qnKG33winV6yJtp3wVJi7wZYR8JEmfXq9uaH29n8ZjYvQOcN2ZzWhDS+9uPZdHY9vroaz6aH
-          xNazJlEL+GsEAPBUnrlEZ2inFjCtXv50JIINqcUxCEBFb/MfhSIsCV1S1QnU3iVypeqnlQNYKem7
-          DuN+pRawUn8u7+Di512wyA43lmAZE9esGS3cuUTWckNO0yVEqikKJA8dpdYbAXQGEunW8eeeBHoh
-          k2F2iWKIlEoADdyQWoIQybDOign4GpZ3UKSRCgLGxLq3GO0eNhb143jjdwd4Au9aAtppiiGBYdG9
-          CAlsMbLvBXIPGEL0qFuSCthp2xt2DTQRDZNL4w3m4l4Kv6BJM6lKoU3Md3YMlKq8/rS8v6yG08fY
-          OC+J9THb8iPB2zfL3+FioPUO3rYYLO1hizZLUUffQYMd5bZ93F9WRYstS4+W/8Yswbl0hVLQMjm9
-          L6HCHVuMnPbQYRgkEDpJHym3bHmQ/EXkc4HZgfEdsjvQd95SURhC9IFi2p+FlzND9InYjS03+XPD
-          Los4gV98BNphdns1NB6i37IhATSGE28JasLU56JSirzpSwkVfGnZEvz68HBwGEVgQy47jASk3zQR
-          QyuQTTqkuSb3800fJerlDg+WKJfOXTEkuRadppPzcMM2S1ccGntJmWV5B7KXRJ1MVqoaZiGSpW3O
-          XYv2kYaZeL1SK/d8PkSR6l4wz7DrrT0D0DmfynWW8f14QJ6PA2t9E6LfyDepqmbH0q4joXiXh1OS
-          D6qgzyOAj2Ux9F/NugrRdyGtk3+kctzV7Pp6IFSnXXQG38wPaPIJ7RlwPT1slK8p14YSspWz7aJ0
-          nilzyj2tIuwN+zNgdNb4v+v5HvfQPLvm/9CfAK0pJDLrk0m+FxYpL+v/CjsKXQpWQnHLmtaJKebL
-          MFRjb4c9qgbTrGt2TXYXD8u0Dms9v72dv6rnr6Zq9Dz6BwAA//8DAJjrpnxVBgAA
+          H4sIAAAAAAAAA4xUwW4jNwy9+ysIXdICjhF7naT1zSha1HXQLpA9pKgXhixxZrjWSAORcmIE+feF
+          NFl7Nk2BXgx4Hvn4+CjyeQSgyKoFKNNoMW3nLn+Zrv+sq/2Hvx4fV8d7evxtur7/ez39+Ac/zO7U
+          OGeE3Rc08i1rYkLbORQKvodNRC2YWae31/P5h9nt1U8FaINFl9PqTi7n4XJ2NZtfTqeXs6vXxCaQ
+          QVYL+GcEAPBcfrNEb/FJLeBq/O1Li8y6RrU4BQGoGFz+ojQzsWgvanwGTfCCvqh+3niAjeLUtjoe
+          N2oBG/WpQcAng7ET6GI4kEWGiBVG9AYZJMBBRwqJ4THEPYP2FliSpRLncsc56NenzmnyeucQllGo
+          IkPawcoLOkd1JoMfHparHyfwsFxBFUxiZAgeWr0nX8NyBcUnBvKCsYsohSzXS95izJ3Z8kkCNKnV
+          niewEiBvXMqqW5QmWIYqRMBeTibuIloyeUwMoQLjsk0VYeQxZHOiZqED9ilel8Ax4EG7VP7kpCyZ
+          PLSZSTugVtfk63ER16JEMn1ZzYzMuWrO4CMLtjyBNR77UrRLvZBX0XCXvN1hrAvTHeJFb3P25f73
+          5UfIKpDHcC8xtTt0X0rgOvjg0e/DBUOFWlLE7/jf9FK4qZPgLxgssUnMpTEP0iCYRjuHvsbiz8B9
+          ciTHCXxqkPE084bqxlHdSMmltgtRdJ5uqECi9tzp/HSOQwt7CTne4gFd6Fr08qaYw/P8YWDeRo37
+          ZxvR4SFX2rIJEfvn+/MJTox2mweDnKFKO8aNfxmuQsQqsc6b6JNzA0B7H6R3Ky/h51fk5bR2LtRd
+          DDt+k6oq8sTNNqLm4POKsYROFfRlBPC5rHf6bmNVF0PbyVbCHku56c1s2hOq80UZwPP5KypBtBsA
+          tzfX43cotxZFk+PBjVBGmwbtOfd8UHSyFAbAaND4v/W8x903T77+P/RnwBjsBO32vJ7vhUXMJ/e/
+          wk5GF8GKMR7I4FYIYx6GxUon119D1b+nbUW+zk+O+pNYddvZ9AZvrudmbtXoZfQVAAD//wMAFBlY
+          tRsGAAA=
       headers:
         CF-RAY:
-          - 96665ca80ca417ec-SJC
+          - 96a9b56b8e8d1566-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -5285,7 +5319,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:31 GMT
+          - Tue, 05 Aug 2025 22:25:10 GMT
         Server:
           - cloudflare
         Strict-Transport-Security:
@@ -5303,13 +5337,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "4588"
+          - "1725"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
         x-envoy-upstream-service-time:
-          - "4591"
+          - "1727"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -5317,13 +5351,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998605"
+          - "29998550"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - 0c72b275-4fab-411f-bd39-d23769942d78
+          - req_1eb51e2829ea410f8e947f1c8e5461a0
       status:
         code: 200
         message: OK
@@ -5332,67 +5366,66 @@ interactions:
         '{"model": "deepseek/deepseek-r1", "messages": [{"role": "system", "content":
         "Answer in a direct and concise tone. Your audience is an expert, so be highly
         specific. If there are ambiguous terms or acronyms, first define them."}, {"role":
-        "user", "content": "Answer the question below with the context.\n\nContext:\n\npqac-56def044:
+        "user", "content": "Answer the question below with the context.\n\nContext:\n\npqac-0779e397:
         Explainable Artificial Intelligence (XAI) is a field focused on providing interpretations
         and explanations for deep learning (DL) model predictions, addressing the ''black-box''
-        nature of these models. XAI aims to make DL models more understandable and trustworthy
-        by offering insights into their decision-making processes. Key concepts in XAI
-        include interpretability (the degree of human understandability intrinsic to
-        a model), justifications (quantitative metrics that explain why a model should
-        be trusted), and explanations (descriptions of why specific predictions were
-        made). XAI is particularly important in fields like chemistry, where understanding
-        predictions can guide hypotheses and ensure models are not learning spurious
-        correlations.\nFrom Wellawatte et al, XAI Review, 2023\n\npqac-fd5c2963: Explainable
-        Artificial Intelligence (XAI) refers to methods and techniques in artificial
-        intelligence that make the decision-making processes of AI systems transparent,
-        interpretable, and understandable to humans. The excerpt references several
-        works that explore XAI concepts, including its definitions, methods, applications,
-        and challenges. For instance,  provide a comprehensive overview of XAI, including
-        taxonomies and opportunities for responsible AI. Other works, such as those
-        by  and Gunning & Aha (2019), discuss interpretable machine learning and DARPA''s
-        XAI program, respectively. These studies emphasize the importance of trust,
-        fairness, and accountability in AI systems, aligning with regulatory frameworks
-        like the EU''s ''right to explanation.''\nFrom Wellawatte et al, XAI Review,
-        2023\n\npqac-fd1b0a4f: XAI, or Explainable Artificial Intelligence, refers to
-        methods and processes that make the decision-making of AI models, particularly
-        deep learning (DL) models, interpretable and understandable. While DL models
-        are often highly accurate, they are typically less interpretable. XAI addresses
-        this trade-off by first developing accurate but opaque models and then adding
-        explanations to clarify predictions. Explanations provide context and causes
-        for predictions, offering insights into the underlying mechanisms. XAI methods
-        can be intrinsic (built into the model) or extrinsic (post-hoc). Evaluating
-        XAI involves attributes like actionability, completeness, correctness, domain
-        applicability, fidelity, robustness, and succinctness.\nFrom Wellawatte et al,
-        XAI Review, 2023\n\npqac-a6d5f255: XAI, or Explainable Artificial Intelligence,
-        refers to methods and techniques that make the predictions of AI models interpretable
-        and understandable to humans. Counterfactual explanations are a key tool in
-        XAI, providing instance-level, actionable insights by identifying changes in
-        input features that would alter the model''s prediction. For example, in chemistry,
-        counterfactuals can suggest modifications to molecular structures to achieve
-        desired properties. Techniques like MMACE and CF-GNNExplainer are used to generate
-        counterfactuals, with MMACE being model-agnostic and applicable to both regression
-        and classification tasks. XAI also contrasts with adversarial training, which
-        focuses on exposing model vulnerabilities during training.\nFrom Wellawatte
-        et al, XAI Review, 2023\n\npqac-e9abf8c6: XAI, or Explainable Artificial Intelligence,
-        refers to methods and techniques used to interpret and understand the decisions
-        made by machine learning (ML) and deep learning (DL) models. In the context
-        of chemistry, XAI is applied to uncover structure-property relationships, such
-        as using counterfactual explanations and descriptor-based methods to interpret
-        black-box models. For example, counterfactual explanations can suggest actionable
-        modifications to molecules to improve properties like blood-brain barrier permeability.
-        Descriptor explanations provide quantitative insights into molecular features,
-        aiding chemists in understanding model predictions. XAI bridges the gap between
-        complex models and human interpretability, making it crucial for domains like
-        drug discovery.\nFrom Wellawatte et al, XAI Review, 2023\n\nValid Keys: pqac-56def044,
-        pqac-fd5c2963, pqac-fd1b0a4f, pqac-a6d5f255, pqac-e9abf8c6\n\n----\n\nQuestion:
-        What is XAI?\n\nWrite an answer based on the context. If the context provides
-        insufficient information reply \"I cannot answer.\" For each part of your answer,
-        indicate which sources most support it via citation keys at the end of sentences,
-        like (pqac-0f650d59). Only cite from the context above and only use the citation
-        keys from the context. ## Valid citation examples, only use comma/space delimited
-        parentheticals: \n- (pqac-d79ef6fa, pqac-0f650d59) \n- (pqac-d79ef6fa) \n##
-        Invalid citation examples: \n- (pqac-d79ef6fa and pqac-0f650d59) \n- (pqac-d79ef6fa;pqac-0f650d59)
-        \n- (pqac-d79ef6fa-pqac-0f650d59) \n- pqac-d79ef6fa and pqac-0f650d59 \n- Example''s
+        nature of these models. XAI aims to enhance trust and usability by offering
+        insights into why a model makes specific predictions. Key concepts in XAI include
+        interpretability (the degree of human understandability of a model), justifications
+        (quantitative metrics that validate model trustworthiness), and explanations
+        (descriptions of why a prediction was made). XAI is particularly relevant in
+        chemistry, where understanding DL predictions can guide hypotheses and ensure
+        models are not learning spurious correlations.\nFrom Wellawatte et al, XAI Review,
+        2023\n\npqac-e0e4c958: XAI, or Explainable Artificial Intelligence, refers to
+        methods and techniques used to make the predictions of AI models interpretable
+        and understandable to humans. In the context of molecular property prediction,
+        XAI methods like molecular counterfactual explanations and descriptor explanations
+        are used to explain black-box models. Counterfactual explanations involve generating
+        molecular structures with minimal changes that result in contrasting properties,
+        while descriptor explanations use surrogate models to attribute chemical properties
+        to specific molecular features. These methods enhance trust in AI predictions,
+        aid in understanding structure-property relationships, and improve accessibility
+        of deep learning in chemistry.\nFrom Wellawatte et al, XAI Review, 2023\n\npqac-b5128345:
+        XAI, or Explainable Artificial Intelligence, refers to methods and techniques
+        used to make the predictions and decisions of AI models understandable to humans.
+        In the context of molecular property prediction, XAI aims to explain how models
+        learn chemical principles, enhancing trust and ensuring the model''s learning
+        aligns with scientific understanding. Key challenges in XAI include representation
+        of explanations, defining molecular distance, adapting explanations for different
+        audiences (e.g., chemists, doctors), exploring chemical space for counterfactuals,
+        and developing systematic frameworks to evaluate explanations. These efforts
+        are crucial as AI models transition to practical applications in industries
+        like healthcare and environmental science.\nFrom Wellawatte et al, XAI Review,
+        2023\n\npqac-314237dc: XAI, or Explainable Artificial Intelligence, refers to
+        methods and techniques that make the decision-making processes of AI models,
+        particularly deep learning (DL) models, interpretable and understandable. While
+        DL models are often highly accurate, they are typically less interpretable.
+        XAI addresses this trade-off by first developing accurate models and then adding
+        explanations to their predictions. Explanations provide context and causes for
+        predictions, offering insights into the underlying mechanisms. XAI methods can
+        be intrinsic (built into the model) or extrinsic (post-hoc). Evaluating XAI
+        involves attributes like actionability, completeness, correctness, domain applicability,
+        fidelity, robustness, and succinctness.\nFrom Wellawatte et al, XAI Review,
+        2023\n\npqac-76cd88cc: Explainable Artificial Intelligence (XAI) refers to methods
+        and techniques used to make the decision-making processes of AI models, particularly
+        complex ones like deep learning (DL) and graph neural networks (GNNs), interpretable
+        and understandable to humans. XAI approaches include gradient-based methods
+        (e.g., integrated gradients, gradCAM), model-agnostic methods like SHAP (based
+        on Shapley values), and visualization techniques such as saliency and similarity
+        maps. These methods aim to attribute model predictions to input features, identify
+        important subgraphs in GNNs, or explain molecular property predictions. XAI
+        is crucial for ensuring transparency, trust, and usability in AI applications,
+        especially in fields like chemistry and molecular modeling.\nFrom Wellawatte
+        et al, XAI Review, 2023\n\nValid Keys: pqac-0779e397, pqac-e0e4c958, pqac-b5128345,
+        pqac-314237dc, pqac-76cd88cc\n\n------------\n\nQuestion: What is XAI?\n\nWrite
+        an answer based on the context. If the context provides insufficient information
+        reply \"I cannot answer.\" For each part of your answer, indicate which sources
+        most support it via citation keys at the end of sentences, like (pqac-0f650d59).
+        Only cite from the context above and only use the citation keys from the context.
+        ## Valid citation examples, only use comma/space delimited parentheticals: \n-
+        (pqac-d79ef6fa, pqac-0f650d59) \n- (pqac-d79ef6fa) \n## Invalid citation examples:
+        \n- (pqac-d79ef6fa and pqac-0f650d59) \n- (pqac-d79ef6fa;pqac-0f650d59) \n-
+        (pqac-d79ef6fa-pqac-0f650d59) \n- pqac-d79ef6fa and pqac-0f650d59 \n- Example''s
         work (pqac-d79ef6fa) \n- (pages pqac-d79ef6fa) \nDo not concatenate citation
         keys, just use them as is. Write in the style of a scientific article, with
         concise sentences and coherent paragraphs. This answer will be used directly,
@@ -5407,7 +5440,7 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5544"
+          - "5470"
         content-type:
           - application/json
         host:
@@ -5423,56 +5456,54 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//4lKAAS4AAAAA///iQjABAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA
-          //9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD/
-          /0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//
-          QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9C
-          YgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0Ji
-          AgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmIC
+          H4sIAAAAAAAAA+JSgAEuAAAAAP//4kIwAQAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmIC
           AAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIA
           AAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAA
           AP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA
           //9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD/
-          /0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//jFhtbxu5Ef4rA36JBKx0cRI7OX0pjFxa
-          uEh8yDXX5FAXwSw5u8sTRW44XNlqEKA/or+wv6QYclcvtnPoN8uSyJlnnpdZfVXWqJVqyS/OXp4/
-          f/ns4vz52UK///Xj9rfn78/f/7L5+8er3355/9cLVak+hq01FNVKXVNtB1aV2gRDTq2UIeqZaP3D
-          9McinqlKhfp30kmtlO4wLXXY9I6SDV5VSkfCREatDhdXSnfBamK1+sdX5ULbx1CzWvnBuUo11lvu
-          PkdCDl6tFKfQq0p5THZLn7/zrvWG7tTqaaU2xIwtqdVXFYMjtVLIbDmhT1JN8Im8VPrmrndoPdaO
-          4DIm21ht0cGVT+Scbclrgtmny6s5kJeGkJkYNpS6YBjQG0ikO2+/DMRgiG3ryUAKEMkbipA6AkPa
-          sg1+scG19S30MWjK54QGNqg76wkcYfTy7uynt3PIQDOkiJ57jORTBdYnin2kJMVW+e5B7pCmTG4g
-          BeiGDXquAI2JxCwHpo5sBOs7knPgRtUO9XpRh7sbBR7TEAlm/RfUi/MLQ83TFy8qyC8bc66f/Xjx
-          fL6ET5dXgHbDcgX5DgWWFAdOuQzUOgw+YW2dTTuod1DII7dbz7btEkv5ofQFfSRjtTCDIXUxDG13
-          3F05ZSbQ7avOfS3u9cspDjrXHxrAcvi8gt8HzoPEcsPsy4A+2ZSZM1a9oRSt5nmBkYQEfvp4Hzgt
-          uqAh5v+gI4YmROCetBx7XP78HnLzJXzoiOn0SI0eamkmRYFDw4w2NRlDBm5t6qzPNCnYYNSdTZTb
-          mkOIQHf7b2HfO0sGcoUpohXCjD2kjnaAkYC26AbRGtTIZCB40NEmihaBB90BMmCufkS6gsYaKn/F
-          UA+cPDGXU03YoPWQL9b7yYzsOKuf4otmvgS48Tf+Eta0y0RJITiwDJkUFBvUaUB3DEkFt53VHVhD
-          XkYlkrLebtCB9f2QBIujCaYOE6BLFB8SqAIS9QrVpq4c7Vn333//p8kYonhRBTy0LXGST2+CIz04
-          jEc00h36ljLLRZa0pazpSAZ0RxurUe4OPcUkNRcg8MKcN8/Oz0fV0I9YN6/0xXwJV758jVPcVRma
-          OlojFxRvvBvbCUPqh8SZDCPTT8QODWWZcrWf4L7mxVjPDiK5gldnexlf0V+360PqiC1DS54Kp/Ns
-          t+isKS9n9yvPenccAJ1t/VhZpHZwmELcQRNxQ7chrhlmtGyXVWbwm1+fMNyoKNBnpzhM/EbNxRfI
-          8xClrAZtLDQ7eJzeFdLhYOzeBqyHzrbdghOuxWIzIXnPwb1DfeysI6jRodeFDHqIqHf5xPvmUoYx
-          ubiza4J37y5fv4FZHsgCWx84WX2PwjzPp73+8+Iv19djclCEWRux7xZFbsfCn4OhTfCcIiYCmxjQ
-          YL9vDXUMzJCQ1yVKDCaEtOsfUGu+VJWK1AyMbgrIkn3Wt2qlfl7jroIr8FTCBz3fjuEjwZRnfKM+
-          iowsS+9/ulEwTOkwurWQXHLxLi3hLYlHAieMSeYWCU35dHFrQt2J6cq3xy8Bib0V8bSYuvF+MYU+
-          WJ94KS7xoSNobOQ0fW1UzSF6GHe5RKkU6zCkkzCZ5liM4YF5i9hlJznEaUnSRwLxyT4Gn4wpuIQr
-          6dqXox5y5jRXHokO5NyuDl5Tn3gJl45DBTY9YbCbPkRZP4TRjSU3EW/vEBmfE1pL0aP7Zb5mBVxe
-          PVgojtaER9aC3FakRoJUy3wkAasjBT6M8KroXs4+Vn5uMhe9V/tjWs8ZuJuQnGRWTSlyhJ7u0DkS
-          y61A22zLxVTqHYhrRT4GpeRNucn6B8D89HZam07MM7dvbDb9wXJHDDWlWyJ/FMhllNOr4/XOWU48
-          hWq2zpSirYdEIxRTeOZPH/IzgyCTMbQITbO/8w99STjvg2S3rD2H3icfgCboIS+O/g/SNRMRMxVL
-          FHtBqYLWbgXhMQwFpiPywYfDHntkiI/5nWwZ43DJlEHaoueIApaMJNMGzZYiY5SNelpXDk1NcSPD
-          0UPeh7Psj6q6R5niWPc8uTgnsY62TyGONjyNMIUDyAcvmNwETBzafH/YUsyb68iu0w3hkPm5/Otw
-          e+y2vPM5Yf8lurDjpougw7i7Fjdewt8mNzXUlBvyWFJHfi8WCYl+iH1ggtlpNN4ny7w6cRuYPXSs
-          kzCq9qDM9sT/YU/66kHWnQr2ZC6j9x1EMW2YS3iHawKWZSoF0TSVrOildR1iJJ1csbrXHel1dmwB
-          32HPK9gMLtleVvswRJ2ftAouny6vxG1C3u1g9L1R7qd296CR6jt8KgotsxhlE3OayXNinlvJ5zGo
-          eAnX48DzFnPc2JiEY+xahhr1Wjbw3ZiQuW+Bo8C1pp2kwzZYA5HM4A16vYN6EIy2IjHnIG/eY3SO
-          SSLeUB4pbTP6+f9tZk0MGzhx0iW8G9HNj1pjQOmM8Oj4ZKasMkOe6Hcf+o6PL+m1hDePumb+pNzI
-          uJkGfcKbA5KNC7cMLrSyebvxksPEpKAjoZ/QVca0v35549W3b/+s1DD9KNDHsOnT5xTW5Fmtzs5e
-          vpKfBaZfLPZvvHr2slIpJHT7fz17+vT827f/AQAA//8DAOX5JnxvFAAA
+          /0JiAgAAAP//QmICAAAA//9CYgIAAAD//0JiAgAAAP//QmICAAAA//9CYgIAAAD//5xY244buRH9
+          lUK/rDRo9c7NO7ZegokvyWxsJ8huEAeZwCixS921YpM9JFsa2TCQj8gX5kuCIvsiaexNkKcZqZtk
+          1alT5xT1OeMyW2YVmcXFzbPr66vLm4vzxSt6c8HPN9cX9ZvHR/6RHh4+/C3Ls9bZLZfksmX2M7rK
+          mizPGluSzpZZSdR6os33wz8Ld5HlmV39Qipky0zVGAplm1ZT4LhSOcJAZbacDs4zVVtW5LPl3z9n
+          2latsyufLU2ndZ6t2bCvPzpCb022zHywbZZnBgNv6eM3nrIp6TFbnudZQ95jRdnyc+aspmyZoffs
+          A5og0VgTyEikrx9bjWxwpQluXeA1K0YNdyaQ1lyRUQSzD7d3cyAjCaH35KGhUNvSA5oSAqna8ENH
+          HkryXBkqIVhwZEpyEGqCkhR7tmbR4IZNBa2ziuI+dg23dxBR9f/+579adIFVp9HpPST4HkEgBk3o
+          jKydvXo7j8dWDtsaDHUONRgKO+s2MPvd+/dzQKdqDqRC50i2ZRPItY5CzFIWdxKbgFHGr4KFumvQ
+          eJi1D6gW5zc3L+jqxU0O8ePVxfXl1U2p+o83P6jy+XOl5gV8uL0DLEuXspFc2dTkyAS4z1Ya1Wax
+          so/3GRiUYCTfUJOnPmVY7SHRTFJj47mqgwc2wcKu3oNvSUlFoHVUshIqeUBH0GBJOZCp0ShZGlzn
+          Qw6dxxVrDvs8ZomaK9NIMDsONZS2QTawMXanqazo68munl1cPr+6fjYv4A+UqmANmRiW0l1JcHZ2
+          AGg87uwMZhHAxQmwMU1YUY1btm6ew9kZCeGExdZ4WVeSV47blJtdH6QKCjuPsv885XN29kvnI0Wn
+          5Q8dmsAhdgVsUXMZn0WkBZSddaFmQ97PT/KdFwD35t5IDUc+OwKFgSrr+BOVgD5l66Q2Ss5bdaxD
+          qlBK7pBsc7AupniwoLU+LGqrAA3qvWc/z4GaVtt9LNzUPL5TtRxYOSyZTFis0EsIbessqpo8zKio
+          ilwOp8qJmozvyqYxnAVWxvrACtYOG5KmGNf99PvbP/VIbtl3qPlTwipYq8e3BHEyag8NtiNmA+VP
+          GmJewJ0BVVPDPrh9HvuBPSjHgRVqWFsHjdUUe1qo3pIL+4Ma57CTfhkLoHlDoGwnBFujCh1qOGQM
+          zBo23KAGH1wnoKMGVaOpyAPqQK4XGDmIySetGDhm3clmmk1UpDXF/vSiBIeLSXdKGEXjcbRAFXjL
+          YQ+OdNqn5jZJYcOBq/h22zm2nQdl3fhajyWd07V68ex5j+UpHV/WqDXFfIaGwxLbIGEeBS/Ylrwl
+          5wmwExqoiSN9TTxsfQFKsxFZN0KTksQ7THVQF88Na3SSU0PBsfKJJX7vAzUYS6n3QFvUHT4NJPHU
+          GlhzSUl9nF11PkjXpZ166cG21ax60ejhGPTmKbV+jkpJ67V1IbXmlgNq6ZHbOwgOjecUQbQb1Iud
+          dbocTkmP2EBNqEOtZAOJhcyWnY26KDRSETgRU9+5hEuU5iieSTjjO6EXYjaKW02p4K0TNgjTHT10
+          7KiJSjk7UdIszxytRcwGd0/GzabKltkfN7jP4Q4MJedE43e9c4owxBa9z/5aY5De+nB795v7DDof
+          5aOm3kCohGjqj6GAtxSgEcqiC2IxjrBMbzvbVTUQqrq3omERkIhcaoAKQ92fv6E9tJZN8IVw842z
+          zTFrl0PL48p24cjMBovoXUuQf0LfI2tPwOeDofb5sYPvRif9rjfSAu4CIDcx3mSDlPQ+2fvgg4OH
+          GUVtz6FT78rh2FTyJ6EW8GfStEUjuj+pXcxg8rukOwc2LQa8tQmNr+pBhPRIEhKaDW4oUvxwu6MZ
+          Jsruf9HV/1lRf00gBbDOUzk0Y026FRCGeeMo+0kix4COJLIfSpTMftxrwCGgEx5D55zgMcyR/mR+
+          ewLHhAKHlBAbD7XdDc0dOZdOluadurqAr+jvASTgqHXkRTt6mMdjyzhdi5QMapyEu39xPMy3qNKe
+          1vXPIuN6dRUdHa27gLumtU6m9ki3SW/+H5GbAB5U9hsAD4P6EelyoDiPRi+QYF697eEs4LeoMdoP
+          KtU5VPsYxGmrFfD6kFy9cI0SJEtk5JMqvOvZq9DAKjZtmqggMrT/UMDrCTQ2W6u3EkIIjlddoJ77
+          GKkwdnvsQNV7EwU1gTIMOT0ofQQDCU7GstnXprA8/vvy9t3TcWz27elrHAEL+ItsfSQycawYORY3
+          jQOLdckBW3RxWos3MenKmM97uzt0FL83cu3gT9RfP5KmpzEWQdn+zpKcp4CfonNE78N+YBhExaSY
+          xXc611pPMDvQ61GoRzE/uaPM8+gpoyLPnsrxoQDNJx2bjRz4fiRAPt2QDgbpWHWBOz/RPdnutG8O
+          hlc1dv5pRxbwDjcEXq5wwYLiQMlE5cI6MEonAXtZk9rE+tgtOY39aBjHCwLUMiqMbrK24p09xcqI
+          uNioo7IzJRq1L+BNZDzKXTjvb9NDPeKoLack0jSdDiJi4G3nlGThrdzfVmxi3Rtpq6l+wWFJC7te
+          w4rCjsj8evPCejT/QT0KuP0Wmr3VgqctxRE99fixvk47DoJ/1NFyiyDH+NWTB4GYng3tmwazQ1uN
+          dbkVL5bZIgnb2oKx0dAPpiAZF6gFkfc0TQtP2SdJnabA2KfCgknIWnJauDuSoh+v+lnO17bTJdS4
+          jZWYKLChvT9kl8QUrHiuCJpcZ8dzhFMNCoTjfNf77dFJgVqZ+eRvcW+yL1/+kWfd8FNQ62zTho/B
+          bsj4bHlxcXUpPwYNv1OND15cXOdZsAH1+NXl+fUPX778BwAA//8DAFMzR2v6FAAA
       headers:
         Access-Control-Allow-Origin:
           - "*"
         CF-RAY:
-          - 96665cc63dcdfa8e-SJC
+          - 96a9b5776a4bce60-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -5480,7 +5511,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:32 GMT
+          - Tue, 05 Aug 2025 22:25:11 GMT
         Permissions-Policy:
           - payment=(self "https://checkout.stripe.com" "https://connect-js.stripe.com"
             "https://js.stripe.com" "https://*.js.stripe.com" "https://hooks.stripe.com")
diff --git a/tests/cassettes/test_pdf_reader_match_doc_details.yaml b/tests/cassettes/test_pdf_reader_match_doc_details.yaml
index ba4d635aa..39ef17742 100644
--- a/tests/cassettes/test_pdf_reader_match_doc_details.yaml
+++ b/tests/cassettes/test_pdf_reader_match_doc_details.yaml
@@ -20,7 +20,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -30,7 +30,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -46,18 +46,18 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA4xTz2/aMBS+56948hkmSMpoubXSpk0aU289NChxnRdwcfws+6UDIf73yQklbGNS
-          Lzn4++Hvvc85JABCV2IBQm0kq8aZ8cN+WW/n5uv3ZVbNstR820r6ub1t908PP3ZiFBX08oqK31Wf
-          FDXOIGuyPaw8SsboOp3Psnn6eZalHdBQhSbK1o7HNzROJ+nNeDodp5OTcENaYRALeE4AAA7dN0a0
-          Fe7EAiaj95MGQ5BrFIszCUB4MvFEyBB0YGlZjAZQkWW0XeqyLF8D2dwecguQC9ZsMBcLyMU9PKIP
-          DhXrNwSy8GXnjLQyTheAaliSQdUa6eHRY6VVBGAZBwu5GPV+suUN+RAdn3PxhMbIX5IZARmkycXq
-          xKtIR45tjcntMbdlWV4m9li3QZoT4wKQ1hL3keIVqxNyPG/H0Np5egl/SUWtrQ6bwqMMZOMmApMT
-          HXpMAFZdC+0fixXOU+O4YNpid91t1tuJofYBzO5OIBNLM5xP03R0xa6okKU24aJGoaTaYDVIh85l
-          W2m6AJKLof9Nc827H1zb9UfsB0ApdIxV4c6NX6N5jH/F/2jnJXeBRUD/phUWrNHHIiqsZWv6ByvC
-          PjA2Ra3tGr3zunu1scjkmPwGAAD//wMAkwZbgbMDAAA=
+          H4sIAAAAAAAAAwAAAP//jFNNj9MwEL3nV4x8blHzsVroDaFdDrCoEhKg3VSJ60xaF8dj2ZNl0ar/
+          HTnpNgWKxCUHvw+/mec8JwBCN2IJQu0kq86Z+bv0w6fN+5vb+0Z/K9z9R7p1/d73hZKfV1/ELCpo
+          s0fFL6pXijpnkDXZEVYeJWN0Ta+viiLPrhf5AHTUoImyreN5QfNskRXzNJ1ni6NwR1phEEt4SAAA
+          nodvjGgbfBJLWMxeTjoMQW5RLE8kAOHJxBMhQ9CBpWUxm0BFltEOqeu63geypX0uLUApWLPBUiyh
+          FG9hhT44VKwfEcjCzZMz0so4XQBq4Y4Mqt5IDyuPjVYRgLs4WCjFbPSTPe/Ih+j4UIqvaIz8IZkR
+          kEGaUqyPvIZ05NjemNIeSlvX9Xlij20fpDkyzgBpLfEYKV6xPiKH03YMbZ2nTfhDKlptddhVHmUg
+          GzcRmJwY0EMCsB5a6H9brHCeOscV03ccrnudj3Ziqn0C8zdHkImlmc7TLJtdsKsaZKlNOKtRKKl2
+          2EzSqXPZN5rOgORs6L/TXPIeB9d2+z/2E6AUOsamcqfGL9E8xr/iX7TTkofAIqB/1Aor1uhjEQ22
+          sjfjgxXhZ2DsqlbbLXrn9fhqW1ddbdo8VymqQiSH5BcAAAD//wMAMSRDcb4DAAA=
       headers:
         CF-RAY:
-          - 96665ccc0ad4fa86-SJC
+          - 96a9b549bd4dfc54-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -65,17 +65,15 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:33 GMT
+          - Tue, 05 Aug 2025 22:25:03 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=BMrt8vqTptk_fnuJI59elD4N43d726oGzmj7aAL2_mM-1753726533-1.0.1.1-IoxN_BcRi67wqB.eGTavtLT1r1KMJkjon6b73yiJiy6f2EYQhS8rDWGN5Qwrfdv8.zONcW31ayk0mj6rFxTZ_qCfyvZlshqFWpQ4rBI8UMA;
-            path=/; expires=Mon, 28-Jul-25 18:45:33 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=o7q.1S_6gtWdEwOfA2LFsA539k1790E7Iyv34N8Jbs8-1754432703-1.0.1.1-0QGMn7L7MB2PcbGhT398XEmXZNEmfXzLoh_eukp8HnUSLKPK8m9GhUHLDEknIvj7JbE8jN.GrU_hMf9P.NnLFvYMf87BUjLc9jTabDFBxN0;
+            path=/; expires=Tue, 05-Aug-25 22:55:03 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=WQ13ZNknZ8XAKyeImjTCdf8t9srvphPzM4Dfc9myDQo-1753726533078-0.0.1.1-604800000;
+          - _cfuvid=aJe.jyzBvSp4Ssyjjap6GFGC0exfACcLK7a89CamCbU-1754432703550-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
-        Strict-Transport-Security:
-          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -89,13 +87,15 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "507"
+          - "480"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
+        strict-transport-security:
+          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "512"
+          - "483"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -109,7 +109,7 @@ interactions:
         x-ratelimit-reset-tokens:
           - 0s
         x-request-id:
-          - e37383b6-6890-4e09-97ba-23915630c194
+          - req_d89ece8bad9bac5e3df4fac78e7999a3
       status:
         code: 200
         message: OK
@@ -121,7 +121,7 @@ interactions:
     response:
       body:
         string:
-          '{"status":"ok","message-type":"work-list","message-version":"1.0.0","message":{"facets":{},"total-results":20205,"items":[{"DOI":"10.26434\/chemrxiv-2022-qfv02","author":[{"given":"Geemi
+          '{"status":"ok","message-type":"work-list","message-version":"1.0.0","message":{"facets":{},"total-results":20334,"items":[{"DOI":"10.26434\/chemrxiv-2022-qfv02","author":[{"given":"Geemi
           P.","family":"Wellawatte","sequence":"first","affiliation":[{"name":"University
           of Rochester"}]},{"given":"Heta A.","family":"Gandhi","sequence":"additional","affiliation":[{"name":"University
           of Rochester"}]},{"given":"Aditi","family":"Seshadri","sequence":"additional","affiliation":[{"name":"University
@@ -142,11 +142,11 @@ interactions:
         Content-Encoding:
           - gzip
         Content-Length:
-          - "441"
+          - "442"
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:35 GMT
+          - Tue, 05 Aug 2025 22:25:04 GMT
         Server:
           - Jetty(9.4.40.v20210413)
         Vary:
@@ -155,9 +155,9 @@ interactions:
           - interest-cohort=()
         x-api-pool:
           - plus
-        x-ratelimit-interval:
+        x-rate-limit-interval:
           - 1s
-        x-ratelimit-limit:
+        x-rate-limit-limit:
           - "150"
       status:
         code: 200
@@ -1225,7 +1225,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -1235,7 +1235,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -1251,1690 +1251,1692 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA1RZSQ+yTJfd96948m7tRCapW98OmScpBFTslSgooKIMVUD/+Q4+PaQ3JkqZFHc4
-          95xz//Pf/vz5p8mq/Nr/868//zzLrv/n35ffbpf+8s+//vzHv/358+fPf/4+/9/J/JXlt1v5vv+O
-          /x6W71s+/vOvP9z//vJ/h/715x+uqZRgTswnmnfV3kTDUXgQdeXYscAfdyU0625PL4axi+fN2rvA
-          Keg3ZOfaoTZK2CkxjpOcnvbJlE1xVkYAIt5S/WActMmc5BoaYVyR3entaVysjSnkn21GdJ17ad9d
-          53dg1dlpWBXxthXnAnRwVtZhmKVjNdPQ3kbouc85okV64k0PtpKhVTr3d59sFEjLAAkfKWAv9+N1
-          Fu/noH++V0L81alloTcEcHkoiJrVc6ONaCQmlB8g5OA1czvGIpNxh6bDUCtY0Fj/nmTs9UQnZn0r
-          vUnF4KO4v0Y0/eSt17+KuAQlWD0DJHFDO223WxNbZ9EaXqaqZnxthBPyrTEhp4O19ljNuAlX5DNQ
-          M9yvZ8YHXwEUTXWoAsjXOKNvZQg3s04sk1vFI8/fV/iw3ro0ooJV9fnZVvDmUjKqOKekpdKkqsBj
-          syDeQ5DisebVEm9XekLPyWnnDZzju/I99Vxih6XZCkEjytjrMpGaOnfM2DmSS7R8J2bZaJVIZKIA
-          qNsD2enxOI9s/viwW6sO8U9p7nHNc3bxvmIJOShrgthW6u+QKlxIvIBG3tye7zXIcTcPbVydqrnj
-          10c4p41Bi9M7y0T+koVQG0pG09SvMhFtvi+c8kCJ/2xXaFZVVcHF2OR0n5cvr91zLx/mYoOHyTsL
-          MbuJ5RfnXCqRTBbuHn8v3hMMg30h+sqRqrFOBhsVr3hL9ZY8W842ogHznn+kFwELcS/rpQ3Sbe/S
-          y+EI2ZR3lY3PrZxRX9jp2bg+rEP4xbvAG8/rYe2E4KIEERXFe2/aZJGOxV2d0VO7J+2wDY+5fE+M
-          N3UUEmidfJcmzHV1RYrt8IjF59dTYHX6MKprcNPGdm/ouBO7ier1eWin9Faq+GMpHgmLh+GJrRNd
-          sMeKjFqvm+txG+0RYOyr9QBGFKJxxIdpMxI7oopGnbivjZQh8w0reutru/1br59bKVIt7SVtPJuy
-          idgtdOhFu5Xa1OymCK+m646mZFBQd7J3NlriTUxXsBCzpH2IIzW1yH5j8jPzrUsO5dMoqKaySyWq
-          UeqjuUCYWGdPzaZrdE/xVHwD4q+bqzan+8cL79OLSM1pE1WC3N5DLHwNidqeGCLOzJwBH6hwChif
-          qy2Hm32CJ/4wEPeqVa2Qb7oStKAOqfJRlJj7xCmA8frwpCgTtWWMZlfQ83BNjt9IycTDRp7wRvPF
-          QZzB0MS+urmASSSTQL3vvClu2RUv/T8I9U31OC4UBOzuygPZgnPPBCt0AzT2zCDeDnXV1HLCBVh3
-          BaLsAr3iCjU08YvszyS5vx7zfJGeHfhEvgzj+XvLeFIebUTfxxvRFXzUuP7UDeB1Z5EoibTNZn63
-          l+F2SwmNjfnpMVvdyWCXuUt922405h/fAs4Y/dLdDl1R/bBFDqpNsxreFi9qY/r5XhH06UjcfDui
-          qVOlL0580yBRrM3eePUMAZ6uKw+S9Bwz8c2xC4ZD3AZCMfkxlydThE/9qSP2ZXXSukTpXuhEKpOE
-          Gx08diHPAS5SwpErS01PwIoaYRTCmbriinhCy61S1PrNi55wdo/72AqOEEprMVgRr2ynwuwB1jfl
-          SmKShd68DvQXsLrWyIV5ajXWvHtHC35SX8m9rJvO3VW+Dx+BmnHiepP7ZoB9b/ulVrsn1dhnU/6b
-          HyT59YcuxRcM99am1jtXZ55M3RFduP2e6nP/RpOah1+84C/dtoodM7Z+fLEO0kSPkO7miT9HCUpi
-          byLaMj+64ZAOYIltH7wOTlPNWDv5UJpFF+BrdEdsFjgVq3LZU/fZDtVs53oJZtgaZKv1BeK9/jTB
-          d9y7NOEeu4rJzLQB7Qp+qV+tnc7ypoYp57fkIB0TTdBc/osgmFYDexyrduztIodQUVTiajdVE68v
-          rcShJSNqCkxt6VwpkryvpoRqK6/Uumf/Bgg9yMlpQn42WinmEI/1Yqg1uHms4pwamp2CiPpyNxkl
-          3/Er7+Kmp1Godx4bDLXE9ftWDzJb8vVBoYJv9eDQ4GErmlBlRIL2lX0CKMjgzTuyMqFD7ECtDAva
-          9A6DELnrS04dfvX25l7b1DASN6I7kEpPvArnAarXpSEWzpRMHEIZwN5MWoBt9zEzwTsqOBnHJ9n5
-          Kyvjv8lexpfgVAejcBsRXeYNUtOmJufPXUdM4l8A2j2/UHKNRm12UKJjTlUSeqozqRrHkAmYOye7
-          gCO3D5roBxiYBS+SE1zeFfc4mwr+zUNVr4psCpPtAE+mv+leTh5zN0LC4evVMMmtLY+I4+qzAI8W
-          XWlKTcFj+3OTYru8uiTXD14rqlEYoI4+W5rsIzvm9aQ0//ar/f74rXiKcY124eVEf/3Of6t39Bff
-          bwsfmuZ8dvE9DPf0jOZnzL736IhdIyPEb4ZP3NjRVYfXAxgpQJqy+XVqhB9fIlcjYojt6lLBhh0R
-          YsePXTU+bhmTd8MIwfemnLU5uVYBVvo6Jd6NvLzZMG8TEuTEoAHaXrzJeG586LlVSpO1O8asxR8Z
-          s2n9oc5NLav5U2xTLLkKR/2rNs+sl7Y6DuMopb64KuIxPl1LOKnzbRi1vphZMTkcLHg0sDSX53Z/
-          cL8QoRME3GFzqb7PJx/h/CZ5JGPO2WPPoZngtCmPxJamumo/hZNuNJHF1PdhV/FDgzjwzC4lNqql
-          jK2jMsELflOXeWU1F3gK8RkZQfAAR4n7Q7mVYLOxK+puXpdsNgRyh4G/JkT1zkI2ifHmDkfjIRK1
-          5wStq7RCQWwnOsPTmPl4tK7qF+6ycB/Y29vO/Kh8G1A0xSH2gxuyHz9Ez8P9RE/4hL2ZY24IWvAK
-          SeBZ67jT7+Udo92g0t2UBh6XGqONSTBIVP0orGUb7eGDtZ5m4mQ73eud44cDwn2fxK6/RjtGl0eC
-          tji2qK3rdTZ/O5dBrYBEciyO8VheghUMX10cYIfNTLxbgwz643EkwROMVsj29IVWrRsP3HY1oJll
-          zRHeU97R4MtabeaUtEQLPg/rW+zM9NKc7hv+QTVqyhXNeiFiNQhfS6KBbafZ+AgvF1AiH6gnC4om
-          bmv5AntR/Q5cFD889gz0C9yM2Ak2d55mfUjm1eYD9pfu1/Y7HiM3c2Ec1SvRl/yzIn8J8v0a9HSX
-          h0Sb54LTIfVwQXOWHjMu4teu9ONzMFRWNVts6IDVB4EufDf7ivF4x73kDwN3q8SMDkrNIDp3Cv31
-          w1dusiucyMMcVlYyVfP+GQpgb5hGHeW6zebryyuhlsEKeOdrZKPIbBd+/aZTQUTduVIGEIpHQ45x
-          8vXGbvXlfvqG2uusiaccxiPW1U1FUgGbWrfwR3RDPk/3K5R4s9MaEkiusB34C/Te93w5+JAd4U3c
-          oJficWhmDvJnYFDfjD9ti+PnBZLT7kD12s/ikURkhTpH8snhed16/GFKdcB50Qei/HS0/lneExBX
-          7YkaBYva8VvfTeA5fKauvrdm0bmsO9jTjgXc3L/nSW6bCFHlWRMvKlfzoMo0RaYRN9RrCsh+/Bv9
-          8r1tVkfEts0pgc/hPBKXeWo7ZttzCOacZOR0HRFa+EAHP/7AHQxeG7shaaDXdEKLhH9Xs7exTWBb
-          uad6LzkVr1ViDs1ORQMvPPh2dF5MxXlvroLJnHRP+OVrmc+EJPy7Zab16tA1228CPHkyaiSD9zfd
-          /bJZ4oMyFpm7CzzRTgg21+0HNdyYBYD54E2TV6Z4QjT4V3SUdrtBihM5phs1jDCnqkkgn9IeMRLM
-          itwHu5wUku4jftGjkB9GY/gMlug1pVlx+It3L+q/5aodxWIDUI6RQ3zPotp4184q/PhDrh/aatyI
-          zYDXYt1R8339aHM67lxIskiiqniv48nj5CNo88AC7lquWrrMT/lObxpx0Cv6zccObI07UUJw3rJ2
-          x1w51MSU2ruOZLN5rFaIZ0igxqDN8XwV9h1wIigkHj/Xql8fmQ7TKtySS9BLGXucTRWqTEip2YTB
-          PDFNfIFMdUJP6PWYe826d4CuEk9svxV+fDlHoj8HRKmsVTu4usVgiGvvv/E1+6QJnC7WkarB9T4v
-          9XGEci86v/hmI/+8TLDoBeIPRy6uSTCrSGzcNoCHFGnTwg/hvDXWRO3EezsmB6VEc7x+DNR2J627
-          kH5ApzK4UXtylHb+5etM6Uh9Kymzft9r5Y+vUvPx/qLZYq8ONKPFVHNee9Q9NmUKNxTwxHecO5pv
-          4bXErx75dEsCv+IVOkowHnOfbFX2zQZktKa84Cc5hXd/Hnd2fMFLPgKmV9SbN2oagvK58yRX+3s8
-          HtfnHDZPeUctXXVb5kzsKsurvUr0g8F7M9ZDFf/yZ5PhjqbKm+6g1TH7e/9mlQgJLPqMOFpw0Ea+
-          3K2grbw4mKn0yMZmk4fI0P2OXnSdq6iDrjpyBrkk6uagohmdzUCOWfchYSI9sklOjRCfrGY3iN9Q
-          mOeYGhccKqpK/R1+xXM/V0eYLmNOA9f9ICYdvRQWPUquDynyxmGjqvBQ3Qux5OfHm1b9XcHsHLyI
-          NViveXzsdi+oMs2jFno90Eh57IO0RQM1MwTx+LHCEj9cLSY2Q0krrvHqiOa77VFFr4g2OlogQBuv
-          7sRLM6USl3kDY3T5EH3g1xqbjsUXCr7bEqXnnvOUb7o7tHLhU+d1k2PGuuCLZnm4EF0osnnRywzK
-          vJ7oFssonsirkvBUGimxL+5YTZsgHCB44zXVTpspnmsfXdHPb0h++FQnLxuRPmlJTKVHPB34pwq8
-          ZoR0uwYfcRn2rtA5sk9/elV8vYQAlVFzJPs3xvNffSghGVPnuv3MzB7FBl5Weid+aRjZ/NPvPz65
-          fWOMRiM/NjBamk+06cv//JM7quxaHTrHwBUDb8sAny7boa0kBy1+igqbfjSppq1ebZc/UCNr69YP
-          Vp61zr6toX//+mEG1TttNs7ShIySK4J0tN/z8GbfBn7+SDCeqTcSVa7RY94diSXS/cyKfBCgXd0p
-          MRe9M4YsTdBXu9aLf7OfGY00CeP4mA/vCXUxa57IhgBdS5p7VhHPWrGv8SU41MSBd6vRL4QrrLZW
-          TwIs7uOFb9RoqY9AHo66N3nbeQJZud6I9x0e3qLvSlSc9IBsrZqiQU6NCJ6dEJKoE5V2hAHpIJwa
-          OtDSovEk3QYdqsIcCOH2iteJ0swAXJoRQ1WadhT9I+Ah7y+DuKnDeOrvig/SGu4kq3Q9Y49P4kNW
-          Jjo14iiMp/24D9ES70A2IqUSNllkykgsDOqM7rblst2+Q9+cU4aOoKSd+M+9BNsye6INol0JJ+YH
-          IJ+mY/DcJ6U21eks//CpXwk7rhrX5kfCt7QqiDq2pkffSaLCYGUP4svcsZ1OyiWHb/6eB/Gdl0h4
-          VY8IdmvFIXv6lts5v5cDdPd0EyB++5iZzliHgRvaBX/VasoSSFGb8eeA7YK6mp6Cdv3pPaqnYGhT
-          WJ1MVD7ljthxcsmawyq840WPEW+pL3ZKsyNcr5ZJjY96qSap50JsM5H9fc6VDvti/dNcqXrYfLP+
-          Kq9rSPAqIOY+kqpx2Lgq4BH25LrwMf5bNzockhwN9Y19q3Fz8pNfPwziwg8okS0F2IQ/NMG+5rHl
-          PpCOG5+6hfqt2M8fbAsJyMniT1q/k3Y2nA2vHNjDCWPBPq8mOOVg0qJa/I/xuolQtM6KQHbltn0+
-          y/sRq7WyJZk8wUxV1VXBOBUqNYN73Y6PW8zwcp6q5qRr8/Gkpvjtl+sBVdJnpqJeRziW3mUgvz9X
-          1N2vFwFJ9VungePsvfE74ytUlR0R7bNeo/HkR9JP//3Vz1MhnAbAH9GlTm/m2UjK3EZ2RPDPv/Mm
-          dNiEsPhfAbzlqhq1ap3DBgcmtbukzKZ9CNKvvxc/B+ZBDLsOlnlJ9FaN/vqXcEjLPFhNl/f8q69N
-          bNJzwJlTEncgoxK9pOJKtGvdeNPT3Hdwn+iFqt+tE4vnVVDC8/55//Vvu1MaJ6D1gUZ2aj5nYyT7
-          EuBe7qnBOjHuTGsYQCemT7d5aXr8tut19NYbk3hW1FRiESglPHbxZ9GzXTU8Be8qO1m2/uF7K6r7
-          0wV+ftD5dOS0Tr9/SzgN9oZE84eLh0IoOsDHF0fNfZS2fPuII1T7m4Ta3yhsJ70xfMisx5uqod5p
-          bBjfOTLskNBiLZ7jYctkATFnfpGgiB/ton85WO91heyzLkATqHm4EbsgCL7vXtG4eqspoHBSRPWp
-          o9p0nPcKfPljT7fOOchmkTsqcF6744DCsZrbRR/Jv3p3LMPQ+De0Jlr6OeACgtH400PuOs3pou+y
-          vn1kISz6ZVj4XcY+mjshlcJt+PGdn96HwQ/3xKEHzeN2p6b562fnT3hWy/xrQHTuErUFbHrz13ul
-          cN2gjgbTxZrZEwvu7//UqrO0YtgCgIW/B2OoTjHbJZkAnzzRyfmQPuepanNT7o/H5zBYDwOJcvcI
-          YKkPqkjWq+1x+4ogVYQwgB1+ZfMpxi9w4odJzFn2Fn+7teE3D3LX/czTCvBFWlx76oVjhWgZJRHY
-          3CYml0Vf/MXH+OieA05dW+20+KOAfaUm2nvXzsNqp32xoCQCLdRWQG2zt3wwsvJJDJMbvM5qPF1+
-          vPnNMh+2Wpd1OISnGm1JsNFaNLtpO8BkVfby/lL7YVWm/vJPLed7aMvFr8GOfoyJteBNpwr7O5wu
-          UkWV74jibtsURxAbuyW7a3Sf2TvdTCiH/Uh+/KzXrKaDD/MDelz6Y4xkXYLDtd4Sci3qmG0qfYJa
-          WUl019f2Xz/p7z5Iyb681v/8vePti8kv/1OozXfcbUeTqlX6jAfH/aRoiccwdM9PO+596qIfPzNb
-          wqPfPgEWPh8MSp61k3sqGWp2nE1O/KhVk6IdTLzsq8hWGyuvQ7dqBT8/RKsDHfGH/dxhqbtholLh
-          3TKdSQPEZn/+y+cYv/1cURxJGjE9Ual4K8UCCEJYkmN4qLy/+63nLQuXfUhazeTGInROvwYN9srW
-          48BzJrSxwgs1To6SzZ8A+ahNFY3klTnEP/4in/h8HlaSGmsjiRWGF77xl/+OiqLYsLwPvUEbzAw8
-          h0GxeZyoEmslYv5LUFDTpGkw/vTyrjrrsiZOMXFd2auGhc8hcPtskBf/5uuvXhJYN+M2fBb/j7+E
-          wxfdQFGpoawpGoJNoKNuO5uBrIXDPB7X+yu6Xd/qUh/ZPMoJS+AXL19S1Xn09o6Cxv3zSrT7iZu7
-          Y9bY6LcvMU4G9piALAY/fbIu9pI3pp/yijfHYjegxf+f41wMQJH6hhzx+VUx/0gF6KoPIea0mVq6
-          2u8vsOhX4mxDE3Va/HHh0Q2fgVs5UjuqmyEHuVU6avY6F7POrRL4zQPv2Tdx+9tn5mehCaAZnGz6
-          6WNBiMrlfBJ3i38C+8mRqf+hqsc9Az+F5robiMYVMaLC93OFZR4RX9w6szgZ4xHyw2wEL1Q3Go3U
-          JofF7w2k6M7N4zQoF7xdmQl1BuvkNd/75Qi5WlLi/PClHEoJT6+VTaxCUSpu0eP4rdTVgLbhSZvO
-          3yrFAcpLYjoNmhsDTzoU/LAlPz4x112honlVsSCzzSHuHTNrgDnji8TgsGo+aTSFw2wO1C6beJ6s
-          e5hiuWjJkl+CmsU/gwgdIPguerRf6h+yyBAJ4QO+muVrrMhpmxZEwVCjxX+RwSmKkurW3GvCcB5e
-          sComYVgbj0Ebs9FTgEiDTV316sRzZ+5qtOxDyX4Tf7y+z+QrejZTT+3onqDROT44bO0hIKezV8bU
-          mucS1g4ziXWw1trHGDdfULdJS9WA9G1fWTXDEgwu0bxGy5h9aCQoLpuaGLdJj0eTbCOAeZNT3Wna
-          qlPa2oX/ub+O+FO7uqJgXRpErVIjE973JIdGmFdBEc2HucmwdsWLPiexnEze9NjHR4zWtUf/CwAA
-          //+kfEnzurCT93uZq1MlmyQ5soNsQUDEGygiKLIHSNW896f4/ub4Pz1zpLQwJt39Wbrh9Tm/nRVH
-          1gIvZ1/7129bTzcthvL7NPhA3CBYvd/hAHe/ChvcUg2zLJ8NMY0ud+zGjZ3zykoXZJSaTqzuwtPp
-          8kAJeKbvFzlTeM23jB8X+F9/UwH/89//HxMF7H+eKOitUcFSJaiUpcQN4ZU7Fth6njSV30RSgJf7
-          exJb6CFYXodtgcGWHeaPf6roCtRuRvycxiRrLSnn9afUoPfwjIn9BF7N/+DGQHz60bnKA1llCH/n
-          wPBVfKJSXcvXXxxDNMlgIvptuwHuPs4B1Lk59KOzRiOyib8HdJguwHm4zsNmFhflxN3eI9H0fACb
-          eBUDSL2YJeqyrc6aMIADrVVAgrv84cwX69PCVtBOJDQNH9BzXmZoXp0fNkVRBjTqWw4KhXQhmXQa
-          1FH42Rn0pPSMrzH/y4ktLiG4RJk9s+dRHWbqORtsqd8RaZHMnF7kYwhf06cjhffUHKa9Tik4b88W
-          X5lr6LDLrangQfEJ8V+xQvkyNzI4l5+YhOwxiPgcJS2c8c/EJqiO6jbHlgjZ0R3JPZS6YT0IvwUF
-          097xOo7DQOTwJMIpDiz8aJKC0lDUDVR1NiGK1DeAXJrRgL0YvAl2nGvO2OUqomjjYyK9OIYu8KS6
-          MOAvPnZNpEb8mcgp9Emr+OKgLmDwT4GGMp8n2GGqEaxLuyXIHcwjVlLk0W3NQgYprV/4W4omQJvu
-          wsDBc7HPPS+UNmEcVShUFZO4udLk/CtdXdR1XUrCfX2MWWUGurT85P8mQPPVddEDNlUa+KfLW1YZ
-          6Zs94LEoRWw+w7SmF5kPkft4+jh8J+ywaA/LhkYTY5+ufBlRWW40ePf7F9FirhqYNgp8tMjbQszX
-          dwB/8QWbKgtI6gVdTj1ZCWDOBzyxxcKsJ072LHiJUptk+/USxDcGOr5fYuVGumF1CzlA6y08E7/L
-          bvm8huV4igb1SrxI+OUbYzYcYgNzxWbn3ig3XN8a8t6wJbngntQZK+GIith1idwekTMWxniAyWVy
-          8O2S3nP+Lx6LQ30kdkOfgH3qjwp28GuQZ3oh+Rb7XANFjy+wHcCjuvnz2kKOti25nefJIbcb5OA7
-          E2RybWZDZarR36CsTPqeX5+orT9xj2zpbZDrOHzpNnS2Bn/YKIhUnNu8Y9+si/Iq++FzKHU1nR78
-          Bh9PKyI3Xggp69hlAD09vGNlZsaBRm9sQU58ZViPeTPinfeWoJ+9von+DLDKhBr4wPlbfIghx+XA
-          y8eUQ6Ooz8TDZwfw1uHbwnx56Dj4yenAmafRAr/nhDGWdVnlpV3RDZ6P8V0x+nzPdwn6jzHGl/B4
-          zpnyCznUGdYNx9xPqlnda3t4Rt8b1jwmGnhghhu0N/OL1UNgOZzRqDbytjeYt1/XAk7owQOa7w6T
-          c9V9omWYHBs+GoHBtqXNlAbHSkGgjgVsTOJcL8OkWjDqNIqV8X3P+ZR3R/GkvWQfRnxVM0av+UgI
-          hWqGgofAdvOKDP6dd+E1AqBBRRIxuRCHOKNQ1svbPYhwgsN9zlVPVdf1LbkwMWE/Y5Lm9RaXSnYK
-          pW4kBmfLA+XTm48UKS3x7X6u6bpZdvj3fSxFz8nhIrfL4FlhB6yuvBQxx4+ioOLwPs5CnHsOpXc+
-          g7mc5zhIgRZx5/FpQPZoZ9g38quz6ORto6f+/vqf5xwC3s0kF3mZ3RFTLkSVLJJbwGN9d4iRqphO
-          fDxbUDjb8gw5TgNTcVEzULI/AftuGtPWc4QRZj5LiBMUIPr1IOvhLNf9vOST5DC/86GE3xfzwebt
-          KtBNuvEzCk3h5v9A9XKWG6eF6DsJCnmy0YUyChRceHmZLvHgVQJb+aYbGrC9EnnFi0rnkqnQ3/pe
-          7xXT+Z3JLYxN8sGe36v5+kjaGNoTjn2eQVW0rWcswd+wTTNXRhVY+bixYTUrR+LUpwtY2cwxgHFu
-          JRxqceQwvhFtsD60FomorkVMJ2YuPAbPJ9YfiqNuXq0sqPusFr4NL0dlv0YqQOPcS/Px9nk6ay6/
-          NkhCi2LXvjv1pvBmg4BTnsgNseHAnkZVgNIjuBLTye/DNFVlikTr6GMrmaWc5b4nCJf8HRDv8o0d
-          cibnVPzDJyPqf2BRUjVFwclbyCvK23wLH14PptfRxsYtjp1NZ0kKEYoLchXDdZhejy1A3iIZOP65
-          HVj6e2WB5cV4+MIs5rB0nZ8Cw/lURP+WisNqmg//4Y00XIaBPS5sjwSDNsTmfKaePHS3QZtnPT5/
-          9bGm82bFqNXOR2z8/ChnTenSIDGgnt9krDNs54t8gMdjafztl8PlrSBA3miSGbgMVnkLn0qk1nVB
-          NIendLUPawL7HPNYCwIA1mSSR9Tk5w/OHv5b5bviBSGdYxlHzbfIFz6PKrCfP7GOWp5vGg1dZEZS
-          8Rf/OcffsgIawH+Re0tSOk4vRkPp05BwZLSOs4//bSA/mLvjwqk1L5mOAjXxeCdOEjbOyrxaCRpF
-          GZB7x1jOauafGc1HTyPXxjpGw0sPfGRkOCLnMrk4i1qLIWQDfZ1huNHoW6owhomqZTis8g9YTtsp
-          hHf9GeLrUfacf/i74sImTgMkuilSF8PP1xCwe9EbSpUTFf/OdxaydBy2tu9DEKavfgZjeAfb0UcW
-          1E3jSgqz1CJOkboEWVuQkugPL+t74yJ+jVVyPzS3epXUO4ReSzPi3+o0p9mvLJAKbg9/jrK6XpbG
-          kNDxep2weW3diLofXIEDes9YuixlPQrKtUHFh76wYi+583e+gGWkgegaap1l3094zrSEGOSXq/xV
-          ummQ0CnE6t4tWNTgEiDe+CQk7cOBrvwphMhwzyx2NUPIactfRWie+YB4yvJx1qt3YuBwWGb/K3y/
-          zpLdUwjPb+GKtbF/q1QpygVa1+bs3y21V1fnSzP49OcNa0zsRTR0Ugi/F2rOx2Mogo1h9A15ru6Q
-          p6/BgdLB8QEmbIglDK8DuQXKiJYX5xFMsAH4gzeO0L22D+JUgwe277F+gO9nZn3qfm2w8FGXwpfd
-          A/+6ns50W2iwQNrLGtE+J64ex9/swv18sXSx72Dd8jr8xxesocSAzbLnDNe7O2CfpS+HNKsKob4y
-          CtHvTQXW/iM16BRWL1/gM6JOqcIfIHCqk78ClwxzRXgG8nX4nU+6nKubb2sMTPVAIZHRDg6x8FrB
-          yXB7fy4rU+X739rD13yGftke0ny5gs6H7D08zCO9ejnl2bsgmuBOSJRUfUQvkWqg8/ZqZ/ZJbvV2
-          4voY5FOV+M19gdEfXolh+uxJcDNP+Qa6PIDXIdx7f92Qd79vuUBRsR/zqvOG2pVtZ8E/vqymh20g
-          XXGDMDwV3AxNpOabwuMP2Pn5DLegUBe27RvAMspAMElB3Z5eQYjyrGmxvp46MBr+yYCQQhknS5IA
-          oj+tBrJHK8OXmZqAP97zBoq5AMlZ+W6UYgpsmFkSwq/VNeotCWkCvXbNSCjk6UALcHjAiLcvRN+U
-          w9BdbpoFw9ODw95s+SrXOGGKKGdJ/+J3YZbHeDpAmGPz9XXoFM1vAfrZ9YT/+Pv20a0H1DUizuB2
-          3OpPho8BMDIvItKHrff4U314wp/lHz/lk85q4dm0Wuxs+JSP/b23wBb+zthNg76ekXQyYBBtHnGu
-          XgLGH/csTp3MZ0QpRC+n9H5MofpuIDYG5IOtiscRFrlywG5rnwGz2KIgjuFlIfJMTbrzAQvpzI0n
-          bmt3lL4bFEIm3lai//HLz6OvgH2Kg51fO4BNJnkGrWCciKx+tmGBPz6F8MfrWHYYfVjAt/JhlOq/
-          f/VlHH+NDxcBv/7Ob1iiATV/eOozaTkDGmSsDwVjbUh8fC2APhsrgKqk7B3qIRo2AGEIQzJfyc5X
-          B3qDXQsbaC349lVVh+/ioEdvvgl8Bhzu0Rjznwd6MDAkV9sZ1VXWfxZoqmONXYF+6i0Czwrqj1og
-          poc/w0b1e49S/s78y59dnxRQzsWQSNq6qZP6OSqwmfUQy1Gxgm3nlwBqsYYvx4cysNd3yMD3a7kQ
-          WeoqZxEPpQU13URzZRozba3ZgrANvacvXI4fuorHjwA7lUIs/eWfQpNK3PXEv8+76zvkoHi8o/k4
-          zDjammaqoHtynsSoJBJNN6laYLGIOZaO8WtYcz7I4P2nrv52ieSB7YZt/Ku//mHne3P31WOw4z9J
-          Fk9RuWT2C2iizcbS+MKAjqFQQPStSmwclQtdxDtNoDrOA9Z49lNPYUUl5FeO4J+aKBho8moO8FP6
-          LVEu0XtYOg+E4H5KXSI/ntKwAtE8wC7/iUR/H+16vQ/vCnmZ1ZGovZ/zdeMOKbyRo0DsB1fke30+
-          AFqyPDa3fgYU0NOMVBQd5v4UqDXbasIMR/g44xf6zjnd6zOIVO+GXwIXgnnQriFCx4vkb1shgI29
-          SCUSv8MTn5+iBebvJ0jQ7XKzCO7lCqzq5yiBzrBvvvi7fwZaeHYMpo4MxD78HvnWXT6JuOthf/oK
-          ubMSWUhE6NgcPh+HhtKvwBmwTcpsZspFBhTJiAEV1J4kExwmX7bxdQCIMyl2sjyji7zKCRRYLvjH
-          pxYtbB7o2mTOPzzl0wPHIXeLBx80oKR0uur9Hz/EsqnbzuI5ywj7e3TF6he91fk733q4rZ6ILZt5
-          7xPDxwDuet0/mko5LKj0QnhE54Gox8JVSTCeJcjNy42kdKnoH97B09sJsJS1C9j1iQD++KX8Sle1
-          n6xM+dMb/jIssbrERIyBSGeAlZ9dRkvn0RDpHjCxI7h3dSmG+wb8G6MRuxY8ldnG2wFOjiZitxi3
-          en6JpxFy3ur85XO9yM8yRpMnP/xf5/J0+Z0PFRA9tiBeDfJo+TjJCIvEnrE+wF5dH0X2gGHP6dhN
-          CVUpPYw+fA+vmBjl5uTsl9UlwPaBTdxaTIb1qMgViD5+jQsLl+r2zs4tGOX56a/qZ6vXnV+Ldfwi
-          PmN0mK5mPo7wd2IcfCVPeeAXyX1Azb/Y2HttzfBPj8btuGDpZp6ibZNjAbLGNSA4yq2cGxnhAGnJ
-          8/gvfobi1O38z1xnlH2qYbOjlw3o12Lw5fze8vk7KgU6tJpD1Kev1+udRi28sybG57yQhgWlkQ3t
-          xVH/+N8eXyyHhvQU+mJDn3RC95cIju+iw/olwXQdTtyC2D60sf+9VPXCz0wJGbH1yFX4ftVtDDQL
-          uZ+L588w0evl7KXiv/PHoadFzLoyIlBD6Mz0cIdgkvv8cEIoKbDMLL9hYR50hijoA79IrnG+zuf+
-          AcLmbeHc0mm9xzuE5KaqMyONS7Tj9QdMcWj5XKpiQNmTGYLdb8Kv5xuCpW1OLprk00S0wPw5HZup
-          GvxOooIlnXtH00fsIJgq8fSHN5Q8snsKSGhTbL/Oa03ORM7gPc2l+a4Yds4vnzWEFkrN/f988gVp
-          mw0ZRwmwVV7diHknQQNT2wr9g1SwQ3t6TwV8dsWLWM3KRFNixQJKEsXCjm4rgH0tefjnv+z1JAVb
-          3Xoz3O45mI/vFQN+DJcCxdvrgpXHT3RWMdIhvOmtjq0HE9MfTlsJ3szm54+VZzt/fhJ8ps+I2Hrb
-          1Is/SB9kakjCWs8lDpXDVQDSLdP3etXS9ZRWAqJ5Fvmj3C/OCharh47vliSarA8lV+EjIFH4Ojg9
-          TdWw0bYQIQ3Cz3zoMj6a7cfCIAaLJfHnc0gpw20GBDcPYguebco94EWAnFAK2Gdjrp7RFEhw53d+
-          tf+fKSgOAfQXLyHxqIk5WYUsgC/3+/SP6DvvetpU/uotTjsP1lt3Fl3IzPVlFgc1oMznUVVI4PKa
-          yH4w0bH8Mgy4/+R1RukFR9vpgipYWkmCNckz6bw15xmmT00it/XRAzpdvfbf/rc7/lNVrxi0JaJB
-          /GeZOovNZCKEBnvCSvM5DeSPrybpG/nHFkbD2kyZAPzz0mCNfteBNtpXgVqfGn/rHebvcSgg0uqY
-          aEGQA0r0LoVfemvIHx5TW+kf4rk4NFg93CFdl1ZMYPn0G6L5EXS23xZnQHpbDcFHQa5Zrywe4FYl
-          8R8fpcuHZ3t4fD86onG/sqaVWISQcrY0q/jWDNMtZ0XoNMoLS6XROutBIAvAHysgkR/y0fKJhQxy
-          mEzECITR6W9F5cOsU87+aTkOYA0rIMHdryB/+z+qweV/601/HJ2Bgc9F+1fPVPtc1ex96EpgmeVA
-          tAImdMXOMwSN/bngeBV0h99uvS9mqjWRuHrO0VzYhfLP/0oDka3HTyyk0EHMk2R/92+b1YXcrR79
-          mxQkDrkXxgJ3PJsb+XmN1unuj1CpwZuc6WoOvz6ZONDMZojlx7Mc6C81JPgLFQ9rMafU29OSOCg1
-          Cjczu17asvNbAYJj1D65fZC6rszbhcB75sQMTguYeBdK8P1pqz0fCR1zy7PBJXMhxuwV7E/oJT7c
-          8we7Z1F1Fn6GJZBRsRL1EhtRq7zQ/lzJFc/cUbkAOsmf/m+9PjyLqspYciXBXL7nePcjAC0A9wDM
-          +pRIMT6aiLZWt8CdT/7Tk0PesS1gw4eLfbAc6j99DmVTUbCrO2ZNB/RjoHG9qdhY80YdxUNpo8c0
-          NVhJ3nrNG7sFcb7Su38bJ5wzp+7WwMUYJKIf5cmZP+a1hT2bXvATnwcw3YSlQfyaqNiO8JZPlgg/
-          II6YEoe9vqoLOToQlE+3IZqWPIb1TJUCupfrCwer70RUTHgFfr6a4IOXnUTtpLUHuPsfPhxP33wq
-          WMGHf/6l/NSflDrd5wHozUyw+8tnZyB3JQN/fowu+e5A//yY+dl/sZmYDKCDKTcIvqWI3DPhDtad
-          z4KPWt7J81bRYbShrqAkE1L82PUeNX6mAEztKBHrGaGIzhD3sFrgjCVBuNXb20IcSCRy9NfLquf0
-          cdY/cIHlY/fnPvlmf3oIu0XQyI6fNYsu3w2uhQbI3X+FKvHjOvin37Hus3Qy6oITx7xAWC+dIxjk
-          Y8Cg9zPnfK6MlL2/4mngfbgrM115KWdThYcQeeObmPdzDca9vv3zQzk5luq1BQUE98vHxy8u6+u5
-          ab4VWo+aiUMmVOnWjXwBRTqCWWDNHHDzrxOg+T24+K+/tirVm4N9rXJ/fm7O5vapgLsf64sd69JN
-          Ic8Czvhr4ly3K7DpQ2vATg7Nf378JiqXEN4OcjIzTDzlVJ49Cbz4m01McO/zhbmeP1BFlwP+q1c8
-          +4Q2LHji7f4tQ5fTKw3g3p+ZeVF8g+3VTSKfgnElBf87gwWeHBdI0jnbJ4gcsJRfhhO3sv3hc7Pd
-          nQmI+ACGTiiwvU1Phwbmwwc5dEyfDmI5jOerKol/9VLZ+zm0PAIfhpde8H/Cb67n67CGCBSPn7/P
-          iUW7foCwLUPgb8/xHBFbqQqY+Otjv580bHFpZ0CZjNnf/YVodWTjA9PlpxBzONBh3fkg3OZRw/Lp
-          VTuLOyMIy+ZozksKPvkiKNePeF2sEcdBACg9tTYHrv7HxW6uGDn/py/vrI6xyaAqX+3DKYHZ6fuY
-          hTNZwZ4/EhRNZfzrJ+SVYBgtXH6tS4q176O9XycAfPrSmRmAno/ByW1huETdfBLhEE1BwYWQWOhL
-          fH+wHP6vnxX65DyDcPXrkc+jEvV2/9nzO4y4g+qH4MglJ19Ih9KZWx0VkMXVEztS/otop2ocvMgy
-          mRVxHsC8hu3+BBhj4yjGvLrWTxKA/fd9ykmsuk59WP75x/MtztpoEY0wRu2A0MzseL/yLiPBuWxi
-          H4gHBmyZc/ABGxbufAnGRCXT47j8u/+rrB2wPP2XBqtX5pNdT1KCrgUDCRRe5OpaDpis17sQ//S0
-          GbkftcPONRS6l8hj70uaiAqyxUG1X87YyWV5WNak26CcCyE5d0yr8tNpK6DunUxfbJdnvQpry8BS
-          PwJiiFGy6xEQQFdaXJylUzdM4cG3//mP11dd5YuR1C4qLDPAfnMgaitJ3/50adlpPsWPlfYazXxY
-          D3NGlFEJwHfPd5QkkoVvws+v571/cqJGpWJ7SrSI2ftH4KzwA7HGmVGH3O+rP38Se7ufuB7mtgAu
-          KjJsXH/mwC/5rYH30j1i/xVXdPHEkfnXD7nf0Nnhg1GWhCwxcuLdqmjYNvkhwCXI3/gcADHaAl9u
-          YDYJ95nf/Y6t+i4V/KmQkFg51sOK7jcRNPjxxso7uQ58cXoHKG2p6k/WVIORnsMZ7n7dDJ0frhfv
-          dR//LxMF3H+eKDjM598+E/aLVnrVXFi3DfFPV1GoV6t2C7ikzkSU5vFzxtL/HGDyjX1iZdVV3ciE
-          eyi1sTkfx+cGtgGTAl7F6EPwp/JrtgXpBsDjmhOsgEDlL4xrwIiWR2J5v0ldzGL7IMStH3w2jhbg
-          snPHwUNoiz645+swnp91D6q76+HXhTuCrTtWPZDiY0A8bhOH8XkobShbwzILJOLy6ck4HGSDjPNR
-          JvfOlJy1DOb+OyMXnL7ByAQHG57YL8UYRQSsn+WmIG8AMXFu+Zky1XhOYERK4nM3/gwm2QA+aLD6
-          Jf4iSTVXmOpBfDwB9DtN48HyPb8b9Ko1lfgVtgdmuGktjHi3wiH0xnyDtTvDJxwMokWXSl3kCCkw
-          ejgCMeYJ0YXQUANltV7wwxFcSpnH1YUfFeok48gnX1Q5S5FwwISc2+aXd1KzBojLiwfGJqD7958+
-          dB9sQc6BGIO1CKmC3LYuiexUN8D6x+uM5rd5n0/0/avXvpcEmIXMA9vkNNPFzc8QhtZqEnOWuPzv
-          czSxaUa0j6IObD9EPbqrS4NDGmrqBPxVQO35sRJPufLDtp7vCVDyicXSXYf1EvK5D8fkCP2N3rDD
-          TFovQdZKbyQ53N8OY82gQvyhfOO7I74HGm5rjFg3a/YJgyH6JicnhWAYVXyGN41y8L0e0PHk+Vgp
-          3grg16kL4Ogfn0QFvViP7+mhAWmJG5Lo8jfnoqJK0OVkUl95NZdhcl6BhHLbfZHLSbSi2d0CDZjq
-          QSYemnJnCw9tCoE1CMSpL6rKVxZjwY8VSVhdepJTFS4P5JwtmSSnmubkZNU9jKgfzYdzHecMPKUG
-          UpUlwOHrbdQ8Tn4j/PyOFyJn0aXenKiEKOQZicjhU8hpWhYJ1Drmg5Ot3QBLv7qNNMXaSHLjO8Ct
-          pyBBymA8yYU/OgP9/ogBexBKOPluV8qepVYEtqGdSD6aZT4rjmfAeVkmkuhGEC0KPYjgqoRn7K86
-          jpZRZzYoBf2V+OeaiZbpyo7I2kaLOHEp0NV7RQwqX+aEz4q2qOsXbQWKm49J0k+aqJxYWwJ8J8/1
-          Xzxs9J5o8Ha+Zf5q/I7qhpbSRvJD+pEnDo711vuBgJyzLRO8ps7AEbsM0eHLPH1wKxlA//ZH//o8
-          lpOoyplLxy/AuknGPIWGTlkhnGNorw+M7/D2oYPh9BVA2cZhP4ubfLaM84x6EEg49X6eyhiSU8Ai
-          VjzsA8kBfC0cINzjFz+fMjssg74c0HmeNqLcG9thrp/lA98d02IPTDRavG6xUU9/jk9fJ49uzgf2
-          MFNfzMwVPBvNeRtAVOlV77+pFoK1Dq42XG7l1UfMFqnM+Tn0ohlHPXGs/jWsYPEryN+cnEgXSclX
-          Gq0PULehio1hQ/kYN04MVvGTEf8g8fm45z+M1TecD0bdAuompQHvZnGe2XMnA9bFoQhxNuTYn20f
-          TMlJzdB6ANeZ3a7JwLwfUgq07AV9tE1SxLULLMDTOjQzvM2NQ9BnNdB+jS+Nz9Ll8Vr3evdpZ9Tw
-          P2dSjaME3h3X+kgKdIeD1GvBpnxOJI/LlC57/UXZ5sszQhNQ11IeLDG0inyORaFVl5HpFAg+NxEr
-          HgvAstTxAU6IfIjp4soh/vVuwyK9WsSOvcfeoYkZWCpLTAqplvP19I4rxAhE9Y8m19LOn20Fks/L
-          8/mX3tT8Or0D+Fy6hJz3+rXnr4Hc0k9IASSHcvReGNCB3JvgP3wIsJDAoY4e2DysVr69bq4A7evm
-          EV2hbtTh2WNA7TYyPu/1ZyW/WkG3uzIT36hbOvC5EcJZzq25Dg9fwG/T10APHRTEG7carGN6kmAN
-          gxDngqLVzGiaJeLyxwOn8tjS9dopD+hcX/l88sXJmZ6VYItcTi2sc440jA+8GOhn3Dtyi+0mXwMf
-          L6DhDIl4JexrqrrrA50PRkasX2Q6q6aPAZy9VMC+U8jD/Lg3HzH8WE+f9+/HgeKEjBDPXU/u8vHr
-          LNUoJ4jHNsQKE1oqrcx3il4/LSSvfLxE2yeKA4Tm4u63TVTn2+sq2vBcZGecj6aUc4qja6huA5V4
-          LTwBGq99Cg1JOPkNrAvA/p3X8YR9v0ykZsdv8QPLqvhiXFUNpfWtipGF4winXTQDHjSSgsb5pszI
-          cfScP8YAwr3+4jvzkP7wpUWZz7jYvNu2ypTo7qL0uCQkt78jJS/nycHNbzqCL2sMtu30sCA4ziGO
-          VyWp+ZdzZdBW/TisjeYT0OYrf5CVX2WSradkWOdbr/3lK47m3hzYC+hCiHK9Ill3CKM9XmLUOGry
-          Vw9yWgyRgN6JoxPfK7qIKYpeg48p/BCnd9WB1w6mBJ2FW4nzyu8O+9DOClpISUmwrhugo5yJ8PqG
-          Gbm6s5DTqIIumND08bnn1tSb1u4zx9to4Qf9ILD82mBGP0WbcbT0JKJ3opV/18RRezfijq8gQHnh
-          zlhlv5az0BKEsOQKZq83x2GdhE1ERvx5YHtlhmHkkTXCWvcE8vo863o5tJaPHmykEv9n19Eg4chC
-          80V7kMi1jnR15GMBDl/u6R84qXXIA9Yb8kVkE1dFVk0knFt/54Hjfp+gqT6jBXe+5gv1Ezmk080E
-          gVi64OCuPwb6MbIC+nl4Is50ER0i/sINOleLEEfvY5U2rFIivucOxEh5D7ATX1oIDvcL0b+vU01P
-          ZiD84dPffuTrtbP/7Re5e2KXb0yrhXAwf9m8jCtxlsvtkoD+GGLi2KW884UThME7NYnJcA7YjF9Y
-          QjXbZp+p3Avdyk8Zg8cCLXKtXyvYLkdgo9ZbR/LKftXAXWT9AL1j4Oz8ADgziIsSBqf0S24rV+Zb
-          PAUl/NtPPfhU+cYlMAUwml4+H+M1X7zBFsHvsig4ZgRpoN8mTsArIvmMCliBzVZbCX7cH/iLd2cz
-          OWtGeLv6WH5Ep5yuPjrAl5MFPnx2FSDujy4Q+Zs1n6qFHfb4bmDBz5PPZK4y0OfY9DC5ooacp3ep
-          LpLuM1CN9BNxPYGpt6cp2f/24/xr1GhNzLWHs5cJxHzYbL1lGSpFdjFNLGvP90CH9S6C0thq4loI
-          AMrm9gN+edkkwcAl6uK5twbW7YcQecfDbel7CwBBfv7FY02Hdwph0sVP8jheXWcFpfeBIFYuMxvO
-          Lzq+H1YqGmPgE6nntpw+xopB9nXx5uKkF8OEvOoADivrEPXsHaMBxYuF2ECz50UYaTS93p0Pe/p1
-          yNnWXlFvmnGB1IY+/JN/FJwVv269GKxkwRhzqUP++FZBxwI/LvQ8sD0kCagQvvvrUSSUvlvRh08L
-          Nv/41eLn+gKEwt+IlxspndPsNJ7ouQ+IXPhNTXqlY+Dlavyw3mdbRJcCaPA3FSrxYDvm06mUF9S7
-          2Jy3e2Orf/wa4Hnofb6TtmHBaRv8Ox9j53+rIIcteLbHv/3OwSZ3QgaXnA2I8n2oEZuz4QKryzsl
-          TvQ9gn4RBw3+1QPLGN162vUVlFMpJfngmjnPdUIPd76+4+s7J20ttvDiaRZWwO0QTZtRSDBczg7W
-          8sslZ/Lw+oHtuViJOQC15i0x/MCnEDvkUvp+RPnz1qDTAl1yr5ZrzaB7KMHP67DNzIxUwLwO9QGa
-          wSUnOu66fGObUynOk36eN50d65VYx1H0oyglZoZTumUZW8GDYbkkAD8t5/I2PYCyWGVskNbcO+qc
-          CC2cROQcHF9go+enBStBKMgzNHSwFP7bRju/nQ+bfXDW7BoIKLCvI77q9Aa4mpkMqPw0n8hedazX
-          4uOPUHLhSuKnO6tLnCgZ1JRngeV7KtBtLJ49NJhgIpFpjfXsKXcDFan0wJbz06Nl+fxGMTCklYQq
-          qwF6SIwZ7vi869UWLG4uH2DDaRLOUusyzCFmbAiGWZ1ZPiA50Ye2gTf95mCnSPp8+9RXF1r8lRBl
-          11N/8QN2ffeX/9ESpT9R/BV6jpW4MfLtZgsbPB/uBnHLzo/oVR5SeBPmoy+WKkvJ5g0u4F7rDftH
-          oA3MEsoh1CaG4j1+csZ9xCO4PW8dtsSzA/74GJgvxoMo3hJQRnz6BdTPJeufuEerrk+2aEBe+DMx
-          P/qJ0lDTNYDn5ICtnd8MH/7jIi8SOf8IfdlhB2vRoN3onL/az85ZUXHoob56XyLfowlMf/lGv0lI
-          vLts5KTILP9ffN65d6RO/HrvAX32r/2h7C6nP1VYYHQrLzgp+aez9uDRIKflwj/+HC3fDIVQ5SwN
-          X8pznQ86eJfgi+Mb1q+fxpmMWupRvQQVOS9vZ/in3y5X7YdDe9H/9K2IDoXdE6tMjd1P6Hu4GSfG
-          hxfaDfTpZApM3awiNuo1wD2sUERZwg3Y/asfU4YbaJRfGztNcq/X5mXa8CQ3P//0GGp1FLzzCHRB
-          zIjp31/1+txWA873e7PjGY6WS3fcQKefl5k558HA8d/qgy6yluI//cbt+YG+P8Xe8dypp3B/5kVg
-          hS/WlG+uLg8sGFAUErzn7xptUdHHgKM2JpJqy/kqHgQB7PUVq96hcubkYfXw1snDvPsZdA2GNgbX
-          4lcRLzS+dGLqLfv7v9jX2nxYm/3NaXzBBViXhKM6pmrdQDvTLR81cHPI9I43mI4Gxa4mvOsuhDyH
-          yope8I4XdKmfxeFfPdWHLFF3PXaArlbG2BWuJd3eTBeC47ceiFfJ94iNwFdDEV/7MzOhUv2r3zAT
-          ed6vf+TnUF8PfLDjIfbkwKJMHj4/sHaFCcubRtT1tTEiShk9JFgDN8Bf9BcHzjPZZuFRQTA/b2fj
-          z4/Cut/pNXc29ArAo8HMp25I1ImhIAH87Zz763wWnfLxWiWE/X7CVlR1Ay0yy4Vd6WfYZkk5kN2v
-          gIJRS/5UQrvmiyEXICNM6oxiuRoofxYbKLWJ6Z/uSzUsJbr4f/XaX25SB1brfRoh5m5P7FxyJV93
-          fwq1Apf4p19TR1NuNCV80Xz8q9c5eY5zD4NEnbGGaDtsLLcV6DJud3z+DFs0k4ERxQ+/d3DM4uUs
-          e74j274z+HxflJp5yt4IRaEHRC9fLF3utEkRjMiLaMUFUpq77QH88UenSOyInc23BVmJVP46aUNO
-          zf7SQrmUeN/b+TuzRLGC7q65+N9bJ9T//AU1KxQfxpRVx/PMx7DtXvO8MkdST0f0XeDYx5S4hpLn
-          7evIzbwrVTrGGuDBGoFJg/dX8SO+X3jOcjGDFD68T4RN6W5RKpG7C424eWB/tLn/xc/48PkSnKzp
-          sPXFnYMXZhOw3R8Sh939L2iCWMNZo6KcZkicYZ9iFhsJ3zgLd2xcUAligXVt7pxOu04+nEK19jnD
-          YJyJ3z4pomwHsLOZP/DPz/rTq5JjL8OGVL8VzeNxIBKq3g6/zUcO9hTwvkc34kx/96uZcPUPyf2S
-          LzPLfyBXrz/sfXk1ZzzgMIB8rDf5lw+7fwJ3fMG2730HKnSwEBWSp9gs9IrOFv8K4F3ddkNWLwbS
-          RqcZfl5wI7n3FXLiJvs7Bk5QwEEFfGehr7KC7vkMiLKFZbQ2Hn3885fuV1EY2k64FlB4Cwn+82v+
-          6dlNCTx85Z1PPkbmwxXP9tz85Ssgt+O5hTjpXHJ5+6HT1zRMgHp4+kS+pF29QLvm0K73scYrB3Vp
-          0/wD18iLsRIsj4izErv/Wx/JzuEIZn+LBDEEDMb2y3qABadl+I//mur9A6aqzEL4OEXhPuH3BZvd
-          4BTmX/GNcbCuDn2JtvZPP2GLG8GK0PsA86egEPtiGc56N6r5z/+dT58hzDf++m6hcktFIh/YtzNP
-          +WIj8RwlePcv69YDDgc15VUQG3vnnC7PNgGHC+cTM7abaB4O+QxRblZ/+VfTW3BNxYiwOZG+nD8w
-          9jQ2sH+J/fyMjHO97vEDdz2BJSkG0YLvVAFHRoiIvgY/SjigJehg2O7MkfYXDW/bF2EnlWdczLpK
-          5/WUJjAdNUpSf8zpalEvg6mUbsTYv0+P0duHfO70+Dw3PqDJewpgGmgGKTrbcWamdYPT+aBl5Fpl
-          tfOH3/DdO/Eef3XERmKYgD9950TfF1h5ZcrEQTz6WMmaCWzHlZvh8V0/iJPkhsq9dVb58/f//EpK
-          rPwX/tMbcnlW8y08lClEHP34GxsOYLg/BQbuftR8mPRSpUNVj/B0DW1i7P7B5oWZxHcSw2Jj59s0
-          Vg67nh9Dn4IvHsbuZnFwUZnNXxqAIsqOhQ0342thM8ZrtL7fkwR2f5ZIxx8dZtMivhguTx6rf/yN
-          57gAcLt3q75OE505vZyBld9k7E3br97jBcJYreGOFzzt/vD79cpeREfGW51ttVWAzVKbWL90U9cy
-          5DjgXOxyBrw5DPTTpBVcHCYiybEgdH1osoQAI/tE3/Nn14v7xH2JsURGSWUzbqoggdbRF16pAnY9
-          ZiGk6hp2dPJSt/Nn4dDtono7Ppwpf7KG/h+/MF/2ZeA2+yeCPz3vGYE7LH94vOsTbL8Ptbr7dwp8
-          OWkwr9zxB7a3jiQYF0KMb1EsRMMDceVfP4V4ZrBE3dFrFsh+tDtOm6iO1tN6fUBGv+K5dDAYZjfm
-          KzCP23M+CpGrLnp5ruBQHZZ/emsxDksP2VHjcbLWH2csvdCFLzOTZzbpGLBewDtAxF0z7ITlqhI2
-          twvIF0xAnjt+r3t/BqC+sInnGZWzTcPJgGyc6zt//dQjAOgDd/7xj59s/Ldv4PPRnnGCoV7v/qUE
-          ZatbZjb/ePUysmABOhS7Wdj9rNWieopG7vb250/8jNasGRtIO3Qi1/76i9qd7wLJOlRYd2ch6kTJ
-          mYG2lreZJ5Kotlv0YqDdmNzOr690++MPq+k3PkTZAaziYREgPUO490uSaNQX+wErnU38dTmW0T/8
-          +pgw/uOb0XAyAxG50FOJr7WgnpoXtv+tR4kGL+cjBgtgkRw82xzRcq6VUAHvteXg5A4fw9J2JASq
-          o2fYOBSuOtykRwH56OxiVbLeOY03ZP/xQR++u01dD9OHATv+zLs/7Wy6Vafw916OJH6atTOOJi7B
-          uN1Cfw1TL2IwwL14zx8p/qtPdNknPP/6Z5T5fvKFLowr7n4xSY/aZ9jkbknB47q/Y2rvZ0wnOSmg
-          NA7KXp8vzvKUvVmcmQViT5xwvR3eaBFNkGhYX4KPOgnhnIhxI1Kfv2LW2Z5pzgFRiDE23/Y9p5kH
-          ZnCwrXw+GtxX3QQmXeDuR8/cPD3pmFdJBV+3qZ8PiyQNDFEjDbrtu8SumNFo55/wH3+X734FFvsR
-          CICUdohtHDL5pgl9ClexyYj6Zux6u9uwgcIkGPj6x2eMg9CKamSesHlaP8Oq6O0DKpLMYOXWKDk/
-          vNMD3P0jfxF+Vr4RuwyQ/vZb/7TrmWVmjx/oPqqNqAsRo9brBAtCc1xnwCvlMC3mCKESZQORHAzq
-          xT+dDOgRUBPZTXXA2kHHgD1/iOsKC12t9zqiP/8Es1FZry2sS6gcjZH89RPm0ZMPMI1zGUtY/9ZU
-          kj4HOPaKiZ3mMtfjJp0yIMXtBycJ36hTCI8M7B4lh+3Pwa+5aVg1NPnXIzFs+exQuWwYWKqmRMwo
-          TvN14lsLfkDLYP3SXZxt8B8z5G1OxvIlPdd//hcsRUqwyvbUGW7BM4UX5uES8y+fdr6FNvYwEYNt
-          i3w77hMraH7cieWPAKzu8SKAs6g5/kl4i3RW07JCb8TxWKvNflhns7NAfvpof/2kmgtea4j++qd7
-          P1pdVOVrQfdTX4ifxUa+uTFfwgxbAr79+Q1//ehyLq5EfZ+u9QjtmkFXJTjjpNw2Z9z9DHHh3yH2
-          Iice5vILDJgRh52n3Q8eJQp2/5iX//wcZ2JYQ4OhS5N5Cpy0/uv/wuTcu8Q0i6M6P9t4+etv+Uc2
-          KodFVSYL7H7fXDfMnK8uzkQYtvboH23wirbwbTJwzYQzPvui5zDL5zfDvb9D9njO16tYWHDJ+QBL
-          qeRR9qm5DCSCtRBNLt3dz3+E0FLjmkTwS2tyMLzt33qrfHgM9CRUPTCEb0vMEZg1OapliAoy+iSo
-          uFrdTnE4gv/DRAH/nycK2E09+/xDFFX6BlwKW+Gw+mw39XSz1sKAlo8H4izJ4hDx8jZQm9FoPtKz
-          QmntVhWaWjYgNn4RQF8vzYCW5CYkurwbdfEukQJPQbL6wqk412zZlRs8SDiYTyP1o2U5UIg2s5pm
-          ARyCge9B6cMfbWbs6WYIFg+FGmSiF4eVIRwB3Zq2h1o0BUQaqyanVmEzUOeiB5Fvq1dTzjJtaKp8
-          hp279wU9iQML8q+vQW6HZoo2Dsc2kkzp6bOO3DnU6MIQVZ/3g5zbwKq52oEiPCVJjI3jvDrUId8Z
-          PBKzJz5WAKXL5RiK5Xd5ENO8WAPN35UBUy3riG6vFGzZIruoD94ulnXnnHPK7bxBcRJm8rIzpeZq
-          vPnwMGcTkV7uN1qcdLIAPnodznWmyVcytDMsDuqJKF11AFNjig801axC7mN9d9Zkf8tZzojtTOSf
-          WfPvYeOQgiybPCVGyfm//fEPLEMs160HrgZriNwEfH3RPba07xR3g+129LEz0F7dOkPRkL5GPta3
-          Uge8cboy0Hh3KvZewTxQazYyxCW+jQ2lUR22/NYP1J6gQaTfUcp5f0xF8AKZPZ++AUfXw+VrQ79o
-          TfKwcTJwvvoukcxvJnGX+etwpL0r6J6hGRuhywLaFHqKlOgACAYPGm3T8GVg4zO2z1Z2FDFbXR6Q
-          Ai0W5++Syxcn/jEguGc+MR9i5oyrORWivoX1fLh+v/v6hgJ63MObGdZSBp4ZHsvfeZL45p7oTMok
-          het8fRKDd5yI3IODBD3HaIjHdbO6nsxHdoLai8Hqi2ed5caeA+jJmk2iXy/TLfbEEGSSpZNrj/yc
-          fxPBRszWd9hkqKZysX5sobaAOzG19V2vybp+kFsJI3aV8ej0UB8V6MmGjYthlsDppL4MKATwQnLr
-          VziMmas2Oh0qn0TxhzrUpYMCJS5FuLh+dZV53tIQ0kw7kWKYS7oocC2QiElAFJD26sJfDu7e0hiw
-          7B/CYctSVYGNxf2InpoxXdOmstBR+ul+cz59I6pp3xiJrVDiwIim3enzG7iR8UeCHxlURmmFD4R1
-          /8PSw6WAfu/3AJLpR/FuUVJ+fhoz0uLxS661KEf8xTumMH/OCjHgfBxojzsbDg/BxMp2ve/vy3Eg
-          HKun4AvC4eWwAT+m4HRJDOInXxpRD8YPiBzmjUOrGfLtO44VZFeVwXq5T5A8+FuGBqr/P9KuZFtZ
-          Xgs+EAPpJGFIJ9IHARFngoigCAIJkKf/F55veGd36NLDIcluqmonOxnSot6mgrSveTkNvvPf/C5C
-          1LAy+b4QuhuCPCyfcJDgwBoT3u12V5cPdqMFHiBzAhqetIbbyWMAP+v7irSb94yXpFecn39iTmUy
-          Ous9ySDE34z4peK6q+TxDpyUPkU2a87NLFxnHm7xAe+yqhiWyApreeX2ItGiKzvMN/Plwann30h9
-          JAf3u9jWCNX4bQaCoHMAi5amyIYgnogxnTlKGOA4UK71NaByTPN1sS0MblFSoHKbn7HTZ02+Z6VN
-          9NRwgfBWYxPepa4KqM0EjbD5K/gYXoDFfcDG9B7dUzj3DEFX+2XSZaiDFb50gyD/3dTNANcMw1lZ
-          KqKKYQa4WrvzYE/rOzFbTqT4Pb5q+ahzN6TzyuKSh+CEgBP0CzqCNm+Wg6CvUDlqd6S8d3eXs1/f
-          ANJ2q5CT1AXLpyIFUOCt8sWQ2PFahLMID9dDTH7xhJ53sSnHAybI2Lmqu37e0U2WlLFFBzr3MZf4
-          XifZxtsgnrTozfBASi3LCjuTixC7Ogt2RQsZz0vJ9Smzw5y92QrmF/xA+v0oAuJeSAXPdjRtXVqD
-          ZpYt05TKnRogtRKu+XLbPRz4i0/Iuj1y8r5eI3hnvzLyz/d3Liiu8pWT0XwR9+az+lgWXgaF/OWj
-          e+DoOpf4Rif70t1EYVHY+ozSjaEW6EIMtL8MdIun4NpCDl3100rpt/lqMIpthaSv90fvDtEVg/ap
-          1OTy8X0wj6f7C+rlxQvGMOLyodOkm6TC9RxULYpi+pEUEdrTHCOj3LUDVQY0w6crj+SG3WVY1P1z
-          lbf4i6LlrdBlmz+o8utKnACN8Xw9z1BWSHDB3Wk3gtVwpRRORVegO98HLrtTAIQXQ2OnTtvLgIoj
-          a8g+c98Rfac/c47snq3s2ufPX7zu9du1lKLMDtAJKGd3Pn3nSlY8ZUD53L5z4f6Na7nIjCQIL1dx
-          mJXCZmEQDSqytP0dzKvwZGWDFyFy3qK63Yr0fsnj4UTIwT7l+kspVF42w1Illronw8LssQiPiK9Q
-          zFZZzPVcxspLd5ZQGLZPl6KrycJoPtvkvit6d/FbtZLLkFvQKTOYnG7vI+7Qm5DidiiH+Sqvo/wy
-          TAOZ7GDkPLT7BB6EcCHh2W4p36eHDlJNTEj25BLAP5zzC0rOpBH197yTujjy04cpuYZC766ZkKcQ
-          qEZOyqxb3dnfFyvk8yALwLNK40V9V6K8VImFkoOpuIIxEhO+q25C16tzbZYbzURZka48UfAq6isP
-          Yk/O5OVNjC1/fb2SreGlRS6G+5MDVjN7mnLTgwcKvNXR10GSXnAopCNJzpDPX+7WFXSLn+Qg6pKO
-          G4FCeee2PEEDrzS8aVQetOu1Jgp/kOK+kp9fecv/mA32fbwI368ib5uw0JlWib6KCsqkS+V3eLrI
-          Y/5prHcgb/gHqclSuPQrfCWIClUmXs55w3xi9jeoJF5L3Nu9jxev9jwoG0KPZdY+uvz9PClwp75u
-          yA0unrsGpHrJrB7aqLie38NcP3sFejYpcL2c3YGtOmpB5hhYwT7+vHPCe14BEQjmX7xtliASCnhE
-          bIUQQ87NnL1hBXVYffEMsyCe2exUw69IzE2hi8EaxJP5+38ks4uPy6FVDKFb1wLxHLJ1+f1os9xR
-          2JBChabL5nz3AqqWGaQMNH9YmTQ0oBncU+TtEjtfhwBUILMzAcVu+clxchC+8ASYGzrQinXnYrha
-          MG7HJ8q38ZLnh6tlfa0CFH9u3iC83o8acqI2ImOzD76/ORBu9kkUM/Zd4UGOXzgeYoIL9yHpvcqN
-          JVTm/IbZt/aO1+igsDKZ3tthK1lxcW7xEXjM2RWhIeVd/NzOgHT52hFlih7xUvlTBDTOlf/GO18O
-          ngiAcjSJk9RKLqwl1oDrJicSyuJRX1cf8pA7s2eiuPQNln1+D4FREpVYIrPTV+t+92DVL2fyWHrH
-          5UPxpkChfC84c8cMLJy6m4H9pD7Z/N3lf7+X0BSSwNwpw0zDVIKPk98SpFwMfT5SO4HPp6Njyika
-          nfXDl4HNC5wwtFGXv/z+FMhUjt4Bt29WQNE14OE2HnQ8fwRA8NPT4FeYFpS+OWcQPpIiQeuTAdxX
-          chvPO8i08L7vXVS+hzpeod2n4Mcnjslx5/7sA9rWWw5+9koerutBVvgKKMgzHM9yaI0Qzd6O3NHN
-          bGaU/g7/tR9iOvu4WaVu2erOc0vUSbIH/qm7BSTfFmFmPrVgEZivKA3iscXiNUIuddIZQ517rkH/
-          UuJY8F9lCCUs6eT4mkyw9IeiAKNUHDDUxoc+83vVgQ8p4/7yH50X0fzhe4L0Y5Wv1j4c5ZB7a8id
-          20M86tOI97ciFZG/82IwuyNWIHiFR6QnrBWzJNi1IKcgJqaQV8P8w6On3S4m9u0bxZwtDgZ0RrcM
-          aOvYLuuyXQdujvPECeaCgbLHMJGPk9YH2/joUn8vLTCFb/vjS83C2esNKMuNEERtDbBfdgig4zAF
-          0q4jT+eL7QRQu8sXknLgA1Y+PhXy4+6NJDeSaVj2j1sLV68G5Pi5N+6qXdRZ1mII0EM93l3hs+1g
-          e31PCYYqbPX5fglDGV2VK3lk80JXLfYKSOXwHeyk4q3T2Pm20G2fM8kGkcYTyKxSFuCCCMoqOIzd
-          TvrCDY+Qk+lemu6lhRs/U/dB38/PfA3DMAReI6oohjoBpHlVGF4+QRUwPZYHHFr3L1SvbU9sdZHc
-          hVugtce+diKG6KJhmtaV/dk3CpbTWV+pvHgQN36NPPmV51RruC8g/s0kxv3h02VyFgNOajcg3WZw
-          M569iwK/7Ogjv5iPgyA9x04UxDVF6izzw/qIQg9yCaegYCiHBt+XqpJmh0LilIckpmIiS3DD38jM
-          kalP/VR6+7v0rYgtmJrOJ3eOBz88URrolRP/lYayY8B9AEX1k8/I1F+wnJUXUgdnculJpitMW37A
-          4pNjAZkX0QBEzRWicAdPx95tF8Dq/nqSuykoYNmp/QscbmZHkCX1YKWnrZOAr5zQyRhKMKIUOXKr
-          s1ekLr5L+ZyVU6CoioTQaTLomjyBIxlQeqCgY4yGh9eXJP/szVitEHActy/g5n/ouukNU7osrfzg
-          sg+xmncNqMH2GM7TrguYRd2BVfscUjArtPJpcXQG9pt9MBxPIU+0oa7dBXSeBuu2MlHWHMR4AYoT
-          gc3e8VIa6rBsd+nBRxQAomu5oC9BtCuljc+irLMsur6typBO5XoKoPuQ3FWCa/XDlwh1PNIFvNcC
-          2Z2qDrmn29WdJl60gCFIJyxveJ3MKsXg1UY52uyz+VvfLb8FQLoGDc2fXwOEUZ4G8HamdFMjTBjH
-          FTfBZ8e7xFTDGv7yWcKKU45XW6ik7NmuyCet6rKv+sICe1pjn3nmk46vyQFD/PVuyN71R0ovogKh
-          S8k+WCuk6OvzqKwyivQncl6ZFS9794rhqtUPPIQ3h3KFeNAg09wgQsrlpS/1rVCkjQ/88EHDLdUp
-          lMP5xf7Z82oQZgUitAlmvoHlCpBJIVxU9/y3vgJ7yUWYS6KLwmnt6Vg/n9ofH3AmAPNZvzY81Epm
-          JJoa9y6dJ2kFatkK6MjwdbMq+FEAd2xicgxeDpiGr32DW7wgih4q+qqsxQylA2aRrg1jTi8GCmEx
-          RwgzqG7jKWu/FpwJEyHNSnl9Ro4lQiShPKi4g+cuN5K3MK5GhiC6z8H6cM4t/OQZDXbb56WS+w4e
-          G80PzmJM4u9Hzlq4vjCPlICPwPeZ2x0seeWIbuXZHta+VCK4vJMUeXHUgjUT4lSmqrjDzPrMhnnD
-          I9BpbYAOQB+GL9PGnfwcnAceb+9eX80XSeCGD4lrvDTKv63OkL1ORMQ/+6G+4VFJ3k2Lj6eX3lPK
-          INeASSDmeByGZJjvlyyUrsEqIVNLfSD4XdzKmz9s+S+I+ay5lSACiYTusBCaVVSOGRTEOSVaVsTx
-          9IE1hDeuslCwZyVANvuRL4EeIitCmiucptmTpRFcienaFhirUSzAR347xPhEgr60dF9B+BmqYHkf
-          g6arRH6EwDM8EgRRTaef3sOQm4a050MfBCZWMmhw+w/yN3z6x4caa7yjfGnpgJ+7wNgfQtNDOozf
-          7sj5jgnb1X8h9XxqG+y0ZieHu0+JlBLOLjmdWAywqruBDMcR0HmZTfh+frXN3oqcO1CzhMen2mx6
-          5Qq+m/3CtOIsvFOOGCyvwTL+vneTb+WuTnSYgeuk8C8+UOk5fgExr8FPfxwWXYsjWMWvA3KXljZz
-          nA/zv/U7S81AfvqqeSvveH+TjsO80rGC9mee0WWz523+b3Bvn0VixY93s/FrQ+7g7kuM4MqAYcND
-          AOIuQ+dBhv/w8V6sB6QErdxMfmR14B2UZ+R9BzNnv96zkC+OciHFLrHjjgRCC5iDX274Qom598BI
-          kF8zBkUrZ8R0O2svn604C9brmFKu6OwSfpjXg1yOp67pgH7devU6eQA6mg2YYU78dqdfGkiy6uar
-          yNomXDGJiJJenXzc9FH4y++a3ty2fDIb8n58voJ5rM2c++k7iq8oyJKw72K+OWtA55tjwIouGbb4
-          HMLzp3RxueXjHoQ+D91DipE+3d18WbzsBjlRGUkkdhNYPuL7C4a9CYj6vuh/+jZ8Vg5GtpU08dzp
-          ogJ/fL2gpZr/4Zfte+K5BMVLehdHePF1EOxtA8a0YccI/PDwpsc1MzvebpB78AuGMjs3RENwlqTD
-          yKLT66jkXFlIGrCtj0yUsFV13rkpN/kyeduO1XGkmAd5AC0DbfIbnnTS46X409sUvGbuus0nmLvd
-          jfz4gxDX3QxHtqj/3p8VTrwHY+up4z3X9vmUnGwGionsoNuD03JuSJUavnSTEN1Tp3huXceBR5uV
-          UYotoI/dsF+hlONLEElWO1Dp1Bs/vRdpgyYMHxEKHcwjfMJzsl+aReXGAnYMk5AgOjMuMYwpEaFx
-          Z5Gpt6d4EQJcwfNYXJHzOAtgKUI3gvHxiXFslGfKbvkf+PdQRsoxzuOln1IP0uvxRdwhfgxbPaAD
-          Qda9iLGIoCHzaReCiE9HZDJYA13xiXlw4E8F8uHo0SWaoxBYrnck5YaHVn6XVdJPbz+Wt4jO/rNs
-          heOg78j295R/dx8P1N75STSHl915ex58fwebBLm11+fvPIlw08cx2PRbsvIvD7atNk+Lhyu6Fq/9
-          CvfnZIcsMWjdNbeYEJ6k3SHgIiPV8TRMPOTlG0EKMi2dz8DWY8uTVYJuoTPwy+Q70pCLKzm+7qeB
-          GsVQw3v4gMGq09BdM3Rq5XLWXng7JJhTqBUJdO3LBx3f01mn2iCGID9PHNFyjzTTMicshGZRImvD
-          L9MyFzzgfz2nS3KKsfW81VDN3knA62mh07iuVvn8vj/JkW0wIO0ahLBd0Qsdk0sSz8rsjPAqukJA
-          75fAXRn+4UF81HAgk3Sgy4EGJdzwIrnCno8xE91meLos1+Dzui/D6Fp9C48OHyGHx0I+yd6iAdyj
-          A0GhP+rLkqXfn56BlPbxacZvRjDo92n7h5/6yMqqP3s4Co2S81t8h+r11RP7JAgxnVWA4c8+Yrnp
-          mvmYuxZofZZDx+D1peMPL0DnfkE/vZayeqzIU/Et/vAI3fR9OF6iawBudzufT8yS/fID2Q5kuBOd
-          LAXOk9wR42jCZo1UvYYOgl9UxlxKtxNLNzn+TreAj4pBH7vLLMm6UBXEJjGXU8Rp23qdP3iGnT4I
-          ML0HcF83FtIqpLgbPh2B1WUhrh3+rtMH9Rj4vVQGymdT2u7wXgK5z10L0/sF63TMopf0YdoH0f1T
-          CNbg84ZQfyRP8tO/cJoOI7AH2gc7EXoxn1dsCR8X3970o68+7+Y5kbt87ohl9OPWUycQQUqElSj8
-          +wXwczh4v/kgdu33YAHQk8B76EqkFi8LrGGYhdLq12HAHFDUUOn5+v7wY0A9dcrppdJmuMNrjdT6
-          srprinsJXFmHCX76PLlH5xR+6Shu+lTVzIplQXiTqhSvX83O591IFbj2k4H76/flLuGn4n/1Rlzn
-          xYvSt5obsN8lNskPr8Xd6ks3YB/aK3JGELgLAWMIAQ4bYrzWUJ99/47Bhs9+emU8H49OBG+X9It+
-          8RQzzJUHrc5fA+a4C8Dyq9fcaf/Tv7y4DouyglIL+EDsu2dDY41AeGMrSk6XxR+aX3x+DtaDaI75
-          cpdf/WzT18hp/Lb6PMrFFzIk00i2L/th1UDggfOncIm38ddN71fkgxAtP723IaKxpmDlgIjbUruD
-          rd5mSbD/Okg/s2lD39PNgW+7OKIjNM34p0eCzM9nZEmlAKbjAJMfPkcqlS1dcL7LSxYAlYheCi1Y
-          SIMNOOanCP3qdfORGApkXKQiMxqVHOcsl4L5uGbI++aLvtiwy+C4K2J02vDZErpTAlP0PRDVx8Tt
-          ZXadZWbKM2LqaeFSUw0recu3yKu6I6Br2SpwyKUVKYzIxatoORrMTyNEWrP7DrNrvEKQou4QMMn5
-          2Qg/vlp7lydRLv1EaZDcRgi2vepKOW69P/LQgkS7nf/G/9NTZHRXI2KkRUW7+PE0ZHPsWJTXSgUW
-          kLgO/OnDSPucAc3v4wovdRDhz/32GqbERgmU2j2PvOxQu8sQ6QFUYv+ElLPdgum1s+AvnxOkXta8
-          nY599qu/YvrDF9MNfoGq3YxffgGrlIjrtgOhRsfk0zdTm4MEJh8jxe1kjeCH92V92XoYv3yY//GF
-          jd+hX/1jq58qcMNv5LDpXUPO1YxchsJCDiG7NmsQv01wSI6HYF9NJ5cGWahAnytE4m7xYhK0KpKn
-          WpgwkPqGzoBclf+nR4H4v3cUKI1QbxVWs6Fh/G5hY3JtMLP7vY4V1wngnDIqsZY2zr+ieOJloN9G
-          dLCzHRiPGs/LjxmZxGxMxWWrixfCt++OJOAmTGeTO7DQeRyUgJPlS8OFn+omiU1wwmMa7ugrwkso
-          X6hhEGWVDcCazTJDp25CLL3PQbwsPVfAbn+pkB3aD3fRn3ILj5ZyJEkLbzmOjiILT09wDvgdeLmT
-          n/ApuIQ7BWlwkoaZLmYKivluEGt5TDo10osm2+mhRNYy7Ycea0kAYr9IiPKtJX35iDiCDHz1xAdd
-          DmhiCTykx5Yj+pgHDVXv273dO6tBJjCqeDqASYLSyLTBoiYNWE7cNQHt7u2gsCG1y0fHmYcejN9E
-          0ZyqWUxRW2WDwU7wca59vNRcY8CE9Vhk1ScnXjpjwhAK7pm4zfMY08/97clIGgsS29Yh54abJELL
-          RPfg2Y/3WHgJXAnLF9pujSDMMNV0X8MjU3jB0o/3XDh+pa9840eM3/MdDnQqGVaKLkWH7kngNbzO
-          3hUIw0jAO/mqu3PV4hk8duuIvNB/D8LIWIH8VcUvCsn1DtZ2ZybyoxoeKODXh7s4w1DCPflqmGZp
-          p68MbyXguC4BuSqgyPkmsSFcdvpAzEd9dTnrBBxZPTBtAECm5qyvTAl86dNC0OmuDZP9GA2YKl0b
-          7KI6ogKv3U3ISkqMMskycj5+eyHkR8nBwKR2s17rdYZluJQk92IDcCSIWxn6sYIOL6oA7ho8V9CT
-          5U3O59HI6ZdrIKhuDSW+qvR02ZEqlO+XS00UuecbLNoThPtGNlBOWSlfqslkobLWF2IrbOpyiR86
-          MLmnZ4LeeQ1ml7+UsnqALYrHmqU033UKpIu2EINe3vqW8TL5IFomiXmNgMWPvRX6bJaiC/CmgfuY
-          lwgmdM7J/abfcvZTrY7cZuNICrEUwbJTYCa5xjdBjlBWDW01PYXDM41IKmpOs5IKKIC6jzsGwsTq
-          i3pmPPhw0BU5g36mlF/XWsb+1yA2Y7v5a31IkXQuZQkv5+s3H+fr8SUL3wAjz5mcmN38RdZE9k4O
-          WvlsOKOoRFnQkzAQr/yumfNENeTPuRWIpVCiL72qtfLIPRPy8AekC/2HKvJxkQosXD+eK1w/zxAc
-          uvYR9HRo3Znbf1+wgK8cHdeGzynXZBo89I8TcTZ/F8gnNyB4MzYKtZMJhEEvIATydEU3XSEuqbQm
-          kq1DOiB199F1/um+sfywF4Ug3dSH5flZLfkeK0+E3rkG2JeX1jLTWzekWHI2CNP4WeHP/8Dr9NZX
-          UKqt/H4QG+93Ry4efdlx4FUmbjCWw7ehXhkn8r3VLsSfoZML+DK3MkFxj3fLgWmocXAg4PyiRl6Y
-          uIDdD1cMw5eTEb8eT8PMlFIGrkfyCOjdHyjd9ekNdq7jkEOl6/kq74cCXnhm/LPnkYghA4VF3iE7
-          NoKB5c9GLa/36IiS7IJcIoSBBnnW14hnsvMgZNc5lQM86SgY/Zhy+MRjiJLogg6i2gzTjnQR3D0c
-          E53OVydfm8+xhcdcIujw3gn5IoUMK23xEymnhz1s43vJWgAtkt/E2cXSG/HwUfWPAB74dlg+766D
-          97aYkU7uRkxvySWC1wge0OEtnlyOu6cOTLMrH+werJpT7yYGkKBTT1yI3vmiSEn4598vLt7nc4av
-          qZxU6xsZeu0ATI19AGHyEYm51igWHtTCcB/KweZPB50ojzyAaHgHJLE1U+evtTSDlj4oCQ47iZJd
-          X96kOCxddJCJq9NjnllQ3nbvS+/jkxJLOlvyS9zOnIFrl39lYZHkTjkzxDnOXiPoe7fef+AeIyOv
-          BpfV70fzL766moTomix6B3fGRUK29WjiWUOpJePu6iJ7FAq6ela4wo/4msihyHh3bc6ghl17/qJH
-          0O/12VfeqVwXuxPRBH+XUzf8bs27qxOxzK6gn4RVWqleGQb3Ou/SZbitIjhPX0JMK0mH5Rf/h7cn
-          Ie16W/KZwYUI29EC5Dr3Yb5etj3Rz/r7QmaZagOr35EBmw5+8cLqMF79z3mWzfZS4F3qKy6f+4IH
-          cgFIAbe9L67G6gun4roSXX2VOX9TghDO8s3EUCwzwBfSvobPOnBJ4PDMsF6OmQhF41mgGx19nWXK
-          NZPP+TFC2nIa8kWxDpFM2NuLOK0XufMxJDMQG++ELrd3qncTYUc5+GYaMoxrQDlNE1+yUXxMvPeR
-          QkeUSjUUlSDH7GYffApOhax2TIEyLmspe1T9ECr0aqPrFl+Iv+c8OZK0AzHXvHfn8KxAufTPAYq/
-          rTKwLEgsaMaX9g8/TNO6737xGaX9ugBKiykBu/3eJ86jqmKex4YGpOm2I6jFj4E9J732y3dE2fI7
-          9Ucjkve0Gki2xUf88YQUvvPMIuXzWevz6amkchIdZrThC5e2pzcLng8vQYWXR67gVp60J9wpRanM
-          MMO8++5nWTINi2QjIc34Wz+Q8gALvQ7iV/2oavA9p3tkz30YU8sWMDxMlz0GZ6l1ifvkC5lrpI7E
-          W7z5qLloyisdffJghUifn+xllB+ZFBBvi2/zZdZEeXS7HXFXPszX89RVIAw+xp8/zkVlFfKWnzGs
-          PiWYi6Z5AU3k74Gk7S0wPtkLllv7XQZwZe4D7Vu7hNkOBkg5Z3q+5PvWgW/+3RHf7lWd4/t7Bc/H
-          QUc+6ABYiuPDgl+1OBFTuFJ9rfDNg3tqROiMO9xMNglecAhWFplqZwAWmLSDLrNTA1YUlHx5340M
-          iu1WAbPvU7wU7LWAoT9G5DDfi2atxUqD8/cWEp0zDpTLIk8C+Hw9E4+vX00HhKspm/G5/cu/vOwI
-          L6g/RoBp1FXxmrDtLD/uQ0eU3D+53Pvah+AXz5z3GefT1RMD6RiiBSkkugxfez5BiM8CRlayy0Ff
-          tFUlM7qFkaZcFDD93veczid0OdTvZqKsHslKw9VIWU3V5ffnKYWE5eVgCet2oGd1aUHK5DU50NMb
-          0FWuIvkPf7DzQFdjXkQpThYPHVirp/Rm4gT+8K65JjVYpWg7Pfh+7jBIrWhYVDMfJazlD+Kh9Nms
-          W/yDzuN5RbZ0EcCiRbEJaz//4h0rRO5CD48MuKvwJS47lPG6P79T+HgexL/8jalZr9A+rXfMDKhv
-          Zt0fWyCVp4U8ckMH8/sQhsAYcEI0aTFjIYynF3zn/BNX1D3lrNK6IdQb28WgeR5zjMiqgDtqT3j2
-          p9SdAxC+ZNe+4t/v42VspQjWhXwKpMGB7qi8rhgOjujhT9LX7goHqYZ3fR8jtdTUYbxdxq3WVl3Q
-          RYhjwI4Nl4BxCWK8U6JbPN0PUws3/I+K+mXE6xO6FqTlIyb6HvTDW+QSCXC+YpOjW16HuX50FfwE
-          OEN6RNoGIzWwIO+kE7HvvaTPv/F06OGSQNFNOm94QvK6TCYHbQxcfp7PFdBdPiZ6hp5gtecrhFs+
-          Ihu+G2Y8nl6QN8sDsZ5xFXMlAhlkh0uJ5ZsuxaQu4he8MvYLE7F/D7RFMILZtbaJ+cgjsGQxU4Kr
-          npUke8ZKznO11/3wJbogTWnoPqsjwCX0tuU/K+eScvRAA2CF3A1/8B9vl4rfjFrEeM8N3d6nlfmn
-          7/3zH5sErbThQWKQ9zIs1zBOpVMJShT4zRNg5lgUcIsPKMjUj7s6uMsAnduMWPzrHs+zqTJgy0ck
-          eSpmzvL6XMoZO/XIztY0nikpVpi58UQOx8IdptkRS5DNWUfyd8rlby3sa7jhS2LkV4Ou3/WjQDlQ
-          KboWd30Q5qWToNWlTzxu/k0rbQjhkoRvcm2NF93sdYbj4sUo3tZv4TzHhPL9cyP6z9+r2V1lgwIG
-          J3r+alYZnDXoYC8gGbPr9HnXKwEgzwxjcBNDt5PrewVgM5kBy8XXfFHEeoShqCt4NAEPaOrPDDQj
-          4ON9kCzulMV8CZZPmOO232cutTmrhIovcOi4lK1OC/+qAOJUiOh7Vs/5+aEW4LWf7ngKzxFdjAky
-          EKpuj4Wkr/WxfnQ1vPjVjPRDegIrIpIGjlh8I/263cvo5Wn24wv4GwUL/dkL+J6T/eYfz2bK99iB
-          z8aQg8/n+HGX9hMFUnaFNYbuwObj2wI3ENSfESGWEWOCe6OTEXC4QH5+nvmPD0LFz4tg30ZLvqRi
-          UQHzob2I4u6CHFcyiyUhXR/bZ5yPBLCmLKSJSa5GazfEI58bGG+zT7JsQfHcQArhs+5eW1fdzu2P
-          8Yjl5fCZkSu0PmX9IA0h+MYO0hbnoK+u/y7h8onyQB4CrRGsXjXleviaxAiU0OVz2BiQcucVFWej
-          j9egtk244YtgX9ybYb5P+wzEaiIjFOXjQBQkjvstfhM/E3TAPaXdCt78pwuW/n1xadxLMzTu+iMQ
-          +EBv+GGKUri3vymmSpvk09iuIeytApNbzlGdSm3swWi4hOj4Oiz5MKjDDaDM2aFNHwH0wa6hDPmw
-          QdpqEX1m5qCE1mExyHF7Pvf1bQaAS5wQfXg5gyAwiwhNTjmhnH1l+lpfyw5wobQih75NOv/pKXZn
-          ETPSSp3sbzkPNz5MlDSd4jVnig5yTYFI+gLYHYZV4eXe2h+Rt603PT1GCJLoOAdSzxyaSXhLN9gb
-          LxM9qtUdusR6dlApjBPa+E2zMrySykOQ5sjt1XoYN3+GO88KsXTszZh75NIK9IVVkEKuMt3wQQGP
-          uUiIHhGz4cb9ToRmVLKBdOzb+MfHoNdxKbE+VqCvSlGGcCddWhQET5fygNw7GH1FASXfz6ehvDLU
-          MLIuN2J6/pzTJpsLGcjkSkzfvcU0PLEKnIxNwes4TOep7Dr4cPwr0dYUNMTQTzdo7i9RIOzg4A6p
-          P0MZPtg70p5Qy//wyveKnujQmvdhhUWmScwZu8FUv4d41vN7BTf7RYWy0wD7bNwvUArzRIL4cQZd
-          1bYrhPwIke6LNSAJJ1mQRNMeGXpuNHzzQS04sXsXKQ85aBbhM2w77psS2U9fdReZhRBu/oZU0z02
-          y9WbA/nelnPwOXPZQNluHKE63gbMzkGv99fgOcv4ZThEB8qVLsPVTv7xqSri6cbXQvnyyhHx+9Nd
-          H2MmXmWQuxCLsf5pFpQ/FXnDA+Sn71DGuWaSfA2v6HabNbA2l8mEcVi4f3hsfTVjCXTM2+h+53y6
-          5F+3A4H32WGavJJmhV+FgfZrjjFFxjXf8JgCf/mm+OSWyy+MWcib3kgQdqucRIXKwvBlZShaTm48
-          T/TBQu/DvJBysxRKa28x5FW7A3J0rnY+ufviBnFg3Inj1pXLc7XxldrxnaDN/t11fyklKF2ZGdn2
-          K266h+uJ0oZnMFDhm9LoFkbyLbE3PjIW8bqdyZMHcbCRnzFfdx33Vwy28RJn0x/oTXc94Je3Hkva
-          Wg9clCJGqo7H15a/+Jz+8dXtVgfnNtd0yffY+uElvHvO3LBkkSdKxQzuSGH3V31UhEcK0CfykV9S
-          gdJBHTKgrMyMTNxzzexcbgXc1ofob75xx9l+m/AelxJBcVWA+XF8zvJvPtMzuwyb3nSDNz48kxxk
-          z3gBbm3Je393wbB+bbc8XjgGloWio6hRo4btdjsLdsqFQcpd2+nTI1B5yMz7IaAZUsHaN0YFf/w0
-          raIU/OnDu7pNMffj18rTWOUY+CcUq68ypvdrmMmZe5qQwTuyvo7AjeBTOR4JerBqTMTikO2/HylF
-          KK4gGPJdpcl5gW8I7fGZTj89TK4+J2LNT93lkkLq4LaexDXHJl7O492CSAr1YD13Mljs+1rLxXT3
-          A65y+uG74X3wHdGFuLfFBtw7v7EQ9yc5aD9yA+bLAsvf+JCSJ5ecil4tQSo7Fjk6ke7SdutC/wXJ
-          m6g849Ju4/dwsnYN8u35GM8gQxhs/kbQA3yGaa8fJHjoNZGo67Fzh/tYauA3nmD/qOMF5b32p9fy
-          VsI39P4dSxjwKsDz6WE3NI+mEcRC0qNLr+f5qnXnThJk70VCSxabJfEzC8zf7Qz6xt8JOrBfOLCp
-          hQ5syseYC+eXNDORSFCxXyhJfZGBeuieyS/ezE6HK6il5Ir3bN5Qvn8xBqAu4IlhuXU8nce7A+ac
-          5Ohvvk/cNQU4YeDG78sBsw0fwNc7K/7w1VIqoQRnHbrkuBahS7PrnEAtjRV07KWHu2QxX0h9ppyR
-          Fr/1eHnnEStp0S0LRvTCLrHsHYZL5/rIO5Aqb7M4C4XT1MYoKBZjoF0OOuC/Y4tYI8/RPzyBFN5C
-          RlqrOnW0fQEWhU1J+shXsE4WG8KA1wFm+HXnjpKTrvDsNgNSxGFs1mOWe3BIDyPy5HPRFPlZ6oB/
-          hyEJxWEcyAWTCm78BBcH3hwmfkCGdFG3CqphJpQbxgCCbn+u8HY7PFiPWRxAf+/OxP8053iOVAZC
-          yx5GovLCYZgb4iqwuZ04ssVvnZvvjAlPRycg9gWs+dimY/KLhyhClFBSKQuWx8cxQF7Ts83iwF0N
-          7SD1AwFpysBNhr9C/32ygvPxfMjXZHE7kDlCsvFDc8D7GGy3ap4qYrRQiknZSSKcBuODyqCLdHr9
-          9OEfPzikS9LQTV+CASZ6sNf5gS7iy2IlTf4IgSQfNH29Pu0AgjarsZQqh3geUuP2xze9s0uGDqVS
-          JS3gXSC1u1ZDd3gPCaxJjZG36QMrKO0XBG9o47fxhDre+AQMcuITZ39LKL3kvSRt+YFYL78c6HT3
-          Q3HDp3/xc7LvUgUio1w3+9Uoy+BCgvz7wCLTTccYv6Is23/EdkK6bb3zUeo761d/IdrxxACcPHsL
-          VnuvQw6GpNn08ACiT+iTx7vu9TlxYg/q8fuFDr4+6dPXVyE8nNs+qPldBOb1sYbyVg/DJK4KOv/0
-          XL5aKf6o/QDoVXDn3+9J8CTBsPHPEQj9fEYm2unNfLPUEXZcKSKnZ97Nhr9v8hAkOTnoX1EnzweD
-          f/kY776ROXCHPUwgfpkOcnK4zze8GMCZCUV058YPeG76AuA0JAdkDG5bV35vhc9PziCn9VZ3PJuz
-          +RevlNOjb3DuCwE8lfdDsDvKx1yoy4MF0e7Bk0DabqXjdbGAvlOryDYmQpft+T89DMMRNPF6faqe
-          LMDVxVJB5Xy27b6C5SrWJBqaZlhY9A6Ap9YPYrDHGszW1skLRDAmqnZN3OGAshUeTlgNhC0/z+ZS
-          jzBtu+hX38pp+bGTX/5A3qT1w2LfpRr6FXMkKnz7MW7lYgazzriY7oR7M6fc5fXHd2afLfKfHgUv
-          lREg76bR+G++lmRkAg4/rfwbHWcWymOEiJInQj53c9GBrR6GSosaOUuXIIHS7WoTNS1vzYJPDIY3
-          ZdJQaS843vSkGxzeQ4wss4MA50zyhVc5XNFRPDEuFe8MA6Bq90hFawcWqfzMcMs/WNjwJfH3sgeD
-          I/smtyg4USr7bPWLj0HVEM0dz2n0knurxMS9mEWzeMfDTRbj75eofRkCKnpfEWz+gtkvnuiPD8nj
-          cb389PmGtcBcwyK1UrTpy/r0/EgWDC6Zt+m1TIyl54hhvUJm0yc1nVji0IKtnoqCzwvmGGGlgEa1
-          tYTLdos+K8IlkaVZQOSn1235qYC8s5gYaBIBcy99CzDnsRaIy8N3uQ1PgtSQP8ixroa7clVswEHs
-          bXQQd9NAuuw/AAAA//+kXcm2qjy3fSAaUidpUonUQUDEHqAiICpVgDz9HezzNf/ebZ5xhhtIVjHn
-          XCtZYYVkh6uJB8MxHt3mzCJfshdyD5tnQUdcJmDXezA+2g93JuesgreeXLCmXhy6Sod6lP/8ywri
-          3t3elZ2AjXW6GSWnbq8HKDNiYPcjOrP5MU1xX4Fmiet5sAgC5PBLM2i+JR3v+VzfQv9Yw0dZliRb
-          p9sw/fHz9TOy+Hqr3uCz1zfBaRVLcspHWEyRGm7/9Bl7qXv9+4pfFhLmhgSHaxTGKz1eM8js4+GX
-          vR66eseY/8uvxIJG1kzpxyr/9memMUDudmFgB5bNfQWylLNgrwfl8CBfukDY9e+tWvTlj4/hELZm
-          vHY9lyK9HredH/+G5TsPBvzNMsWulw/6wtwVFu71r1meObsYl+tDg4K+/YiepZY7equzQQcftqD8
-          hakrfO6TB2Mh/e31+cpdWmdkwI/6HNE2zBXj0zXE/09HgfS/OwpiQdRI3ilnfQIIJ5A7Qz7gE4m4
-          lE2fNbTw+4dd8/6JlzG+ptD79Ro+ftZPvF1MlkfiQ3BmTpxUwI884IHwaixyOtdbswrjPYc1/V6I
-          EUanmL/uZzR9is7YDwVDX2XdV+A5UySCL/3HXVaFr0DYGBMOP6FBJ11PM8gGvoezJh6Gta/ODqoN
-          XiSmr5TNpg0rRM15POKHX32GldmeFUig8cJ3JN8B9Z9BDrMukkhQHUFMJJOrUfxNRex1BxKPo2Qx
-          csNfcqKeQp4uhtI+oK6NfsDep/ew1ui3yD8jLnaFR9WXhnYbwrPsz1veAPq7XUEPwzIISXCaooHj
-          LhML39n7h7PDwLrrdn/wIBZkbRZQ8mvWl1L1yGhhGXDRb6WrZt4D4IX+Fx8HVY6Xp1FG0E+ZI3EV
-          osTskKYWMk/eQOwnJO4WR4uBEvaRBQJK7IbPsy2Q0Sm3Z2peNjDXTJUinmU2ok5V7Qo9bDu0xLFF
-          9Ers6OJXCYT0o0vYbVlFZ/k1KVFs0XlOuuEwTO7PqFCjDC+MaVsU3OmRBQhmMYe1tX8144aEL2rv
-          9TSXmowH1jttGozGGhP/eDzoq7WgCqZWoJJ4BcHAP81zBo+/veKV2nkhKOxvn6u2udhqniZYjcbJ
-          gCQHiNjHh+1+x/iawOVFZ+IPr2PMatb9AX+/Q7Lb36lg+/2EA+U/OjEande3Yr1bcNmMN4nip6GP
-          pXeMEPwlN+znPjvQ7X1uwb4f5M6YkjurfVfDkSCPhFHiNMundzKokykj99+ncclhvbPQ3uQ7djEV
-          wCaftAyZUGxJ+GU8lz0B2YPT1JrEb74bWOTjMUKzhmwcHA9uw5Ofv89JFu7E+X3zZjOee099m/vE
-          vcetvk34wUP2Uen4WUu9y/70hYedOdyIMnaXQciK0YGYUxCJCqVy2aR7GjDBtYvV6qjTZeI1BzFS
-          n5GrcUqKrRHrGcLcMIPDrOFmcVYnA1c9Qli1C62gsiLIyAsPF+Jc7Hagn1Oy35qk7B0UshkTd1C/
-          6Ms4A1Y/k6/z2rxsyLtdXeLdswvlD2d+hOipRVj3rBNYLp+6QmK6z61cz9dmCYt+A+yk/YhhrUO8
-          HYQqQZVRS9h22NvAtoKkSeaJybCTC+5Az5daQSavWNhquevAprqXQqK0OjHkY6BzzHatgAXtA86c
-          xhsE7lpDBHymC/j0MA5TZ6UOovxb/4tHxTc1dQeVmXAKtsf3N2zNVU/gk8lsglucg23fbxQm3Qkr
-          r3MChPJ3jUA0bDXxvNcvXj4ft4Oew76xQ3GiL/yaPKBH7jI53Yqjy75eaonul+JONNM4Alb/qiUc
-          oyAJ+D3+8CxjZ9D5tinR3G4dluJiz7A5z8f597tOzba4oglzxlcDvk6meNNmcYGeOmkzgOA5UFMe
-          vpCRvhmxHVZqSJ0yD2j7MCS6Qj8ulZWDDHu7euBUHQpKEvPDQM2bSuJueuPyw2GJkFA6x2DRr0rM
-          FpEjw/t7TrAqk6O7zMHkQLvpHVw2ZgkG/2nm8Cy0Or4sLCnWk7Qy8NDBAhubVwwCQmcD/ST/Q278
-          NlByNS0NnlXnjtVTmAISJmkLz7zxw5b5YOgv/f5M2Da/+3zQyhhsQ5Gb8Bez75n2Wqkv7UpztAye
-          RRSv04eVKRQFJZcvwA6zhK5QhDBF/cG7BjOnXgZKJqeFOZkLcjoELuWFgOboEk5C0E/TN6blmJqw
-          bjONhMGo6hwS7w4YYm4iIXs6DytbNCWSz/kZ+8e6ibd2/JSQn40Libvt5043KVbQbyhn7Mu9A1aG
-          e6SwqYIhiMdbTbfC28eSm7257/cbcD2zjnCgmUAU/tYXm7BEG2gvjoA1UU4aAUomi4SFFtg9PNSY
-          c0c3hZv/uAbSMfLB8um1HD4WtcFhlPTDuDWvFH051SLm96QV/LUaAoCOH5FogzA29MUUJaxWHgbw
-          nnF0tS7xBgXx1pKj/DFcztc2iEbVXWaZaRPALvYGUa6IB5KwTV3QoQx7tCLJx/F40wAPtcyD7PH0
-          w6cbtiktSCKim//wiF4zRiGE13sOGpsa+HiyLnR7U3G/5R2XxEBMNQjaKRDhrcDHmYt+Z8AhEDNo
-          z69EK2vPZaMtT+CtgBtO6ZXTqXUNRcS5vIkfULZc7li5HYTnMSSROEvFxiSogm+hMLHvdYW+XKav
-          hbyDbuBTzLoxt91TFo0aowboYX7cOX1eN8BUPoOtce7dzXGeBuwPjIFTrq4aAT2WCi2+leHnrJGG
-          THHiocXtmXlxiszdjCczg/FmLFg/pCOYJFv7CuYpGEhm/fpiTE3XQvWouAH0LoErCC6ZwfFVb+Ro
-          x69YeN1QCHNOcDHe7Zu/jz0DradDAxCPfbFM133q0WUqyTX+LvqqK88OymggxKEFjnd7fKBFfhSz
-          iNYzWLJSHqFol1ZwwELkUqPiItl/lwO+ZdVLn+KzVKH4/ChnOdrulDpBWYKGPBBR9XLTv6+H08sj
-          89ywuvHRsP3hLd9v+kD6ufstn+fNQ09ei8nTKAydKrHmIDU98+T24V7NMubfDNmbeCfPFst0DLMG
-          IsdWPez4pR8vtTRmcGTuG1Z/+OauIWONyOE+9SyO8jv+TPutZnYx2WTHT8Wof9UHuvtrHEiiVQ5b
-          8CxC8JPwZ6baK9DXYrpu4HJ4xSRtNbbZsnzqwEM9UIzPN8Zdf8/jiPbnEWeSmmYryWtB9KNK+Fl4
-          Ed281i8hVz3LgCuNsFgvmvCQoeaaM40zeVheWiWik0h8EuTgTelhviqwHc4dtn/uFq/T5ZfAwiMf
-          sq9Xw3IU84Ck87bPrkuHP/tA19ttJKkqXF1+OU0s3PHQ3/frNGK1FnlCBshlt38h2Su+Wk89bArl
-          7l+3z97jbhHsscHWjK7+WSCPIxH7roHdLWcLCO3eJwH7/JzoghVbAdZTrnDQSAcwo5w6CCSCvuMt
-          UCzCfO9hlzIh9g4kKMZoy1N47cswuOz4moVKYf49n2jnW9dsVponcs+UPQ5w/qNbI/YjmCB7C/j4
-          2ui/OapruOdTcjxqWUyPyi+Bl1/WEmOwdMqZ3GeEvRO5WLnyabEGr6KG2+CHM716h2Iyxa4C7qhb
-          xP41FqVLazp7R9uRHPVqLMaAQbVgx9d45pqKif/ZKzELQFzzVw/bX34sUksjkT4d9OXiWQv0yxVh
-          39vSYqOqEv7Lbw45qDr5fCkPuK8QzUx5Tps1OEADLjUbEeeKa30pMqWGfGpfsL3bM8nFWwoZnXkS
-          /Ztw7p9/A0/IQVC9nU8xtyN5gB0vYsf+foe1qtgWYqv8EP12dQBNxlsH//hCcrRfdHsv6wzdJDv8
-          578AnVJghX1AdCev9vykdTDyDkcc1OULdOX27eRpbGqijLkcrwp5t7Btbw7JgdU1xPVeBlQ/SU7w
-          LQqH1T9KG2THCP75UzE6eFBgVDyHP7xbrDENZYhPs0w8Dt4ajirq/C/+ylVfDVugrxXUZ8zMdI/f
-          yyFxeflOn/5MD6kHlounbFBVvy+St+YM6DwcLEhO9hefdrxIAkM0oYm7F1YcXAIuE3ELzPs1JzYr
-          2g33rb0a/OGnp8C18TZeyges8uEbjMY3petL+X7hlqTsLD4HzeUO/G+Bwdde8N/7L49hDeHJLRxs
-          r0YL6FXRMvhUjRNRfX2g2yG1GAgF9kBO2qoPnOjwGhSEQZhv0TTqpPS5Gtn5oSRe/xHB+kNzCJ9K
-          kpP7a16a5aV9Zfgxonj/fTPM6HBOUDcZB3w1b5NL+WT8h3eI4hSZvjycmwg5vQxJWfA22NeLhZ5W
-          EhxfyCte+vqcoz9+EqOQxMvdPPXIraU3Vm18crdyqzpoXeueGA1e48U6P3LoP9qAnIynQ5c9HkM0
-          6o+A4+e1mVR8+/7xzQBAcBjWENgyfHyLW8B2gky/CicasKmYbGZ2vjr7ppjCXnW+OBi6fc5liC2w
-          x6tATMOGkvwmbvBsbd3MkyaMV8nWekhfnkrOao3pxHBpCoSfzOBgWrWBu0mFBj8S9OZqWyuw6UbC
-          gECvC3z0gEOF86XWwCWqLRJw6qVZ1y6O4J7/yC16hZQqoZsCSV+vRPmUlbtCI3xA8T19cHA9vfTl
-          xLEt8v1X/4eP9W3mrHJvpAmxI2k//QfVc4+CPBcIrqg3bOvnxsPr93fY/z02vdceHyhqspHE34TT
-          V8mzS8gsNTcDoznqRPWNCvYHaOAzmz2G6vj0NSRaY4G1TK4bfuQpC//4in1qu2Ff7x5whMyB8M1I
-          TNFjqREU+APBAtcWCydnMmD0/ZZgFazx0i2WBpnzlJHjWzi6/LG1IGTWqA7g+2wVdECiBi/qsyVa
-          JmsDp+uPHG7ejWL/mH7dBbRr+Bcv55c6D/qy85s/fEeUy5JRaq3Eg4M/IxJoV01fCF46eP0OB2zG
-          V90dP9agQZFfXJzX2Wegf/jxL99psH6401dvGLhR64wVQGS6AcYXZSOoOXKqhszdbvKtk5+oL7HP
-          woZO8VRE4CZvOGBkAgdKnzFEgn7PgoPUDM3yntkOCrFWk1O7HCnFjJyDJ25HEpHn3rHFoArs+4lN
-          kZnBsqHDl7/PhoN3vUEn1jWT4etacwFTlypdpqTewNmuY/wX/+egjkz05KIOq7WrxsIobDm6m7YT
-          pNlLbLaQPeRw4oIQK6A5NGt1uDuA70pvrp7XLd5YoLXodfIZcjJvkz7BetXQI94Mgk3jHv8OtIeA
-          +3IRTv2M6vRVssY/PveH7xcmVx9oVMGdnDyvdil7r1p0OxopCf/24yzYuexurhgsQXWkgjMi5Q9P
-          7FOWgqaXfEWDm1fQ+RI2SUw4Xq/QBI1x7nxPoL8yVkuk1Kcrtn5mNayvWlGAAXgxGCP94s5nyf5C
-          IV4wts4HGVAh2qeU7PjI+PKDPi3OWgHbe/ABWgTeHdsVZKCq21/Q6SgZauPJz8AFbYTPrv2LR16I
-          THDNgYkD5dwOm16jCnytcsO4nma6bY97Dse8dXAqOb27RPG6QPHKzMQr+nmY5NfUQgx/5rxs72PD
-          /elnijcm//QUPvG2EkXGZ5+S4b71mUvbHB3t94p3+9KXZvlAeFouV6Ls+XMJi3qB1/4RBi2gnEuZ
-          bJNRqZdHHHRhN9CgxC1E4+r+0+s2cYohTMfTL9Cv3iGmupUn8p8+E7hbVdBru0+BqdkTVkluFNxq
-          UBE9Fr0h1pkemnH5FrXsoEOGvRs8DUvGRAayoDRiY3yww0cc4vHf+zssAsXoN8UGMRuHxP1+W8C6
-          9pT/4Wt80cddD+TSFIL+sO36w9A8Pq9DDyrpefzj681arrWGbkczxWad+MWOLyq05xt8z/1kIGA/
-          wbHrGzP6fXSXQze2Bbs+SOKr9yy4r9lv4FO9tp2fte52vtYR3PNH8L4zUTMpZGr/+PgMrgkFizrS
-          HhjHW0R07lrSdbq8UvDHz83twDRTGdslUGTtStzzZumrrj+yf/rpjs/iJR7LGaxccyX4ZnLFEmYD
-          hJCZip3v1fs8xTGHWDHXXR9L3VG3vyJajKWeD9fFaHY+O8PnNeCxr94v/+kXf/wp7VVjoPa2fSFq
-          hi8OyvuZ0vv0q2DkjxvJ9ny/IPn+nz//2fdsFl4Nc0ep8Zm5BgNpL7qBIg8diZGYo759XkL/T99J
-          s2m/E0yRZ3AUfq+g2PW1qeWWDe34I2CbYmwm9+dV8KLeW2KHn2BYy1D7oj3eBi+70GL+0U0JnIaV
-          +6e3Es29ObDqvQ3b8HYcZumTphBbjw/RBM8vuGXMRHhhaBfwvWo0K8/deEi4CmOv/2R0+/6sGR7t
-          z0rs9MI1E7M9a3jr7xnJljQeFv6yjWC7mK+Z7YZnQ/+eN8TCRFR1Htz9/fr9zs6A6FfvGY9WsE8F
-          +7Asvr9rLl68MQvA+8NBgrtX3KwXR6nRay7OxNZj1Eyiw2gAz1aFvVl8gel9zju51OaMKAUwdOEw
-          XzWou6aOT57SgW8eHkpgXasem1XuuiQoT+0f3/nLdzFF6GYC3FVXcj8en+6mRHoLubVS8DGv25gt
-          vWMIys91Im489jHVf9II6Ne44+x1+xVrWg8PaEubNK9sewO7vs5AGgIHu0bzdodLVzCglus1AN2s
-          Fesc9TVcLaT+8Q1312860FTegJXLItIZ3+49uLy2gpi+d6XzS05NsB3ZAGdFxxbzANcSfoww3k+E
-          DM0fnwMHvJ5IuOO3TUikvY/pfQqYg3IcNut2s+DhfVfJbbi77ga10IPT+KqJv61ZQdZvmCORyZJA
-          fA61TkUsMaC8tWwARoO6Y0PnBYaL0BHNHnE8B8o2yj/XFeb2F6jNuutJkGfhRnb80bAtK3tQX7QQ
-          7+8PNtFlO9CoRbDPzXb05TPZ3l9+J9pFDfX5dDFFeF2TCVuqNcerrlw7uNVCTo7JV9PZrftaMvx2
-          Iw4eDY63/op5mJjZd14YU224kRx6AED0CyTfrQpOfr07yCYXSI61rhTzn54+O58Iq/B4od/mFj7Q
-          n1694w8wV3reQpWMIX5K1zBefIN6Mj+bF3KqniVdPu6P389s58GhLVJ9WWwZwk9cd4G467F/9Yl/
-          +legHH/FjKR4Rq+nOBBNn576mtN7BvWzR7H/ri/Fri9aoGVHFWtnfyhWuW0zWLpmgA22qWO6fDwG
-          eirRiFM9jsWq2YkCi/b7JeHON5cDNwSwM383bJz01l0VF+cwvu18r3q845nhSxmGLrhhNbrBeK6Z
-          bwr+8NFtNQz6cINXClNwhUThfaJPf/rfPfktf/WCmP0e2wBagQD/8d+tJL8F/PpJxU72jugUlq8E
-          wuzMYf9hfvTOf4f71OcgnA+j5cdsObU9FOINBxTUX5dudbTBvR4STOo5L7a1z77wm6KMqCT4uYMz
-          cgosw+cZn5bzi24/puUR55q3YP3piH7XdznDKJ8Y4o2bsZ+49CqQfbQuEF/vdqA9HDtYfi4T3vNL
-          vGGZ0+C/+PB8xDHtD0MOK8/V8F6PctcmPDlI+LnnnY/0dGb4RIYnZ7gHHHyZA/v2Lx5YlY9LApzb
-          dF2bsgQxwX/68Fvf9ZEUULUX/qunDZL8heTkfufVvRjN+x0UC/jTp2+vB3CXutBZuH6WB9a5KwT0
-          vnwTKHmyPAPpeY6Ht2f2f/sZCFRy9XE6DxDoTpoS4xe8mjVf0lb+w/vC10vc+ao5PZQ81w8ktqL6
-          WPNTCyTZQzOz1+vmXb8FKt9O5A8/ztUDJuBWMBvWUYjjDZZdDbcgMAJh53/r85izcCvHnpyetuey
-          TyMJobYAiE/AZtz1a9YLOlxni+iGnrvj6VpqcEjeOU7u07tZ0nYY/+wlEP/qd2ndlDCSYws7rqwW
-          wsTSFpZO+gpYZ/AHYQrvOaiP3Tvoamtzt4PwTeDzQkqinRjLnf/sbWjlcxCt/r2gvvJm0ZfbOHKi
-          0qAvdeZrMDruNRCnEN2//P+Xj3e9QWnWb21USNpaFuM56Ony8MMR7PXI+UDhCkgRAwvmoR0FoneO
-          6crfFhNZWXjF+q6PbaslRcCf+Nde32D1KQ7kEa6hdcFGYno63yRGhXb/mWVXfhU0Yp0W2n7QkD99
-          fNP3LrpRle5YXWBY8Ar2NmmvFwfNbn9LEWkiMtOSwXZSxMWidbyCQr1pgoddSGC0zo8Mzm5yIOHr
-          nOz1CwJB62M+qAul0v/hxT98u8cXdwRmOEOhjjiiGU4zkJ1PwMh4Y2yvYaVz1eHiwNxqLKJO1tDQ
-          Oha+gDxSQJy61eg6iNAEym+q8JFPgT7/4Y9ZABdiB1+m+H7cHyv/xKeMjabwhlk/FRv8KgYMADeM
-          w6+YnhsQ9GcWSKdsiClFbIVarJ2w/wnLeP52qggjz89mgoXtH1+TK516BGtfou964ihz+iPExzsT
-          Dcsh0Vn4KRwSyI9YAosVPDvY+j6PnfU1uU0y3lqw13Nw0kOrmd90qVF8LktyO2Vu8aenw9uR1fGd
-          5zudrq9TArpHe8PKni8Wno0jRL+8Q7CSpvFyuGo94tZaIf6v+ADuPcIUVMPs7frTMqyal7Ny4U0f
-          oqa8Mqw5vWT/n44C+X93FNSm0RNT437NsvZhhT5CfMPaO22GRf4OPDBTnyXHd3fQF/sQRggJehh8
-          ro81Hht4DVB/SESSzkkHftfTuwRrYJzIJWHPVCjO0QLxB074NPBBw/HWa0R0s2usaQc80KM/srA9
-          YGkG77h2yZr6I+zNlglE4errW9SjCHDr/YSf94V117jbNpiqpjBv4r0Zft3TyWBjMRjjY6IB+rDn
-          DSYSgtjF/adYmfsthYPV6+RY2G28mnfbQV/1ZuBc/QQFd1PtB7xnyhLw9N3qRMyABku3Xfb3I8O2
-          IcpD7mN4M0v6j0vX1J8hH0oCtubSBfQ8TBtsAs0g/mZXBfed5BHq9X4mmldcwMYfHEEOWiopnC7S
-          x1ZUApSG4ZVc40GKKe9oM3q/WIzvYgf0LX1sETLZliW5pVxcyj9yBcn77LX7U6sKOnT9LCWhExHs
-          yueY8yv6QLWqcUTrRAjmVy/KIKj0OICG0blc8SlMNDNbTXz/dAdLpvcMrBm2xfbrPekbG0gevMGz
-          GzQg/A4Uit8vTAqjxHF0zV0hVy4PVBVajo1kIcU2W6uIwGcasBmKSbwG/GmD7gdRYnbgQSc9IiwY
-          LCkm+fUeguWmhBAd4XAjQRH3gKPnfEZVoeTYenoCWFykhwiqGk+C9pjFW3I2IRzgJ5v71KPNNqCi
-          hkn/8vDznWcu+5tPAQRjFRIvuHvNChuUgk8QBYFYWDrgLr7xQAbJpHlVvH0OC5U1GD7FE7mQ+Tds
-          sH2z8LPhluBJsXS+9ZcW7fZDsir0i/nVLyIy2AfBxxC9wXot4PLf79HKFgtTVixk8vIYrKGs79+X
-          MUiArxSHlfOL2RKrIfII8InizPwwcxWt0EVnr0QFt2cz0XetIaPyc3wF18fAsYEyQ5xRn2Ba1wP3
-          t56rHdfE2QIW0M9azcj9zA05utJSkM7WTGgV1oPcf1HsTvKvTGFd3Ahx1ceXzqZdJ3Ah02n+2ikX
-          Ex2vPBpzmJFLtN8iyOLjDM6T85mnObWKebZ5Ef35S+kdPvHKCW2KjtZokEvwfOpshJoW5WH1Ijo7
-          z3pvbgpEQn0biVUfCB3bsOuQWwOGeJ2EKP2s3xH0IZJmqJwDl/2zt+c6hbjgjyRe+GdUoevn+yLH
-          8nEBwnNVcnT5VCXJ6gOmfDU7Ivyzt+v7/nUXS+sTmLBfAT9ozeqzkx57xF8OY8CK72+z/tn/5z4z
-          wQGE8rBIB8VC9CVHAUTkBbYBxfWf/2AzePoDy9iPCtzv4Ej8a7nEizZMPZQeqouPf+tzWL+9TLIU
-          4WPXVTF9bn4GGVarCIavzBUAkHtwz7QlGPb1XpO9rwTwkks0czsNwk85m+i1dkd8UpzfsIzSu4bx
-          J7/O1Ne+dOPtKEB5WL+IUqMnoCB0FAgPJ25mUm0altPaMjC6mBo2za86bNqP7dEFci+cSPINrL43
-          OnDuG3/mb/A3CMzppkGyfi1sieOjWYh1kVHE9CU+IUmKp9kOcqBJ+ROnhd8B+koUE81sL2IV3lR3
-          e5ZWBx4XviLapJwGwRPFBep10RNHYlvwY9itRyd9S7CV0LVYPqUnyqHZImxKjAMW5rnxcKi9B34G
-          5ajTNT3OMuxThXitU8R/8QIkzPaahferdKeDiRSI9dLCfi2Nxeo11gNt4W/CR5aP3W+Dfx3MzYwn
-          z/tY6UJ3P+eIueRhIErSia626vPyfPEsUlxfNKbXI3xAO1B0kg6gcre2S0X4zQ8PfKzOPhgjNHSw
-          0DOTuO/2oS9O0n6RrIgr8X7BMizd9gqQDkODeBq1Gp5sl0p+BTbG927JCkEoFA2qH7knplCG+rb7
-          C6ilheAUswxdJQOPsBAuBb5+Dlqzxy8GHleC54uCx3/2CB/pacFWxwP6PQ+vCJ3GKiC3d3Z0OWHz
-          E/kWPAvisAseSBkoG/p8TxI2o+jY8F9ta+HVVs6zfLBTd61uugGXS+oT91X17r98je4Pjijwtg3j
-          cosDNEw3EbuK7w7C8cVaKFC6ZgbdeooFXIjlv3xTxi91n2NvaEBxHz3G91FxV4hnBbx50SC+2BUu
-          z9hyBcX3+Ux03ywBR6PrCK7BJcXOU1cAHx28Flqm8cNJ3610MeQ7hNd7zeMTZv2CzS/vCMmhecZu
-          cdLjdVPumXQJ/DO+/fCqL9l1C1EUKj5RxE5u1l96zpAt6AEJymsV0wK5OXzHTxyIZ0UqxsPKOfAj
-          aDZ+bH5NhRZ3BurDg4RtX/lQosNPCQ2SS8QGERgmaZwT+JnsGYfdcgUCCB0NPuOTQvJuasHUM6KC
-          VMF09vyGC+FU3h0oM2NFckw9ytqDBaEmZU9yu8+qy5qWmkItXzDJ4joGLGOdawT7RCFRa8vxPMhq
-          DQ6T2hPc2jKdjx6bopehvrGSTFNBiRnmEPiigq9PQy14Tix5mEtRgbO/fChquYdeY4XJlV1DfXUj
-          14JHXXoGh9QM9WVmzRqW7BvgE69NxRp38gY99kYDgPtTsT6cfIEPqXqTRxWMxUDMLEeP9LiQtH/N
-          YJSVo4cS8lZJ8j5XDT1ZRxb50UZn3g6WeGYcoQbhZbtjf06tePmptxYJn+BMsPNW3GV2iQikOnoF
-          wh5fNmafRCdKK4/1D6yG5WpCBtLpOBLntjXNT+8ODlwY8UNccWD0rXnfRKi5wmF+nQ518Uva9wLZ
-          XCpneF+Thlg314K/SJawOohkIIvhZjBrJRUbz6MK2Mz4QChZszmvYo9c4ravEvpB5AfC01DjWY+K
-          BOCP5eOi40C8EXMUoUV+OjEn3W64qPc2eK+tgoT7fqw3zXyg02ef6xYTAP7hn7/4iCXwi8e2EUV0
-          uiczuR9vR8DCgKvgebI+M8XcoVhfF8mQI/WIsE67dthopHvwFo0rDuhD0cfUtiE03meM/Xdxcrfb
-          /GlhdlAG/Lxe/WK7WhcGyIq84tN0Gt3VVixe3v01oGUjglX7GQ7M1YtKTIAtl/1pjw6+6vxN1KBE
-          +jI/DwF83RNCTHg9N+S5+blUh8GFuJL4Kebn/djJQx08CP75oU5f18gExBJ9cu+bMObW0dRgJfX9
-          ju+f7gbwuZZX+1zPC/283fVwXDSAg19N/vIjfVivDTmmN2PLUA19CtZlg0s/ROR4jF90ZZyTBnZ7
-          IPker5b8B1sQf4YP8aOs0DeL7FM3ePk4Q1lBYCELW8HsUp7nzc7SgvKOM8qdmbDEe7VjszDVgYV6
-          UCbza4FvfdLfsgPrwtnv/GF5d76ldgTY3oAkEHu0n6myaogC5YgTxXVc4WNyJtzxNo5pZwyLIYUM
-          EqxOJU7ci2DCJ9+Bd+FZ4X/xx86IAz8f/kdOIUwB1WKdheyluwbD5ATu+hveo/w2YhIwar/ENDyW
-          FmBUB5DTzoeW5nuSoSHVLTm+b1ednvvvAtT7dsD22QBgk5nEgacYG3PZOkWxKppcAbw2FvHNXnf5
-          a/kOgW3YP3JKFU8XggqX0P0c6CzFXRqPohZ5cMcXMyfJVjxeCy+ERi4nxDgxL0r7yzsDsa1fidNi
-          ptieFh+CqU4arC+Cpo+PTZFhqzoyPgrZ1Ozdag5kpajE2vzw6GIEUwL7Nw6I4uu/eElqMYLTJ8tI
-          7JuQLsvRlaHhziy236Xkrq/eyiFZe4uYx0IsaBv8DPicnnLQie1tWGN4Mv7hTSZksoJML30GN9tK
-          g3y+fsGW3YoIeu1BD7j3Rde37bLf+rlOZ3J/mmwxt10qi92lTohnezhewhlEf8+bua5T4l8e6ws8
-          w/cVX1996Apbf9EgZzdjsNyYviFv0RgR+JCB/PFDkmn9Q97CYcJW8Hg0m6rcFMBe2iuJtrNBl8Mi
-          OkhmWx7/40sH06shHwKBGB1509UIMh7s+S4Quvkdj54kheCWfcwZEXYaaMduIzzxkk2cBQnDqB1Y
-          BiYHdSW+8FibGYrIAsXIVMH1FyouZR5yAl+QPmeenaZmSW2bAayZCVgp8AdsnqhEsJfKhsSpE+32
-          VC3yTcApCRTP1YlgLiM0XV4KpGPagD/7A7zbrbNkflSXa6AzwrQIXgEs3Del9nx9/OUv7OnK0my3
-          axBBd9K/gRCKGpjRq+Dhy+IdYsNGcsdTmX2hd6ljrL0+z+YNOieBD5LLWE0136XNB2cQXkIXG77p
-          FpsRBqms9Kb8Lx7981f33okkcN7nhj7K1USq3fTkD9+uj+fHgGZoYBLv8ZPg1VtgxcxP7PPKGI8P
-          h0TokYsXYoCTrvOS8vJApgYQO8U+l3dLxhnok9HidNJ/zeY1C4Mehf0OBPUbAOF7lTMYmZOGz8DS
-          G56/v2p0uedrMBQnvRDSMlvgfTLf+AjwbfhbD1Csd3Pe7WdYOFcP0M7vA6mf0oK2tZiBMd87pBXb
-          1rk+MfJ/eN298aZLcLGUENjHgoRibTYLcofqX370Ej6OR851PdkzE4qtVn8X69HTWfQOLJvgsybT
-          DbYTDx3zyZCdPzSkY+URwEtTkGPwPLiL/G1Y+H0fbfzHl1bkJRo8ZZE0869uaqbDcVEQd19V7OHZ
-          puNp7zCd9Dkmlp1eil2/CdH7M7fktLlqvOjow8AodTdsxo0G+MDFX/i99LeZeUevZv3p2IRx/d6w
-          /vrM7qxHcfrPnhhzfjdLzbQB+gnFI9hncfzj6/Lxodszd9DOzWLrnwREBf4EaLf/9bMaIzzGxoG4
-          FR7pUsiS+Mcn8elgas2m7XfgrLYSkSsaz8XGPGwW7PwQ51uQUI4NrPEfvtSkQ0m3xzPqIH2JEclS
-          Yx7on/1ovXeYxRAe4y//zGuQhLcaqy4fxKytkwSWh4gjbhTyOtnvMAOH7VTN9P3I9aGRb+O/fGq8
-          z8ogwPaWAM2sZOx8GE3/acpUy6dVSGZhXz9hj++wCo0QZ4N8doXogEx4qgeWmPQn6mNGik4+fMY3
-          9t5XY1j6nvPgg30sWIfZpi8JSDtI+IOPfXC5N5vMlBaU+6eLfU1wXcqKVvIXz7B2X+Jhs8cfA0Np
-          2vCuZxV/+AZMkNxnaWDo0Ge3OETBBXyJVYVTvDjJ2IPrveLJ1bcflLphZCE6ncaAo2+u2YQiqdFy
-          ed+DGnim/ve9cA6LhthJ74Kh6bQM/q7MjO37dnEX7Rv3iH+yJj6WuQqEQ/FN5fLJH3FirmLzac7A
-          kNuDL2GcfXu6znRTUJTDKwlCuS1oQG8bxPX1i9WOlYtt92f0GeUHwU91jUmNmvEvP+Jdb3RHYt1F
-          KARligt1COmeL0KwsWZN9KCa9YXhsANtQQ3I2Zzf7srYjiLfPkuHi7kYimXcb896PKHzh08bym+/
-          FG5qgP70mT3eqgyqqZhjBfNPujboNKKUwhInHb0BOmI1Rzuewn/4Ys/nG3yuD2Wej2FVCNflawKh
-          tkqc4sWhk60oPAqf8gm7x7NCt+fdb6GXtxYxs0dbTKP8UMBVOJ6xriSkWX0xWeBPf9gk+NM3Eund
-          wT0+YavbPnR9UCaDusXhmccsA4gnLhviYHEOarSxdK5mTURNcE5wqdZuwZrPukdKb8gkWOB7+IZ5
-          AxHOVn+PPzRex2DKwVA/Z2zveIKlZ+JB8HkU2GiVJl62GHrQOhwpOc3ZL55e6m+BxdcG5O7vUzGq
-          vDUhCrTjjI63I+Uf9CWjHf9hpXLsePzTXzwi+fN4MKi+vRtP/rNPovFWSJddLwJP/RtiNwpTndSX
-          o4l+Ornt9uLQ6Xb7LGBfP6L/zpu7WsaThTt/nNdBJM2+HxmiMPYC+oHKsDYIz/CnlzZ2u/UTT2I+
-          aiDmudP+/nbBCne/gmTsolmKzscdf5MOuC82wP50dN3v2y87+LM/Mwn+9JcwH5h//El9PX/u9sd/
-          n5/0MsuWwrkT2300cPts3X5HxKuYHpTPgaEsPYniSqGTLqsltFSNxaqdULqmtvyAe5sD9p9aFW8N
-          pAEc2Ypix7eY4qP1kwWhe1Cw/c7e7qqjogS+nurkZKmXht3zL6AjX+76Xxiv2eKZf3oG1mjHu9T9
-          gPJPTyFOKGjxGjy6Eg5C08wSiIqB3gbVQBEp4czu+JArPrEBz58wIpmECjA++Kcs/vmb8y4/w1bK
-          vgwT6QCJU2ZRMdKvW4F4XSNc8vYZkHc1ZMDIxYQY77oc1jx2F7DHu50PtO7mgjyE+jc+BLPiDfpv
-          x4fgWYcVCUP+VRBl3ByYSvyKdZeJi/FJfRHueB+rQWa75JdU+V98nxl455rtUe1nlC9NMdtz3BW/
-          9t1l0GpniZh46cHWWbdWvvGMGcTZ0wDLa3zJ8E/POPXvNd7jzSLzbrtiK4THgmZzYQDyqe+BoEme
-          vvD0x/7TS+VQqGNyfEEHfKKFDWgyepTd8xOIpPZDlM2PAWdnHwcul8QnZ3hT9aXKohaeYXIiRXB/
-          6Mtvxp4cFf6H6EqCG3bPP9CUvj9i4eFEaY12DOiOPD4VZzveWl9swa6v4qBVjZjNGV0G//jPnz+L
-          Wh5AIXikxBB/ZUPjs/AAW0+vwYbmCkyJsJjoydNjcGD70aVDVGQwOAQb9s1Z0JcHfxVhXey3rrP9
-          0EycMCZgx1N/+ou74wUN/ulrXSiyBaXwmYI/ffK01yeWtc9qdK0FSIKd78xq8k0gf0EjPt2FTCe3
-          wTbgrtdg/VrgYsZLEkI8xdsMJFjEK42uM3xZrEPu72fh0gf3zGARpCE5HQsTfNX0lv7539zckOMu
-          FD1ZsPOdP35GCfOQU+gGtwz7J1DFK6yXFOx61B8+p/QteiOkn+1DFLOvh9UAQQl3fRBfWutGl4vo
-          dZDyzIkovDLQP38GSi/lM/f0s3iSL0cNHgT3gq33+zEs81PwkHPJA6Jj7hlTVbWrf/jHis7p8A/f
-          GU9oE02cD8My8UUt//mD89QrsPgu7eGur2C9hs94lQ5Lj/7ivaJcQp0qi+fAYf0F//D1llK7/dMP
-          Z853JrBVxzGCGpv0OHu/mWb6DdMMdz2PuAc/a+iuD/7xMXytvB6sD/e5Qbq5NbH2IaX0e91yuL8/
-          tnb/Xp5mFoFKAjIJ8M8E81+9w7lkAY6L4BPv+qf3h5f2/+/o8vaT9h9+NFP7PKy3G1mAwZaE3ObY
-          LLY8ZkNghpcnCXhjBqukcDMsgk81A5hrO15REgnYp2IWWp0H36MbJ8Bi2GaGwd0b1iHccrDXZ/Z6
-          hUWnuSIzWF+Lj4N3c282+/bsQdvyB2JU50NMpctPk2eFoeRPf562L9wLEHqInUXS9zsT/Fm22lEK
-          uPhZxauYjwrAXx/NfGqvA00ttYSb5FyxIhtPsCl6oyBZ+g7zH3/bji+3B2rweeJAEjq6sjKTAFPq
-          f8GvfF3BErB5BOtL5GB3xx8z/TxruNcvgx1fFvwff7rpkTp3B03RucP6/cJjZrGkPBixvi16JIPX
-          2h7ngwR++/uzCjrqPSanp8XF5E+fWFRT3PV2g1JP4hb4to0EB08n3aee/wx4CwIXe+LDaLgwcx7w
-          NH2YgBEunrv5rwHCvo/1YNr50urKuz5DqDX3uz47rGOtwUViN6xoQt/8q6fs/HBfb9RQz1csxOfb
-          mVhZZ8c0pZuB7LGPp98xF8Dyu1QicC6/BzadkRvW4W2Pf/mQ2O3pREnPeAr8q08U8Qzi5bskHlp5
-          ViMPNG97PapP4M735qVGTzrz68IiYsk+9hB50WXnJzCR5nH/+5eGso2kgJd9zIl/Y1x3iz9NBF3b
-          jOdF7aZY0EH9QLv+R453Euv85LkiNFPMzlxHrZhiq7P++f+5iPv4C/2lhjqMDIIN/d30URzDv+/F
-          mqEf3c2+WD08W70W8GjVwJKAR/dvP46wEvTppzkGXA62gM1a6OkDlwYEL3274CAo78P4p3fvegKx
-          Wv0Y89hcvn/Pw378Ngouab4LeI01Dq41mAo623UOJZi5AXJeWN/+9IA9PgX8+/ZuRlZSeJSa1kiy
-          +7w188kME4gnrf5Xf52EO69AZPtVwMKsGmg10xxZhfMg5gAUfe0PSIP+h/1ivRv0QiibuwEDNvwS
-          Zdfzql0/Art+ScwdT/7awDf+Px0F4H93FLBmo8+w4j5gbARzQYqy9/TA431Ygqveg8idfthvH1+6
-          nWXJgcuvkeZ3V12a1XTPHboJoUAefNTqA89eO9gKgkx8ZLDNitCthY/5yQZQcfp4PRexhZTXtAag
-          COxh6wq0yFUIDezXV9b9/Z63BExcbGPtGJx0djQMA/7I6YMVq4AFPby4FhLCUuJXdwio9r5Z0FV1
-          NeBcVnV5IdEdFI2vR8BtmzFQCJ0vgBTUJLgLqJnajMywLrUy2K4T4y6r925hNkU9VpfxpVO6BDxM
-          cz0NiuaquluYhw/kRZKGAzdRBtbsNxOZdgVxWUn+sE3vZobgRZ2A3i/fgvMSWIEocSD2NOM1LF8k
-          5hB/fSaAzqVtFv+ibEgWu40oqJX1reYOPMy8ycN+eTnH1GWDCKZqdieRm5Bm/32GrOR7JNcAHoYt
-          fvw8+F5P2szjktc3NqQ8Ehy2wV70Lob3obICqNJLgYOv3rpsl1s5eiuxT5RsGoqBBuMMH112C77S
-          k4kXnIUmEpqtJYri3Irt8OJZyMudF6DqXgJO6y4mag/1CQejSwYa7re0o3Y54djiS5dk5n0DxvXK
-          Eu9QW+6qveMvOlr2j5xt3tZp8nxlKLtrMTHl4j3wQLY1dNcZORAcUQJLMT4zGLFjTkJif5vtXMb7
-          XJFLhfHvx+vL8bf0cM2tEz4v48vlW+nkwKNiBvPBY8RmVDUphZf1Os3QilUgyPmlRt8SYWLa0QkI
-          x99Lhsb1whKlkq1mm0+hCAXXkMnjJdoNn3WHEYrvy4OcZfByqb0wENwO1oBz/WQVrDnQCmFspOSm
-          jmBYBXYyICcF0sxsZ1FfX0nKIE/LE2xjeGq4KioN2BJfIY5YoWIWOTGEHp0F7KdOCkhhLya8NquH
-          r3RJhv2+ygzG76UjxeV0cYXFiRkofw2DnF22c+lBTWQUvu7m3mGgF3z+C1toghNPToF0pBPz5DyY
-          IZoFscMddeH4rUpooe5IdA4k+rL5m4ZWalpEOTSNvvyeUQvn7j1ixbqmw9oVVotOgClw4TGo2VRd
-          jBAWlSspXuWdslH5ZZGItiV4Zz6mhF3OLHoc25Y4X93Wt89dC9C1od5+xs2K2aisePjq/4+0K9lS
-          ldfCD8RABCRhSN9LUBBxBogIqEiTAHn6f2Gd4Z3d4VmeqkKy83VpNqiI/1UNb28dixIGxh6hNOnP
-          EcUdnuF8Uzyk3U/1wGIrymB44DvizuE7WhUdY9Ddj++//4+JfY4l1asTFC7RVaNHLnalS1oWeP7o
-          CZ0pPmMpEzqX2JOnDJxuK7Iky8MLHUXR9LgwO5ZQ9aqEPJb1OPBDE79B/MlCYvH7d0RiWXXhJ39B
-          5KTOS/urf4177lBYUkpH/oJ02OlJQpD82vpCZ0YJz3DZE20hNZiZXTfCz3DQiBIq2sC2h7iXprFH
-          mDVPHVhC8ZMetsNumJozjeasL2pIPnBHNLq8huXyXOZ/zw+y57CKWbJCtmQumHUvwUDP96iWahab
-          wT6Ut1urWxTA4DhpeMXmE+y17rol0E8NuRaidA2u/FsCol+gs3A/DrODhjckSGZQqgd7b90vRQ3j
-          hF2R+lgRILJav2H+xQk6rns2X9g5jqVwmY9/+EO/lXGWmOP3SgxJ0QHHzrkuvCpuQN7Nf4NldioV
-          Fk1eIMvnNY1+7rsUcn1kEHc9f7U1VnQWfJPgiiX8bbSFkkyA6phsZ76HdaCJjkRYuGRFgWJUw3zr
-          gh6GKeeTuGW+0f5xynp4HiMr6HSkaPvUbM4wJqFOnCpRo20+vKXmtTsQt7MCMPIXy4TD6Wmh/OJl
-          YD1LUglr9jARBA6qRt3ZriSd1ATvhmaK+qxHKjyOIMBMI/rRXjVgCp4iDFF+2CHAZ85kQu+U2MjN
-          lwMgTHk4w41f0Dl6+nTun/cC9vN0RnH6rZqRediqZEWNg6XuzeZratYqlJ9kQeZj7Sk5bF0WCkAZ
-          opc0ADy/b1VI78lAPFeIwQpuVQK250W2HZl070fUBH4etcTfX490Tt/eDPMvkIiivN7eit+zLZkn
-          iEgiPoYcO7OySvXYuMic45PH8/RpSyNoNRQ04hgt96MjSPtwOwqpI8Xbzz3vQuZ4wMjbi6VH8j5d
-          oTjvLgHjxdXAUZz60BNRirlOCzy+uVQiFO7ojKzHOc65oY5TYJ4YhB5A+lC6HocQZo/RQSqbArAo
-          bRZAmCMB77X+3tDFq1LwuAYBimy+0770KMzwtRhqMJrhSuerMWzzLU42/P7Q6Z3vVGixCYMUY+Fz
-          YnRVKb2dCaBLbJcapxrUlgLuqyOjuORRd2QrWYou24o8kBYwehGMgeUKCGWtevf25vA5S1lqjmTr
-          MO+tzuzM0nFvusTQBmfYJ/o9g9aAdVKcnxc6emdRAEBRDYLu8TFnudjoIFjNE1Gxaw289j6/pXf/
-          nEnWWZjSvgrP0qDIN8zL/Ysu53uiQqVuE3J93QOwzu4lgPl6DZH/NV2wOOGwgqx7ZejqsrHX+exd
-          lmx/xxPrkYjaH9+C3hsDJv9+tzMwtSv96kkPMjln2ZBL4EWOnggx2jwsxfh0gXcrI3Kqyitdxnao
-          f3qL2H6+5ovP5jMch2dE0rh1KCsKdid9DUbGVzqzDVV1oRDR46Uj+aWy2hpwVSr1KMPIiI02X/vq
-          VME9ml1kDZ+7NjMPpoTM+6IgTW1rbz2XpQ/ukq+Qwk2ZaPoUrgpPrksIakrQzE+qyJLVcCHywRNq
-          S/pWK6hdzj6Wuj7TiL+3MCDX4oSi5jp4sxU8bFCoqYxORVCB0fDQtsbSaiTa8zj/XlSrOtixUATL
-          xperObzVww2LMUL5iYnWxyx0Pz4K6P2kNux368Pemq1FTsXZHIbzPa8l8ogN4mLnNOydWZMl6E4u
-          8oKDTEfNtDvpltxVcsu8D5j3c9fCMjHOxNU87OH5CwuIrsUBKVV5BfPYjjMIuEOO20NFou/QPHRY
-          4vMR8xNe6HSC+/Cnh5Fdg0u+ykqdwV99XvC0UhKCdwK92jhiTiwfOZ0MZ4XuPPcB03DmwF61KoCN
-          ztyJNfW+xhqexUFhOqt49dJLTnH3niHPRmd0VMakoYHwUEWFXnOiWITNl2+FZrivuCvJ9V6m+/vU
-          b7dg1gIxo+cHzE/qyBAbC48hCQ/ewDBpDUojPmL+W8sNq2jrKskhHJGj9WmzyrVnglda2kTN+zaf
-          5MZYwejeLfKQp/0wqiZvAv8MVOIqxmuY3rncSqP7sPCtvB486lGfhZ7j+ThAxeitQnJSJTSnIrHy
-          AA+UBu0oBcdiDETnXFMyNMUb1pF4Qo5zEMAcpEcdfPE+JH7YCGB2iMv95gty59DM2WKsUxDpzILk
-          yfOGVbcdFX7yz9bHuK01YvlpDfsuIci8jWq05/fbmQd2UIlSZWcNDxrzBt0rrYNpq89FUQUTHJJA
-          Q7J7iZoBCbUMrlG1J9b3bUZ0drIaMG+9J8heP9EkFbMA91Uro1tbR8P+c/dKCOG2FeOsZvls+vFZ
-          3PCY6OlXHvYrDgrAnTofWYddTFn3TASw8UkgSbaYT9vnEOOjg9wNb8gzLgqo7xaTGOkRgdHsVx0m
-          K5E3vdJ7qyjAVpSkXEbHd+UNi0Oc7U6ms418cOei5SwxIxDn8oO0Jp/pmvdnWdrwIej2edrMj1Mu
-          wkLtruS64REBWVfD/Da8Axjc3Hx5g2cosh/9iYI5VPL9dyfK8FlzBtH4w+JRa/p08Ke3nDp9eRte
-          FGBOZxbZ7sUZvt8qWyFXD10wx+Mumo+JG0B6NONAvU6lt7R8WIOu2CEix+Mjwtg6itDBQkeCVhU8
-          Ynq+DnwqBSgRH17O7iqnhFNb68E04h2Y0OEFpQK0V4J8Jm3IWWL9nz5B1rLyzTw7zzMs365KCtls
-          G6K+TFfqqwIg9MmfYM2+xzec7epO7EPRDhSlSgK+ge4hEx+LiGz+BkqS+0VHpZsB7ZjGhA7uT8g+
-          l09vrfdOCn/8sOn3iAzadiIkqvZ4X1xATp5RpMK0kBqik+M3Wiw/kuE0im/kCDDwWHdJMXSFzCLo
-          +wqj1RoWGRbZWhMntg/ea52OMryPfYNcfLzm9Ps0R8g+HLi93+/Qr9jwYfDOJsw3V1MbEjVU4TUf
-          ZKRs/ozfPVYMnzVrIEcUPW22jxWWLrZ3IapxHsFULTT+5/der10+756qADd9G9Sj3YJZ/9SzpN17
-          HEiSHg+TXPuJOB94GXPMhwP4lR5KyDA7kaBH0DaLrE4mTMSWJXFJMRi5+Nj9+aVSrHNt0/8j3PQR
-          0q6mrX3XKVSZ+RufibxcIZ2z7tT/6Q+PsfJonsyAgeiRTMipU0Pbl6EZgGKIZRKIGRlo7hEIxMoP
-          UbpEk7ZCqIkw5vsPUg1XyVlYfFzwaTWEq6+72/KI/Qokyf4Gc9zdBppajwBql9BHKntRvbXbboYK
-          w1FBxU+PTUZYS4m49376SVvDzCjgDsk9Cc3TRMer/u3gZbHN7U7ExCN18oiB0kMBxZWoDayzGKN0
-          KJOA6MRptjNSuwTqn7Ql57cya2TUt65J5DEQ76dneXZggTELGrrh6Qh++hzm3zEJIHgW3iCrffs3
-          H/RsaCLay4sq+e7K4QUvUjNhMxx/eUCwnyloiPbO3lAoIEA6PRFvmczjCFZYcSRUnTpaHvNBhJQ7
-          j8Fhw+sZbTuOV1hzxOy04A8PpfnQfrHgPwSNbngknuE7QBveNxRKugkV9YSQpXY1xU+WSSGygjro
-          naRrRrlO21/+sPmPOMdX4/KG4dMn5JJ5I5hrzoih2uafYKd+RrrlLSMkH9whexvP5aLcTHAgRkQu
-          HOfm6+aPpbW1Y2LTYxvNv/kz6YGETC6hdG64NAaRcbT//OBqfAcbTvuTE8y7RtNmO+hteI+rFKnw
-          y4BZ0aEvZA/sICvcBc183H9LUAzPD4bb+HM7ucBiKA8nZGc49Xj5aWdw80fEaGuhmXZPpZeUHj+R
-          pw7XZn2DuwzDhdyRPeg+Xaf35ML3Wl6QE+pRtErlp4K5+MiRddLifKXok4m//Mty97LHSlLGgh/+
-          /+b/Vt8M2PQw5k51See4wi2kOnRIKAc7bXvfGVThuyaazUFtHZuKlXhWBZhe7EH75QvwGIeXP/7i
-          NZPGcPOrxPXiV7M+WS6Fy7nuA9GKDgP1Vi2TdL4HAbBXP19WbJbwy4gyOiXW6HXFUEH4y4sM5fXU
-          JmF/qCAbfAMiB5mRsxvfQTY4vNDJWbWcFxJ/ho8PhkTf1as2X7XKl+JPGqLNzzdzK+Q+bE2K0TH2
-          D2AQroEPX2meEiND52ba/LC4fU6CLY/ZdiCKom7GKTqFSjOsHTPGQONOHbI/sqXRMD/2MMDn5+95
-          ovV4gRiCmx2So2Rn+VKMiS0FifFGys465+OP/2q4PANBqtx80d6dCc/5fENeuHMGvmMGExwCzUIy
-          n3nN+360e4jmTNzyHo5SxFctHJ5ssN33esmp+kYtOMb1iPdbvZLJ1AJxLY6E2GHVel+HOC3wlNNW
-          f6+DRvdLHsJzvt6Ikli+tqDDi4G38nMM2OrAD99qhTN0knuCwn1qA64kYwDn76kmxzTcDbhEWiU1
-          l4dH1AZY0Z6L/RT89LqevE8NH2a+D2c7NNEVqg2ducj34UNQNaLziQymaKfZUJqMA7mjHYyWZXBd
-          sPEFsRf0icjNqgMJ33GOdOJoDWV2Jyg9s3Ql2kk7RLS9NjN8O1VMHj/8X4/TDBb3gpFuHjVvkdWX
-          KVmamiNN1B4N5S9eAvRP1hKUn8p8nN16lHbX/IgXBE8Nm34MH4Jvx6BjnVbefGQ7FTaTeA/mjGCN
-          ZC6ugD01FyQv0dXb1/ETwpodTSLH3a3BdVzhn/7DXcu00ZT3+Rn28+5LzPUceOw90N+/z4ODsRxz
-          7iNdBfiYLxz+Kt2Frrq9K8ACcpFYe/6bU4o1Bi7Aj8iW5zX19J5sqG/nz2U+MbY8RnGl6msiZF3s
-          t7bMvY3hcR8/kbW9b1riZwULWvjolGjvfF79uZW4dUyRRcaAcvWlt6Gv9gPR91eeDr+8RvL1O9Gu
-          7hHso7Iv4RMEn5/fyQch0WeQfw8S3l2dMyBOWIYwV+SJWBe9iljt/Ywhl8mYhHq03bLNv0WABPWK
-          sXO5DoTOAStwJ6EhJtNsN+znrxGk++8RpTVPmo6L9OCXhwXdSym1Uf9OPdwFc4QS4wQATWphhG+H
-          AEyhif/yEIATkyG61F689ZSfRFjnqxEs/WvMl0/hyiD1iz2S08nLNz++it0ehMSjM2kWO7gL8H0S
-          eEw3PTkuUxD+9B1BF63Pl8cJ9WL5tlWk4OkKSEl6CPxcOyLjKH4p9VgzBJ1UT4EQtw6YO+bAwdZ8
-          XpG/qzttLjw5lCinXvD6kvq8HVQngJqBioB8n32z6eUWbnxGdHqJARtvXR/ucZ0Sb88JEZXVugWH
-          Mg5I2WmtRru7zYAN7/BSBDJlk2Yvwh1Seyx6113+wu9DBvfh+RCIQsE3c1febGixMUOsR1RFvzwC
-          kPlS453Luh7/SudA+vGtcxRb+vPjcP9UmODECmu0xLLfQzkcL2TzLznbK4wvlsciJdrV7LT5yFYq
-          NA5qQLShqptlv14D8KRmgYxtfYQ9Q1kXXaErg/XBXYbNz+mAz1OFlHd6pTR926p0qrkbcW+E12qf
-          3nTYlswX+Vk/RSQEOAb4fbkga8InQJ8UjEBMoyHY3eMpmr6Vtf78Kl5W3IAl68JUOvWAJUcN3/JF
-          TDsXGt/jnnhQWcHqRwsHfDWNSZ67DaXd1pRZLaI8kMoo+OkRAZ6y4Y5hsdRaB4DPiAE36EQ7QNnj
-          3fAbQnwsdBTd/JAuxfh1YXDfZXixCBtRw1U7uOkVJIdVoM3W8Y7hj2/Oa5KCSdGWSjqpIdjy7a9G
-          AuEhA3euIuRYpM/p/gRlIKJuwUAd+GHlImcE59iGxOUSDFbTSxN4HQ5XIg+GSnn1dXL/8ipNbVVv
-          FkWvhpm1vILnlmfNJTrJ0mVxzWApw42vpF6EK/tVUfyG72jhIpeBpkMjotvRsPWJnzDw6cijZMsH
-          5l7d9TAykE1++e9iuqkLBnlWCQrgpRk7BrPglQ6EBGINvHlb34O/+ZKHO6dhJenMwp1lc0Q9NYQu
-          7unlS2afLET+uo+GFGPUwxO4n8ldEJpods8fERaAxcRsRC5fT2BtoWP6O+S05UQpixYb+pFzI7pf
-          NBFV9FcJ1XyeiO7e4nzd8Ffc8q6A3fzOn//+6QmqtqrGsyTMfvktOfbtEs2ZLaaQeZs9sczRHtYf
-          Hh730YLcEWX5Cm5dAsV563pgTH7Dvw6hCEKbFvjAHR4eTXRLhNL+vP+tX4J5v/Ty33pisF4kun4K
-          e4XVYR+jKyNXdCHf/Qg3vMJCOufe6q56L/3yTMNY1oi0N9zDbQs58jMceXvy3WOo0NeKvEjxI/Zd
-          eGegaqlHcnCwtN96HvAURUGqF5PhRfFpW5wSPaL88N7sRRMc4/OFyPaVa+ZQBB0cm9zGh4Mc0Ul7
-          V7o0uqRE5o0UWyNc04Q30wHIVu0xJwf+IECdnFPi/9ZXry/FlHSTN7E0fqocMw/AQHgWTeQ+IRst
-          zfUdQjllK+Q+cKXNH2lKYfC2E5I/ziQaxVvNSgwjiZs//HqzeHB1qLYDs+XB2w5bydfhwbz1RClD
-          rRlLPIfS5rdQET6oNg+a20HYuPffekZDxfTcSd10lwLeBzaYh/op/NZDifZ67aI5zDIGSiNfIxWb
-          T7pITC7DM2RfRA+rr0dKpIaQZ5svZjZ/uD9L8lnaXYcWb/6s4SQpjWEoGjvMpl+5+eE55KPHHVmv
-          ZzyM92lghbu03eHByDIdFZ31pXOhzsjevVew6I6HoT4rJRY1L/AWkDXqL88nHjhY3qbfO6gZDxqw
-          g30d/vLT6sDHgbTpqyn73lh41RAmPz5ZvzthBk7ySALmvt1xjm0rAd/3rgj2zl7wqOkbNfxAJ/vT
-          +4uQZOFvPAOeLlXTR4xc/tZnSexecFNt+SNczmGErupQNPP3kbVQkm4y0nVPjn55r9QwLrv5HT1f
-          fnrgi/kw4G4EepSjpQ20e4eJUsZOtAiXkIMl18/k2L3jnGomiP+fHQXwf+8oOBU6g8xKn5q1v71r
-          GBuhSJCcf6OFfQETjq9CJmdskW3FN+lhzsEH8tCLa+Zofyslc9EpKQ++Newb57pCemJnktVXIaKG
-          cAph1V0aZNBX3ayTf3rD5K4rSLs/D2DWRTpK/ZMLscguZr5nX2EmfqN5QPrNMbX58t1umeDNHhmf
-          aPBm7TqvkpGdS6Ks+RHMZ65koRnbY5An24rhRX+KkJinY9DULRqWA4EBIHh9E+3i1NoIn6wozSPn
-          Ieu49aW+ma8OSK3uE1vKp2bOqKfCsnLzYNe1lrZ8/PkMV62riA1OfoRHMJowB7kfdBa4DkTQGxHe
-          nU+BqQM9j6/Jo4KX0QhQwrB2zp1cx4X4ounBfNeX7RIgQ4SPU4OIaSRStI7HkgHynvEJOvMK5dfv
-          +S0JZ6Yk8VVBAxUdZ5ZcIY5IitxDPgtioUIuE2WkJeEUrV1UvSV6Gg1S7lKaj8QpZuHko28AmVed
-          c6+gU6UasB25nAezWdl8z8C1aMRA3Iswmp8z20pSqFTEqMN9NPEQF7BJQYbMZLrk3LFRGQlb8Izu
-          eRMBMmxXXE2D7CGVuxfDvD6mAshvnSfOwdCHcX28Sqj7ekBinw8G7shKM+QNpSXn5z71eHdxz1L6
-          Jfw2vm9vbXIIoRmfrsS6vbFG7aCWYVlFJl6pP2sUjjqWwiSztz6DZs5lfQaB/j0NSHmpg7a+0CuF
-          4pFxSFAW/sCzd02VJlIKxGMkzluVsUphKq0NKZfdPJDqe2qllFEOJLRu16jrPLuCV7L/kkfXwYHe
-          pDMLnre7gmSwbmcid+dWejBxT3I93YHlU8oZrLprQ1w65x5ZxR5LsmPPyP/YTLPVSyiJgwqITKo+
-          GkP1XUitIx6JcqfffAlooELr3TEoUroO8Eu63UrPFAaJVisaWFaRM0n/RgNx5fidL9MBJpC9hiFe
-          XqqnfdtbX4k9F0qkYPPnb36dITMxN9zYvNtwVFgZONE+JEqKOzA37UuV4mmOSPYM6ggvhtrC8TCr
-          RIWfO109cq4k7/Kc0YkhocZm9IWlx0WCRBeYZ8Ne73cZ9k8IkCGybU70T8NKMJE/JMg9WdtL17sJ
-          +Sk7E7WYr83+xoa61OixShyXyz2exZEreZdzFnwX4ThwfPCUJWWor8h96E+PT/P7GX5COURW3AjR
-          bHhPHx6nwEcn9xx6qxqaOnxyNxtd40bIl8LtOAkV/QuHF/qKZk94naX7vbeDZcMDjsjjCG/EuaKj
-          Y4QD348all7FsA+wYH087jy9V+j0+Y0c31jW5s6Ta+kkiBmy3tJbo2kmvv/qy7PQt8G3tLKl5fyd
-          CHpnBmD3LVbFUxgmSB9xTFkr/M6iGcsKOmYlzamMRQ66EAzEi56Kx82NDuFBf0gYmj5P5z5UTGmb
-          n8SaM2ZYj3UVSrxM6oDd2XUze4L7hourdHgF9d2bHwOrShs+oZv2wsOqa1wNf99HeLafnPtyMicN
-          A74H4nFyNNbcsWdp3Xu7APpxkNPHzeSgUfIcurLJ11uPdXeGtyYhwZK0n4FyIS2ldL3CYCq6eGDt
-          5yJIROEq4sawB2TDU0kWzwFmq/QDRuHrCfDTzTt0hLmRz6gyXREU9g1FvXACM7ePCsjBM0sMkdVz
-          7BxjBkjd0OLFDV85ViWhgrtoCJGlyHE0nXknkLJ4apD8Tb754h7EBPrN50Kc2zmi/G1U8a8+SMmc
-          Om1NLqiH/c33SGD1b7r9vVIqKzsn2VS3A060mAUbHiJ9Z6vD9v1ZiP1zTPwpcCNOXzoIKT22SPFu
-          IF+e+iOAU2DqRBctvsFf3vRF3Ao3dErnaNgXB5DA84V7EmVEqrYW924UCdinqOTVd0SPN4OVipmd
-          0F1qQUTru6vCQyElSFMXBuDi3mFYCsD9q69JHThV8tYHT9CGN0uneBkkeH6T21EuciohNP7+je6d
-          OHrdIhkulHWsEfW51MPcM7MguW30QebLavL1+spSyHCTQ8qTtcvX3d1Wf/yEglkw8v1VmXRIGfDE
-          vI0SjXLPTpDMtMTI3upnKZlTAN329EHG3YspHm4eB9ohH4kR7y3wx2eH1ycnTvW8DPwjVUrppprv
-          gDXeqscf95UNmTosUNHsP9G8ewm6dLlSH7nPcPJWy1qwdHi9cqK4cUbH4QlEcJIuFiodtvOoWsut
-          9JsfrdH5lOWvdJYYbOkk8OVYm4OAU6HCniyk8GUzrNwoj5JRiw0qpvODsu7nbkOrlA8otK7jsEhX
-          fpZ+/JOomunNH1EaoZb0GB0n6Qn45OO68NuhlJyxk+Xzrb60UrsrUpR+sJ3z34y2UNmxN5Ka67Ph
-          5Om1iiNzyQganiePOzavDmpJpZHLnh7y5Y7NUirZl03us9VsfUNb8ze+mBsipC1olXsJtMDBLZo1
-          j2tc14ZjAd/oxExGzu/USyGKgc6igr372jjQ9yj171uHF77UGp60cITCrIWYqbrtFuEDk8BdVL2Q
-          My5qvu8UrAPYx18kC0Dw5lMUCVK7+2ikdFhbW9uxz6R33xKSRmYD1qn86BJedh5x6fhqqErTETay
-          PZBTKcYNHtKnC7/RIUEa7zr5THLUQitw2mA5PU+U5I+wBLFDPnjPWKDBObAz6aiPNKjM8JnPczYK
-          oCws/Jv/EXW1Kwd7w3sQSzvYlCsbD/7h/e3q3iP85ZpKYjVzO3NS0oayRRVLn27dIbk9dWDlg68M
-          L9HDJUHiXrzZuMcyNNyGxcKztfL1yO5nuBvdL1INZqHz+/tqxZv5ZQIqIFfjeguMov84nYkrbokq
-          yRcILZG7ELe+CjlRa/stqcVckdtZsXIutcQRfqzMw/HrbkfrxpcSqzIBsW2ogn1VHSroP3IJt8NQ
-          guX3+wKQFcgMmzhfaCyHcNUOHOaE/pNjxj+38MfnfsXcG0LaTIZvBQ7o+sBt9HWuZ1Pa9CMy9kQD
-          vC2dOImX3zyyxJoFpLQ4DLNuRsQcIuLNjn4OIKOoEdH9/ZYAm00Pw53qI0tnnGHRYFOCo44pMb53
-          P18jJvZFx5FfuNXKYVgZ4GZiYgs9ud1CR9tDk/Tg/m6aQACnMR/lKBNFH5oYuXvOBDQdpwwE5zUl
-          +gkijde7oIaGAYvg0H89OtOyM8HGN7i7nAldr69zBqNxqUhxrkSP3qBgA8KHMvE3vbJs+gW2rLIE
-          63MveGS/s1u4Vry06dEinyOLqOD7KLoATucd6H588z52HVEG95mP36lawaZnkclw0fbzzRnOI+TQ
-          xuceH5ZmBi+f/okUZnrlxHDfHOTlwkdpZGqAl4JJBp5yxMQLSi9arObZSz/+DjPBoOutlzoQnOd0
-          8zOfoXt3lQ/9z+GKN35pGt4w36C/BR5SSuFLf/MJzt5bJZrN982UkDX407+/elvs9FrC2n0/kO5O
-          Uz5LzywQs7g8EuRPFViel892JvR9JX76foGZlpUJz/gTYrHw/GY7NplK9G2vCLkvN5qbqBDAQQ8+
-          yPeOKl0avhVBXvd3gt5lHhEhSTCcyKcheqRl+aI/vBmmL/uBtF440UW67maghzVEf/WUUW/T2ycT
-          c3fh2VBrHFPYPJsH8c9F28ztPivhIa2/ZKtHSq+u4cJN/xM7uxgDt7B3E9IkugfcXVAGdvo8bDAe
-          xCngTSMGi20qzA/vyIa/dDHbjwofp6wlqsGcADUEwEDR2XZHcEYHlr2dZ3/4e54jMMwj1WVIGv1L
-          4r161WjbIAYqZyHCogg+ObWRP8Ic3Hx0+vG58AhWuFbtJ1hofdA6XTzJkr4sBnms8xnMyZ0f4Q49
-          hIDah0u0LiZiYHcT7iTPDwKd/Mk2f3oSeUFRNXPcW2doxu646Xk3Wo+stAKu6Wfk5nYy/PybOAwS
-          DbJ5D8C60DUVWda/Yr78NmBhX2EqPo4XnthyVzXT1Ck1XJz5TCzfkwf6BH4GjlPCB4KyJM3kY6OA
-          6klv0HF2JjD7zr78w7tDeY2jOdfzAHKPZMISyESNlty7EscXSYIdUN2cHt7fEnrcsnUd44C33Pb3
-          AojnjA8Oetc2C/b0CtJjPWE20sSICkk5ioeCHYixX/CwPj9PDl7VVQha4117Y+Lr4h++epdp0kbr
-          pBdSbLgIx/7hlq+yK7sH0emtPz9HL0mBYWnRkTh6TyISpI0LU/uqEec9cM13b81Y2vw5SVRjBvjn
-          3wjfJEQ5dDqdsaC/QfIaTXTZ/AadrbsqnZVlT6xNv7DhkoTQelfRn/+cvnI1gsqEBF3MG9IoO82l
-          pNHCJUaXKB4nLAOEKyt8UYS5j1f2251em/8gR6nN89nand7ghkUl+Ja3nUf1j+KCR6925OgxhscW
-          73SGow3E7fc9vb/3NyNBRyFYFfrDT9gZxAwwofqwtPY5BbpvBsQZlzpa9Me7BB2+uMjMcRXN85Al
-          8H5iPOTsncyjnwuVJabOrUAUdKlZx2PCSKl90QKmX2/N4lwzE/a3mMP87tUP9Ck+R2iV6gGhIhbz
-          8eA5Nai3HXCIFZycNphlpWYyF4Te2YtOR0YU4SlsRvLDSyrL0wjXvRZv/q3SKL8eZFAK6wMde4XP
-          F31m3oC/mWHADgMDPmAkGVg16iGzMoVoyb+GAHYf7oPUOcqbtd52ULuEO+Dx+jS0WevNM6xy20Su
-          uIN0/b1vVqhvSFvUSJvFgo2h099uRCPg0yyHORGk/eXjBeJ5/Tbzd8ed4c8/XurwEq2cBKo/v3zp
-          kqdGbfTZuhRuJwyMI2oWDQ4ljA0bkRQgZeBlLLIwpDEihR/jnNhBrcJenHpU+MnhL0+SNv1OXIrn
-          iLrq3v57v88U22A5zKUAH5cdDA7rTgArA9QUGqZlIT/uQU5u0tYl4tmLAbfc52iWX8YIl9Lf4eVO
-          nXwvyl8MDVB/yJHok4fV0DShOfYGsUzN9Zb6vStBeBQx0jZ9QgXdCWAi2hTJZviMlnsuQCjF1xpp
-          oa5s/iXgxG2+IZTZQT6/r6EpaSRckaJpOF9ubGjCJl0Jsjf+XYSHOUv1HCXkmBX2dqJGtMF6ZWqy
-          PX80//KISq3VQNz03x/+VV5ebHr9qHEH9VQCc1xnzDDRGUyndzSLy5vXka8xisdeicVBl+AA6T6P
-          h4m7NTV4U9NB0aJSbZEFo/3zp89tfFYij/jwKhoX19LI5HOt1CZ04WFAaoV9wFvM1ZciWLlE/eo3
-          Omd0Gn/8gDwex3m/GG4LP0IXB7uH12qzfBb8w1JeehTcL89o3WfzG46vKSHuw/QAJ7YLhhsfIq2u
-          WzBPn6stvl4eCn742QxSzoFR1FKiP6KPN0mJmMHr9fhCfrkuw1gFcSzZV2Ah5dxJ3tJmx+CnF4OB
-          +1KAPeWYHiqT0wk6cglghzp5w+cC89/399YtHxTRoVSC/Z5olOJobOGdEU9E15sxXz8qdSEx3TrY
-          3ZRGW5+dXsFiL+tI8YtTjjNozJABDodsj6+133yXVPvlEH+NXEChFHOwW+uaoOdNpBN2n6y08Q+6
-          KGcN7Lf5CyHuBxL0uNWINVm61DrCEa8D0Jr9zjNq6OYVj06nZuva6E4sZEpEkC0kR23T9wU8hbkS
-          0EXu8vl2OhbQMA2LhDHWwVIyt+CwvCWCd/5U0dl+LqK0+QeUCONZWx4hnmHVak+8eorV7G/mq4eW
-          0/okK5wnxbrG1AAqfIj89SR79HyKXemxLimybu/AW4ebxv35q/PRs8FeOEZQut9XSnzxe6XL98z4
-          cDcnCuZ3n0u+vI8ZAw9PGyN0Yn2NHD5vDkTjdnYvCephu7FbgPZolugok2ZY35VUg/C4asjL91qE
-          d+UxgxpxLYKKOMvnu5Ov4Dj5PtnEAVgk04fAOscQBayz0PWy+j00x84g8etBveWBnhksxuFC9KtC
-          GqrWwxlEXjoihNd7M275CHxOXYoPkyhEM58mJeyMySRo+n6aGR2F7qAv1EBWp77B+kJTCt5Ud0j2
-          SbVo/9NPN7OJiJN6drNeeSKAaxC7yNz4c1ZaOYb0JOXET7u5wfWh6KGWwTnohJQHVGVLHX68XROs
-          WTs07Mvs+8Pv+aPT8TpMoRZy8BXuBOQzLzVnGU2sQARrl6g/f4yCPBH9pujR5Xsfo6kR0FvEUn3H
-          i9Do+R6cixp857eF/NMt+IeXy3uvY2Hzz+vm98AvH7Q2Pfgtd8kbRo1gotuP/2zpxgL+pofIW1oa
-          katwDmF77iSiBTuZ8q6aFVBaoi9eXvqeUp7mG95ee6L42XtYXsM9hhtfoi1fjbBbuplo8NcQKenK
-          RAs4LexhvcIaKaUYD/OmX2A2XT/EPU5fj76NYwt36HoMyG/8FeutwlSaG6I0khvxJaex0DDRiiXu
-          pHtceaIsRE8rIvJd/WjUb8Xxh+9BiqXewya/7dDZ9IH6m2/M1xLhlq8SZ5WwRnZKqEP1dAz+/Orq
-          QH+FhsEUyHSvdvSXT/x+/qcPudwP3+KdEU4bX1rDsq5uB2QxPaCAOdm/+bZCqUsDZL13IaAyTm34
-          OjxOyJWTRZuJE8/QtXT3d0IDLHlnV+BxvPJIp1ISzdq32bp6zQUyJBV49E+vfcsnZtq28X56XfoI
-          fUzcrR5XCUw6HII5QddtPGk/fGV4Cm9KsFyHFWCUhCOc1MwKZjurGypVDw5u+pNs+iefxuSQgsuc
-          n4ipnyCYDKU1JbhyGjLz9QLotj4A9DA7EbOgT21W9mEp3bKxR/6urykB+TWERzlbUKAv52gajOcZ
-          zF6rkru8BPnSEeD/8j4kHx+axh8EjgFOb2uk2GEhmr/YK+DNfEbIKHSnmcH0XGEOzk/kkFFpcG8n
-          qXR3EhOzWsXl0+ZHwDCMd+J7xxosZ5s1wZZPEATMB1gZfx8C/zOP+Jc/k7q00kMb+AhZ5+sYzSiI
-          EmnLnwJYMVLTJXce/8Yn2B0kB/A//6LxIo8UA4k5XbYz6GFSUXTvhGc0bnk8lFrTD3jt0NHNT2TS
-          Ly+6fE5yTl31XErzyHrkx9/jF79duOUH+FWH+/yX74hDABQSIObr0ahuwz/9dH58lIg3DmUPrN1J
-          RUfhkdF1+C6hZL17higvuYowkVUVnpXqQe6zJzSTcOQ6kPp1hmfI1wNllg+ELxvZwbzl8cOGvz++
-          C670cqfTI/R18VSYTFB/9QNYhcdiwy5TbGQ8nd5bQOMwUFMDBSlb3krukS3CG/GumEsGt2G1ryz8
-          8aMziWlElZPEgkmtV+SMdNKmXz2XOZcif416uojydxSoc0mIc3WlfNbFmwx/fm2rB29bv0mlH/+Z
-          5NRGQ0xhBlEaN+g/AAAA//+kXcu6sjASfCAWcpOEJXe5BwERd4KIgIpcEiBPPx/nn+XsZnkWHjFJ
-          V1dVh+6r79UtFS9QgzT45chni/2GfFzUkv+0v396cFyXyLTAH/8wU2Ebt7a4x8BChzQkO94vQl6V
-          0oj0iQT8udNX/qawUv4rX8R/HkswwrFR5FZxR0zl1G2n17Or5MS7Txj6cTv+6QH4ObMKFvxg73qP
-          vrm8+1kokJuo5T01uIOzH/yQoZc1Xb9Va4E/vnBKnxplP9dtgwJ/nokG6riYBkaUoH3dSqReZ7nd
-          6PHRw5MjNCjgE0y3Q2Xm8p5/iAmOHSAVj2u41+fC9TrLIzGyZII7f0Cmde2K9fliJGnX8+HxeELF
-          +oFzCR8BqJCz+0l0DDtF3vGcOMNX9ZZbpBhyUT1MsuO9vimFBOEKtCJcnvInWSysVnA6vw4oiNK4
-          IH4XKWDg7ZiYXmkA+vKOEhRuPsB89Pq1vK8omrzXB3AbWfv1ozFL4TKbr/CTC3xLdrwBf+fl+fQM
-          byHF6QM9rDokezuht/35KYz5UFDE9Qpdjae+gVMTnsPu09h0fuk39ng+DWeCrC+nT3/1ZZvF7b/1
-          pbNWluD1SzMU/TIn2XxFUSB1wpGYu/7cft2CgW/9QqTOuegtjPEIpfvs796Zn3oc80MS3KYbQuap
-          Iu2OLznkn88c6eGhBv/w90OvT6IHUl1Q5PMRdJhu2f3AlK4f0BtwkLoDQixk6FwxQILiokYkPE1r
-          O/35A2/9aKGTZ7sjq528FB7t9UCCoqW7LD9U4MhlJnH2ehx1NtT/PzcK5P99oyC81neiYNy06/sW
-          TrDOfwZmfe7XknG6aNBg0UyMyI/asRCCCs6bVYac9/voqwqPH/kXfiJy0isPUPZ7FEF07wgJKcDt
-          2rmRDcO4iEMZvlUqPPDvDs+TUSIt9GZvNnsnh8bjBQPmOrqAK2DZAZOoPxxlB6Ug9enewURNKPI7
-          RIoNhtFHrofZJeZmnkb+8aKdXBdhT9wb9wJTKOkNXPgAhAyiWkLd49xA13YLZF3c/Q6Uwm9ysLAD
-          MpqIH9fVdQbIoXQKaSRsCWW/qwQZ4rF4weZv3DxnGoDweYWkMDaDbkx1zAD/Lg0s6c65WG6WE8IX
-          p/6wmA06YNHTZqH1Ujj0jPTCE6JY12TlWsTkVATYGwB97V2kIUe870lOpo67ifBaQjukk9El293R
-          KjhOjEKSQif0l02nWmaev5hc1VEs1izRJngwuTfuLCWmvHAX77KsoIlYjrfQqajFUApL10Dad24A
-          h9e8kv1P45HwOg4UL0bEy3k8yMibcoUKWQA6eFaohvyMbbzV9MoaYmGN0AOlzihM7PsO89NmIu0t
-          oWJVrjCFhyyykaNcWLDVBtXkGt0IcW0v/Ts/GE6JnZCUsV4jP/64Ep5GdiUxeY+FoJC7JZNLj1Cu
-          l7rHB/wWwcO5r8iFdWG71GmUy/dX7RFD4fOEP4ZdLG+ZZaLIiF8F3zlogNX7dMYv8fWkm9HkHbw/
-          KwOLxbvyhEq1LThvRkkKvXe9yadHBY6waki2MTdvQVoBYWiWKam+UC6mIOwNeR2WgVTlHHq4YvoS
-          Cry+IMTev+NyTEAOp7d/IylMsLde+McC9eNtIdZ6f1NyquFdZpzcRE/BviTCA79y2a1Yj8T37qpv
-          iUcmOFuriiXngrxlcFEGoDpZ6K7AtycYFGNoPFpITHsSvSWzSgl64xQSI6vNcbly/R3AXwvwojfP
-          hN4vYgRedueQwlfZYpmqyJW9+bqGQmvy45wr0V1WrrcYBdHBBGvsWZpsdu5G1C8ZR/L3eY//WsgO
-          W6CvzwSJkISwQrduUhNep6Ym13OMiSPc6oR9qscFDmv/QQ9MIZhnzYcwB+IZb4fyrfdaUklQ3vCF
-          uK/ZTbavDSBcX6pI9OFme9waER4GCz+g0/XyHhcShTa0yeOJNE+4AiEZWFe2Eo8nNnuALX7CpoO6
-          z0ToAd8q4E5iY4NrgjLkanVe0OkcZ3KXNy9Mq3tGV7f4faB+SToU6Bex2GhMJ9m+2j4xMqkqlsMm
-          aLJKVyGkv4V43KKuEYCnQxJKyc/QKY/9O5Tq8BZe1EffrmIe8bLBBjN6DqFD1+gnLjB7cxwJP7nr
-          sZX9hNLYohBpOLklLMDDHZ66X4vc8OmMHGsFErRNy/p3vrZTVbJAYUyN6FdWSdjPz9zkTejncEsu
-          p3Y7XykLZQdWxDzZRstd+McGib93lb2KK51vdr/Jt8boUHT69fo0VbkL//BOn0+ZJ4zTQ4PJ99AS
-          +4e8lh1WrZIBwhJSxdeBbnybVjCevQ+6f3OroP6qVvJaRDbKP4AvNl9VPnJz/OTEBsl5FF5ogPC4
-          nFNyufkD2JKt4WWd2i0yHLEotpuyGTA/fCVk17zabgTad2hVDEZGoZ68TcpgD8MlPhO3hzdAXrfV
-          h23hKOSknN7eZuRaKh9L50TsxWgSzHPA2t85uiDlmbsF64Z0+xf/lwkEOnf0fQypImnEunR3gCPz
-          EcKtVXW8CCe/FdrhzMi/b3Yi1XAcR7JYYg7VdKMovIjXZP1SeQJ14ffhrW0+7RJ8JxEKdaSSIOFG
-          b/vDkze0UuLCEer0e/ltoLhHKkpqM/G4yXB5+MJXEQtaut+Z/WZ3AJH9QmcmTYoVKydD3kaBRabM
-          ie3CdHZ5BGeqEicZJ0D1XJgghmKOUHaN9En90rvM9gMkrpxy3lZvaQYS+6GSq8SKgByuWg8bxtdQ
-          xshDspjMp4d6MqhEURnfo8tT9iGd+UN4HO9IX1z4MCCPI47E4FIW1FkvroyoeMBsJjHFksztBF8w
-          FJA+VQ995bNPL8/i5Yfs3ly99WncMghyBu148aajaNMKegGHSHjZu2az7ziFJGQqLP2d7/qpZLJt
-          U4IU8oNgLWDawdbrKxSlbUKXl21EsoUqP4QH5lwsibAycllcPeTbfTeOenyrYUDzEGWEOC051Wwu
-          /wyhwvLruCbbDSe5jDKokmo1nWISvVGD7j18hcW7HTwKr1kKj/W2ocJX04TbuLchX0ZVRGpyjYpN
-          75kFyqe7RPSrdyp4YCm+zE3nAV3DpBnZKv74MCbEJUVxnQr6kK9QFj5tiNzraxrX+/WawXAFLDkx
-          cd+ultRr8PkQTliuGrfgvzaF0E67ngRdYI1CSo1G3kKJJyGiTTGF+x30VomdULw3frLMhzmDosr/
-          kKskkz46OmXlY71sKJkq2aNEEiTYGs8IeQx/GCnzFQy5HoiL5fd6pOtFmQy4jFcdi3Vgtxiu4wCj
-          wE2QEorXEftdMsAjRgFyljIuFhwbCpyL+0hi+qCUuNm9klv2fiHqZTRbwa3rSY61JSDq533WV0h8
-          A15AVZDwC+UEf989D282h5C/hHVL2U2xIf2hC9njM+n5hUJYWuSFwpdgt+v8Uyy5x96MzOvt6lHf
-          tTJwaKcO86/W94SkniVYHn0Xc+7Ta/lGxCz8GLceIX3ri3V4mjV8eFFESh7ihAp8ucC8eGLk5YrZ
-          rkrjdn/rifwWzyO1f/4C5e4iEmOMjy35FTCC65yGSDXUPKGd6SjS4wxVpCm93bJMp1RQAW2FhScm
-          LQ3VIQY7XyBKjKqR+mn6kR/kXpK/88MOH+jD6P4hyIsZLWHfNFVkJmNfpHKeQcFe7msu9/fSIEHi
-          b96Gr7oN+3cGQzDDVF92/gckltcRUrOCEkFpDSiKGggoZ52S3zqnEzig7oouf3gfK4cNWKj00R8/
-          Wlt7ZvYK6oiUuVzG7YaLO/jm5RFv61EZhcPvFQNeKQSidrlB19/Lu/99HwmVh5tsrppGYM9fpMr9
-          acTnK2BhTGaXeNN2HLcItSWo96lijLSYCe94QwPl21Mhpk9rQLFphXD2RAepfRQU622f42n2W0kc
-          5HrJv/jvp8ZF5jltEpqitYRHSbGIWQkW5RT25cPPaOZk5xMFK5X6Bnc8ROgu7l3fv48cHkzhTU6O
-          Y3u04tdF/vv9XmVKCcmBGsJQkxbMrLgp6OHq9pBXbgJydn6yYgVZcHbSNzIEqQVrwNQQZrYohKua
-          uMXmHbIMaO7BRMGl/ha0heIHqosRovQdGIBVpVgDPuLvRHMGr11BOZdANO4BOh/pma62snXwtaYK
-          SUq+HZdcXvYpao8I6flF86adn8I6WPxQuDSbvkLvqBwFXl2I1vdyMmicJkKtxRAvXHrTCV+8Nyg8
-          Lwdkb+f9RqA5TjDj4xx5h6fY7nychXcGfkjsJsYoqHbFSOdu75nzyV2d+7ZqBNgxrJHSHj2wxOT3
-          AX/5JbSap76dUquDTRjLxHu3rreww68CSfBxyL94IdeJ/eOX6HRcv/rM06QGbnjr0X3tOm8rb60P
-          hcPWhExVv9qtYFj777yRjAOnhJ+NqQFn8z4R/zycWi6ErSWl/ongrfXVgu+L+wc+CQiQu3fZ50vI
-          RJDzVxsF6vr2ljR+afC5HtLwOKyZt40/uQQuk83hEorC+Ke35BGWDXG1WiyI74YZrL4DR9Ixvo1U
-          j881dIZzSWxmENtVq0IePtla3PO7UKzNwy2h4HTRrg/Ugmu+0b98/6c/x74q2fIPb0lU1CrdcvHX
-          w12/hdQ+/CjFZhj+w58LsY/J0nSrK+c/3kPKvn99x50lGZYfeddXEV2TRY/BNxEQMs3L6q3Gpdn+
-          9BpybY8dl/bQ2/B028aQ7ToPsPxCGalKdI+c+tMRTODMNZBD2URcz94dV+slysJJmxD6CA3d8cSF
-          Uu3fUEBKDlC33t+pHJoD8Sg5tcvSdBusjhgQ23zl+grjZygJz+shJC9/r0hgQ4JFYlFk10E/bqPV
-          5JAzvxIetDpP1l/G2UBEZ4MEtSXTXZ9gEOLQJebexYP2S7TBc3MdQmkgQbJoWlOCF6f/kLbz/e3y
-          XGrZPkgass2X6C1LM23SpBoBMsA2Fdvf+c/y4kZMsm7FhgbsQlH6PEn4+97aaXboICut75HnXTyP
-          6482CoBthMkp9ka94/qxA69kv9FNkeFRPy0/8tVcpPDt/SydP8LgIw3G2dzxvR+Xp7pu4Lf5B5QX
-          npmw1fGVyfip3pBHrb0L/qFL//gnepxvTEuWJxfCz9xDZF2vYSHcj14KuO1QIF9KLsWOp7nE6fAa
-          0me3FpR55RKUtOudeKJ9BkPnnHr4bKoNGcfsQOnnuWiyZZI7cp6q6W3JRYbgLz8UznzQt0b88KD3
-          3yZC9PimW5dcJFhcpwd6qKvpLXWa5/K+PridDCNh4TQ1cMO3nNiPY1IsdD31kBdqlzifiB0pt2gR
-          XFNDCtkftoof7ZYeeE9w2Z+fK+aNmw14swX055ckw/T8hn/6IFzPruHRL8YdXOITIcquJ9drkVjy
-          XOQjugXXK+AjNJbQzK9HvHab3FLNVV151y/IZoa8Xdb6I0K7O7nIOVaXZMNU1GBTkFvImF+Drqxl
-          inAITjwWElIlpCzdDaaXq4GQN0R0+vyCBWLDbbBE1yXBx/WQwls62Cis31m7toXxgSBwMuK+jLWY
-          rHnzpZ+WSAgFl1lffv1mQOZFbsh3uYjuPRg6GE0Dh6VdHy4ddxblM+6Vv3gu6DLoFTQcNkH3KmuT
-          5W7oqfzWs4rsfoc+VaSwJM9UKuTX1w+gRifa8PFIMxSE2nXchFfW/Z1fhJpF8v70NGScu0m0P33/
-          aZke4ORnIO0GKo8XrUKC7xBTpKDXsusBpYLhtbmHyxC8PAoj0f0Xn3/8dEoihoe7X7Dr1zxZDlPO
-          giIxKFLB510s0ffswz3+Qrz7WdgO+jvc/TVkuQ+Nbu2jc4/cJhch99ONdv//trw6vkb+9PNyQqcS
-          jnmakfw5fehK700u//EL+8AmHr43KIbx7HzCtBhqfTVEmQW7ng5lxuE8Og9PW9p73SN1eNzGdXgG
-          NdzxFWnq+dUuI/E76cNLZ3TH3wtdFs67gyEweeT/bg7lfD7vwCyBAIWqpRdrrRUVgFzGIq++kXH9
-          Rk0vb+7ZJxp9JHSrkRrBa8nY+3lZ6DQJKYTPMWnD1VDFZNebOWgO3BuDPb6my+v6gY7+nojjY4Zi
-          MRpsUFPmiUKdPXvrzq/gdCtqdKrRmmxf5r3P1aURXn6drc/dPIZQ8g0fqVvXFNPNcnxZLOeamLL/
-          HJcmKgdIi+2IkKcKYLs+pAFej8DCTOI6Ba/TfQ6ytgQojx2t4E/AwX/5Elkbc9TJ7TNNkF9SSIxa
-          H4pZ28YMRMwE0L/zs+tfsPt3IVcW92KMynP056cg7+Yq+qwOMZQtO2aQfx6+7Yqkdwgr5WBiWeAO
-          lDKvSJTb7JUh517r4z89PNdKRspT5YC/eJXPZj4hfz1vCVWFoQJ60qso9oQrZfOXJ0JxtnV0+dOv
-          kTulcH3p4szN3cdbhtUt4ZwGD1Jsj8s49+G0Abjs+iThPJ17JidJ3v1SZBHyaxddLG2odALG7bc8
-          UgqZWwW3zDBJxZz8ZPbNUwjPCz6FqyRPxZzZbCXv8Y58aph0gs6t+vMDiOfV27jnZx+el+lErre7
-          1G6zlPlwf++RmEvMesuUqpK8809khYxRrB62ffC+5B1Rn9lrpOas1PKn471w168tPdUwB0XzvoQi
-          gO64XO7HO4i9jiHlVvh0WyS3AeR63og2FAuYDrfNh8M6fEi44iZZIL9t8DTyK3FlGXsL6EoXjNbO
-          HxQWtTQL6AfK2hohf/zw3vjHZ7f354uPz0KmVM8PGOY/1kPGOfP0VeqPe4+w0EWngv/qv499jwBx
-          hIwY7itNuOKmNvKuF4gSaVtBK1Wx/vwVZD5mv2Cr4y/78ysw14QBINEJbrCfape4DtN7K/jQCbZk
-          GYjDL9+E3/WxVNXPL3FSH7f9Y80ZSP34itC4yd4yVbkNMsWwSKY4AdgYzdj9aP2AvPjyS+Yv5SYg
-          O0xFdEsD49LQ7wTYx2L8+X066R+SDSEe9ZBDSqBzDt9Fx0rwrsQfZFnfpJcmAYfWK3pu3wTs8diD
-          nQ8hVU3cZIumKIPj6Ujx6l7mZMGxr4Gd74SH543qZN9P+BlPOTFO9FPQ17Z3gcbGgOziXembqc/s
-          n7+HtOgI9K1i4xCyrwcOD7nbAsp6gfKn70gAiq+3VPYV/vE1Yn9iucDD623Ag+FrxD0sr1ZgN9v+
-          h7eKtU9N/YAuh97zePlvvWTKfiWQ/VojqaXEYGnugwiEM+iI3dxZumr9w4LVIrlYrq0H3RYjZ//y
-          MTKOB5EuDutjGOhPB+34kNBCMMs/fUh0P1vAyoLD5x/emizlWoyXEwO77Qj2CpdSsBvaKhncvxum
-          kRAX61ytrOyrTwadfuc04ZtLsYDUrXwslpfa226encFxggq639G9pfNtwFA4H7sQ2r0x0jHJQxh3
-          GUcMLNaUblIX7hXCjaTui01+x+wOYVt4CrLYdaL04MMcClt22nsifwqsHm42fHhxRP7p34o/bjCO
-          I0JulB71GbuXFFo/4YMP0rMsyHNvxpIplhWC+PIr5tgL9x5y2oOYvnui23U5x/JenwgH6fP1ZqD3
-          PZS7q4j+/Jm3hxUfHv3hTcJv0407vpVSs08FUbDptGyUSp9//smVi8yW97ukh+51kPEfP59KP8EQ
-          /bTTns/UcfXNky/rl3NHFPIr6fZ5RCV8jueWeBWmxatlrj08XRz1Hx9cPOnpi4c+DUMi2JfiL58C
-          hYvO6AKWdiQj38ZyDdcI/+l1Wp9rXhYTiez+/ntc/vyG47V9hNxRXrz9fPBwjxekki/rYfmB3OMg
-          Pa4YfBc+Wb6hbsGFRwDpUqQC/jGpvLzrVaQFDatTyZI7wD62vQKPjYLHPpRAogtXYqEjMw4nM1cg
-          kQ8NsrQD1enE0EE8lSkiesm37eaVXvTHD/HhPffF8vhqGPpBxJOrdK/1hduOLPz5yoS0kLV0UqjJ
-          AhePff3Te2u/Tw2Xq85G1hJ5yfqTxRg+bDyFDF5ZHf94p5N2fypclWEtVnWIGfA7jPcQskTTV3Of
-          SsMQhyWGI4JkNUSOhUbdGH94034OX3WS+69YI93pG7p9E9eHdfuCyKPk2y5AdQyw+0vETLUooT/x
-          Zcu3gLmGAgPUcU20H4Tpj8FI4++azjb3Zu8BHdXo/pqHZPF/zQbF2dVDYQq1cfVf0AD0M67oz8+m
-          xv0nQdu7JuFA16VYA6ZnpMMyNpgZjmO77fUsaUv2qcZ1Y+ob/r0W+Lr3iKD9vE83n+R/55+4O1+i
-          hVpskHWpTLyLLhRTQ8kEjql8xEyoYIB3fISsevDRSbZDnftGw/DP/3TZOPYW/5k2UMeNiMuvybb0
-          HAUuVLj4vOONTBfkN6E83W41MtCI6dL7gf+3v8jIJCah10yH8s6/dz9w1ec/fveXv0zf/YLl6dvK
-          UTgpEzIifxm3yLyEcMeb8LDXx6Zv6BlQ24eI3wFSdJ7y1iLzOOaIW7aILnN5jyA3JQOWeNMc6WpI
-          NTzQPVv7PJuQiX3nEKBJQreIO7Wb9HIl8FjLFulfTfWE18YoMDX5hvj58aqz54fLwM88QOLVN9Sy
-          lmiXf34ZBmz0S/b1j//VO/70XxPcgzvsPMMkF/7z85Y8k3s4NT2HMvqNParaGQSHQKEkMH9zuyLk
-          bHCvz5A/fbHOfuLLu1+FR/hLCvqW9pfG1ihHxf79kzVL/j9/TJFhm/zj26wcf/CgO2tBe+cxSGBU
-          Lru+sICwwMEAZfep//ywccpadZHdH9GRm/wMb7FHmMGMj3Jy3/Fu44QFyk19NVCw1885XQITlJVg
-          Is4v0ADHNtsm73qZ7Pkg6YfVreD5rmRIr+4Z4DrhWEJhnsrw/T06lERcZkFBu2l/fKDYzPo+/T83
-          Cjj2f18pcDaShB33vFPsdusArOgWhkCaw4Qrzk8XHsPTmTgVXlpqwJ4Hurt8URA5HmVDOXfl6xZ7
-          5DQRE+B2q3qQdLaHwrBh6XpLuRzifNCREpeyvhXXSyfp18pCSDwu3lbOai1j99ES9Xh7jWvyzG0Y
-          87ulHRbJiBuzY+Exqc/omj1Eb9U+ogt/j+BEsvfwAFsklyIMnQ8l1kPkwfouiCSxuXlBYTGO46ye
-          jVAm6gPixR0yfd5fr4B03goUlNJWLK/jV5LbpE2xfF1BspKDxsvzspyIdlx+YGsvJQZIswJyPzpz
-          u5BRyqCD1xQhJQ1B//zsbTiePxXvs7/AtnmZCM+KwaCbvBpAgNCKZe/87onq3XC7zUj+SCWt3sTT
-          T1NLWUOPZE7ZLeyarvpKDi4LTPF1Jrn10z3qdVCRb8ZDJeVanQDXXOIMTtcxQprn7CXJReNlG0wS
-          uWtLV0yBtKXgZjxVfGVOxch3b/8D6/cykPzRKB5Vg3MIJZU6SFsZjXIqq/lwGr8hOplmk2xc2/eQ
-          c1KEskbgC/oedU3uuauHNMvt9Y0aiijn7/SNnFLqAZ2LxZLPNlsh3/964DNt4CNz0rEnil1VHms7
-          vgi/qcwRZWjikZu1nytHPVJDqS57sMWNuIHHjatJcF1BsUaLUu5IeMMC0tWEVTxzk4O4K1AJjGxk
-          LX90IZQNjpxDP9P56YF66JTpFWkhCEeuaicW+DHbkDwsXF0IQ9TDi7cp5D7cv4BUECyQe4U+0Zub
-          D4S/9Ze/N0ISJ72CLUHDACCiR7x5zmnkcHxT4PpVXGIo5prgq5q70vfpZ8jhkFwsvTxa8iO1NKSC
-          4JewM8k7+HmWD3Ij50pfL9thgZ5VuOGxAD8dr9WsQUmPXHSTb+m4ALGW5DMQRZJfchtw6zn/QKbv
-          fyRldT1ZUGAbUIyTazj36JVsuiN+4KF9K+TM5KpOwdu+y1n55LCs8XLbo8HVIGsOOfEWQpKJdhz/
-          F5+YWfJlnIX2sEiTGM/I8uuXvp4PYgqP70OInJ4/6RwrTTa457ZIiid7TgQfEx7aT+mAtOPiAHoa
-          xb1NtWzu+z2CbdpoB93t6WMmrVNv3cQwhRCtR2I9p2pvElDf5b6CCJ358AfWX93WMs0vJXKjehjZ
-          fIUMrG8nj5y6N243Zf0sMMvwFSXjAYzLOTqwcMcjZHyWi8c+HvgOH6mhhUA8Wi0/v38NrG7vGwpu
-          RBrXNf908nw4myjgWTTyRtLd5VdNVGQsdVGwVD2w8DpxHrHj9yvBurN00OiEOzph75EsTg43GD2U
-          Ft2fD8fbJCdT4N173LCoxknBG9Y9Ay1rryiILy3lL7FSycYnLVCZBVJLfqXYwFcV6yQ1S9XjYPHK
-          5MPZ1THjXWV9X6+PfHHjQ9imvw+gDXeIpEevKSj4RmrLsUofA3xQGBQQzW5nWPwyWfjCB8rOT6Vd
-          tNO3hkNVqeHKKAsQRhjFMkOFUwiQmOhcckst+cW7Cgra+dhOar1iGONDjG7i/aJvr3xj4BcKK/KG
-          2wEsnlg28C4uF3IPZt3j70M7ybHdXckZlw1YxWOLZZX7EmTKqgQI9HhflrkOofhqvgGRC2yAMIti
-          dJHVO1iCK5vC03jpiM+bV28zfnkKXWs8E+863L0tGE++zL18nzzmUKFLxhUhzDie/sNPoSsSV359
-          5ZoUJ1xQgQlLDSInnfATp+rIFezFlW9voyKu0vkJN8O3ciyl2SDpgXM92gCQSY2aVUiD7ZpQ/CQ8
-          TJ/QJLeL/aILeTMRFBF7Q9aX58CCToYGg/xyxN9xsXTemU4D7PBzQO73WxcCb70s6N7fiGRZpejD
-          XZA3sCI3QL51Lovp0TYlfIY2i9JTZFNuUevpyDuGTaq63IoVNIcUPC3pQhRx+Om4/AYVdNAgEnU8
-          e94GbqYIgreeE8RbfbtIJlVkdSjOyLx6T30b+TMrs4Z3JehadCMl+ZmVSeXdw4XnX94Cto6HKFM8
-          kn35CxA8ldHgwag6PGh56gkXP1zgiuwAV/LrmyzN9GChhxMLS3VpA7oK60ceFb0IFyG57XiwpJJ/
-          N1USShlqx8AoeCljrxtBwLQ9zrctXn6EPw25J9ICyoSlAoIl4MMGwBftf8tZkVk4LOF6vgaAraIy
-          g9EPayj0caAvfmg3kK7THd0Ze/bmvNdrGXtMh5nbM2n/4TUZ2CMp22xN1vvwy6C/MgxB6+nnbT8K
-          FbjncxS9gZ5wxmP/PRLzRsZRa8B6ja8f8ONMg1Rb17c/60l7Wb3CFMUP/e2x0M5Z2AHP+RefPD+n
-          pbznE5T13GXczz+GV62OSdw9Cirsvx/08DuEwnTlxmV04HYsfJyE86A7Hl9+zQrKr/5M0iDs9UlQ
-          DnfI0H2QJxwNfXHNzwBnfR+Mm8HEW83ZnGRbRy/kHAJYUKkZPjCAxxSlEu+O602gOSSZBJCbx2tL
-          FZfmoIj6mDjDsEuC+uXCGjcu0tJjADjslwzoqDQg7zpIXscYUgc7OLjkgVxfFwiXbHL6GxviV6TR
-          V9NFE0ybQieaur/0op2tDui/o06QeIy80h/d6hhHZYryRDsmFGhaDgtBDcLO1ZdkJkdrge++MpEz
-          c5/25/VfDVbizO370RaL8UURcL41F4oktjx27UsLKl4DiaZQMaHLAX0ko+PumCmHsaBPf5xgvZUO
-          iQNIPCobtS/3j0cY4ifXgPUTXtg/vkV8hzUKnntMHayPU4/cu3VL8OeF0794QnodsJQwa6xAmTVD
-          pKoFaokiv2IZPuANuUQV9LV9az6Ik40j1gU/xy1InUwWTOFAAi83PT5UgA9I+7kjpNl8SwFKS5iO
-          drufJ3ukyeVlyHIvLXgOsq5Ye7r28unc25iD/tbOWg0NWB9xHwogTT38XSsfeAXrkKQOUrBt5zsE
-          2qsYkKZflZYPxlMIEbxZyBQ1sPM3KYMfpjuGt7LJWxpFcwW7UpmJ+jOpt11+VgYix3iTSBOJvm1s
-          DuHf/nnSHBazcsrv4C9+ThU66ItFfjZAsnTCgn7SPZoV/AC/rX1ByuVg6tyyLpu881GE3HgGlElz
-          C3IS6MPZ2we70kDKoFsaEUl/wKfCpu7PkyQBcoTikqxIjmuYvUTuHz+in8cC5fdHNsjVf948VsH6
-          Bq+i/kMuuZ2S7Xj8GDC9zB2xJyyDbfi8oKzx/A8flx57qwSdBv4qbBOtraC3nMpEg69cMHDhnVv6
-          PFSeKAWBV4SHZzCO5El+GO75Ea+m/izW2Jl8QNtWClNsdt6SrywDvZPooUeetN4ySD4P1Ne4Em9Y
-          Fw9/3HpvqiU/0G0rOYqXFdjwEY4aQX3WjMs7LEroFx8NqUn7pv1wtVjovumMD470aLficevho1cU
-          knXhJ6FxcPIh0RiZ/OVv3szcEED2ekG2YAQeVYfEhofAN4gbXGKwTtsaQirCIynxWBcLLrVKrrfK
-          IRp7u+trx26WpBe5gRyfcce1w4wCqsK4E/1xnwsybxwD37RvSXzMb+24rnINeqeNyB8fXmc4KxC3
-          7AtFx5/dkq0bS7h9Fh1l+fFa8LSwFpjoDcbsSNJ2+c7AB0nzc8nz9PwkGxuwGBQeikO6vPSEI6XP
-          /+FTyB864K2X4ijBhwkadBoXy2P1R1LKUI0BOn3WT4spY2xQOhgGMuaZAuKzrQR2/MBre6Jg1fTN
-          BsVhmXf9p47r86MNsvwtCHn8RhFsZltGcM8/GL7cW/JPjxzNy4pOZ74HeO1ehjx1KUHPigPFzjcy
-          CDJGJSi/d8V+fhT5pykDuX2Fb7Gt1azA29muQ0bKyEgebiDBCNoHcnPoRldnUzM5Yy8b0q1j3dJ0
-          wgx02DAl7nD/0uW93irYQumM4Vp9wWI7hgjqMz+F4s7flzLga6g9HiPmD9AreGgKG3wvmYZXEDgJ
-          9a8vBn68+kmeZhIm+/+H8PfyLEx+m6Tv7R8yeFiiJwpWbtn5Z6WBnZ8Q963n+rLiyIVvMnUke07V
-          fv7OA1CYdxYy3+8P/NMTnLJbPG4zFvQx6yLUX06L4xDglj5/cw+vPQgIoslrH0tXSsCarQlv/DvW
-          +4n/fI7b+6yFa1UbLQ0FFcMa1y46Q9/RBdPPF7gu8xehtzeALVi9GNqj+cDQYbvk9wVoAH4otcRA
-          d8UTzp/zXcYe7Iju1QjsfFoD13OPSfbRsLexA9+DjxEr4S3ctoI6ZhkD/qFY4UGz+XEqF9jIp0PS
-          EK9SnJZrUBWBna+gkzOt4xaEQfSn71BxHSSdSIduH/NzFFFoQMsTUEv3tqwgR16iu2DJhVWT7XOu
-          /vkR3iLHfg+vW+ShTFwpnUI5ciHzKD2k3cSgWH/3bwWNCJTEJOe0XTN7XCCTTiOyP0vVUv88S1DT
-          ziGWVy7yVtM9TVCI24o45/2lxU6zPpA/vU8olF+nYs095wN3/EI+e0bFVrDXWuY428W/++mVLKdj
-          ysgy90HoL9/RI7fgP/wIed9eKMkMcYFz8/6G4IQLsAmykcGB3GRkLKI3biv4QHgwyg6lprHQye27
-          AS5vGhFlkBqKnQn10uvStEQhxyahQpYz0K7rB3oQzR7ZQ/i5Ayf5CESx+TDZcqwy4E/fGLs/tBbS
-          LZT+/IKdD+q0u1wtOM0WIuGwqa3APzDzt3/kedXZ5A9vYZjSc8jo78zre7m14GHgnnh2/Flfam6x
-          5YsbHZDtlEO7qHd5gFHNNOFhhmtCf+XSSCf90aNsOUx0oy+XhdaqjciuL992bd9uCNm5SXd+dhlZ
-          bXjXsim2Z8w958+4Kl2cSe33EaObUHDJxilD/4//WO9Bpr+N0XJ4b6/Zn76lvQmBD73E+oVr4l3G
-          Wcze019+RcEcKoBtPbk/Jn4UktNFrdvtEislZMDpEx54yx7ZPT7lT8gpyP+lUUudPo5hXNYX4rou
-          8AgzXhjIinhDCvpd2qmOJRfm2DCJC7bHiCFeeeAGx4WY7/ENhtud8tIfH5IdP/CW0PiEcPnG53D1
-          c5xMTzm0ARvxPJaQI41TJ35L2XrzAQll56ovQjam8L0PWvadMtWXQ9qU8rmeCmR0p7dOFOOQwUJn
-          KNJwIdHtAY93WPp7CWpIrJH/4yvG/JGQLj21Qpja0IDP6Lv9lx9pD8cAfMh90LPPmnabHqcB6DRW
-          kX3tarrFzbLI9kEvQvP7DFos1FoOGjWtkGu2I1hmP+7kvmIQ7s6cUHQCfUcQN9caufJSgqmF5vR3
-          vpC764Xh2PThn38Xruy90Rc+NRvIejee6NkDt8u+P9IfnwyOh1NBf0epguqVSVH4+LlU6IVygB2B
-          I3HvldM23dv4yNFv0vb9WtrpT3+l1/CNdE9U9K0VYwwTdw2Q8zraBQufGYZ9FhJiziYcqZHgDYqI
-          v/35EzrG3PkO30uqEUteaIEPlVrLOo1Ukuz4N+74K+Hpo4ZiuR3a+f2tFTgSh8esJ+T6yteKAqq5
-          vSIrBqL+BVOVQ2i3d2Q2NdZX9arE8ObNNT5M8kr//FE505YbSrFpeOxMoo9cdJ/on16jk2XYMubP
-          FV5/Dz1hG03l5R3/UbgsCt3jPYfJIevRn/6kbf/24clgfv/iZ+kFosCLrtSkkMWPh+fnXEol9E1y
-          MamTUJvrMjm9kG7Ht28yx6iP/uKH+Jt0SOiOB//8s1vW3ZK+5MUP/DyrR2jWk1UIf/zWNCsBabzc
-          FP/WT9M3tOOn67FODheoa+qJOEUrgte5lFKYMZ8EC77gFNTgu03Gv/cBnRRLAwt83zfw8ZonZruT
-          6VG+HRUYSftLnulS/ONf4G1dv0iRz3mycMZlgOvm9+SMjbhdtZq1JPb2ycjf+s/lbVXg884mxH40
-          tb6U/PKRlcm9kfDIHAqKrONwjMvmQk4ZZyerzvca2P1jlOhZ3m7nIdykfPoa+CZu74RCjw9hwNwv
-          WLxf+2QbnF7789dR8DqewZL81A7Ir+FMrF9gUoFlkPvnPyHbeIx6zxjSB2Tlg0OnXY9uzPtZ//k5
-          xCI9pJiUKIQ6Z76IoR6PHs3Ve8NTU3yG4tfGlJsQDcH1PGBikKNHqWIIGWAnPUG2OrQjjbJfA/7y
-          Q3gT9H/+97/409qq9BYklAZgnDrY+f3Ro2OShdL3QS5Y7tGrWNqnvkDzlnNkx4uEPWUVIymLkZNn
-          rqX6dk8PC1CGMUdunMYFftxADu/e84b81bq29JrWPVAycg/f7l1s1189NrAtYxb3xsPTf456jGDz
-          C+9ETZv3OIVuE8HtfDGIgc1OX4TjK5WOqzMh30egnbhvMUBptC1iXBTiLcebasMgvx6J8m4Tj34e
-          IgTyY4px2WJ15FvxjqUrznwsLUpG2faGpv2GQhNK12QF9MPy7D8+ypFfR7dZG2vY1j4lt12/0PSZ
-          SmBco8ser6E3cXfnA7/PMMOrE2QtNvghB6dr8UHaGfP68gsKDZ5JcEMW63btH5+S9vjFzJ1vd33c
-          4n96+UCTl7ee4l8Hfu7iovIcAm8V/FX657f7NEbjv/U9HoGPgovtJWwWP0LwuesVBvppGqnp5xtw
-          bTtGj5VbdCrW/gL2z2P2ByZA7/qSwpAcMDkxzxHM7vnlypeSQKIlzQOMVmn3UCrlT8hwpE7+9AKs
-          z+yE19vepEj8FD7c80/IllvmrXxta/AZG5Qkh+pFN4991VB5lhLRpoNeLFl+i//pkfj0ebb7+bLg
-          KDZsSD2x1umf3jTCqgoFOBrerLJa+IdHKP3Tm4oLcoicbCLBI9ZGvv6Ymrjzk3CJ2TnZoHnYgGFi
-          BtPFSfTVEysJAvZbEjS4Q0GKpA7BzpfD155PiR3jFPpYOYS0ufmUdb0B/z0/Zh6/gdIi6cM/f42c
-          rOPgLRWNM6h6EYfcL1q8JWmDGgRZfUPXE46TubWVWOaVrA9hN7+KPmmD5nhWLAYfrR54eFVHA+7+
-          IEJRFHg4PjID1A4fB+mkaf/qA6LIQ7f587foismdgcI8A+Rnbxfwj+SQwZC7YxTcTwPtdz0CfzLf
-          YZqe+XHLXksKrpPgkZ0P7/6vxADnLmQofFNNX19Nd4eBeMhC0Yu2hEJ2ieQd/0LuQk7jWmXxJkf0
-          9iQpOY5gcThmguciKXY/+6Fve31JcmsN/ePTG+Zud/AJrgf0V+/CBfusQdszCeaOt9HDqpph0Duv
-          iDjlPuZWlWkPeFbUke0YUkHCEA1w/5vs/ljyrx62/759f5uW3rCRAW1wXOR9XIWS64hrUKaZg2xZ
-          m5I1UBYemmYpEBRFs7fUsWQfH/FXDZfvVQLLRZtTeHv+KmSeDmrCHnndkk1/zlB0VvZXbkTNlav1
-          sa9PsQKiXU3tKKfMrgdfS7HraQ0S9QnDRROJt02Ws0FeSXs84STSqSL/YvCMLUr++MA2tfoEvXf7
-          wZ/zsx6XEYnsP72v3cQ52crZaeBACpl4WaiNrNSvivxqQg050ifw1rF1UngQagbd34WdsO35iOFL
-          7d7/6o97fciCdj7U4QGnr/aff4J+0Pnnl70L6RzK+rW0ULA+DnRVcWhDzXnNyHIuTiLkXVTL+TFq
-          MVTGuf1xUWTL9GzI6LTXW0kRdDYsM+eH7L3+sY7SKYTCl3kgb2sFb979NBmoL+1ffaKL+aSX9XVV
-          iLP77SvoQQ5uzCohdDzXyfbg5QyMtzxHe36hlBkfUPpCbkXOe0TF7ge48HdhenQCU1esx8+jg8d9
-          YNzOt9r1qD8/8PYcK2IPyael79Ax/vQbqTiiJHzt6Bb843MPQJFHb56VwXWiPQnP8zT+q7f++M5E
-          KW6gPkT6WYM7PoYLJ+xNGjWm+vOb8eEwCQm9jp9avrDvY7h91U6nubr7XUN/QVmY9nTzvG8Jcyt0
-          cCeMG13l8eaLWoZ8dDIbqSAy0xmw2pgrMjI2A/yMuM8/f9EHpQWW8H63ofidMhSz61D80zPwZwsk
-          5t+bt3CvZwddlK//6sf0xiq53FlfjJfDIyq2+f1qZKnJPyTY9Sd9TXoJdr0f4prVRn5ZqS3/+UWe
-          +OKT/XkreGu/D6Lxspash6/EgP/nSgH3v68UNM7JJOH3WCerPG/D3sfMJaasmzp3ufxCONyvMQmk
-          JE8WkwcSvPFjHAr8oCeCbWUfmaw2Jq64iaB/jZwIMlmRw9J9q5Rt6lEEobANSJvhVedOBzuCjVna
-          KDSsfRLtWixQny8llhV/AtPb3Cf1nhgHg8tTaqeewhpuEzURuirzuPigx3A7wjd5msoC1uAyLPAQ
-          hDpxn+/Mo1ZWMcCMnS9CdCbFco+dASqL4RH0oB1dHpKqwdlIF1TxdGqnDIsf+QFOUlis5ptu4y3d
-          ZHWQKuSJBzpunp9s0C9WhzzvtU+XurEreKiebChqVQUWUcY8bKUoRuqTHIpNuUgWSG7qD12yKPW4
-          482s5DhYVHLlhzZZb+5XgquVqkhrUayvn2dSQ2OdEuRvPqGLUkHpeAhZmSS5VnvblYy+XPQ9JKnO
-          cx6FoqaA4ehNCDF3NqHaL6pkh3ExsQSjKr51fqpB/nocQginYGQD+cLI5eHpEKXH3fjv+cHA/ZAb
-          VRrlXi7OpDru1VBcKm6cMgVHQGzUDQXng1Zw3eRa0HxeLeTOntEuv+0JIXd9K1gazAQsc/wMoVpa
-          BrHkx1dffu8cy09+OBGLMUrAXVO/h+wDxkTTUV5wsQg2mZwOaUjF0i621JV7eLncE8yRCBU8448G
-          tO6ygKXUcxK2nuNKfqmnGKFX7RUckH4htL6pRBKYZToNCm+BDTVguHFOCXispiHc1O+PKNHlqa83
-          RSqhtKgZOTcXUuDJXSCMZZUQ85up439Iu5JuZWEk+oNYyCQJS2aRKcjsDhQRkIdMAfLr++DXy971
-          +p3DE1J1696bpIoVU1GDkdfpOP+Oerhemu8E9anqkHpDJmBuEpRgfmOxd0xpGGeKLn1wVaMC2wWv
-          2jTqzpTY7r6BwhfuQhaMyIenh/TEd44yVWaapEq0WHjyYHB/FvvlyS+QIa2A8pMtF7RyUBbwdTDO
-          69EAjDJPAvB6ysc284rC6SlcFZhmTo48jbAqOfIN6tW1xXf/Ztob0utcvNsstZCBPYMp0h8DZPb1
-          udDYL8jaxDdBlM8Tj5Ogke3voLMZ7NiYQsYoDscd+y0S21T7ojudJYTGgFiw1EsBG8TyG0b9TAus
-          Wh8gVLph8bVXxoQqzysoimBNPjZtUFAsnDt+/E27ul2iywrbDlfLWdFUwMaKWMIR1x+k22sSEj0F
-          FXQp6r4MJVUV7PYtWsHqTwrWkbqEe3xqWtDS8xOVYZWDVSppHlL7CyBljuWCLiy5FAPfvSPtcTWb
-          46SeB6Xk3aOLmZLieF9e9J7FsIxRptn0cnE9WH/Aw+MqsbI5IFmBEI3Zwyu8IAKdgMUBHvmP3Md5
-          abbw8+5EceB8dCHpGGKLOBHUpXpe9g5mIyPvYQCX0RKRJIOwodPB8oDDEXVhzs0H7I583eFnmUN8
-          B+Wl4NbjlOxi18ZCx9dB3SoULaKZn8pl3p2mIBabmMB0CbWwyP+qv/wEzPWjenTzqIqV52xPjM7Z
-          HSXacC9Wy36bsHyXtAeXaR2ZqCe5eA0sBjmXKQs5rXFreBJeBwUjT7Imu5JDyuAsVGLi2f1YAAq+
-          5kuLLmXLFUTQ+AHybxTj6ylj7DWosAQLm/7gDIzA3m306cRnPnQeJXUeWMTC42FxZ2skhWANyZu6
-          9eD4vkiJ4s7ez53kwL8g5ZeV6/1xuo/cAz7ZoVr4srUAjfl3JK58ouEo+17DedpbS1yROfzD750x
-          ZEU0HfGNEzN9ES67fnxhudfqsjpyDfZFWGsRxxaNvb5iQzJu75qnwcfAt2scqbtBPxY4b6cWuX+T
-          3rCU0xjC/jQ9bNe83DR/TUqBW3zXvUW6/417E+yLqFwmBufFaIZ0p0IKyGVTIM2FbcF+TlUpOo4V
-          4ICGTDOdTjILFUPIkKp+hWIR3jdaXEoioKtObc36xZ4Jjcuk4sdFr8ddincDPvRX6LFIfZNNfcmG
-          +KtXRjIkNtEebwFwBWqxgf2akL/mBeEcnwYvW29LuMaC1cIlTyrsvcN1XNPBcmBWfWvsjZ0BtswU
-          ehAtDsQ+5Ukjwwx1Lv69DHUhf9ew4N4OU0F7/ZsXmrmq5MjnCYZ162BLKT9ge18eNTzz3GmhkDEV
-          2HkJJqSF/bv8AW0D5A7kDMa2HHo9uDkj/e5bQ+zZ+IMM1pcL8suP0EIIX8L0Nvaa/lph4XwVj82t
-          uGCWh2SJHOeWC2Gua7j9+ETjzytKWC5Vl0RgHLEybzJy4okJ9+83NsWU/Y44ndoo3D6tLIhesa9Y
-          4jzF5i7aOxJP5ZNGJSUFIeNlsIVKzQfoTtto3NcT18JclWL8jKceHOs7iedB/qDo6b9sxhSbCd4y
-          9ME3veiKVdPTVezzeETl/V6DTZlbHjy0nsWvXChC7pcfOVa8Regyt2CzgabBjCsHX1jOGrkqQ/Uv
-          n7DSxQrB0/ebQBTfy4XR/DScjniCKKAux5HIupi6i7dCr1hXdAW3yt4E5VtCt7nT6DX9UWB7TGov
-          Vp9LtsA4MAiXhJb0yzdv7cFINhGWJcxUaVkmZkI296UyAzpfYvzwtdjX06mFzEZJOC8pKeReD0YQ
-          /e8eLdzbuhSDGfMVzLURYTvb92Y97w/6XGr9E2tVk9rrWL86GHhtiV7C7JI9nW6ZSOhi9Xpqs8H6
-          1EMNjsel2sd+T+3hbz8LMP2TaJQKVj3uGz138IffSsHoNheXDwVydE4tQNWxuhpMGsG7Mxb4+qHs
-          YvfowoTZ+3XCkiMrgLWDUwbFcZw9mO2dvWq8/xBZv7gtlN3N9ubG9Qp5P+OPvm9ds+hnW4NWbvXI
-          vZyJPR34Bd6TvyCdHR7h2om3QNSnuvuH5/v9VdNgI8ettjnjiwk1vSXejHbET+lO2/uQn5NzdQXA
-          Iw0LwtVRnAwyG5SwD2LO3oO0dcSCD8myAhOD/Ql9R9SaJUE/fNmuGsrACBcdWQc/wgQecxbKysfX
-          77iB3UZzCw48QMrtAZvd2+REXL3ERdbrw9pbrXMUlJrZxMrVrIoFTQEFl79KwjLXazYZ1dSDZXJL
-          8fN8/x78R13htlE7RlPOjZMq+dP5qp1ibAQibubzI/XhweeW9MhPAq7cAp/FU0U5/dka8mgkVvzh
-          2eWzPe2NfWU5NL8bjyWOSY/Jd2kHVRDfcUmqqCHp41TC29rV6CKnNJnV11U7BzeVxZ50vzREu/ua
-          GD06BaehZRV0m6Lj4BJ1R5evFqvzfrqZ4naxSizzSLDX1r1N4qu8WcjSjz7jUwhWYFfWHTvN3hY7
-          E7U0vFBiu3RCmoT7374JwHNXghxPl9U9spgBynIMj/UqChK/ZxZOo0iwV3X9uJtvq4bGXqWotO86
-          YJa62SE3qBT25DAhq1/bHUSdC9C9y+5k05E0QWBlMb6Z4DOuKeBo+A3ilyeAsbZ/8S824Ox4bDKw
-          KpGBpwn7/FdjOQQNIbv9pIGRn7hldTxL5Z4TGwBX3yR04dp9XL97CmGdXpSjryxV4Jct+eBmuT1+
-          VPe82As0lFBqmw92713SbH3M7gBv1uJdP1t71LdogVza/WFbPZXhd5qk+qcfcDg8xpDIwDBA+tsC
-          9/mtwMuYtJCFhYWUmmmamiTXDFLx2CJVVbSRXi66A5993yx8fdvC3T1fK3AmrYn9N5jDnYeTB4Ql
-          L7CqIpswN+ZLwemw7LcQNGBfnMWDPMw2dDnwhQv/1lU88BKZyo1r5rfEC/Cmrx6yv7xkH/huQcB5
-          E1b5R1/s10dkwkRxBu9s2BnYaU/2xI5p3jhBT7849MQEv1SaYXvYcTEdfEY81hs5TCCNuypuD/FU
-          eBJ6xFNccMUSsHCyJgVbyjjb23a5ZWJH+D/8fBKRdOHVzGBnzS4yKYqoG/vyc/H23k18U2Km2T/X
-          mBKP+oUUvZdV5s5UOTz0B74c8Tz/pUUCEuVoKm19p4J8JnqCu0Qc9NMTtFQbE4ye+WvhtUEqWFbO
-          ByDVwujBv0BtSG0iC46bXWFZieNxTvWehh9lPiYrjsk46YK7QHCcaDJMKxvJ49s+IFvuCbatPT7w
-          4prAL64p7Gynvtn6YKJgMwUxcpgzW+xnTmQhK+w+0m5fC+xop1qopXGIvXs3qLt7mk3hfqcBlsrC
-          U1fLE6BAL/cOe9jo1GlyfROuQ3LB15TywvWaOxR4MrSKsrSdwlUX3AmmDOch5/La7PXhhgGAS64s
-          HOL2og0tW4JGbufYvvJdscbsuos//RfcFE/FJnwbotPDL36xRLIZpGQtvGomv4DuY9rb5y9c4D5a
-          NpZf+FVsNwkq4hZbBb58NlHd353G/4unh6fL9krnTxrY4ZagUqiIOuLHvouXsvaXEyddCuY5YA2q
-          QgXRw9q/9mxYQQKbl8Z4Z+U22tuBj0JwXLE59HO4qXUuwJOnVIt47vwC7+fKEnm5K7F5svNwisVH
-          DmKbWZHx4+v1MzBEaHCNtzVKWuxhwNDw9l5NHDdsXazvURRAMubqIlZONxJBWwf4LF7qT78X+yfC
-          EObB1/EqypMa+oxVRXTrvxRdRO5qkzXTcnjb8pO3HHwXa4dlvZ7NN7J8/lYMslVpwHEovNR6/1Zn
-          X6ytc7mHNyTTdThOeTWX//Kh9M4fdT/8FhDd1CdWG7YIj/qei965aJF+SS6A/vF16OcGNst2IJvy
-          dw2grSYKUhsWFN+ff3HgHbpeb9/mH99J7wq3gGHH4c4eB539Yk+xulw3suosEcAvn7zDbxlHNXWg
-          ES+S93cX5Ya1nXAVm8mPkV+YrHrgJwurJKC8ZmOzkMQpn8NIcVWUtoPRsO/O4eHBR5ANA7th++E8
-          8R/1zS37lTELmj/6mlru9vGYQHzY89G7BZp29IdCt65DIp3g459/4F7OoT0VBafBAx88wedvIUG3
-          PwniIImWLS0ae/OFLIJq+f69f1H8+CS8f5871jNWU+mKQzwMM9dZ9uGRAib1BRZe1ITGEtf7zWJs
-          igDhyXstcDhZdvel/yywOEmDrJKqQvKypQCW1ZB6o/txyPjDkwMPf/k9jkaom6AmBkRI6pSfXzLA
-          29rWSK5jHOJN7jOY3FRtoXnWKna0s61otHW+nLg8Jkf/rBX+/KfnMrXqkY+r+AhfvleaIROSUX15
-          UB14HWtYpZsV1FEJEQUl78d3mZpfB2DmVxndVEUc98ni4S8ekZ3twUiIpNLCLx+ylpPDXXpr2i//
-          kaIwX0KIZNPw4O/eejXP4TS+3BV8ZuN+rN84kjsVOmJ+o7H3rccwXOVw4P/h2YVrg+ZfffBQMXvO
-          sNdk6x7yAs9p5yEz4dpwf3iuINRJ8FnggYe9WevWefmrJewy5yvYlD85+PlfyIJBptIaGAIoT+8M
-          23+vtNgS962IB1/2TpO7hrsdcJmI3usbOXCaR2wkCQXp18dBts9rgLkyTQ0jlRGx1dmGTZO6VICW
-          1CzyPi5VTPfxVIKfv3flqF7dpTbmxeIu6fg+bcq4U6ExQajk1wWbQtlsNc8PwuGfLV3LCeEq50YF
-          fQUxSDnVejixUiWJXcrPSB8bEyz+39TBPyDkCzdtxbjCPPL/+TdPEt4B/auPd4ZlsGffjyP/7PyA
-          6iDo3tnnRUIA5HOYYekPPR/TJWTi1pXg4f8ie/50zY7PBQRSpN2Qkfh/47a8bhCOYlXjZ0JOxdea
-          mwioctrgC9A2snFPwoqfl/ZAL1OgmvnQo7/4X4SXfWk49e0H4sHnsdo8HLKex7MDyPXv4XmJMYAN
-          +XQOT+fxhlT+i5u9uKcCTDD3xI4XmAVt9L4mfuPrCbtBl6mkv9YmzIPRWahJYkbSfmtBpC7fP2Ro
-          qtqMMl2X4MgPbJppB/bZsFpY5MPJcw69cvgbys9f8+rbA44T6/75P/2L7ex0B/s6rCbcn9SIjTx9
-          NPs68ObPL8MGflrq9pd/LCgPhozlN5AAfeh98CxvDTa04VxMacdn0GDtCB36Ipy/VrvD4k9RvFmo
-          Ts3GpZkHqVr68079fFP3pwIEoJhNgPSBrcDY3YcaaL29Yqk26WZ/v74BlBpsYsS5XnP4vQ/+Vx9k
-          vZ/HHbRjD/lXm+P48E9Xxqw88assPdawD8BUijkNOHN3PcZusMpaUQFB5LU6Tp8cPvwtLTpTq9Ii
-          +dNfCuYVqRrgX12OtdujaQ6+YEH69ecgJcL0uATlhRUO/nj8XR0ZnN534fBfsTm1f2A9Y1v6+WML
-          o7J9s8t3WwBzEM34RapenYtPJcEuwxYyJDEAG9KHDF5u0YbuitYQvBaJA2CtSh4OZ6bZNmv0YKYy
-          9s+/biaS9S2MLY891mOwt+VhmvDw87AurZm6ymHNw1c/WMhKqTfZlRQnAFWUhq7FzSi2VK9ouI5W
-          i53baQH7Ps0Z+PEjxcDcSKhl9oEb/eULG6alyuKdj+Cxn+B1ocUC4kaVCZNgpLB2bvZwe+D4AVNG
-          2vCTJWd7O1vvHoqu8/BoeFXC9fpMdgg4Z0K28fGKNdIfPYwzjyD1psgN0zvXBCbnXELuER/kz/Nr
-          iFsyLtPdocftfkkMcLsUpkdv33bceucagcPPQn4qyzYdWrYCB2sqcJZSOViC6JiDuXIyMmxxU7eL
-          3lAwXNur19weqz39/NK7Lb1RZH2ncP3YogmfDKti86v148+fBoe/+PObmo0P3qxoueRz+C+12s2G
-          0sK5Sgbkle4JTOv6NX96FLmXl9b8Bfg4Un7U4/egMIDsdszCW+Z+cPx8nsnWfU7muQ4fOtaDxh23
-          4ZQvkEtWC+WvT6LODa0EcId3iFX1m4e7flYNeKvbDzK5zQ1Xs/jW8G9ZBGTIlqAefpomMlnxxApd
-          X2yaU54+8M+F7Qm0raqHf6KA80PtsRLVc7hLf38LPPxarDdsHeLP6tVAr+zWG7Y9L3b2GHdTksty
-          +HUyYYp1V2AZ9Lm3fZllbDEAFrgzSoblqzmBH/6A+Vb5yPRv9Ljo8MEDtbJFpNui2AyqEyfi9bT1
-          HhM0brOdvo7Pl0RfsGXYToEBXDORIZ2ANPXrAqK+swAefBFp2C/ArjpxJAgwyxYRE6HZxWgMYAad
-          o+e8rgPup8eTcyah1yR91X37Fh2TU6O+MFUzjCS9rNLv++J85Ue1v4bfBKC4KNGv3v/jd0d9RWj6
-          KwmJGiv76cvlfPo8ii3nd++fHiiM7M/+2PU9gg0ADr4UWqYSa+qnf/tN5KbII33oAahcU4xc76Wq
-          Ix+dJRB27hUhUvX2fsFdAiKU19g59kPWoYEWEDkv9Hb18RpJGUqlGIlqsnDiIBPOYhML3PtvjLWN
-          Hcn6/uS+oINYQpaRheO0EoECFpPoyD47OVkn/0tB77vOi3h23mAJ0SRBxioqb78pns1p3SuC/qOl
-          kTe5kOyPSx2IS7kJqMSGoXIJ4j24I7ig6w+PklBRYNsGV3Toe7BG/jLA6l0WP/1o75bOWzDMkOM1
-          BgzCffuGHcxvUoDTZPg2Q8f1PrSck7l8RpErMEOexxVeO1nOTnvw34DLodHbKdZvrD4ywCl5mN2U
-          0aNvbKBiBUMNnp7fFN0jOyaM/RktqCUV+/Obxu2muSyIt9xdmgj36hIrTAmTG/P2hO5DNxPr4gCy
-          yj1C3sH/N2WeePiW9QDJtWmqx/7kQ3y55Yz/rf/PbyzDB7f8fVoGrDNyMwC+Hkbo0Gs7aWcPpoyy
-          IRtOuTpXjsKLqsfb6F5SDsHMqmaiEvExfiyT3+w/v5EpkLhsNBYA9pJnDl5laHnU41zbc/HpFUAj
-          GKOnd741JMA8BAYrbsg4/ONtu9wzKETMGXuSI4Xcj6/S5Y6xZkCdcO77L4PH/htOD324CnP3AFy7
-          XlDZV0k427QHBe/x5+Lr99YUG9SSSTz0JkoxN6nLPn3yn1/k0a2ahOTilTt8BAPBF/RsyHp7xo54
-          6CFvLT2DkM97mMD7+/BQPuiEbMiH+Zl/uzE29lVRufZb8zDs0BUd/H9cnza7gNBoAyT7pl0w78uj
-          Aryf8x4Z2DtYbbfMfvttyImDDuzV37eEGWY0fD/8YXKlyggcJxgWeDstZBWaTRIPfoyVTfkDa92I
-          C/yKNcB2pg8jed0aAVrFVqNLnq72r96JB59b+IlKQsJHmwIf0W3GHiae+qXoMhD+nyMF7P8+UlBG
-          uYAvt9AgdN1dHkL88jVkhEBp6OTRL1BEXY2dY9TMXFeuBVX+A7BqvE7hFvm3AEpq2i2QMyKy9oaQ
-          we3uGUtrvrFKPNuOoF/MOZLfDqtuFHBaAOOHh6U4cMCaT00r9o8+xNfwUtt7bpotVIHzPRq1ajb+
-          Um8Tgghm6KkE+kjvd5gI3nyMHl1FrWCHSdfg/GwRTnH4JOttFgLoS+8BXdGQF1/2qyTw6ecONkv8
-          bcgcbrWoCwLlbeWCRvL+u5XilFAEmalUNd/k5lPi6iATSWiiiw11fADS2JDwva+UcK/uqgme6lnz
-          sLrZNv7shgAWYvkLC/zCXrMsh1Do3RiVfvwauehdJFA76xvOUyZvNk3zpLPcg24RJplTSfOkfLg8
-          9gXZf2+52MRwVGDRPgYcf125YBi5X8W0pwm+NapeMJP4hjC0x2Gh2bQOV09SO9FwnBg/mnC2Fzke
-          cvgWYIdztBWAYz5gh1or/yG3TcA4A5or4Z1RfGR5La+SRDhn8GUWvUdC99oM4FK0YFIdCblBrtrs
-          rj5YWH6f2vLpj1sQbaFUYt+AE1KZ22kk10ckwPMpS5H8ZuWGNI26iLe93fHt7xTYDMO5NBTzNUSy
-          3HUje5VsQ/zO1tX7PW895bMEi8YOlzM6+fZ6TYb293vQlQhxyFHJuxM3+/6HCppVACvkbQ3zs0Th
-          fNQtdfc/iQ+vjFsihUsam6bsNwsfXAGX7Tp8CrzFKy9+h3eBsySJ7ck5qTRETgUW6n0xm/1MFF7c
-          /OeOU5RlDRFYNoLCqw9QNG2WzXDxlokds8NlhB+bbLdnbADjkhtYj3to7yo2S5H3+g5lR3zRlnrv
-          YXHqcxwLrkNogb09REB4iBD3ytT1/CA59BxZR7fb0trsVRA7mOEHwumF0gsW8dgDv/xKLi5fLGXA
-          KnCyiYLkWZBCWj7lLOTwpOBYnZpxuSeOAZfL1CLjUi8qlp5NB4/+qEsjDrT63Sc/Fz1H1RdwuXUh
-          YT5ghZZUoOUPzUuI6+vcQ/sijSi+ybrKSvJtF1vyPOP0VSUqrWY8Ba8MKpcNL4W9p/mdFZi+XdF1
-          RXax3zSZFblMtnCJ7kGzy3Gdwbfev/FNVB/jRv2B4d/7Pt/NK8RyfUAidSVHPBb/8hnQVzPBfkAb
-          IaftQADzO1ZRzq0nQFr9xQNeW12UHPm19aDuIP4zKeTmV0fliBMv4oE/3taDl72PpdyKXKZaSO7B
-          yd55IxTEI38XMS/fgI7xuYeI0lxse+O92f4iWRHzYHyiS2So6na68xacMt9CN0mb1XUJcAWd7qV7
-          /smsbCYWTR+8ws8dmW93VbfHVTRhJ5XcMlvyMpLzLHVilPU9Lv34NG4pmQ1oclGC5PGiqVztkWP0
-          8G1G1+oygzEMMpPnKeHtzaksqOTy4Q0II4tD2uVT28ReP7VIl06FMrUowvXCcQqwHnyCfTaPAW1i
-          k4JO99SXv96rGkZQK0s068DB1wvYwr27/SVH4yiMSqnmyNYWpxwOW06QlTLCuF78+AGl1fhiiXF3
-          +3h/E6InG2AUMfG4FhIdiFWcd8jhimEkjPvMoNXt6+/3qaS/pAE83yUWSfWfZzN1pZsgcksO2+cd
-          NfjG75IQBPSMVf5+G6dz+oEQCnWGNYLW4h23b0F89qKHNHi+jExLrYN4PA8Hfv8mLMdlPeeGjomN
-          qL6DTXG/HkxaWsfJN6oJV4UmCwdrS/Cr6LdxcwhtAGdNX8jWsUPoAx9BovoESSPvjyQoDAd2tXRF
-          RexSYAvQvYOhI8pY0TeDMAMcArhTq4F8J6ls5hn5yi++MArR0967h+XBbRxrdJHeIlkk+baKVq9d
-          UGxYlvptLt0OfNXMcO4YN8AJUF2hGjEmenTGN9yu80LBGVYE337xqWLzcTTCqvHVunNgZza7gu33
-          88IucdmQvN5OBPu1uGDLYLC6N7CEsN5aFevva930nTT78KRdLlgZGTgS/xVSYhc3MnIfsC/2e6Jp
-          0EyNAsvlgpv1KQMI2xfWvTN1NO5Cj2cOw/OieTRBfsG4ChdB/fz3hzyXu9u0ypx6aL3hG8vfjzVy
-          6j1mxcd9xwsQTqRoj3oPrG5dkYv+DEDCwDdFxowYhDorARx7urHicllaZL+e9TjddUUBGXFVFKSG
-          qa6+kU3iOLHlkvB7RNh3b0Si32QhvuEFqLPkvGvR1udoObmGM3JtbDpw85T4hw9kNZUbFK+PvEJW
-          ESgFo12LB0TxhcJuo/Ngp2spEj21VFFixUPBUH9kgCYenzhC3wrsFjJ8sWiuIUrr3rS5oAoH4Qr7
-          AkewvREab6EEWZMekeLagr2uGRzg8km3Yz2QTR/vA+9osQ4+M4HtOncUOL4nRgsv29jdZg+cdV9F
-          yUkwmvUzRa144DuS+Li1t559mjylfm4oigR1XMX7JxJ/8S1F+lVlrvMCYbadJ2zlPUv2TCYlPH36
-          OzK0/FHQ7iYl4sgrCKGlsYu9ZypfNAfGw6/OYgmhHGMR/97HoBTuqTXro3/SsIynG3atpA23t/F3
-          NEq+uYs/X0lBru8hAb/3jV4Ps2EeT0+DoGQ3dK0e6jjAauuhM3UjkuSXrhL5dFpg+khHJN/3GpCS
-          Wi3xQ07Y29mJG7cXN7LwuzsRkjjtXSxbARyI85FdKFrwi3WTEh96rJ0vzMF39i6BK8SBQSOktH/F
-          vpAbL94N8bgIJ4rFL77gQ7lwi5gmGCw2i2nY5bHvsexXsxeSziW42Bn9L/+nV+J0kJbx3RMiy2/2
-          Lfns0O+9GqNzboSL+gWK+IfPPpaP+sb57TsDv/jz0v1PJb7BtOIcMQA/6fJu7xpl91Br1T9shrw3
-          EuI8J8garrhQ8HsvBuB8JXhZj8EzBndVSecLJaxXf1r+HKUthvPMPGD5fWnIE+ydbJlre1BOHhby
-          P7iy133Pj/PiUomP3w/GnGsSoPvrgC4u+7JJ8dIoSGl8j5N24MJtEt+UIOkwwrlJncg0qW0k3s/s
-          C2tvAdskqMIeXli6Qm7UELB5NFnh48XHKKPNK6GnRAzgUmo6zv7GpJh8w18geS4Dvj5OAKx/vcFC
-          XvxY+M6Hg7p1Qi2Iets8PdY1u2ZezdfjLFz9zTvpiguWDocmXIpn+stfe6cYxYfSlklLB3xg46a/
-          m5Cl/RQX39MF0NvfiYaDm2se77vtuLLfk/erj/i6tm/7qFcZrMB+WRh7VcKVqeQSqsxFRU78TWxs
-          mloAFal7eecz04zrOf1Q8HVdTsiTag6s+HouIdUwDlLXOAAHviVHO8EX9oynTrZvf+vh3O8S+tWT
-          zXhOGoCXEWM9Wz2bVC7JoQ3U+yJGXmmvw0XQgNFFA1LKCoRbNL8gpNzXgL06msIZBE0kcm0VYAP4
-          QN3u9LSDQ79g70/Qxn0sry289Uu21FYmgv21fatzbVgpRuqDKbaThVj4tBMZa2l2JQzqVv/f+lzZ
-          eWvm3/OWjT7jrI1r+1/9ncQowlmJr82ae6cdKAklLcz6F41EtvIeNnrEoldp8DYmydr949dOaKNx
-          +7iAhzcl4ZDR3KOGSOwVgs+7CrFf/y0qcad0h8l3iPD1+b2ErMKxFTjqrbd4b23kxuBMQ004LM/H
-          dxr3Uni1cHc92ts5eQ5LjaccIFitiR3DblVMk1MgjLyEcBLcv/YmfGAArrK3euKdPka7ZisltuR1
-          xukz1ZtR/ArVcXP3jN0ZErDpXiWIh/7z1khMbTxfyx7Cv/OAcl6PAT2vYQTUZX7jX/3bQLqu4rM1
-          T/jgc/b3VtcVPPE6gzw/PjW7jzdL5DxzxTeA5GZLC2P5PX85qduozgynsz98w5fZcostN5kSnJUx
-          RvpN1m32q40VoD5mg8pjfbchNhfhKqgzsrOhG7dDL4MuQsNCf7+8PVmfeQdH/cK6YQ32PH/s9p9+
-          UuVmb8YMbIeFfDXwoafG7SpxFLhttwJfg2fSrCJ3s+DAIuWI3w859FsuKPitIG2s/oq1Mq8G/Dil
-          6Z3m8wmsYvTigSAzNNZ1Jgqnd6r58K/Flder66Zu+ioqcFt8f+HVAhR/TDDQQqHXBnYWq1f/xfuP
-          P7qNnoEhjpwJ3pd+WciorfbOZrP5a5yLHX6ai01nrQfsL+0V23duKMiNtB2QtlzC1rW3ixXtkvLT
-          j1g9vdWCyQK/B0vAx1iv6Ams9OJQguF4sQejC2wmIj8pEdL57lH7JVK582x28NV/GyS9XtDGpntb
-          IMX9nZASn7WQjabGEGvxK3pscL/aHPfuJGjc/AqlV2Ta03DZjf/q12tQjNOjj2lom4mCLHI1wUbX
-          oiMYzQyRIUMGzO3bYaH+egYepQ7XcRqUtgfUjD/ImZk93PX1ODLpMHeEJl3/x+eAxz02XHLzFM6K
-          +3bglgoYKSaaC3yq0gT86oPThK7NZK7qiAf+e+TwU/bLMZiw/2O1ZeP2O9hdRSkhpLMdqeJwDNap
-          0wUyF7X0KkcewukvVyM4m7WOjUqtwj1YeRpO1nNceg/KZKsr3YI2jFZPuBdEJSHd+jD7Rl986Ini
-          mxisIRbnmkbKn/cZiQslTTw/mgVfg30KN/LEOzj4kyc0iaxOzfOewfMxCslFfx3Y79+nAr9twaBL
-          He7h0nwNBTxNHC7c4ScxyLsJEF6++FivhzpPBGkQh5mNvdfbVxkGFj50x81Ceqej8Kev4fH/PYpX
-          6mI/eVUGsRdJKJFhTDiMmhX+9GfxetbNlnuWAKBLYuwdfIJtvoYklsNa4J8+JV/aC2AuOTK6Rc7Q
-          TPIpYMGWfLIFvMHXPlwdC8wAPRZ+cYtx5yZxBbUvbQt7Erpxm1+UCU+nal9O9tGFbxMKHhqGhheG
-          0wWypue/YwsniZB5fM/Dr9kBx9907Ky91+CsyRcYcpdk2bN0AERt1gUeeICcAnnhRkmBAcHlekaa
-          +cZ234HXCv6UQPNI313BsNG3Dh7xj8yu3Mf1xPusGE33Ah3+WzFTf6AXrnAoPN7wuHFRnYsjwkX1
-          l/WZjM1+tr8ROPgS9hjUqes5nSH8sO0TS3P5VddnHiTi7/uZHTHDjX/9WYDIqME6c5Ns7sNFEfSm
-          UEKqk7sjc8QLfOfTiNJjvbef3qHucEc//GQs/krDysXzQrSxbVZfVv1/9VyjK2Lj6c1Iv3r1X/3F
-          iQEU97KycVDMJxsrG3SEl3nvlzsRmHA9ny4VTJp68cQ0QWQacdFDFdYSRrFlF2u5DBTsTxfsvd/v
-          kRBdPlngp/+dz0MEO6CtDrq8H6Db9zM068z0O/QG8Ysk2cIFfj+oDmpVhZGq5jL51UsRXPe/BcfW
-          WCzU+5zANeiGf/hIKmHIoTDQKkZ6qo6cWo4BpK9WglB1jkJ84sUFuKFnYnfK82ILtnn65S++xp9+
-          3L5Pdv+nD5zsEYQkXssKCPT0xJe7gsn28hUfCifBRZoBkoboLZ/BuJH4ZROzPdyptCwFLwYvbH5v
-          C5gs9daLty0skAyQPNKtZZbQtPkBZ+b1A6ZLvkbimJQiRtrMq+Sk+j28RqcG6XbhNFyWnWj4tKQB
-          GdSsELoJ2ApKLc8jY8vkkTv8Fwgt/4lf57wLJ/AkGryHJYVdZoRkDdvUEE93+/bzcxpyt2AELxNl
-          YFu46CFrel4H/Yp8sNnDszo+K20A3w92ljqbqnG9cCcFtt+/FzJ07UI2s9k8KGiPGvvWehmZq3SC
-          sODd59J/bx5YFNGGICk4ZnmL2V6snBon4uH/eufq0Yw7XZsRTPfZxYotpAX5wPcqTtZrXG7D7oX7
-          rkYsqCIdIVsqrfDnRwvP5/bEr3JtGiy2swGCx2x4+5m+jnuzhA/wPo9oAZV2JvPhx8L2mrYLI9TK
-          yH2ptwUnKIlLfdQf0jvtBNegHdBlfjUqyfenAJdPvHktXRGVVCuqIdY/E9Jgu5HJEygDVFWbooJ8
-          KvvHV4GhkQzLXz8i3DPnPPhc+wt2i1EfB/YpdPDnh7mH37Rb96/5j0/FONnATPfWAFnsONjnNLnY
-          rJYXgDPzCla+06QynNI/QMesEKnkGLabjo0AKyl6oOAkCvYGOwDh7a+Qsb5CK9wunGVBo8EQHetn
-          7wlYAsiFDsDy32m3CU24AHLzJfIEcVzs1ejB4ffuIXY/0uXnd2lAKpvkwJur+tM74OS3CCXl80KG
-          8yyWwukz3D3aAMm40jlUQO0r2zKC6KtucY4USAodLtzxPZv+Y2vAj/9obJZ5CHa7byP48+flHrzU
-          mZ4zDU6cs6A0faRhX5S9Bi4TNP75uVjh5AHK7KTj5+MvaOZ9yjKYPkdl4QXwAWu4bBCKQPKwErde
-          w4qqS0NLNG/LoKUmYEOX68DPT1HUsAo3vqVbOLiZ9uOr487tqvNbP+8ePNmxAWVKQdP6hMg7+MOG
-          K5GGVXYuEbq/ZJVIK6zhXMkTkqO3FNLN85aDY/8EW6Gn/vPDIKIMF5tHfSIuuw9AN3gfPTykAALK
-          FwXlsY7wRXo/wTDDC4RQyjWkdqYasuul7n76C6NyPYpP8FfzR3ygvA08dc/gB/6LJ0TJYtGGFz+H
-          zTil2D784glNQgexl0hYbqPenvJpbKHEOgg9CrSEe7TxFTzXwPeWYa4AZoKaFlOz5bFSMjd7FTZH
-          gSsbP/GTjfuQZE0+wV4TC6RYczj+9muEqqgc5Ofevfn56TC7kHQhOdybdX/JO6xviYm0B1WH07VY
-          JfgQmxpdoBir//ydRpMPv6x2bHpQph6Ci332lvyjhuzBp+D3MzvYKVMppLP1HIEokYXf+o7zz+89
-          9J8H3JIB63DZNfHCspW36EQI11i1Big64nHE4mI2MwFF+/Of8CV9cAWJ3mHy20/A1952mkVezP3n
-          z6PCJeO4f0+JIh75jz0/fjVrfg+GH757K4G4wD8/8vas/hbx3bwKck0vJaQ+VoM8+D0Xm8ooO1Tv
-          3xSrRz2deOa2w8B3T//46n8AAAD//6RdS5eyMBL9QS7kJUmWvEXABAFp3AEiCqICJkB+/Rz8Zjm7
-          WfbpPjQk9bj3VqUifQ7KBoqv+EKsMMzM+VLiHPR/UUX27ayAT9XEAQzmt8SSlG6K791hDnjn5pMY
-          MApN+RpwCpkb5CRxgNTw8WWUqBfAQIXFGIfFl14SjP1Pz+KNLAO6bEgCDFFtmTv3ZjFCuxxVxnsH
-          ixPsiyWOjwZwH2FArpK2NJRSqwKuoZ2Idxo9Xy7aPPvxNaqOacBpBp8buPKnX7xt+E8vDI79/qeH
-          8rnt8hAoygT+2fckF3wE+d91ZL5W9QU7XhwBaNU9pbuC7swZM5rs5sapWVCLHZg6X3PVn307DfUK
-          QZUuJQS1e8Bqow/x8h7sHN3wHZM/sdG5/HeAvbquD/ZyFAyzk1XvneQQRHB9vfrzZbdgFDjEI0b4
-          voN51T9++gK9r/Uv8Q0eLfzhv3z9vhVPueDE1QPdGT0Gc3/WRpRns8nMqXcboahqB9n5YDKvfTWA
-          IyOD8FeP0sXPJ+ZjyVq48nHye56UzMoDut4rJgEzJH9Wzdr91VcwOk1/zT97naS/K92s+tWc2Jsc
-          Dj2ZyF7R3Xi1rxJuLXtPsOOGpvTaVAl8HuScuWkq+uznn2a+oXR62sIwv14XCD3knYjXpHefp+qc
-          wxiH2b/6yz+9mr+VgsRrvOEKvlowCsn2h5fNyT+SDDZ+WdLtsTCKFS9OYMUP//AyLcsThqveSuJV
-          n1xQ/8pAqIs6wdv9dlhQzzLolWrKnImmplB+hwiaqXtjv3rGZNpvDb6ArbNC3cYxb71nAq9ilRJn
-          k33NRfLu/a5m0fafvSyaJAmonNwTqU5JDabj7KZwqR4+vruuzydj1FUUonCk8126N3WwkzK02+Z/
-          xLs413ihb1mDecbN9QjXIx7S17sEu6tss6OOGF+2hxxCd3SfWL1Mb/5B+NxCSeYEyyIOfKmuslZV
-          m8Ihnru5AY59M4FU7zZY2IdjQUFwNxCRPiFLJyr53GVfAa58mKp/otpM25taQmsWHJZobj6seqiH
-          YCTXjKTGrqB9J/eA5R+JuMFUgsXFuIXb7WOhnyyzwHQ9Wyo8wZvPyEa/xtOH3/Pd/9NSIP/vloLn
-          cNux4OwJMU+AkquNd2T0slSHeNGPFYYsvrvM/rt8zel6tCPU9lsRw74XG842RILCwTIY/lpVMftO
-          /oYfkM3ssC+dWOgTxwUSKT1y9dS/Rm7NXQkUW4qZcyKnghMcWij+C0ZCXi+/mF9rF7lspwQvm0Pj
-          L9rFreHyPKXkGm0UfxYKlELxZu1ZpreGL0mipaFD2enMlgdSjExWFvgI3iG5PMAGjOyRLfB1UJ/0
-          ZBqneD5V3EBDoEnEyV9z8xbDE4bjEm0oX9Yu3T9V3yD4kDrao4kWSwEfITxHG5ElX+Y1bHRzA27b
-          khM3SB6AalSr4C0Sb3QxQhHMpvSa4N69RCQV7/Egv+yZoouFbPb3zepmXMVT+HSVK9E+NeHLMd5q
-          ULibN2K1597/upZaQbjAF7vybTNMpWy7aDqcfFb5JQBTIX0tcD2Jf8T6q7bN8jIaB/FNoLGr+Nf6
-          dMyHHNzL6EipYxWD+ECJh+rj9kFhVAn+crzdU+i/K5/oVn7iAtxcWni46SUhX330WbbgBChSkJL9
-          iZ0LiSRbDwq0M4lNzmPD8084oeOnjEnodUcgaRftgZzjCdPL/XUz6bw0FCXLyFn0DDJfLrahgJJH
-          t+CNWt98GT6BgXb3NKe72f0U8/C3jrb8y04/eynEpI4DuO4n0eYubCTP/XhQwIlOLtbiDIJoUQXu
-          /PGP/WmnJF5sxxXWrk+BBXQ8A0m4+RuY7ixAjJORAnrpPBdWtWew+ElhQa3JpaogODbz3uHXnIak
-          C5CqtyW78lCNl7T9QFib5EFMvnMKqdF6CYRvHzBzllDBnPjowsc+Ctmxf5q+VCIvRQ25UWIly60R
-          4MJC+CinE7tC22qEUQpb5H8FHy9SfzFnkGoedKvLH6na4lXI6Ul5QNMfdHbUgDYIejzkUFNrhWVd
-          oQKuNOoEw3N7I/grvviyJSP92Ru7aJPze98NnHP9RMydeuVL8JY2UNhINvF7dy74/XuK0D67x5Rk
-          Yt8sx1jWoNjtGrq+j79YhTYh709wyJUhMxZSwVqQuE0OLOrbLBay26jCpr/OZK/aik+9lyRAkDQA
-          I++s+3M/CD283L42C3/2rPhjACEaX+xYcMOXhSCvVXGMm3U/p4Z3OxJAB31O9AWDOxAa9dnB4wAo
-          C+BOi+dLE7fQgp8DKUwXD8JXmt6wfdc5uX7j1R+CJ4a4HTNiVeAT8+dl94BDpyjMZaHFJ2doRnRs
-          cUJfu0ovBPHSSUjFdxG38sH35X7eCSoR0ytzvNvU8HlLNtDdzDkJcB43MxDQAhWtisnxKewaXlS4
-          hvcLeTHL3buFaI6qA4/SwIgJ9X0jj8sOQhTjD7EiZ1ssAZJUeIr9DSMt8gfZ044VFC3sEHe7hc28
-          uacaFJqNxbB8dE3WNnsXnpEg0VFxJT6282Qg5d1MxN4yBqbl2/XIjg4+qVzfLiZza5bgdrxWzFW1
-          5yDuDlcKrdNnz3xWnxqBBcoCg50TYq6JZiz/vZQWdpfHg8SW3DTtpQ428HxtBZI47AV4HIYJvGy3
-          e+aGaeOLQRmN6B5oPguFaAKTf7EcVPjnI/Fk4wjmrZxa0LleCDnLA4sXUn8q1VMclxgMZoNk4DQF
-          v/3y7MQDrDxnEPJpKBk+vzqfN/SUIRk4EcP5tiuWZJQr5NImw2ItUn8aa6WDr+37zf76sY6l4HaI
-          4A3kDjvejwbgxWeS1A5Pd1aMwqmRnNJVFGveHVgsA2fgClYFyJtKI/riJI2Uqb0B7dcnwOIBjTE/
-          pt8Wss3dw2Co5mLenTUDJQZmGPTuKRbW9Ydb5izEQPkFyGeYlTAPjx4r4U4rxu6qYBidkje52H9X
-          kyddmCE5agZCauDHS3y8L1AIQcVMdN4XIrmqKsg+W4MYcmE0M94vEkJu+WZh7bKB341TBOvi7bKb
-          aI/N1/rmLgxoSrAA/CHmjXnJoHJsvswUTw6YVeE6wt5/LJjp4Y0P41XV4B3fOuYWg+pzdoswMtPG
-          xMhR/gpRLK8ZvPmGSUU/DmJBUC8WfHZHlQXxwgEvBacFdXxumGbj7zDBz+X9s09CAvYBSwY8C/ba
-          tLA9DHQgp3B6oFNRNcTN8mZYnmMP4W4uC/InfP/49OZTiKgBTHKczyYQ07Wlron6lOiSdzbn7V2y
-          YASTnpzpifK52BVvVPuPPRX5C/N/8f1Wvx12ndq8mB8v6MFHZJxYcLD4wP9eU4t2XXchOdeOjQhH
-          N4BbYbXn4/Ud8+mzV6AqBxVLArr1Z+vOBPgI+pBcN3M7SBfQuxB68YZV9eXlz1vTLZF+zlSy7kcz
-          ZO/sDXSiKCyCvDHlRYoy+Jy2AE/qJgTyeF0MeChbnelaNg/LLrAiKCbrqU4ufIrB6AIJij6o2d79
-          O5vL6byNoJeAO9FWf1ouYe/CflgEvDG4BeTHY69BSb6/ycWvnEGY9K5GgVa8mfdQxmYKLq2L9nGc
-          rP7jmMPmVWD4wNueHP9mHQg+DQxIlcEhh7v2Lrj+bCcUNB1h0TfKOG+bvYd+62ea9ybmMQgTtPoT
-          3XytTbxgXy+RcVY3+LHvRM600FrUvEceKRVZM/l+s+Y7ZMS0N2Ff8K+rWoDkVUGsVH7z797K8//m
-          279DXUwSoyr8PgYdT4fwCORSwB0cAkOi22jfm6MMdiG09MbBoHNmv7Ps9wTvMXqz4xBcwHI2ti08
-          XIaSHJh54y2ZrBD5WeoRrw6CYc0XLRSq0iDH/tmY0+eWZTDvtx6Vv5k2LH/mloIGPf5wE26tgYpe
-          YAB1hjdydLe6KXTXCcPyfkuYdXke4gUcgxyRvitxeQ/PvvCqwhB+fINi8b1OrXlEiYvURJdZeXRi
-          MAntbKF7YPhUtJ6BPzF5mtAaj4ih7h+cHcdjp5rDziRk+QvM4ZXP/S6JkgnzZn4Wa758w2Y/Jf/s
-          f1LRqUL8LWjkpoc3wMNd6CB0lN5YwvcJjO9DmEKwlA3dZOdumPE1l+B3mQ0WhhptFmtyR3hTrndm
-          p+bXHDf8tUrQjrT2t2cNZwcVAwcNJ7qxpgJMfLfTYHMNzyQqG92U8Z/dw5vdm8yMP23c41rPkTXe
-          XXbcaIU/+8XSwx8+uvXx0Myqtp7anaSUaCMOi2kXnUqg+DjGW+1zMHkcZgk8XXcbtj7f/+0/+Nt1
-          W8yZ7PisVYgF45M0soM2OYUUK/qIJDHAzC+SyV/zf6eqZVrSjSxUYD7kS72Omb/iv+aYFuNN+abQ
-          7AaLHO6vrT/C5RX+4iPNZvj0J3h9pSDIJExMjzSAh23owcM2roihzAaQTrLewrtw7JmbxjvOxhfp
-          AcOHPcmFjJrTZb1r9RmyHfvl8+82c3P4QFqMl/BaF7PsXtsfniReJQUxgyGwYCD5Azug0IuF4UNd
-          GA5OzXBIbC432ZxDnO5m5lNvHqgsjpaq8Him3zBRijkWCxca4XvLysmSiunxtDuowq5iOImdWP7y
-          XIO4d9Yh1xUyv8+kzIEyCSn7Q/V9mL0pf8Ad1mdmxHvNXNL2Dn94ie6W6lAIu4JZcFvBmfz4HLci
-          24Dsox/IeXMhYHpXSgK73dSu+HXbcAhiDfyZAmQnqIjNoj1SD5bz60wFP90080cTeij3Ukj2m71S
-          zOL+WoKh9GRy9FR5mImbKzCJqpm4GlcGKtejC5/hd4fFVsbxrApnCm3kH5lzPM4+/xS75Ze/8DR3
-          4TBc+2mD1njBqhKi4mtnUQCrftJZxaYqntoyNJCfJR6VFU9vFtSMNXQvh+xf/JRnT01BraAnM/4m
-          2f9en4UAubfL2GX824IvuCEBog0eyBEVbTOBkmE4J58Ty1AMG26e2g0STGoQba9847l9wwgGvv9m
-          vlUTc9qnwwSeZ6TQTdFCMEV7q4Z/VNiQfEq/fNweWK3Kz+608kMDCEcWePCcBQKW30Y9vBMe9NA1
-          P38/f/Xn6YIswHmr4+/UqvHXoVEEi+KK2SGP9kD4+spbBY/0wfz3coqnuj1WUFbMkTlHP+STQHYe
-          vIhn94dXB84ltYLPj3sjh1zeFJzyxVHPrsNXf+viWdXeHqil3mcaflNzjV/Lf/E6/rv/8qcKggpj
-          tn8mmr8YTOvgLr5zsiem6L/dcRCgf6xcLDtbYZgEs5ugGuACc09Sh6X7mhrcde2FHY7FdZhuyjOF
-          XRs16/e8+GuNhwg1Sc8ujuMUsya8Hr/8x/Z1kw5cuC+1uuoHxK5FajK63g77WMIYq6UXxHL0DEMI
-          bXOL0eP+GCb3z5SAX/sd8YQMm+KtizfIuRaEAvdBCvlWOxP8nCdG4sf12dDo8aGw2S8JOTburljy
-          XFagefJNYu7ivTkm3Hr/+BnzQmsYltFuKvhVfIO5/a0o5nIHRjV5STPBW2UbLyk5bOCQTjIzwXbv
-          T4On1IBVscssh7342N+mEhwsv8AN967+DG5I2i1kiIlG4BdMYdFmyDekA55We+Nxjg0of08Drrfa
-          bDIaE2mXROVMt8vf6I8rPoB8OOQkWE8xjuPZCUEwXHpi5/ax4F1LQniPt2+ilQ6JP4uNIih7HWbY
-          vj3NJVYOFChFkrC4TEt/JtdFRVgQLiStPjOYJp3WcCeNOdPeQWhOP32khCQh9ubC+PAp5gluPxLH
-          Ij1hMBs1o5BEB4kc7L+rvxhW/f7hfxY3BjT5/W+O4Lt/D1SmtG6mzy3MUH5+egyr9dbkH0BKeB9z
-          TA4lqfiE7SCCnWXcKfekfBi/PQ1BcrUmUrHDvpHxcdJQaJcVWfGdySttduD6/ex4nXb+1MGngKat
-          iSl3QbtOUZUWWFf7hhzpy4wnKocj2EErZ3Y2fwf6+/3KD5mxvTU+i9R18O+utJlxdf+Gf3whjAYV
-          b/fnjz8j6Qwhlg+c+QpcGnbZhCPq5uuWrPEeCPbg99C8xxbzt++oWVZ7Azt64sQ5kTmml7D3wFkp
-          JUJaNDTz+AclgJPuj+BAgPHA71EJmz/lTI430SgWu6iFn15CSHrw4p6Vfgd+eo0V+m8wMypqsGHf
-          iRnXne7P2TGuUSXWxjp1oC8odYYWHudiYq67Dg7Ww1ZAD0ZG/HEFh8/6MG8QvM1bpinJyeTk70nh
-          v3ijhlEh/vSsZrtdiJV7W8C5r5aAxY1LcHLqTK7gRYLeQl2SkHxr8q+7WLB5TH9Ei3oa85jZEKpC
-          vKMbz4U+o28ygaX/pMSuDkcgRgK30BpP2c//FtgEJezvrkgOondo5l4vR/hFjUGMmKZ82kWXEtKs
-          sJidsG9Mg/dmA9d4jev1/82ERyrUz7n6w2NgnqyrAx3ldPqnb/AUKg+YxtWO5mjC8ezPeQ+DeuiJ
-          tep7a5EaovB4UIh3zPziiQwt/T2fHM2nVsiysozQXhrMvMwNfBmdYg++bmvL/WSlxbIv3xsQHn0F
-          w+4txm0pH1242gfDTw0Bfnkr4w//Up5viqaPdbmEy21OVrz85fxEsgCueh45whHE7Gx3LuzvmsCO
-          aaFzsUuEEoLtHTGPa9+B9/uTgv4uKSL+LQ0LYcWLyHuaj3U/Z//jT0uLxr2rMl9QFpOfWgVD+r7b
-          pKDeqREyMVD+6ZGXVa8V9WGGsHrmOuZX+1nwvyhs0ekKNswAVwNM83FS4E8fOr1pEv/0JrTqOaTc
-          Npr5wwMQyFKz5t8/cznwdSpfW3J2Per3YpLtlwZ+8dKI97W/zMtA4V0gPSGkOHAmBksHO58BpsUb
-          7AsSOAVIcMySos/JjGfbjELQXpcLFc7NwResvul+/kd7WNKGWoW7gOOuS5gv4cWfzKSp0YQHH0/i
-          OwT/9OFVb8XzH444Z4clgJtPQNl5aGj8bSKI/6ufrPoOC01bACs/ouzv1seL59499edv/lElw/T1
-          lR4GcYUZnu5vszfzmsKzIQnEsuy3yd/FicJdIWQUbogWy+NVNeD6flTues1//fbjh0+tSHQHfomn
-          5Z8eGuI3NgV7MHsotpVLjp9TE0/uOEhQPuof4hB5Auxp7VpYhYbOzCQqGjraQwUPcmuxqPDrYWam
-          qMF33w80EflgCiztSvg+RwMzYVeZU7dMHrzq35REMZXARLUug+/J+9Kdx64x++WjXBQDgl3/WXwd
-          U03hq51edIeDl8/X/PxPb7T9OCiEIji6sEH1H8Ns2sTzcyOkQLw5e2ZI9qPhr+n0BpEEe1I92cun
-          J22bAt3rd3i5HWs+oc2uB0qsM+ateGrl5xU8hUJLnF0Tc1ZLWAArf2FXd6v7y36vQWgFscNIufNN
-          Prq5Bn762ar/8qV69yU4CsmZ6LwUhlFQLw5Y8QM7/vDD3GQKtE7Dnjl7Hhaznakh/IQ5W/nBhQss
-          pSUU5kNN3I9iDVJ22S4AZd2Hypv1CN1Dxhv1t19nkwyA4ybU4LT7YIoa+dHMIVchWPETc93dpZh4
-          Rkfoq6AjxpnX8WxnEUa1E7rETM9WI696DLxF8o1K6fNULFvSUvgB+Yz5LF1jLh8XCkAw3kg44qlY
-          ZHF0oDpvbszSN1WxmOz9hpW8vRJs3+yV72YecA/mHzPIKDU/fgp3hZQRPOxf8cQe4QStsthjZWxb
-          PlXHb6KGT1oz99g/+PjcTx4K7apibuE0w/zD09Epfa/7PcbTy51CdH39ZRj9aUr8XXIUwJO9edDd
-          pdkMLE8yCZLIl5iLyS6epsckQE10FaqUZQNm3WM96LtcIIZiXBqutosKfbQJmBFe63imtNzAwKVH
-          CjrntF4cqWLQVcqDxP6z4Xy4XiAkeVkwz9V8X7LsegIr3yb7rtdMKQOeA++vVeDzZDsWRHV2YUb2
-          Gp3OKR7GJ7SX3aqHED9oybDsFw2iw/ZUkf1oaVx2N0/806OpkJFDzCM1DKBvn1pGutNQULuoJSgA
-          fsPAqplJiaSXwJLjA9F7IzHnZ5LkaOfyhapG58XT9XHEMH9+fWZs6GWYN391Am8PVK/47Vus+WGd
-          Qsd74j+sCVCE2w6iTTCwVY/jj+r9qP7Zz/F6MIp/9SUpxz7br/hqNtLFgPBaPoke1tEwP5MyAyte
-          Z9age/6iUa2E7VfJSWJfXnwp7O0G/PQjY4j1RthwlsH0vVHx+NbE9WLqO4Wyoo+/+kix0GKooTqW
-          13/+Od6wncExqFV2sExU/PYL/OLByq8B9Q86/umheDPd3/4cuiqEPoIBVld9eKHO0MHXQXnSLRsG
-          0AU3PYJj2QNmZPeaL+H98ADf/bJl/iSdAF3cRtjxwc8x/xvadYofbHfj3lPxLT30Mf+Lsu5Xv6GL
-          BSWfldtJgX/xBhFL5IPPh+sJoty+HeiKX7mgnOYcHefLRC7319Zcsou8wMeFDfhnvy1yWQ61PJJ+
-          +qU/edlowNCTKnI0TtogfQuzhNHovIllRpX/+K0vIe2FZSB9Nlx85C0gXpIzd61Xznu6JGitD1Jg
-          urgRei+skNFpW7o0kgzoGX8U2E43GV+fxDR/+vwPz5HwGGFTvBx5hx6jGhHfdOkwn6UihIh3IzOZ
-          3JnTDR9zuNYHsPxJ63gU2tn55T/ivgIr/uELsNYzyMFxnFisYaKh2q/3rNyPsbnsyxoiwmDJ9GkR
-          wIwXVYA3yXDI4ZG6hXD4+oq66j1EF4SzOXqZCX/5hFjKzS++qcQTtMY7Ok9LAqbDX6PA1tQuzNL/
-          XgONBjkBqz6yrtfGfF/lZwCRab1J5ZcFZ9brqwHpYb6IY86TPzuXPIHl6Ryz4NGdTfm0UzyIL/CP
-          ZWNrcflsUxdUtWuwlU/xabPdLeA0dC3xgIXNf/rsXfYIO671o0/s8QCZ3cci4bed4hGg3QM2i5rR
-          50+PG+rwgaRb5zPTWHJzWf0ZaU9Dpqsex9f916AC2pRETpSaU+IGb/DT53G+dWLpKOoJLOExWfWp
-          mz/D+qyog57LxFU1exDOlTWqt/dfTezmdgSzm+clnJPhRCyrr5plWsweNeRK2emVcsD887eEUTq+
-          Vz2oi+c7hBY8amCkPFILs89PBkYrPmc4D5HPvmhPf/Vcpl0yDMS3lrTwoSshOyR+VfBj+uzgQ2CU
-          mQFq43mxxej/uvhA+d8tBds3mZhzUHo+XYkGESjgiRjeIY556sYb2LLwTR+PWOOSUZsPdBNcjaRy
-          6PPZHqQcBRu6Z9qhvvv0pCcQCSwaySFeDFOUjEsN5de3IfvroTQlf4kzWNKpYC7dub6ck2EERHQj
-          Euj3t88nk2ng/XowKhxZM3xH5ZqCaW9JpKxGNiyjcu1gQkPMvKP8AktoswTKZ75nDrh940k2tAj5
-          tU7Jfl/ei0nyngZSuupIrCbIGlncZB0aCnhkp3rCBfu87ymSxf5FjKEp4/b39/NHUvAswAasg+sD
-          uPOvgBX70ojfO1vpoah5DbFFtOFT4ooYan9OjxN+GIe5uSgJrF/Gg5DMbX2RbTcOZN1RY4l9d3x+
-          xQqFn+LywlwUQfO5d68SqN/DH9mX2b6Z/GVQ4anfpeyvVpaBS8ZBQ3mOryzCb7WYf3//ftWMEfOo
-          cyGPxRQ5EfysEN8pZI82EaRG9aZf93kDgmDwCF2F18S8tEuK75+lO6iTwI4ca/KIlz/rkMDhZt7I
-          fj6fBkk9P3t46RNE7M8NAL5/GQ/4QJ+SHUPDBcx+jTV85GNPvCrpwGovG9SPLmIW/8jNAk5FBIsJ
-          Kf/WV7hXcYlq23LIRQA6EB79rYO7/mwTD2cS4ChZFJiXxwM7vysdTCU5QRSqek2IaHamdHZbCmE1
-          NaTcdsdBPGufEcRl/WCZXTfm/OjSEvJtEeDZSkExG6WuQvvkfan6iB/8mcB3CLF5d9gZgRRMTnPK
-          INufJeayw7HhVfSxkFk8XBY53l88orPvATSVGrkg8AfEVMtGuPOuT7wTywl0B3k/wYQUFt2FHBbT
-          JPcOavWLSIIqaBt5MY8a9Pxwy8pt+jbnRFNHuLWjDTER580ke6UGFn9OyG1+XnxpHLUS6j2IGAm9
-          wFxKopfIR3BmRej55lzGRQfwqVovIvEexVexFRWqh4CwaGhg8c/+OL05xAHkBUbZ0XtwnrXTOleg
-          A5MzfD2YXT4y5WERNbPXfzIg6ZsbbYfh1ny17FGjBT1Dsj4v5uPoUqT3u4hFud3EAjetCOo5ehCr
-          Kt/+96sEKbi7mw2dusL05U1YP5CR0wez39UdTCEaDXCzwppZFZL9OQt8QaVNRkl5Ro257I5cQdPB
-          X2jLL5Evvr8Bht9oyNgvXswv8W8DNu3FJXll6YXo3MIQgpvZEMuJ1Oa7b7QeEgkrFCW6xaVt1hhI
-          WE6MOLmkAEZlLqDzw5KxbF6fxfekJxQlVVoxA0xnIGORG2D+CAozy0/oj6InGUjtLZsUzz+14fBS
-          RZDG5YZc5yCI+aZEb5gH8YVp0S32hZP+rMCRv1LiVInNJdkxXJh1jzOxJ2gB6ayFG1hZTc32Ysv9
-          5aQnG/BdaoUdr7rKFxFTFf7sRTfTp8/dbn7AvydlzJ9F1Iz9eN+gr9aPxJDrsOCheDLQJspKkuZ2
-          G38PlE9Q6/cj8/floxDcwfSQbCw2+8U3EYvLv/3AWw1Dc7mlfQDX/Sa3XbQfpAS+H1BXJocU5Rv7
-          S9MTBx7GQmQaDk+DcKCghk9xjxgJxq3Pr39WhR6wk4i+8Yti2WaDpjbHZ0BOLeuLWWu3PZROKiF2
-          8pjBbLZBD8sN9Zlmz94wa+WgQO2916hi16Yv1XnewuOtPzKna17F/HwOFfRrkxI/KJVhUU9jBIOb
-          hJgG9xcuU/Z2YPgyKhJ0/WuYFGzlkFlhyzx+T2NZ8KQJeL17YFlEyPDt7icHvs+1iV/WMTWFzDUw
-          cu6qSTDX3HgR8FaB2zOmdJHDHnCrqHOkZM8LVoCSm/K27gy45gdy/tyKX7xqlU2cLSzpAPEn0Rne
-          YOffAHHB3m7m3RG8oft+Hf+9f2mUeg/t2tlT9fgyCuEMaw3dXvGDWLLrNjMRPw8Ir8uR2aY0m794
-          DKNihOwsooov769VQbntT1hYFNOU6vQSwGforIPqL+PAn/dTCw21M5hROKj4yp4fgu+9XZgDycaf
-          b+l9g1Z/x9tJsAahJBcI4cIh3iXRMZYiZNToUtw4VtMUA37PdQluxD+MR+A8hvlaHFP1qn8xbg+1
-          4f/yH9pY9zcLwGseFqqRHNqh94fFD5GKPlYmF/0ly4doF1Sb7Pn5JnBSzTf97pPaFx5dheHBLb8k
-          NjdxzMNdngHaXixSdmA2F6rts1++xougbxv+53oQnh+OzI5nowJLWaQdpKfcoHDwomGmMk9Bd1J9
-          EqXOPhbS4NvCbD/4xBliGYz27W7A02EbMiM/u0BagA7R6U9NSE7nJxC4OlXQTbs7SyurLmSiywG6
-          qOGX5Gt+nP0+y2FB9gPTM13nQkEeEF3eyYn4rUyGqYjbFDrR5oMXWbML6VF5ED7UzqPbUkma3nnF
-          IZC775Ed3tSNl2E8VUhcDEa81R5ET90q0LqlE9ETwwaLeVMfu/NsnEiRfSJfNhvoIkErGrb34Rhz
-          ozYnZBSjssZXHQzqaXzA+2GbEH1CL7Cs/gT28URIypYNYPBsKmg8iQvLjswauNa+PLQ/T4AddkXA
-          F28ZM/iB5y/mqMH+0n42FtwVlwG3LE/NJTOVFFAhEUiYV8Tsb1EtIL3YKpTiPS7olykCNAqqEJfl
-          sjnl/r0Fkg5vJD2+7sUU6WUFy4ULxMjtpuDZfreByRYnzJA1O2Yytj314l2vzP/b6cUSK4oHBmxf
-          mNsN/bCIzujCIAQtObb0EH9DMazRfNhmVMBhNIhau31DYkYO05zHzVx2580bLnlfEXs+0mKuol0H
-          s3ejE9+8OvHcXmsK4MQ0ppf3kc9cVSjMuvrMUvvzKZjeph4yLmOAgZjg4V++TxfhSJwtW/gkOs0b
-          PY1jwS7zVQCjTx81bMhLoeqWrrcE0SaEonG8Mcu+d+Y8KDQBk7I1iF7XLpjaD2oB9x8JsQl/+vxA
-          hVC1jnlCDMeT43neJg7YGOVIzmLQF9wovRx2EZxJ8PDrgR+6+oGK70bEAv2sgwWLlwOfd+vJbmt8
-          nxmAoyoZUGSXq37h8yZRM1i/tAe73ENQLIl230A5oXdmBsHJl5/XLIDnu3MnBosWPsOTZ8CrzjAz
-          E+/jz/m+rlD9NpR/8X9s76oElb/11uVucJupjOMUzd6WEr0Mm+Jzi94tJElYspPlnPm4f0UTSlN9
-          yyo704GYhi8MO1PXsbrFFp9lR1eQm32u7Pym+0Lw+3AdJAq/7PA9soK3974D59rxmZVib5ivsdf+
-          8BzxRWHXfMqiSmFRmfaKry1/0crWhfjcLJQng+JzFM41rFWzJZb85sP8/MoYegVmRDtMziDHGsBg
-          p4Yic5oEgJfVNC7MnLdISjkLzX98wt+sxx7PogbkOs8lsP6Mu9y2ivFWoRJa19Sim/zsc64coQfW
-          fM3seNHNJbaPCXyGFqFzaB64qGDrvbPNeIc7FnHAmqsaQB3IGVWHpiqmXtla4Ld+cfv3GHj9oN2P
-          HzADEIfLCawX0G5xRiy6e5njL//ZeoBYWKHUfL6+Zwylk0JYHJRSw5unsAFVk/fMTHI8iAmsQ8TH
-          zCOnun7zL9ElCtsJM+LJ4cgZCnc1lGLkEw/c1pKM+KmBOZaYpdFLjue/MMW7H56xwJ9nypJXLvAA
-          zoTcRtEdJHCKI3Csu4QEkrwOYoWzhhzDaNb8+BzG73RI4JGMV3JJFqWY4FHL4d7Zzcxa+dLSXOfN
-          P/uyXYvy2X61BvpOj4D9fZzJH6gMBPjjp8ernvPFfOMIfvLiRF8bUzXZ5W9bwh9ejABxwLpeE1zz
-          C3Oq1i/ExILrHaLajp2sFMRj5hoBMq8PgWjeSWymR+VVsMQKYPHONE25vd4xsrY4xKitTE7XZmCo
-          9fZIRQXUfDkuXgajgD+w1PEA8OjwB4HzUC8Ey1MG5lD/jgDWPMSdXM8Fz8ypQ7I1zcz+4qhgkTLl
-          aN6rJdt/Xjd/weIyoV8+VV7lzeRmscvBUnwtKlTfeVj4Rk1BZ5tfZm3ls7kYtf+Ay+XtMrN077/4
-          16vXW6cTgz2U4mePkH/LiRi5JIOl+15L0MVozzDAT75cY0OApx6kdJt9YnPpx74C48wfTNf+tjGb
-          QIxVP9xvmB/4xFwUHChw2xGLHMOIxrN6kSe44g3iJ/k0TOikTyiNkUnnDxEL/qjMAKWNwhkJ83ng
-          7XPYwN0+j5lbBc0w3vPIQelDvbL9JBiDEB9woM5QVunXTdJi3oZvA4aXQ0/0mZZx7wxljzaPJKCg
-          VngzPvKDA8yyrrCSb0p/qVIQqCa71yS2qk2zNJepQ/F0HfCnrD0gWS8x/Pe+10+142NZMAfi5/VB
-          bqUi/OJDAM+C3RDj0fYFT6yhhRd03mJx5b+0zqMOdY0qM1NMxmY5axmE6vEkr/kh4uKPTyPeRFhG
-          hubT5+drwb63dHYKEh6v+DIDp48wEVtzJNA/Km8DjnWbkIM4SsUYivcQPcbBwc8RvfmMjT2Fuyy9
-          sPX7i4UgVEP9TwmJo/ffeFxkzYGOuCvw7kn+4vdx86epOzf/UDkYr+ZykMkEk/v5TrDh5fGst7QH
-          tzYIcfeI7+CThixQAd5fmS+2iT/bt/t61y88kOKebQa+ST4KfHzZyPxd3DfL9S/A8PSSLIJfprC2
-          4F0USPbFhtjZyY258acY0N+gPVYfwokPVSVhSPUyYvEesYLhnZDD3Ts6kENo7Pkw8ayE08vosNQk
-          gE9Ge86Qpd8Q0Vvmxax/h4l6gvJCl67pimXZmik4R4c9SezZb2T3objQoqFHLmE8x/Pmvn+oL/Hd
-          MMznEkx622Xoeu8o08OYxPPnvZOgLA8xBegxmgtXFQgNtTV+8QS8uTqVuxVvUrDyv+e6PkBx2i8z
-          uP4pJnQ6TP/4TVCVns+5AbV/eCUwDNmc0LVSoH81BqLvxmRtkZs8ZE2hzMzVHuksux1cNh+JGGka
-          ADYDD+/SRuVEe5VXk/fCt1KjoR/JfhcfBnb5k0v1p1eYz6wcZnfwM7jiQ7oZIhHMSuQYYPYQJfvs
-          M6/5eRGgeHI+dLOt/GIuSB6ggE1fFn/TZJjBhU478oordiT8Zc4lOQRw5cfEHJ+1OWlZH0HyDB7M
-          GMw5/gibuwOb7S1j5v57AIKyXmTy58hPEmAyDdM4Renu0b9zcg7NHryPkhnB9K7uyQ09viZfzKMB
-          m6LUfnqBua6/AKtHnuA+te3in/9Yp8uXaPIO+vOoXBOo7zc1c/UlbL5p8BXAGk9IsPID/vx2I1w2
-          g0T2yNBM6dDVEewW6cS8l8ljZo57FVKeVng7W575732ntwZJgaJuWBi/Y/QYPw5JD+/nyj+iHMnG
-          ZBNr62iN2AtoA7fngJJg232buYrmFv3y6Q//8cQVN/B5d56Yl5/JX654qsDk+iXbb0zV5xEyJrjR
-          q5QcPjgu3q/nAcNKKkbi9IsLpkdllOizzwmdXoE8zPHBqSBAypfpu6Ln3O129Q+/ssPO38crf6+h
-          B12RHJ83xL/NZUrQLT0FBEtbKxbgUcvgRwntf/7CZWy7KJhmi2DPjQrZp1GE0ueWres5Ddx6iTVq
-          SeHidps6pgBP0QJObyEn69VIjVT4JwuphTuQ4zTqsXzZvyHct+OZaSwXfR4sxhvlwIrZ37r+s+h9
-          I/V46lUqttuQT1+YVijsNHflL4s/56QJYBEYF+Y4+ZVzrX4usLjpBQvk97nhn7GncLrsHxisfJnL
-          jpHDwyXj//Sx7/PeO3D4bnYYrXqjBI+ago5TbbHsRLVmEXGnqJs4X7D87p7+6u850nIf0mwoCnMh
-          8zuDwwgbdshOn2Kxm4eCCmIPWDCMs8kW0zYAjsqavsyNEL8DeqpBGYoZ+/t2hXkVvWcE1WMs47D6
-          Ts23+9IA3vNdzuzE6Pi8O6ghQPdWYsbKV98n/VsBfL4vzAFkD/i2bN8g+rw1DNrOMSdvgSr8Lg8F
-          o2S5xDMz5Rr8NdQkB9TjeCniQlBnZYnIjx9N8zaWIJoqjWnVdTCnUYjpL1/jjf3ex+8q2qVQRfs9
-          3qEerxdndhrsLPPvx+8AlfBZhc+HAxiWax7PP/xj31yNZY/iaQr27aP93of5KHKasU7vFM6D3JLg
-          tNn74l9QCvB02WkrX3sW02rP//RKg+uHeO7fbwmsehjVlp3jL4KDXDX8syix8us6TbX2J0jaIKB8
-          xVNie+cOXO2PHcyzaU5aObzVi1vOJJ4EfZh2B9UAK78nhndRODNHokLjOgurP7x+ekEO2zraYsUw
-          Un9h8qOFF7eamd5uGaD31A1gDHODCnCv8u9pbUFy7opJUl1dAA02fzU87TctIWV2GOjKx+FeYV92
-          uJ5uMbdunwkufYfXabOZP8NzkoE1njEyim7zw5Po3osehhErBsm+ZQ+AH3nIdDW4xaOenUJgVUnO
-          StZfh+lffjaMBstkeHBqD98e5lsf0OmCan8JxXCCYWpl5Pr+PuPlHHQWXPkintfnTWYxvX/6GJUT
-          r/e56H0XCI/Clc6ufeLLYgaLCtmyW1sg3UHe2e8c7v5Sk+33355//qyDBal1b7F0lB0uds+oAgmL
-          Wrzo/VgMK9EByqvmxIn2Y0FfT31UIzgYdHIeN38SjT6CyTZIWNHKrJnt5tGjyJ8DOs8UFuPZ6oV/
-          8T2sgnuz4pURlVt6xxu8p/F06KYJ3jOVMpfvNuYULJEKHt6ksuQw7RtBMEAID271JZ6d1YCnwTnZ
-          jYFkYsne0bVj7ZXDHlh/7Boab76gK85hspxezMBvNabP+yWBPz5TRmy9ac4QNKjtdId+WhYUosus
-          Ba56G56o2ptLsLnV6s/fFfAKmzm1IgfWrp8xDJgA+OupQ1VMfEDsbJELvknuCtIztaM8Aaj51Tfg
-          UJu/esH33/eq9wNK8GvN18Lqnzv2xD0VusLy++ufheGrHa/kVu41f17tH63rz4ITvTfTzp7eYDoc
-          FhY8hmEYVj6+c8+dj8Eu9ga+bw4e/BP6A9nvzNan6/eDn95vi0Efr3qwBlHzeqx48FjI91Sr0A8v
-          GsV6inoBhw28/wkuCeXs6M/P+8OCn11UsP341M1pZytvNYfWh/y9Ja+ZbLatgYE+CX2s/i3FcK5R
-          g7KSaEf2GLgc/T2gqzzQWm/bA1rlsIUFVAnTcio0o3MLNXizzw4jXxv701f7U3/4l+jZI+TLLGsd
-          1C6Hjva5KIJJPUsunGzHw7u7duZfr7/nsAi0C530BTdzoJ5VuOJ9Zosl48uqN4F1P1nADnRg9vBV
-          wGcXFoS0UtB86zRbw/VXwqeE5/F0hrXxW39mJ9EY//wNXR9PC4tl/QHTLJ8sJHLTxs3+e+DSvC1T
-          eB92E8k8DZvypt4a6jIOHpYGA8fzvnF72GrFjTgrfqZfNknoV486Z/cvWBRsKT/+TwK7bv1+37gq
-          7HRYsEySa3P26WmCYmjFjIjt2R9Ttw3Qim8ZIb3G5dc3oHAC5gbff3xBzw4arG3HwSrTh2IQNvcO
-          isQNSNLnmilukiVHEfwYJJLnHeBfNltIFjqdBfk3bii8VMZPv//lT85lx8vB7R0rZD+KXcOXrd/C
-          L+nPxH5vAi49P08LLe5osCh6pcWit1flV8/AyxBL4Gcv8G13MTFR1A7TcdEVKD9HmejZkhYjvL4U
-          +IF/X2Lur1YsBEukoG/zqpi9gzmfwCHT4KqvYiXF/jA9n00Af3q8voMXTt+zLQBZ/sQkqwJz4FoS
-          vtHjQBOKWnooJHSxNdTT444dxMCPpb/kJsE4aBdWGZ4aT2XBLKC/pZT98qm4rTePXZ97J4Lt7M4n
-          sgNvMP9JAdvXytKMf5YmoDw4XUiylUWfvj6mA1NxdyNHdKhj7knUgNmu+WOBFwrDc99ob7R9ehWV
-          +ImASdjMAtRdKJFf/ZkpEa4BGJU/dp6PNJ6xQUZoHp8WBrN18Oc6DTfQnK4vuhXjYVi+8Ajhp//T
-          iSYphj9XN2f86U+MEODF06+ec98rGUtYXwzvdzut/HTf4MdDYHzeJp8MVsk1ZPgou+D742+rf1Nl
-          UVp/uVfx5qfn08m7A39JXHGE7q6J6fbjRP6iZ+6i3h3FYBfXYoAev5P2f7UU7P53S8ElqCLmpOdn
-          PGuJmKFwt2dkL5SjuezGaYH5tYWEIGHho7JMJawPh4Q2jauZAlLMB8JEsph3h994/OSWsxPUSqK7
-          EEaxdFFZDeldOJNT+DhxeTzaEN2Ws00MbNwHOTK4BU+mCbF4yAJ/1vZfD4IrGTCo/+58wm83h+wT
-          7EiU9Z4vvg3+BqBmMn0a0svnYX6SQJVnPrtN+jBwnLkP2BSVjuHJ2w7zc6vnsDInk9j22YvlTdFI
-          yNqXFXM2x82wbLWghSfylQk+2Q1YGm3qEaHBlunZWyzmo5V5cHv1ZLZ/h7Y/5ZhtoHVUbuSkG10z
-          reuBDoWU4dksHF96KQaEp+D5H66+pHdBZ9t2fj7FzX9KTmhEqngzBEQarUJAhOTmBhRRGukLqOR9
-          9xf8ndzBGxoHQDVrr7X2rtoSdr7VJ93dtlPerneMp2lPHltLYpED5SMqJ4+XJLqArJ1hBA0dI9GR
-          aYEiMwfL4SaROGlDZ2HMQpTL4swRn/PGflHD0AWOsntM5MuX/jzu6lDWdt6dXFCH/VavdgJ8Her2
-          7/lCnLMZjHO7IgendFIqfhRbpq19xMatMLQ19gmC7BgwRIfJ1ptnylSYkhOHdfD5gFU8zpms6PFt
-          yneeQnm4Xwc4+2OKj/k57ZfDbRTgqylZdDtNFqXX1PZAUUWIBFpL07WVAhcW03zEys5TgHBTy1U2
-          woqfJPRQnR08rwPkZlEkkcNGdL0EiiIHt7DFF2+8alyP5RpyFzXDDjBxuqvHJICcm3fk+q5BsZTL
-          3obcLthj/RjeU0GSlgwyrXEnJpfllLjfjw5D+z4gwd8XdLpGn0Yub0xB7LOmFjv9dRrknowTwTO1
-          AX9vphDeXM3D2m+895rLybF8nrClad90mcjuAaP1YxDLqgaf9tXNll83ncPexf34wt5XJqiUtCB2
-          dSF0LJ23B4k3i1hj62u/7vduCbvrqOIs/Zo9r34EA1b2QSPPKLk5c0OdYb99L8kSrdPm/TmvIXDv
-          LZJDGmzrW8lFO39tVdFq1K/7PIKwznoRnxbHKnZfqaulWBUDjKTy6U/7vVvLXXW5TKL4Vou5GOcO
-          WC/7gU9z3PmkK4etN+RQ4PQNFF/w6iKUf3iQzdyd8qUtctCI93usSIWWrootIykPD3es+xbTb+Om
-          Ql0P3uTB9QdtER9RBtvSKIiJXZAO8i41YG0eZnzP7UOxa+UhgIyXYYx2buxwErIyeIu/7jRL+btY
-          RS/yYMa4JX6qmZnyl65RAX14Kr55/lysl1NzBuVraMj5okONFFtJw9Ww9sTUJ6Ff4OiJ8rFhQ6ww
-          9dHhylO+wnrascQ8PANnR1w8wEqj1sRs40O/IFVBrhIfO0P+1dqIF2pYSXKNk6feFdt+MeWsDCNi
-          gOXp8FEbrtDfSddJVtO6+EKqlrKewK1qnCrpDiqeBI2iZUkcczdn2ykZ3N4XG2H01QQIkxBOyFSx
-          HWGrWAaf0WHAZpBY1ZUpBrFqzpLg0C9igtBI+dvKCXKrJNcNHwdn8qpVlN1m35Djp0odYcNnOUCf
-          mJjHnKd0UncZnIv9Bxv+qSzmoa06+NLFEcevdwQENj6jv/nO8s/F53y/SGTn/irR/uFHKfWE9Axt
-          w6REP/i3vnXHdpUzG+lov+HRHF2wAXwm8HB0LO79eq3UTI4fTYwDaTwALrzfStl1ni6JGEfvl6q2
-          VpA9Bzy1RmBRymqPBiIr0vCJOS9F77tghXWwpzh7jWeH74ejB5u2PpC4Qg/Kb+Oz3/CToLOr0t21
-          SXLYBRyLdduJfP4Wn1fY43tOji7xnDk+vCbIRIOFD/fZ87nyHQnyvlHe5GwdpJSCag1kyRElrJ7T
-          Q8pJh7YBE7JVcmLplc6jLjRQ5EwTX+v00XP3RTBk/BQWfIifgj+J4nZKVyotHLbi6g++X0S/+Z3o
-          cnAcOtSMCve83iOpVmewlunVlOl2UVzaPKu03T9NCYqL4hDvc78ALjhBFWzfTzBTe84i8pEOsjKI
-          sNG/vIL/HpIZMuQTESss1n7R7sYA9Yv0IvreKvwhQQTCRmgc4s/YAULkMSssz+cnDrPM7vkDeYqw
-          Si42Nj4l0nb6dffYJA6PHbgu/VIU4gMsCMOpZOgx5Vtt0KHF5h1ajqbtCFcnSSSgXV9/8UOQTnEI
-          qsv9hNXe8/yhWX0XvioPYMVM9sXs6pYAOO2Rk+D6YdL5vjAGjOdgwHHkhgUfJlW316cthf+2kUO3
-          46Dw9zsTzYe2wtB+QG3uMMHto6TDlwEG7A45wLp3Z1N6O3ICVJrjF9/wkdAtPkoweZ9MbFS824/+
-          3OfQhXggSJSfDh+iIydLzkBxHHO8s+uYxwrxd2XI7/lC1bwjOeM+r4m5HjU6RJeTITOqcSDZMzj3
-          tJ0fOjSZ+oQRZcOe9wgw4Ba/ySXbR85ymMYH1DO9Jt7FVf3dB8+6/OMz3MHnCyrdLQSl15nB9pvF
-          Dh3RxIGZu87TmipfMH8Zqsuc33yxel+/zlz4kSm3vvQgRrAYlLuzRiA389nGVoUegE747sm/eHJ8
-          BVEvhE9+gjO31RTQc5cu9UU1ZTTXbxJropnSb5BMckD0iVyuYAFrLXQ6bDm3wxf92NC6m0pRttCM
-          8CmHCuUvmXkGW/87fEUiLnhSvBoYpGuHHdZZnHUvHVRoscOC7bOSU6qLOAStEl1x6O8LQHFtIVha
-          Ro5P7bF25ubsJGDDe2J1IgMGgD+BXD7KfuNTlbb4pTZBZdmNaP/9wnRNptMAX2nOY+1dvvsVW0su
-          m0x5IieGcQGFiifCZlY04n2xWywk3HGwaD4Oxh9T6XkV1jYkpJPImb0c6dq8tACmPZ2n0p1ZrcbW
-          8pElNYDkcdl6SSfgzvzwDePt4jlOSt6u3E2LQ7RSKTTqtTaEoGYXcqprlS5e6CiQvwNE7Pm762fm
-          pYlQnDn9h0fpetl6x27x9Te/lDYfP5E2vkjQV7j6iwcTCZSmFOMTmF2NYhpnYIv/+Op8LxofrOcE
-          3sZTOn0FAzn0doQCZFr9jtiP/O6Xq6mGUnx+eCSjbFjsXKVdgfSKj/jsFy+HTvjlwhsdS2LAOktp
-          xwTzjx8hj/OTfofKJgRHdqmwkcsjnV237eAPH3WhUxyhOaYuALW8EG1grPSH3zBAUz5JWW9ToeKX
-          TvavuoqD3eut0e/7AyGpDUgcJr6C4cdvf+vt+okSSksx9iDXuQRlqP2kVOb6B+Svrwsx9pHcU8S4
-          kYyZ1EBk+741AS8GCmbfTfPR2jl/fJDWqolPo2yCdayQAANeYrDx2HXpHICHCduxeSNarmq69p+L
-          ATY8xOE+evbrty9sedMjiPbXY7oGWK8BcyvPJO5Ehg5tctbBfDl5+BLcfZ+ybZnDOM4fxGsnHyy9
-          OBpS8i0sbGm80a8MU3RSh5Cw9WJvtI2/r8CIyTrVzCXrl0u39YpNbxz+i8f+My1BJsEcazlRes7V
-          LQ4o9zEmVtIKDoWx94Eb3yWnXXfuhY0v7+uJZ/HFCR2NR9c8gtXXDxA7jF9/40cubLu8JGHcWv0W
-          7x+wETpnWt51WlD9hQeYucJl01+rP4xPwf3FBxxjPPkkPN0mqSQSwcpBDAq+Wt+lHBzND75lr10x
-          d2nJgJdIDazt8kmbtbchAqaCCclEWdZmuaOeXPDlDZ+41wXQA5IHgPuTivFX0dJND+Zyb+5faDGI
-          nQ4ViXOYOjidOP0dpnTkkAvwjj1O+wpBMNxIuQKkMirB4Eb6Udq1CTSMeCJXYzALApxYAFd1x//w
-          XFvHyhBkQh0DG7k5UOp3PQck763h28aPllrpAljcVAWftv0jhG71kLuJOuTMspUzFV8zh9ZBOhN9
-          w4OlpokK6/4NNj4tFvM+ngXIaV6M9kUp9UTvvx+w6dVp3b2Bs9yR4MLtffCJfm3AtZfShbfw5WMk
-          8y86eK9khTcrtbBl3/t+wfurCXHZHiawXmxKq701QXEW9I1vLGAVM9GAycjaRG17TyOlePWgYRU9
-          +fHtnXuxaljiMCGXRAh9eg9GCDa9SBR15IsZC5YBjedBwIcNr5eJnJS9dYvu+JntRec3//LGv0ha
-          00JbwXvhILN/VNhLeXPDewVB6TPaaC0p8Zd7mun7VNxJRMXtAay+kj7ATZXNLf5A0LuXVQHtrp4Q
-          lG6FMzcVfci/+KggQQLLLT7PkH4kBy3xU0jnvpERbH3xQdQ6IcVaP7sMWtGQEn9/Qj1l0oKDWi1a
-          5NVrlbbWu+QDg+ozI7jT860RRWbISvB9EH1vHIpNPwuwLJQRa2vT+wTGyQduehl7sxvR3QeLujRb
-          7A2jzOjB2o3wAe3i6+Dj7RuA9fZ4cODHJ0K3L9JFDnwPFqt3xVu8AMutvqlQkz9XjED4KMjhMGaw
-          FD4iOYcLS2tpujYwH3Y1PhQs0MZWKw2ZUzY9tb+7PTnwxxK4DWgIQnVezMEnyMGmRzHmhKGg9Ul9
-          SMpUNPi064Z+EZvZlQ33Dqbd4rTFEjjcR9746PSpn0s6BiW0AfPpQsTIPEu3eCdAoDgfxHiD7qzX
-          ys5AxKIDamtNpjP74kIwP/14AtPp64+lLQrAWjlCfnp/8bdTQ/s+qLC+ex0cKtltItsPWcXJiu5g
-          rj7LIG98jdx/7wPEqyDfjNMRGxvezEWeRzJ/3yNsp0/bWVgIM9D1LSJHrW+0EYn5BPE9macdc74W
-          9C7JgtSfcTEVdfzpV/+gibDNYY2kM3+gGx6X8r4Uy58+oYQ8vh/o9woicTbswBzOTwRf+lDjx6EL
-          /GXfNgNUUu2JSpnRfYGpBlXSivlMsp2f0NntroN8FA45Ma06Tju5dOzf+2DrrJNiDu1LAPh3GxBn
-          gqb/m29QntET7RjGpbNY5We5zp4LMS6JAnjodxxkNdQj6eE5KdU6djuFr9vk6Cx9sQ53ZEob38XW
-          Z01TvsFpBrb1QXBBcjpr4zRB/+1aCCoir1HtOQkiCh8XnLhG6ZNmoghuftamh67Fun/lufxa32dy
-          NOsBjIWtJNKR/Wb4EPV3h++6YbMUzSt2qrMKhP0+qmF3OxhoV3FawWG1gXI5uwyJa+1JxwLOhhyr
-          UvDHp9sNH0Fushdsx3RrPOa+O7D5fcSRS/cXr2pAjsNErsUlBcsdMS5kH3DGXhacteVqSiuIDeWK
-          b3fHT8cNP2TndH3iqxmfC2F0whW46HtBourdnBVujR8EfOqnvX13+nn/PejghnsNm+xabhe/Npv/
-          MBXkWNyldKjAPoBYNwYkH0lHaRsOCsTgamG1PMza2uYlAop4naY1+prpshdEBlbfazDdG7+mghpP
-          AcRXU8BpJ5R02NdRB1oLYmywXEY3PLZl8nrQ6V2qskavqepBUUpUjClj9Y0qGbmcdbU3gfHWOMMt
-          rGy5VJsXRtwu7vnw8kzAb31p0jL3M5sPKywh3GOFjUp/+On9iQ9ybOTOVDT9dk1fdwg4rMhcBf72
-          78ZPN39O9RfdeQUgVofLn3/CTyZr/vFzv1zV7WJaaILMPutoPjKiRqPQEGAlsfUE6kMBBvhYN/9h
-          /fzh186tl04mpJGI9wJDutybOpApqQ1y2gcTnT3DN4FU8Sd8khPP//PTtv/xwSl7f/XOwPvhPRId
-          VgQrAmYAkh1xyTF53MCmhwwgvRCDnug7Fls8EeQ//j3TDvQe383AfrAqsdNnp60QegFAeZIQy6Sj
-          0yguJ0GCvQFfwtmjq5jNhlxUCcKbn+M09bwmch3h3SS54JjypLg38MX328XRqPap+DFtsPHfH392
-          eO3O2vBmxdafP/Pjc9AH6IYPB7Xu58J3TVlcVAcjIYs3P1mJZEVPb1Ol9ab207/Spm+I5cI1XbUk
-          m0EXORq2vh2jDcVonuHenl1y1kFI57E3csgaOSERqxz7OZSr7H/1wL1fZUf7gJ+/YITRyfnjsxt/
-          x3ZgGRrXoccAZku+Tb/4umx6UprH9ovVqpnTuWH7DzwKly/iDq5POTYfZqB/uQFf8AP1o3ZHEyzb
-          x4SP2WsrUdn0E3dRMozZaPEHpX8mUO/tmRyDlVJ6l3hB8q+GSo5aU/u7wm1yeDywCjHB8E4JTGwI
-          WLk4ohE/pv6PP97OhUGM5+Ckq3g0VWmKDIwEd6q09S5m6OdvI4nykTPhMA6h+3l+sV25Sz8HN92F
-          1bNAEzvdW391i7YGP31xOBwbv59aHEpGDPYYbf7iGnClCgNeZBDnzi9tucX8GW7+I9oVt9ynqIpm
-          OIH1Pe0zsjUqvawlDNK5w1Y/smCu5eS8D0zv9fMj0kbuqCsXTe7g+NJjOlfhVYSpk36wpYlNuhR3
-          t4N5HVdEEaGmrfKSRrCYmDO56GbXz/LO3/RD52DU5rRYS7QE8s/vWzf9xG14AJGcFRMoPkrB1bMU
-          gfrNyuQ0EdNZ0LWJwOZvE7T5WZPKCgjo5USwkV5Mh29dLoRx4X0Qn15DysnPXoHH41Wf1qCsC7r/
-          iCWc7pKJj7bdp2uR5AMkH8iSy0UjWh+OYw1yHCO069qyn6vbO4NMJLokzmNAh7YuJLB/1znZ/Oti
-          PhxlBoLnpSe4ijVtcJ97CSbKh2I7Tcz0z+8ygadNO/awFjl+Wg/4w7fnI4/6xQ0+thzAgP3T/0N4
-          f5bgNh5T7PPCWVu/3TrIEd9o08b3/C0/AiXnCSpy+Lpy33qho4LEYDhiu6BKJ5dR9Z/+xvqJCGBE
-          XsTBG261afzxW0jtGsZ1WZP75k+vfgIgoDfusOEXonPX6DNEfuhi8wR32hwrxIUBNEXUXqelnw/i
-          2MF9wtwx3vjkn39TcfdsymSFoTPWrsnv+4kOPipYWFtfoVMWE3Y6QQcL9D+cjJdCn6QtPlMte3kw
-          2KsASW8Wa0tVXQNIYeOSw+nd9ot0jXMYj+oXQbmu+gXu10lKocVMcnaK6DiZrA0Vu8d//HE4ZHIN
-          2sQOkXBnT/58eE0KZHcnnVivtwh+/gCctdsFW2dhOz7WqQL8jZ+64fkQliP3W2/Y5F1lyzfZDOQv
-          7GHinDXa1qsiSrGhXrG+4cdffmmWyEIsTTR9orbhGSTu/oADeNc1TlxNVRI520S7UW7o3KFgkONH
-          FxPrswK/ZzGXydxYU3K8pCOlqHJX2D6zhGx+gzZFRsbsN396kmLBovQeVBCyO9/BJiNr9E+fbPj4
-          Hz87TKoGdBW+TPxzFP0ZNC0DOaN8YuOS5HTdS5YKpcE3idZ2lbbxMwU0hakQe2ck/Qw1iYHdIfzL
-          T9FNL+TS5peSZOOT3LVfazjxYY5+ft6ifgQdKpmVYDx+mpSicrsVYsuPWZ3BFqs0XTuZNYIGZ+F6
-          KnaYXbxfPgOBN8j9ATVKIos794iTM/+ma+hWmRzvxQ5n8po5y95tdVjdjxd8ei9vul4dL4Fgd2fJ
-          Tz/8/GKYfvD6Wx/FELxLXcadBAg6X97Ott9rmEbXiqjZtfMXnF9zmZhcT+yA+TprZGx9z5LtFgQf
-          PKnAlKCGWVd6ONr4E0FMFMHKAAY5Kzz157hmbSiR6kwwm1KNfKetqM0sBWyjx0dblbJyoZvyGT6U
-          R4b+9geIRHgiyf1e98uVqI3cx9cRo6b3enq5Rzr8+YFaTvJ+iOJagnxzvOHTNy3AHHwe+caRT/jn
-          Dyyodx8yhcKdqGsb+WuhKGfZXIYjPo3SySHSvLg/P3VaNz63yFJvQ8eOE/y4wox2wQEGsDzOj+ld
-          Z1m6euwH/vwJcv+mxd/6k6VaQ0Qx5p3/y/dK9nkG5D6aJyoAYhvQ5FULG0na0JVhHzp8OfaXqPxx
-          DzZ8FGVWO/fYp0nXzyLretAazjWxPgrU1sbrP1CTooD8/LLpel/P8rz7cojywtkZ2svgwr/44Ppl
-          T+BZGkDKj/L0cJKYUvnrn2H8mkNi8m7ujJed6IH80b9Qe/zY/Q5V7gxf7OWE7Z0Y080vPUPbj3Vy
-          2vB8DQ4whJseIRfTOftU+fLBX374tOHzEpTQhCier0Tf9u/sb7eaW56X/PwFwG/5UJAszxAfH1YM
-          VvVifGC6pntUb/yX1rmUQ0TFimTTjeuHMlgU+J+Sgn/913/991Yg8E/dPLNqKwwYs2X89/+WCvx7
-          9++hTqrqV1jwzzQkefbP//lPCcI/bd/U7fg/Y1Nm32GrNZAlkf8rN/hnbMak+v/++tf2wP/7r/8H
-          AAD//wMAAEU/LroFAgA=
+          H4sIAAAAAAAAA5x6S6+CTLfm/PyKN9/Uk8hNqurMuN+lEBCxR4CIoIjcCqiT/u8d3V+600mPerIT
+          lQ0UtdZzW/z3f/zzz7+6vCmL6V//9c+/XvU4/es/v9/dsin713/98z/+459//vnnv39//68jyzYv
+          b7f6Xf0O//1Yv2/l+q//+of539/8n4P+659/bafG9HdX6QW2rqlqsFeNGuvC3Yq4VzDVcGfGJxI+
+          nGO08r5bQvGIAHaUY6DSiLdr9CiLG8nT6zbQs6yEUGLOCtFcOaabdXM6eOFkhD3/6qrsYp8K6OnX
+          Aht76RUN5a1Y4DPbn2dANnng86jQIPJIOtOAezYk1B4ZiNDGYKO1Y3dTJX8HvamySSyE3rB8ArBA
+          VXwIPkru3TCBpGih2pYFVrXqArYum33o3ASR+D4+RGvhYwcetCLA53NOwWoegx16nVt/HvmYpSvL
+          1SLSUaNhVae1u2hcUQDLYEJyXcwhH7uwSaB4LVv/sIvnfEus1UCOs+Pm+egrOTu7sgJc8ohwoR/3
+          7ipt2obU+LMQE9M9XUSjF6F38S1i2JyncpXj7qAs7XV8PHm7aFNQsEPw9nJImed6NLeeJSFtxxGi
+          sM9omK/6aQfVQ1ZhY9WFaDkhpUb9YzyT2yP3AJliUIpyBGyMh8jI+U9/EdFlpDxxzkriLodHXwNW
+          owfsR7XaMA9tL0GXZc/4yD/WZtsd7ABWieFg5biWLvOuVAe9ZibCgQ0woA9pekKPWzDWMToNFFtS
+          BZ+n02V+Rp9Ls/g8SWCijjoJNinPWS4fAnjgYE6uF6ZxOSL3PVoVZsOKB3ZgdeAJIhUuN5IVnwfo
+          SpELoHfcKfNKVS6il33YIym1BHxPDpXLGfC9wfBepdjHutAsdjNB0N1dmdgl/xr43/GFEl/IaS05
+          lcj+FsDr/uyQHL5gvt1aaiGrUQqiHkcVbNoLB7ARbg65TtQGc28cYnhNqh22DulpoHIRashqpIKk
+          Q+gNJARvRUwe8pvYqMTqZGTChsJNeuL4vD0ifm8BCVqn40Lsx+2mrjB+Sb/1EvtXL9eyVtArhS6+
+          OZPu8gmtM/SpaE6sAjs5431kB7XPsJnXe3pyKde0zkF1gpDYE7Up4aI0BZzg7cgtEK18fWjhDkkn
+          jie+C3i6PJ3MB4hfHJIwtI62uahDpKzSkRTMW8pngT964JF+L8aX5rBa2ylA5jM28YVv2Wg5NWIJ
+          KbUrIm1c1rBab1lAU+kOywJWBspFQYr6tMTYmYoiohEv12gkNU/0YAwp974sARqwzBH5RgPAv/PD
+          jAolufjLZCgDa7mPFB1KdsKqZzYD51tjC3/r0wtPinihFwR4fMssTvWHMqzBZyhgo8M9PheelHOF
+          LHJIZLtlBuCpq0y2TQ50bqKIrY4/upQbqwK12lzMHPVlwHiVwaFdFp6x+ymrnL1HfQKUR2Fg2efH
+          Zp25JIPc8ATY+Zy0hhMvlYHuj3OGgw980HUYzyNUKyefF1DecnYsWQGwplJiqT4nKiulcIaXceWx
+          1F3kYWl2DxH2aYFJ/Hg+wfZmJw4Sd7OJ/RI7deHpm0N5jz8Ee3wMXvfUZODWr97c3xM++vZnBWxB
+          2r79tQ5bB4UWmaWi4+SRUZdyq76DWHMO8xLgNWdWtSrRzqe9z7m5F3FZH4boUPITxsk5oaOOxxbw
+          pWnguH4BsKrncwtTMWZweooMlwlSxUetpV2J2XjYZVZx1sBSxW9y9ZOHOrrWnMBH2Ij+/gbrYRPf
+          iIGsNN7wbToGOV25uIdKzKj4pvZKs83YKUBZDzoxU9sDU7uOqfg4H3hiyMjJF5pLEO38tSdGvAXR
+          IlhhiVhTKvF1f5iHBZZqhuTj2yT4/lAos6cwAU9LPhHTUd9gcU9Vj9ikfRH/0VnNwr7kHvnMcyNp
+          Jx/p6kVbBQrJ3DCuh6aZDLmbobK5L7/GWhctD/5iwe/v/gF7Vb4x56eCtNqfiGS+52ht12cGHdHU
+          sJMulcvL0WWDAT7YJJ1aX93iLvHggSEcdtlABfStfDqIe1nGWfSMKGN5egn0aYdmsTUbsKX2voTe
+          ESoYFxcl4kqX1gh0DiBHhVHzeQgfT/FwMiLiNLSOiKLzEB5Hr8Q/ftuAt/OAxcJqHpXDbVgXcHjC
+          OOoQ1q7o4I6P45qIfh6P5ISnAWykC2tEMPuaf/tFj/aiISokNrHtt6Sym3NnYK9GrY8GMrvbu/MV
+          mEPpTLTjmYuW3/4XJ7EkeL69XSrmnwrubn1IfnzIaeZ1hmzjd9g3F8ll2qmHkL4y3aea/6C03rEQ
+          JfH1if17bLr82z+J6KC/O5/LjXUgUUVSMLytJ84+J23YLlsiQKvNrkR3/aXZmJnRkBo8Y3J+QSFa
+          szbg0JevfYb2n2GNn8UCw+oo4LI4vBsOZImELkN/x6b1uefrrV5n2PLem0TG5xGN01UTUPN6GPh8
+          7xLAS5+DCNF9KMiJqlxOv3iKzAo6uJwDN+fAealB1VzHH75GvFyEBnpFlw+xvvjPXszpCXZVeyGq
+          8ZFzbrDfIfr2K7mRDtMFhJGDXrV3InE1vxp6zsIEZUqEsS6Kb9r1xSjBL37hkAabu5Q3a4Ycw8r4
+          MugLWPCrltDBd3zsnYtjswpeLogPYkP/ITzSZlu3yEeErzJsSq8XWIoFtWDeWwbBzZgN9IAPAXyt
+          W0qKlq7R0uw+IurvoCPuButoLdAaI66QGCKdJxpR4XbSEBHFCzleqnuzSm1RQ6WIihl9Py+Y2gyM
+          HtY2C54EaHcQxRY+4SJjS/LTqJt3uo/QTnPxCQxXd0OiwEEyhzE+HuW2+fTmOT0wpXciv/Xw5jx8
+          9zu9Yv0FhXwzBCVAV7UVieqmdbRSfYt/+Ov3xUem88Q/BBj5XUsMsclyKlX7J6wADLHtlly+DYdr
+          ARf3xWGpWlg6MrwPgeG45kz8glWXg7i18ChwjxlMF0Vlvcl5QoERbCwny5wv2+SH4AWKhGRji9z1
+          HvUxVNXd6VtP+4gk3lYh85SpxMCx73752UJf/CXHbVsGehJWDUq6z2D/hNR8rF9XBl4Y8YVxZ+tg
+          YSQ5BjDYW8Q/mU+XXtJ+gT89EwR4jbbLVorQcSA37wrHGLjfZyTpyd//s+WL9IB5OMnMheUMFnvs
+          QrjQcCBWc+0p5cauBBZvCvN2Tm063gQrPvCuqxCzTQiY41Qq4GLhA9FfMM2XV+KUP/wg5q6UVF6R
+          sxJGz5rM/P74cOlyihMYYGD7rCaSfKY23Q7snulJeY/f6gLC3IGDHBb4pw9oVL0z0X/0MzHC0qfL
+          rtM0KJDpTvLXkuSMV/mcQMNSntnD2Ww29CxHGEYvlpimi9xBzB8VWu8FMzPBnXen+BkvEOVPiZwK
+          fxmGDIAKXltdn0F63ZrVYCQOKrmlEZx3srvy77yGL0MyfXCw9WGLvNSHd5/3/+qLNOPSw/5+6PAt
+          pb273ZgMgm89EOXLBytZ5AT1xqfB8bCadHoMCgdAKbAkf/DxsGriWYAWrwvzYiiz2wu87sEL8Npv
+          /QvRgoqIgbPA6cTzmB4MjHTOYFBfz8QZrIyuWrAYsM8ZD2e3u+wyCy9Y8KI2m7+ajKOOjr6kkN6i
+          C3F8NgTrXlocKAT2lShYNBse9niEt1s3+3vqdepyE6wEzPTYYIcTdxH59gNgbLUj7lffrrZ/T4Aq
+          NgIxhXMyLPWHj+HrcFyx/sbKsDDtwYM7WuU460MwLNbu9YT9x519npNZlZaaNsKECD65nZZ3s4mL
+          4EB73kbitxe74fTLu4T4YqD5oMUsWAc2UNA1cmX/kDZazl/roYDsKLbY1Mr3sKmSAcGzLSWffRkH
+          dyBjBA+xNB+Ids+BSyv/mMHieuZ95mANoL9Jbghb3n+ThJwll1/6MQUbmoJ52TixIT4MfHTB5dnf
+          ajINS0jqUfRC+Y4DdfYG7mLaNVy7UZxJrvHgc/IbBnn84U0sl23AWs4fCM1q52DJWYm6ZvtOhJIv
+          bOTkWEOzuL0wo/3rORH/uX2ixZRuPoydjCcukz2bRSBOAuvMn32oqLt80sedIUriR8UKDUKXS8PD
+          Ag9P6UJ0MSkHWnIjd/jtr66JON8YS/OBOgGWyE+ORtuoywvMUCzhQIgLSr58DTPaKfjLf/mf3iv8
+          3ZV4mes3Cy+bM1SPHSYnsa9VEohVByNG47DUiFy+lS6ogZtEPrbibjfM08Nc4EGuPOIuj5quetHF
+          cM7yhOhNUtEtvVxCeD2o9nyIWTGnO+TMkFiph3VsbtF7X6sGGO2e+Igbwmg7Z7sUitlrh1X1Xg3r
+          +F58cJ738dwbcFUn9XxrQZG3BXEHVRrob796lq7EPlUPQKZZrRFzHgaivoIeUAEYHRSz946ogh4A
+          ktA6hfVnY/99/i8fIf598Yjs817DcMsqQLQzXPzFP0Bi5cCJ4Ozx+PaWjurSDTRDAGh7XxiXP7wM
+          oKN1LC71c9Ws1vNQQmVVjsT/PJx8NbJlEXnXVrC3vRiwLHqgoGonJETVtYdLlayuoHXCC/nqyWFo
+          Gi6Gvl4+8S/foNx63EGXNJEv3rVHvrjWHAOWs0aS9AKrzg84auDxnhtsmYYC1o4wnKjdgw6fK/DI
+          KX3rAco9wZy5eeXo9uLPGVLKWiZ2LbfRTw9D83y7ExM/PmDbniCFy5HJcZ4xobvwjqJAhjEyrNbw
+          k1N2DiT0YJIW2w/3rW4hP7Vwe95d4sGqdpfcRQGsFpUQK3JgtI60KhF1LhF2v3jHSepsAAF4DnG9
+          N1bX/ONvkD0md4zjUmr4L9/Ape07bOXbXl2//AQ9OZa+fPWiWy3CCk477BH1aIkNPUwzBx4yl+Oj
+          OxcqeRrWAv/0xyaBiJpzI6BgL1+we8YrpT+8vh9EiUjxbosWGAxP8K13HAKBoev1bmigPj4HHHSX
+          R7MYb92Ak8hi8vN/vH5xU5g0oUfsSFdcphS5EISKkOAoFXfqd70Fuko+Ip6lf+jq0EsH21N8x+6P
+          P27HQYMhMe7YOpfo3/X5rT8szzGb053TP4HpavJMxhg16w3KC1Tdaz/PA7QBt7fvBtxOD5Momty6
+          E74BT5Qn1/cP73WXD86Jmf/651ifRnVxNYsDlsGFvnX23nSqhL6DabZMP3xylycUY0Bvpwt2HONE
+          V5PbbVDbMQR7F/6o/vQrEI2qxljpTw3NKyqg4+iX87ztx2hbr64FL0JQk3i+3KPV3skV+vGHxjdD
+          NH3zKrTL8wmre3yK2JLiAuxSdfH3D1bLty//wwMKCux23cNdWuBsgJTjERvZZRnm08Am0ENGgCP1
+          Lg3bugcSbCuhneeDR5oNZKX0w8tvnrRzewZGAkTXS46P37zuxw/og663mdW0QF01NQggq3YPfDMb
+          LV/bvebBefQUcvzq66++H8F52WR//61HTnlKogi1i058R5MHTgWPFLClupuv5iseKK8vLbRfxoxl
+          FVgRG3owhIdsO/mdcnlQOkeq+Os/P3nPTETf/lVA5vl+x/gzG+58dGIHJhf6wPJ2vLj0p5cW983N
+          HI9qwHz91y+PxAmagPvLqyCXdvyfH1rcfhlRdGoH4h99JaK24RXgubClv2zPZ0MzQCuoUu1ODJTr
+          KhXDtw9usjJiA025O/D68kSLU5YYC72Tb87bTSB0BoP4cZPRZZ89A3QxAMHHlDou8/PHl7tUEOvB
+          9MN8hfcn3Dmzj92RCnRpLqICv/4EZ3xpNuzAphK82iKYH/a1b9Y2LWLIiGn+0/tg3tsXDeb98UNO
+          +6vqfp/nEza71SX2iPpmkUDjIDAXEGdllaqzcEMWnCluZjQdg4hTjvMGv/kSyVi1U7dL/EnA93n6
+          +85+gZdeVAlSLE/CZ0ZB0eTA6w7+9Ksfn57gW38LotStiCGMmrqZpzBFev+WZlGAHzqq/dNHM/eu
+          /F3GF+78zRfBlF40YuQwANtJPxZ/6/Wsyx5sRa4IcP7sb9il6JAvQfGeYWBgl2BSlu7iJL4F2oFA
+          4oztzd1U2w7+8E0r2IauJcUl/Gyt+cWX2t3q2OtE9773iPY8oGh6o6KD2iAO2BvqMNqIXPfwkTqF
+          v5/eb7peqdD99IAvelkUkQkMCfD0vMDK8d65q708RuhhMyOeU9kR99nPNRQE9CKa1Z7zCQdRCue6
+          1vDXv+UrBIUA//KL1ucpOSXz/G9+y16Gy0xXT/jLuzwz6ZpfXgS/eRHxknCk8+oOsTg8wR7L2yrl
+          /FfPw57priRpbUadzb4vIXgwIj4NDtOQs04WGPElJdInTQdu26klkLlrTOTCD4Yfn0ETnN9ExtNA
+          1+3C98AUPfzLgyMyOOIGkof6xtrl8hjWrE0Z+GkkCd/GEuf0rtysAyO32J/NRVL5vUUlmMDq9M1b
+          ifrnt1hYTsRLWz/f3p0hwVFN5nmfbQ/aTw9zExMi+tj95oNsfRsU8KuPQ/VGgOJH8PzlV8Tuwd4l
+          Y50HEM21Nh0y6zmsjOZsoNXGYhbdIVT5X9796qoQOy1SXQ4S6wkv6ZiT5Du/+PJfB/V4FMgRF4a7
+          dUcjhYkYjcTTPEvdgMf58KimIfGxnjZrZ3sQou5Q+6I7bCrNzYGDz0Oq4fIqvei3XktxWLJ+7slL
+          B/xLO/l/eYUk3tp87o5GBpmLGPi7wmmHP36P96zxxXM3XyQwWJANgie+TvRD13yTjYO4jSPRrkPj
+          zsfm6cPsfIzw+evH1slbWngJs8xf4GqCX34M2/75xPIxHKPp6Uc9SpeOIwmDD24/2O8A7hOlxd+8
+          LCeXVWTE9q4fsJOLUjMpy82DX3zGapcNw/Lhhx4G9tsiqrpyw7AnvQheCWmJZWfP/Govjxl98y9s
+          QB6A0ffl6nd9glsORONvnvDje5n56dXq2oKfn/npwxla6QhNKvik1N/JsHa2BqFwlyTsDONT3Tgj
+          7v/0u2cmVrPBR7X9zYOMV8RQgoq3COPrbofttHnm1FiiAvGlbhCny1/RZD0PBfjqlbmO6CffvvwO
+          PLOXiW0fOJeOhGrwYUeM3/dhPtDLujEArYWNy+98hbocq6BTptnf/LweiGUyCjyctIgcdUsDzIdv
+          OsSerhArWHwPy4bTGd4HOcWeBj7uNlR2CszLqH3XKzWsZqIN/vg+/u7nJnO+D2HLB9hg8FX90+On
+          i6MT5+s3OW22W1D48EpUWZRyqm2uBMxTquIibedoqa+Pp/huE3YWfTak9GlICxKOx5j8+H5rX4EG
+          88izya0/+HSpPXuBn9hOiR+kjUvx3rDAYekyf29PsbsN4ecvz8THLHWiadoHM/iuZ96f2jP45B9j
+          gUZ0Pc6vcdIAm3FzC05OpRCl7Ug+x+xOAtuFt/ydqMz0O1+oAK4uKtbqV06XtQ9iyJbyDkvDQ1Hp
+          a7A1EIBPhk2OZ6Lpy2dAVeHpi9/I3bKZXyDqQD0Lki/kKxXCAgVe482LOo/NepPeCTzorw6fmbZt
+          tmzeb/Dus/5Xb63gl6/CY/MRsM60xjD38dWBg2h08w7rwrAap10Jf/2F7YlRly5sYlhu2CPOEr/V
+          wXp4Gtx3u95nkW/n9OePf/mDZvKxSl4g4eDlcxWJ/vzq1f3RK2D/secff+S/+oNLzEOsfPMd1tpO
+          Ifzqdf/tdp06v520hOj+KXyEZobS+iFl6CgpZ6LslfPQz6yYQHGfzNjC3zxyQKGATtFm4a8/a9hP
+          zHpIS+F7Fu/JJdocKSqQrbQN1hcBNkOXKRb8+hGsvoc+3y7PvQGUFNz93BUmOt9f7hPm06vF37ym
+          odJ8T2FUfvko5WN1+eYb6K67AfGFmz98zP2qQQpmwx934NF88aWDrXLjsdHd2GZzPXUUd1xxx/i5
+          f7nrWLLiT08QdYymf/MhPnL8fBDIrC5vA3zr07eJtul2tInsrQNWVJzx9SV27vibp/z4EeMwzhfu
+          8mDQ6/z0cXn062j+8hecBUbH0usoUQJ8e4Z2WgzE5vJxmIDFLMhrOPc3Lx0WKgsCNOfpiaVlr0Wb
+          asshTFy7/OU/dGr3mg+Hm1Bi97DXAPf1y+C5n3Xs57me80v/zOAlcFn/dHZ9tYN8VKCkgA6+DMwK
+          ttVtElQVqUsKsFZgk9duhHfIaVjRokLdjs3Tg4fwOPjit76Xh7ZXfvkh1jumcUc7Sn1xvsjXP73I
+          nMtmQb95LM41nk6KPSVg5l4Vcc7COV/0vGDgv35vBfzP//z/eKOA/X+/UdBbo4KlWlApS4kXwjO3
+          L7F1O2gqv4mkBHfvfSOO0EOw3HfbAoMt281P/1DTFaifGfFzGpOss6Sc129Six7DLSbODRwb/g03
+          BuLDm851HsgqQ/grB4aX4hOV6lq+vuMYokkGE9Ev2wVw13EOoM7NoR/ZGo3IJr4L6DKfAOfhOg+b
+          WZ6UA3d5jETT8wFs4lkMID3GLFGXbXXXhAEc6KwSEvzJC3c+Wc8OdoJ2IKFp+IDaeZWheXXf2BRF
+          GdCo7zgolNKJZNJhUEfh7WTwKKU2Psf8OyeOuITgFGXOzNqjOsz06G6wo/6HSItk5vQk70N4n54f
+          Uh5vmst05ykF9nb7ItQ5dNnl0tZwp/iE+PdYoXyVGxmcq2dMQnYfRHyOkg7O+G1iE9R7dZtjS4Ts
+          6I3kGkqfYd0J7wUFUz0RZz8OA5HDgwinOLBw0SYlpaGoG6j+OIQoUt8CcmpHA/Zi8CDYdc8541Sr
+          iKKNj4l05xi6wIPqwYA/+dgzkRrxNpFT6JNO8cVBXcDgHwINZT5PsMvUI1iXbkuQN5h7rKToSLc1
+          CxmkdH7pbymaAG0/JwYORw/73O1EaRvGUY1CVTGJlyttzt/T1UOfzycl4ff+GLPODHTq+Ml/T4Dm
+          q+ehArZ1GviH00NWGemVFXBfViI2b2Ha0JPMh8grbj4OHwk7LFphOdBoY+zTla8iKsutBq9+fyda
+          zNUD00WBjxZ5W4h5fw3gV1+wrbOApMfgk9OjrAQw5wOeOGJpNhMnHy14ilKHZN/PSxBfGOj6foWV
+          C/kMq1fKAVovoU38T3bJ5zWsxkM0qGdyjIR3vjFmyyE2MFdsfrwL5YbzQ0PHB+xILngHdcZKOKIy
+          9jwid3vkjqUx7mBymlx8OaXXnP/VY7lr9sRp6Q2wN72o4Qe+DHJLTyTfYp9roXjkS+wEcK9u/rx2
+          kKNdRy72PLnkcoEcfGSCTM7tbKhMPfoblJVJ//bXM+qaZ9wjR3oY5DwOL7oNH0eDb2yURCrtLv+w
+          D9ZDeZ29sR1Kn4ZOBb/B4mZF5MILIWVdpwrgUQ+vWJmZcaDRA1uQE+8Z1mPejHj3sSXo7awPot8C
+          rDKhBp5wfpVPYshxNfDyPuXQKOozOWLbBby1e3UwXwodB285HTjzMFrgfZswxrIuq7xUlxYYjj7G
+          V8Xo82+/S9Avxhifwr2dM9ULcuhjWBccc2+pYfVj10MbvS5YOzLRwAMz3KCzmS+s7gLL5YxWddBx
+          e4B5e386wAk9KKD5+GBi159ntAyT68CiFRjsWNpMabCvFQSaWMDGJM7NMkyqBaOPRrEyPq45n/Le
+          KB60u+zDiK8bxug1HwmhUM9QOCKwXY5lBn/7XR5bAdCgJomYnIhL3FGomuXh7UQ4weE65+pRVdf1
+          IXkwMWE/Y5LmzRZXSnYIpc9IDM6RB8qnFx8pUlrhy9Vu6LpZTvg7HkvRbXK5yPtk0FbYAasrL0XM
+          /qkoqNw99rMQ50eX0iufwVzOcxykQIs4e7wZkN07GfaN/OwuOnk46KY/Xv7zNoeA9zLJQ8fM+RBT
+          LkWVLJJXwn1zdYmRqphOfDxbULAdeYYcp4GpPKkZqNi3gH0vjWl3dIURZj5LiBuUIHr3IOvhLDf9
+          vOST5DJve1fB1515YvNyFugmXfgZhaZw8d+gvrvLhdNC9JoEhdzY6EQZBQoePN1NjxzhWQJb9aAb
+          GrCzEnnFi0rniqnR7/7ujxXT+ZHJHYxN8sRHv1fztUi6GDoTjn2eQXW0rTaW4HvYppmrohqsfNw6
+          sJ6VPXGbwwmsbOYawLA7CYdaHLmMb0QbbHadRSKqaxHzETMP7oPbDeuF4n4ZXlnQ57la+DLcXZV9
+          GakADbuX5v3leXPXXL5vkIQWxZ5zdZtN4c0WAbc6kAtiw4E9jKoApSI4E9PNr8M01VWKRGvvYyuZ
+          pZzlXgcIl/wRkOPpFbvEJnYq/vjJiPo3WJRUTVFwOC7kHuVdvoXFsQfTfe9g4xLH7qazJIUIxSU5
+          i+E6TPdiC9BxkQwcv70PWPprbYHlzhzxiVnMYfl8/BQY7rMm+qtSXFbTfPjHN9JwGgZ2v7A9Egza
+          EofzmWY6oqsDujzrsf3Sx4bOmxWjTrP32Hj7Uc6a0qlFYkCPfpux7rDZJ3kH9/vK+D0vl8s7QYC8
+          0SYz8Bis8hY+VEhtmpJoLk/p6uzWBPY55rEWBACsySSPqM3tJ84K/6Hyn/IOIZ1jGUftq8wXPo9q
+          8N1/Yu21PN80GnrIjKTyV/85x1+yEhrAv5NrR1I6TndGQ+nNkHBkdK7LbKjZQL4zP8R6c2rDS6ar
+          QE3cX4mbhK27MvdOgkZZBeT6YSx3NfPnjOb9USPn1tpHw10PfGRkOCJ2lZzcRW3EELKBvs4w3Gj0
+          qlQYw0TVMhzW+RMsh+0Qwqt+C/F5Lx/dP/5dcekQtwUS3RTpE8PnyxCwd9JbSpUDFX/7OwtZOg5b
+          1/chCNN7P4MxvIJt7yML6qZxJqVZaRGnSJ8EWVuQkujHl8219RC/xiq57tpLs0rqFcJjRzPiX5o0
+          p9m7KpEKLoU/R1nTLEtrSGh/Pk/YPHdeRL0nrsEOPWYsnZaqGQXl3KLySe9YcZbc/e0vYBlpILqG
+          Onf5Pk9oZ1pCDPLOVf4sXTRI6BRilT49sKjBKUC88UxI2ocDXflDCJHh2Sz2NEPIacefRWjafECO
+          yvJ01/PxwMBht8z+S3i93CW7phDaD+GMtbF/qFQpqwVa59b2r5baq6v7ohm8+fOGNSY+RjR0Uwhf
+          J2rO+30ogo1h9A0dPd0lN1+DA6WD6wNM2BBLGJ4HcgmUES137kgwwQbgd8dxhN65K4hbD0ewvfZN
+          AV7PmfWp93LAwkefFN6dHvjn9WDTbaHBAmkva9/EkWvG8T178Lu/WDo5V7BueRP+6QVrqDBgs+w2
+          w/XqDdhn6d0l7apCqK+MQvRrW4O1f0otOoT13Rf4jKhTqvA7CNz64K/AI8NcE56BfBO+5oMu5+rm
+          OxoDUz1QSGR0g0ssvNZwMrzen6vaVPn+vfbwPtvQr7pdmi9n8PEhew1380jPx5zy7FUQTXAlJErq
+          PqKnSDWQvd27mb2RS7MduD4G+VQnfntdYPTjKzFMbz0JLuYh38AnD+B5CPMZ6J8h/7xf1QJFxSnm
+          VecN9VN1Hwv+9LKa7raBfMoLhOGh5GZoIjXfFB4/wVefz3ALSnVhu74FLKMMBJMUNN3hHoQoz9oO
+          6+vhA0bDPxgQUijjZEkSQPSb1UJ2b2X4NFMT8Ptr3kIxFyCxlddGKabAgZklIXxfPaPZkpAm8Nit
+          GQmFPB1oCXYFjHjnRPRN2Q2f00WzYHgoOHycLV/lWjdMEeUs6a9+F2YpxsMOwhyb95dLp2h+CNDP
+          zgf80+/bU7cKqGtEnMFlvzXPDO8DYGTHiEhPtvnWn+rDA34uf/qUTz5WB23T6rC74UM+9tfeAlv4
+          trGXBn0zI+lgwCDajsQ9HxMwvrlbefjIfEaUUjzmlF73KVQfLcTGgHyw1fE4wjJXdtjrHBswiyMK
+          4hieFiLP1KRfPWAhnbnwxOucD6WPFoWQibeV6D99+Sz6GjiHOPjqaxewySTPoBOMA5HV5zYs8M2n
+          EL55Hcsuow8LeNU+jFL9/Ycv4/hufbgI+P7bv2GJBtT++NRn0moGNMhYHwrG2pJ4f18AvbVWAFVJ
+          uRFrG6JhAxCGMCTzmXz16kAv8NPBFloLvrxU1eU/cdCjB98GPgN212iM+WeBCgaG5Oy4o7rK+tsC
+          bb1vsCfQZ7NF4FZDvfi+gXHEz2Gj+rVHKX9l/vrn609KKOdiSCRt3dRJfe4V2M56iOWoXMH21ZcA
+          arGGT/tCGdjzI2Tg476ciCx9ancRd5UFNd1Ec20aM+2s2YKwC483Xzjtn3QV908BflQKsfTrP4Um
+          tfj1E3+/f86PkIPi/orm/TDjaGvbqYbewb0Ro5ZINF2keoHlIuZY2sf3Yc35IIPXt7r62ymSB/Yz
+          bOMPf/3dV+/Nn5cegy//k2Q5KiqXzH4JTbQ5WBrvGNAxFEqIXnWFjb1yoot4pQlUx3nAGs8+myms
+          qYT82hX8QxsFA03u7Q4+K78jyil6DMvnCEJwPaQekYubNKxANHfwk79Foj/2TrNeh0eNjpn1IVF3
+          tfN143YpvJC9QJyCK/MvPu8ArVgem1s/AwroYUYqinZzfwjUhu00YYYjLGx8R685p198BpF6vOC7
+          wIVgHrRziND+JPnbVgpgY09ShcTXcMP2TbTA/HoGCbqcLhbBvVyDVX3uJfAxnIsvvq/PgZZHJwbT
+          hwzE2b2LfPucnon49cP+9BJydyWykIjQdThs74eW0pfAGbBLqmxmqkUGFMmIATXUbiQTXCZftvG+
+          A4gzKXazPKOLvMoJFFgu+NNTixa2BTq3mfvHp3y64zjkbfHggxZUlE5nvf/pQyybuuMuR3cZYX+N
+          zlh9oYc6v+ZLD7f1KGLLYR7u8sUT+PXr/t5UqmFB1TGEe2QPRN2XnkqC0ZYgNy8XktKlpj++g4eH
+          G2Ap6xbw9ScC+OlL+Z6uaj9ZmfLzG/4yLLG6xESMgUhngJW3U0XL50hDpB+BiV3Bu6pLOVw34F8Y
+          jTiNcFSZbbzs4ORqIvbKcWvmu3gYIXdc3V8/N4t8q2I0HeXCf388ni5ve1cD8ciW5NiAPFqebjLC
+          MnFmrA+wV9eizAoY9pyOvZRQldLd6MPHcI+JUW1uzr5YXQJsHzjEa8RkWPeKXIPo6Te4tHClbo/M
+          7sAozzd/VZ9bs371tdjEd+IzxgfT1czHEb4PjIvP5CYP/CJ5BdT8k4OP960d/vxo3I0Lli7mIdo2
+          ORYga5wDgqPcyrmREXaQVjyPf/UzlIfPV/+Z64yyZz1sTnR3AH1ZDD7Zjy2fX6NSol2nuUS9+Xqz
+          XmnUwStrYmznpTQsKI0c6Cyu+tN/3/piOTSkh9AXW3qjE7reRbB/lB+snxJM1+HALYjtQwf7r1Pd
+          LPzMVJARuyM5C6+Xuo2BZiHveTr6M0z0ZrGPqfi3/zg8ahGzrowI1BC6M91dIZjkPt8dEEpKLDPL
+          e1iYgs4QBX3gl8k5ztfZ7gsQtg8L55ZOm2+9Q0guqjoz0rhEX75+gikOLZ9LVQwoezBD8M2b8P32
+          gGDp2oOHJvkwES0w3+6HzVQNviZRwZLOPaLpKX4gmGrx8OMbSorsmgISOhQ7d3ttiE3kDF7TXJqv
+          iuHk/PJcQ2ih1Pyu55kvSNscyLhKgK3q7EXMIwlamDpW6O+kkh26w2Mq4e1T3onVrkw0JVYsoCRR
+          LOzqjgLY+5KHv/zliycp2JruOMPtmoN5/1gx4MdwKVG83U9YKd6iu4qRDuFF73RsFUxM3zjtJHgx
+          27c/1kfH/eVJ8JbeIuLoXdss/iA9kakhCWs9l7hUDlcBSJdM/+JVR9dDWguI5lnkj3K/uCtYrB66
+          vleR77SLkrPwFJAovFycHqZ62GhXipAG4XPefTI+mp1iYRCDxYr4sx1SynCbAcHlCLEFbYdyBTwJ
+          kBMqAftszH3fAwgk+NV3fv1dzxSUuwD6yzEh8aiJOVmFLIB373Xz9+g1f/20qfzwFqefI2y2jy16
+          kJmb0ywOakCZZ1HXSODyhsh+MNGxejEMuL7ldUbpCUfb4YRqWFlJgjXpaNJ5a+0ZpjdNIpe16AGd
+          zsfu7/l3X/6nql4zaEtEg/i3KnUXh8lECA32gJX2eRjIT68m6QP5+w5Gw9pOmQB8e2mxRl/rQFvt
+          pUCtT43f/Q7zaz+UEGlNTLQgyAEl+ieFL3ppyY+PqaP0hWiXuxaruyuk69KJCaxufks0P4Lu9t7i
+          DEgPqyV4L8gNe6zKAlzqJP7pUbo8ebaH+0fxIRr3rhpai2UIKedIs4ov7TBdclaEbqvcsVQZnbvu
+          BLIA/LQCEvkhHy3PWMggh8lEjEAY3f5S1j7MPortH5b9ANawBhL85hXk9/xHNTj9G2/6/egODLwt
+          2h+eqY5dN+x1+FTAMquBaCVM6IrdWwha53nC8SroLr9del/MVGsicX2bo7l0SuUv/0oDkW3GZyyk
+          0EXMjWS/83ft6kHu0oz+RQoSl1xLY4FfPptb+XaO1unqj1BpwIPYdDWHd59MHGhnM8RycasG+k4N
+          Cb5D5Yi1mFOa7WZJHJRahZuZr1/aMvuhAME1Gp9cnkhdV+bhQXC85cQMDguYeA9K8PHs6m8/Ejrm
+          1tEBp8yDGLNnEK0sm/jw2z/Ys0XVXfgZVkBG5UrUU2xEnXJHFVjoGc/cXjkBOsnP/ne/PrRFVWUs
+          uZZgLl9z/M0jAC0BVwBmvUmkHIs2op31WeBXT/75ySH/sB1gw8LDPlh2zc+fQ9lUFOzprvmd6L0Z
+          aJwvKjbWvP1OzyoHFdPUYiV56A1vfCMI+0yv/mWccM4cPpcWLsYgEX0vT+78NM8d7Nn0hG/YHsB0
+          EZYW8WuiYifCWz5ZInyCOGIqHPb6qi5k70JQ3byWaFpSDKtNlRJ6p/MdB6vvRlRMeAU+X5rgg7uT
+          RN2kdTv4zT98OB5e+VSygg9/+aV802+Uup9nAejFTLD3zmd3IFclA788Rpd8b6C/PGa+9S9sJiYD
+          6GDKLYIPKSLXTLiC9atnwVOtruR2qekwOlBXUJIJKS6+fo8ab1MApraXiHWLUERniHtYL3DGkiBc
+          mu1hIQ4kEtn762nVc1rY+hMusCq++dwz35xnD+FnETTy5c+GRafXBtdSA+Tq30P1+x5W8Offse6z
+          dDKakhPHvERYr9w9GOR9wKDHLed8roqU73zlqIHH7qrMdOWlnE0VHkJ0HB/EvNoNGL/49peHcnIs
+          NWsHSgiup6eP71zWN3Pbvmq07jUTh0yo0u0z8iUU6QhmgTVzwM3vjwDN187Dv/naqtQPDvaNyv3y
+          3JzNnUMJv3msL35Yj24KuZVwxi8T57pTg00fOgN+5ND8y+M3UTmF8LKTk5lh4imn8nyUwJ2/OMQE
+          1z5fmLP9hCo67fAPr3j2Bh1Y8uT4zW8ZuhzuaQC/85mZF8UH2O6fSeRTMK6k5N82WODB9YAk2Rmx
+          +tAFS/ViOHGruje22+3qTkDEOzB8hBI723RzaWAWPsiha/p0EKthtM+qJP7wUvnOc2i1Bz4MT73g
+          v4X33MznYQ0RKIu3jyLhHX39A4RdFQJ/u412RBylLmHir8X3fNKwxZWTAWUyZv+bL0SrKxtPmC5v
+          hZjDjg7rVw/CbR41LB/ujbt4M4KwavfmvKTgmS+Ccn6K58UacRwEgNJD53Dg7D897OWKkfM/f3ll
+          dYxNBtX56uwOCcwOr2IWbLKCb/9IUDSV8TdPyGvBMDq4vDuPlGvfR995nQDw4UVnZgB6PgYHr4Ph
+          En3mgwiHaApKLoTEQi/i+4Pl8r95VugTewbh6jcjn0cV6p3++e3vMOJ2qh+CPZccfCEdKnfudFRC
+          Ftc37Er5O6IfVePgSZbJrIjzAOY17GbYtIyDoxjz6trcSAC+1/cpJ7HqOvVh9cuP50ucddEiGmGM
+          uuH7esKX71feYyQ4V23sA3HHgC1zdz5gw9KbT8GYqGQq9svf+e9V44Ll5t81WN8zn3z9JCXoXDKQ
+          QOFOzp7lgsm6P0rx56fNyHuqH+yeQ+FzF3l8fJE2ooJscVDtFxu7uSwPy5p8NijnQkjsD9Op/HTY
+          SqgfD6YvdsutWYW1Y2Cl7wExxCj5+hEQQE9aPJyl02eYwp3v/OWP53tT54uRNB4qLTPAfrsjaidJ
+          r/5w6thpPsTFSnuNZj5shjkjyqgE4PXtd5QkkoUvwttv5u/85ECNWsXOlGgR850fAVvh/xcAAAD/
+          /6RdydayMBZ8IBYySZIlk8wSBETcCSICoowB8vR9+P5e9q6XHhwwualbVTe5DMQaZ1YbMr+v/vxJ
+          fN79xJWZfwXwUPHAxvVrDsKS3Vp4L73DfuKuostZGtl/9ZD7DdmuEIyKLD4SIyPnWxUN26bkIlyC
+          7I3tAEjRFvhKCx+TeJ+F3e/Yqs9Swa8GCYnVQz2s6H6TQIvzN1bfyXUQiuM7QOmPav5kTTUYqR3O
+          cPfrZuh+cb2cX/fx/9lRwP/vHQWshL5EJ8M3Wuua9aA1hsRfH7xYLz8rL+DIHiYiHx/fgbQ4ZuDv
+          6/nE8YSrtsrzq4fWMtozB9EGNr8gBbRMsyEntvVrLrNFHtR3OyNqlAUa382eAR9BcCAmb0zaUryq
+          BlnmqcGWoFpAMLQ7D2Hti/7RXNaMQFC3wLynZ/zitwNYPGdrQXAWAmIf7lJGKm/xIXe4TTM8/fhs
+          rGaXh3bZ875wkzowRTf2AY3ynZFMlSt3ii+MAw2Po9icbgRQc72paEqHmHhRYVP+TLsEuiZc98/b
+          YL7lmQ+cirZEH/c9PZSwqmQqWPJLWxYAXRylRd6l1IgcRs7A9hf9B1GnVzio76O7ro03w+meGUT1
+          3CpaUTap8GkdRKKuGqLbfAl1APE7wNdv6dH10Fw9GIXpicTnV5Otd+DESMw0Qk5f8QP6dr4EaHL7
+          HLvkQOk2UORAuVFeRO/TeKD8O1JRD3FJ1Ky9AVY2rxuyeu0+U13+0q3KFxFO7i/HSo1mumrxEUJe
+          W02iKGc+W6N6EdHq/+5E16E2sOUn6hE10xY/tVjX5tF6i6jgvZWcmbswLJfBfoBU+XDYVGxYL87B
+          9aFWYOizwn6O98vcIbxI8Y08++ztsgk/VKhc8gpfX+/3sJ3pO0avu9P6vT5OUeOZIIfXTNQwFjmd
+          Cmz2ZtByfPrYdWoVcKeyC6CpkoIYiwY00oa5DpJi+ZDLR/9k/O0VJghqgevf5+ECpqhdZNRZvxe5
+          C5ul/b0GNscr++czl/J9mv6bj9P41SIe09iCaUJlfEI2yeisByWq7EAh4fql2ZxNtIV5rkYzOH/j
+          jHNl0UG3+Bfg65gZlG3pd4RxrV2Ic9Yu9fKNA4hu83CYj2YqZovxnhO4DGODb0awAc5pTw6Ktngj
+          YZr0LrdSOUHS4D9JzKou2CaeGDA+zjJ+XB9Xyl4PPwnwenkkWRSW7hR6ZwOWz3IkUToH0Xa/+xI4
+          XyUb68LOStIu3iDq+Jgo4YuNqGtxIxJc0SRal4t0Xb8ai3xVmLFxGhdtM6OwQo8oMEneyInGRbUo
+          wuHZrcSxqTYsyy/RoXsimS/Nw0FbXmvgoPE3fsmtvh7qVcGyiPxMVYja5e7A3twgRD2BD/9oNixY
+          /saHaL2A/fuxyoTT5zuCrR3NuXvYJ8oa7zmGNqdjnKf3mg6J96hAmKk8Pp3KNpsy4zgjGeQyvvHG
+          WeMyGRSQnKQzVl3dBXxtHmR4FuIHTg2WG6hhBAzyhvdGtGfnuHzyCRrYPuQf1iSJRivpAwdtAXX8
+          g0V9jY7j2MPvY2Bn1CxcNFuVDFEsFL1fR4cQrBz3MWA3NLF/jPlI4yd16KW0qXuCsfDKtuPqVxD3
+          ICOeG6nuemeUHCDroWE9ndFAdHaIQessD4IrXcgI34QW1IoznNHW/sD6AaUK37pjz7AbFMCCYZOg
+          rAg5PncDdkfPpAU6jW48c9U9GYRQXlLwt96YwZUjoUKwAuf48Zm5fbzm5vM2EDUfLU5qm6PL+f5W
+          0fP1+85MT7/Z1NZEBpo6dz5yBR0IH236geeqH8mtntJ68YO6R+1cKDPcraYN81kgBZz/mtP59dOW
+          VbMNuOMhtpQjAFRsdAZelNeHONnpPUxf5e5Awp0sYip2PiyPk85CluQRCc2fki1lFVfI7TLVPzwu
+          v7p/wl6F34N59g9D3dYC/70EcHjZV6KU1582P46Lgb6VkZBc9VzKLr/CgOVSVMRewX4GQf0lcIzq
+          HGszsFyqfaC47/E7E6W03bpf/fMC9vjEpx1/1huJVFTE6rzHR6/9yjwJoYFNeya1/AHcPt/oH97c
+          qnrYmPIuQ/4HQ5xlnR7xq2WW6A9PbxP50e3tqzms7242ryue3Nnjf4XUrNjG58Y7gI4PAgO9tHNH
+          Inhus23oDwuAiS8TN5z7euHid45aZ3sQ45aYLk2MPID+Tz5i+7IpA/mgbyyJuMx8cG0Pw9rcDiN8
+          F8ee3C/tx11BsSbo8DYQljPVpCvGlxQ5tyAkd3O5RDQx4gCtfn/3W06qM9rg3oGMt9n4EjtyxqrR
+          SUdNAzWiKMoRLNPQp9CbS8nvzaUY2L/5qu/g7LeT22rrlPYNtA5+gz3p01LapmGMrOcS4dgDM+Ci
+          dlFRHFFtRtt2yvhzP0C4VomC78NXzoRg1H4o2fIzPqero7HUPQboHaU38qp+IyVf78lD0d+6P77h
+          bm/L8+AXSyF+3d9JLXy9K4uOr4zD3ug+waK+lAYlz4tC9usDlVtHh4UwXvFTs8yB6xw7gYXPVeR1
+          ccKIoiCNkTdpyR8eZNv3E4nIbLMT0a5GF7HB66FDvVcbcvKxNgjv0YDwcVFXoge3eyZAvVNRUsiU
+          /OE5tWxHgn/580E00V3UxPNAFx8/Pvu12npLqnsOwVu3cMGsCKyXTJ7RpgYzfsCJRJuy6CX6FulM
+          dPjxarYr5QAVsTxj9fOz3CWPQAjN/ZDaKSCHYT3LFYOyRMywLbTDMHH6b4E69xZJeOjqaGt+Px/J
+          n1oj1v1Tax22IwsFkZyTZ+Md6B9fAyKuMp89R18wP/l6Q8C1HWLNL0ub3oVrIesAE5wdO1dbUTFa
+          cFg+X5+uFLlTadwSFHDwgm9/6/sgPwrYguJI8PMquTPzCDe4WAshyiLE2moTtURzUTFEOW5nwMsH
+          2ULP5/1CrKk+1ltgySL6VqyFr3v+XetOyuGBDyzyDPwuW51OD+Fjpo9ZnE3iruui+CA8FJicN0ah
+          3LXrIHQkzySm8XPB1sOthMrwmHymxxe6nPjSA/X5Z5KYmitYmiNwECyUkdz4vho4mb1KME4Wl5ys
+          FGREuBYl/Ajsh6RVXGYLdysbmIoHjZyDtsq2e+UFoHm+3/6yRGu2PY+9BG7XVMUJ08kDLTk9AQU9
+          ZDPvHypAS92SYSXfIN75JlgiXpxRmig+1oPbMVvU4MxAxZl9X9jvhzB6xMLRUq2ZwQM3bNtnbeH1
+          3k/7dXWg2ovv4fa6fIh+skqNrqK/wKS/HIn2PrE1BWbpwFkpTaIeOi3azOTSw9PUi+Qk+ly9/x6U
+          nsvNxJb1eg/baHUSEJxHTfCcgWG9XJ0cVte3SRJ5TaIN8EILrTEg5IzNEKzr5OggGdbnXzzWm1VZ
+          ELYjfJKkCbyMKo9zA03QBrNQX181aa6/VFrfDSZucdyyjWdDFp3sw2V2dnwhlq1K4EY5l3jocoj6
+          0S8tlDPQmlcb0np6ZXcfDtvJJTbfP7VOUOICnd/Z018oEN3lxP986bceFqyp7zQb//gW+cYFvqyZ
+          PQgH8RWCA/mmPgUGoTTveh+6D7HFd8+80jVtrz9g+8VGjArdo1k52uIxO/IBsZ73tiaGdmfh+Ju/
+          2Hhdt2jtdFeHyrPXyHl7jNlc3dcFaXVmzfTZOdqyyjYE4UkY/K3Tt2G5uWkAP/R6JBauN0B1ocpB
+          e8Q10d1fBtZT/XvAV3IMiLsWWsSuSF1gcVlT8i9/YDuz4EAbgr3O9CIiAXGDBp+n5D5gM+NQ/+vh
+          g8KW3Ezvnc11Iv1gel8s7JKMiUh39WV4KK8u9hjlknGld22gYPfrnt+0mjtMYQPltHTJJXv42ppp
+          W4vsUPRIarDXmueFC4T7+MzHIdMAy6w1AwuKMnL6jl22om9XSpH3NGfokLGmrnrIpS3QUoKpk9Il
+          PV4L+G5jj+z5M+M+4k8CO15jrU3NbLvfDQmypIiI1ukvsJnO04La51eQNOZPgFpYcdAfP1rPKePS
+          7bqIyP3aI86T+QaEm/Z0IPzk/r94XK1bMULB/q0kCM1ZWwe87eNrF9ieI5FSPn32UCzhRLJTPtaz
+          Am0VxaKX4/O2nSJ6EG+BtDX5Sm6mrAP6tZMZ9oR5YB8cf2DV4pWBQVzK+JafLmC6c7oBxdrXZm7H
+          r1kPfi1k4cvFJsP12YasjwfFTCHEMhddW0EKdBDMNflb/9GyVDdeAuMnw96mGi5NwnSD+3gStZz8
+          aDODLIV4fBx8Hr25msTy4IEtUFLs00of+Pj+DuHxE1NscJKWsUESj0C+mD22r77nTtexM0BzCXPi
+          nmlAua3wC9jeA96nM/hp24svWvAtHjNxoH2k9KZdLWBU6gF75fLTOk5sPOT2Dusvt4cMhLSQdQjA
+          SfCB8utcqr2YHu56nuDBnYbJcMkD8D8mJMbkGu685yM4V15Lou0W1tOUdTMwlKQkenHvMqqj3wg7
+          e3fMR/B01y+Xt6jq5xD7Nx5Fi56hEDqprOPcjSp3KBllBP15uWFt51PEq8sefa6/inhl4w6bx5c5
+          +kXlF4fz7ZQter9If3ye4KE2BmrIjxl2/JX3N37uBmp5jgovt0dFTmg/Q3UMVAk5Wjtgb+XtQThm
+          uIVKd7SxA+07XbbZdGB8Db8+utu1Ntmb/QNvrn0Q99q+6uU1X1QI7l2LQ0fE0c73NuBgbpnhWgTD
+          33pDcwVT/KffeDfjKkQtycG7fqonWYodmKO8wdgsM23lg9SA0sBgrKnHvcfN6xGD9d1ickJHJaPb
+          VRTB7s9giz7eYMxysYfH7TrMki7a9E/vAj6MauL+SKuNebI9oOonFcZzlg3roc9yIFbFBcvej6GT
+          y9IWXlPb9kVT3typY/QNJoVKsTrXbzpQJPBo17vYZR1AlyKYGfi6Wy3G4Jhownv0GXhFcYzPWlnW
+          NERHHxy32/APX4U9vlDBkvO8+MJe8YdmD1Ply/nvbPu62873AXsOA6z8RIsKf/lk17vY3xiiLQ1l
+          JfTePiGRbfkG2KV68QCFGZ1BcYWAVPbR+POjsOeap5onp2sLoKJyswSvSTTTF0jAh+WePnPjUVae
+          728Z7fkVW4dPN9DKEz34DB4PrD6fZUZm5ZL++QN+p1tOzXWKC6EwnNSZLdYqWzNNaiHqW8Pnp1s1
+          LG9B8aEypJPPneIObDDuRshY0ROrnqtmdLPZB0rvauIzsl7XRLgmJXS/7ojxZ2SysZmYHjJHMmP9
+          yv+GJV62AumQuWPn2m4RYZR4li7H9oNloX65VHnqIzowJxYrE1VrDmA0wv3cNtGDmqMbUZMcrQ/w
+          Iqf73oxw568ANmOOPenuRNxXuVjQc8y3L12ZIdvSQfnBicSabwn+KeNfh4hBK7wt/nv3B2m0sjMs
+          8ab66/HB0j99Anc/YRYxIdF8i64s/Ag6Jd7rlLu/E//zBB/5OnZ0SwBLCycdrmfpu/uBZ3d5S3IM
+          y9WLsTkRi+7x4sHfX8/RiPLZKt6t4B8fM+spHTb7ZfMwuFQitgU1cYXGq1XoI0/HoWOhbHUFaYbW
+          r+awMn1bd3NTPgXkmxRYh2vnDvj59OGtdBsfkAMF02dsUiSIJwn7VvYFW7e6/j+9qmvdMmwnMP+k
+          TMYjMd/N2/2LN2gQ+vN1gSfuNOHJh9GWbL7APC/ZzpcaGE12i/d8ngl8kYlgaOI3sTRY1n/+A3xc
+          5BUbRvAZ6C/zQmnnJ/h0GitKDuIrgA/KtMQ8OsXwh5eQky1KXh4Wh9n6WDqcHpaI84zz3a1wgwre
+          TwgQq4rLaIN+naPv0mwkR52YDSz+FPBb6Qn+82u20e1k+Ke/Lvm7GSaS5aXkPqT2b70Ou98ywk90
+          90jyuobZbxqqB+CKzifGeOjqJbhHPCofdYPNEDMa7VvQwB99xnifb41TbKeH90wUSVycRzC9woiV
+          2mnEWKnPOaBWKIcITNczdo6XBkzvtxNCZsAh0a7hB2zf+yuFQ5O88Tm9r+6uVy2IX6mJz/tesvXa
+          vRkY8KNGcC2fwPLzw/nP/53ZgITZEtG1gZnQSETJp7c7M3XgoIw3E2yImxj1GgD8Pz9R2fnhNt/S
+          BIhy6JM9n0TEMtwNXvVXhe3r41f/+WdSLJ/yfYezP3DXcmzhEPTSHEHFpotesCnc9cSO10Bbgjxk
+          wI1nI6Lt/ut8gXqCWrfyZqYK2vonuoUEOXU/weZjrSahLD5gnweUxO4vo8vGnh/wb/78nn6jP70D
+          bffVY/nR+2DjqimAHh0NEpZHB0xHx/sdmdl7kHsl1u72Pp10uPtFOx+oI37394DTXBxyAv0LrEfz
+          XElcMfjYZLoJrKPMbzD9RDkx+cLQuMHgVFj8chOnuN7oJClm+F+98Y01d2vXIIUVj1ufP4nj0E3F
+          j4We9UWzZB9Kbb3W9QgNwXeIrh6Lge56nhcGlsVuMV0ppRpjwd0v8OEPYDA+Pz8e6l2w+ZtjoWgV
+          QeFArb5b2CHZqq2PzYNA/9Uy0dKKgnn3R6Wdf2ATu2M9Zw3vgVjIe+Jr5URHoQw2oBJNwc5Sfev1
+          y1wglCIKiZU/RW3g7K8Pr4/Hi/gWLemkhJYKirB2yNloNm2jEc+DsgiLmaHOMNDD3apgnssRKcCb
+          0PUZvuU//UZ869tkI3tkCsAbC8bGbZY1npyeLcRjevBhxKtg+9MzNjlq2JmHl0b9MeBR4tw8Hziu
+          TbmnPfTwjq42Nnc+yZ5FYQNcfheJrCjeQNNQa+H7IlFsV2KtTarkM5A9B8EMsucXrIOB5H/59m6m
+          YvR7HIxy70EgESvnVu3nUGOEcSLf8e1wrqOdj+cw7N7hPBsLAPPtfGtBOzLPmQk2T6Ne2VXwE0sL
+          McTHK9r8Q9BDz4sFHHBt45K8qzy4+1czRBwLNlFTYjQ/lQdWwmGho5o7BSxfckCeW2vR7RoMEnhw
+          hUMwaip33fkozOLhhM89bur56kw/WL7UgOjF50N3f6mFvV/aeOf/dFOgIsPfipaZeYvnenPfYAG0
+          a7u9x16d0Ut7ShHu3c4fT8Ez2j73sYVMsh5Jcku+WpfWRQVQZ1TY3senu9bDDHZ/aEY+OdIBxy8W
+          XirA+4eAudItDI4GNE9z6/OXFwPWj1iKkG9HSPS7kdTzVkk5vJXHxIfpp4z+5a+Pkcb4FhtGPZSK
+          LKFURBoxHwyIpiV8Of/0n3mbz5mARgLByS24+ZRsesZHCirg7yS7+BW3OVhgfgiBe5oyfNrrQ7+E
+          jg+YOFcP62z7zlanPTvQf4TY31R70xZmbliw+1Mzew8jd+VAHUOjkg/kFoe1O/sFyUEq4osPzodz
+          xPIHUknqO0ix7Qlfur6KIEZ/9TOqe01G87Jhpd0vJ88eN8N6tssY1Fm6Em/PJ9NWGAUM/K+Ktffl
+          4i7NgjapEnOIT8WK6+0WPVnpL98bzLPRxt2vlf7yI3eJuL3Hw7CByioxdq7+PdsOt6wHtI+z+Xhp
+          PxqVgLhAy46lWTKX4o8fVbC/X4b5uPuHvIg0GYa/tcQWX9Bo558QXmFbEUcJKkAf9WqB+5KEWM02
+          NtvevpPDP36OrYNTb5EOW3hldAM/m0akO/7HEpPQvYMtbIbFPKQpnI5XinHtqRlvVRYDgS/qPsv6
+          VvbnJ6Ew234++OaRRo8ZbmAO541Yt1mqh0C0LHhn2Gnmma4ciNh4ED5XfvjzV+p/8b/zD3IOHqe/
+          +icL/uqVJx0udHlW7xHt/syMx6Ksl4SvS1g+q5H8qydUxspAxqYKlpX7p16tZ8zA+tyb2Le0mU5m
+          dy+ArLMNTg9rqxGKDuyf3sN/9Vm++bx1dMjsA/HB0QJraBgslLWbvM9vmi2vNbWg21ssNnwcgCVI
+          8hn++ed7vqzX6AZTWJpfgv/w+Fe+nil8XiyPGLYl1V0htwHqbtJE8Gcsss2ZxAA+g/RB/uoFyyor
+          EKisd/b5/f3zUsoVal6bgI2b0w8bL9118JV/p796Us2dyneIuDwTiRtKX+1Pn0K5oRdyOpVGtkSJ
+          WcKVxOI+nyldxuBQwDqXrsTZ+f6fn4b2ejUORH5ziVe+S6m9reFeT43ABFJgQGxidv5OLs3IDDId
+          7vljPral4xKq8RasYnqdPxczrXd+sfdANTxilaeDNqVdvEDPSwT/MBbl8OdfgEPmHuaGRHO2pOpD
+          gt4rHH2q2q+IUiSw//Dtj/9yzU2Y4a0ECVH3eiO9274H3d8QYId8zzXXMd5+KtxaiCaEXrSIcR7C
+          x2WpyXNTaTTK5nODj+2lzuQPP/7q5YnF/YiXns1o+upBiE6vvadLldfaGj23H/g/dhQI/3tHgXr/
+          2v5hPUnaunptDo/mtvos9+jpBlnfgN9UG4jc9os7kuhtIGt4RTPHcypdgL9VyIEoIGrdEUD7R2PA
+          pvwlJD//Wm1hdoa/xcnqQ9TatfDb5A3+XIJnIJ/8aOMVXUZckkwzHa0AsLoehDA12xmb1ykEVFpD
+          Hap6xGOrOoyARr+0h9rxHRAn6NqMXu8PFirtISdn2ztTenRNB6YVfWDr1rfZr01kC7rz2yDPdz5F
+          6/3COoj9wtwXDnbn0uuzCtE4fjKiZMiKBFUeJTidjBhb82d118/51AK9FXrieyeorcZCQmn/PNGP
+          gzUsHFYNuDZSRxSnpIDO6+ohnzxd7KaCnbGde9+gmi8ziRNDrdkIhQ5sZ2YiZhZ9IjrFZwvw5Njh
+          +AjbbD0Ca4YHFUvEGGUGzIst5WjbJpU8TSEFy32JA6SgpJ3HlZg1d5hVHoGpdEi43tSMO71/PbyM
+          NkssGNUDf9XfCbrJ5OOvmPvVg6WPG4wV4GOFwl5b7kqoo24kZ3wKvRMQLHhi4aaeNSyr9Tws0mo8
+          kBMzDva2i+YKc0lzlD1Zg5jLSR64+JtKAEyVMwtTxtMVZCcHPqbYJGHFJwP/9S4lijrJJPJwaICQ
+          97aKrPQyY+sJOLC18ydFYuID4mQujbag/LAwdVnHP3BiFHGXT8Ag3Rv3LvgDn60faC5g/Rk+Ud7o
+          4c56jirp3vX1zJ6mj8uBLyjgeP15M3j26sD7X2+BQcLHpJB9SRul1UjhhpUnMRnRjcbre5ZhkbUt
+          kR901tb2OhZH50dYrEiEc9dnZwewpZ5DLt5P1bZCdUIgB82JXA6Vn3HP9ecgX/M77ASSrgklwA10
+          xywlzjN61zT/rA0iWjpiS/4esq68QhW+DpuNbxknA1HoDyrU1/xCigw+gfB+ag7Kz6FPrulM3eUF
+          gQofCUQ4Z6yTxpahFcKj4B3J3RrfGmWEd4HmYgiIrJ16jd4UxpLK+DNgTTiEA/WiSIVNUX2J31kx
+          Xf3HZqG3fTP8MiKfaE3Ma4zELi1xmgiTtjEr00JzCX7kDoRB46X810Cr3L7YPUh0oAM5BhCulMWu
+          /LUpr4TJjNhW/JB0yZWabYpDCiXGUAmWtcOwFU7nQMFtTGxx3R3QawAgNNtJ9NeT+3J5KMMc1Mpm
+          EGuOabQMop5DQRTfOByjIdsexKv+/Z4qNVI9yeP3gfhNSbFt/mzKP8DGo+ZZLNi4LQeXZjFlUXyP
+          MQ7eOtoZOJDgVKXzLLzVe8ZPwLMALGTbX9hUrVlB9vYzb1yKnSl7RxubLw58gWqaj72d0lUeSQor
+          7KTk/J5cd8HP1oFndLxiTQyWetFlmYdkfO2dWrg8297BUiG5vojEnEZ2WICqezC1+g/Wjp9T1qeO
+          OEKetQ2fmj4H5t5RZcQ/5AvZ1wMlAnk4UHjdVp85KTT7wzvghGyO08Kj9aztXVlDS7KJP3su4Cet
+          NqARxG9fcEy/Zos6eQDfSP1ZOLJsRBv/nMDiUREcP0eDLt+Pv8EPk8/YHmlVd394s8zKixhWlgKO
+          CxEP7Cp5Evcmi3T88XqFOEV5YC2XV3cqqz4GzPVww7ZSZpQWCd0gFYonllH8dIX01ftQ1wyZyJnu
+          Auqurxw0+eN43srcjjasLyJUfkpE/NrSXGrgyEF/93cqeMWlwzd8IHViP1iN465mtSofpZg+9R3P
+          tPoXOXKFLDKSPT5djVMU2MPjAybk9jfeP14v4VX0X1jPGnEYi8ehhHt+wf5l8OtlGBJH+sGzj3Ut
+          vrvrT3358A+f3Or6ymY2vocwiiq0z9cnY7to6RF88A3B55LVZv2Tp/CYy2d87SdN43Lc/JA+nAx8
+          Fy1bW65304GE+d6I9jrfBtrYnAju35LDl4O80TUuJBVeVEUmzwq39c8B3QZOhH2T8LadwaY8UAPX
+          YDj7XVGzQ9+fpFa69/PDn7l3GFEGLCJ8+V6EjcfQDqs9vxZoHy8juTjrOqyQfW+IwbKJI8aUKfWt
+          8Qdjrt2I6aRjtH2DEqJzzKfzR7mMYOPPUgJfZyvH98j3XRbvXenNEPrT8XxDYBXfjY4+zfFAtLfw
+          zviT9G5RnXy+5CQXtB6M3C6kt7j6OBraq7u6n7JE77wc8D0OPxlrjLRCugxDP97nY302RxYGl1rB
+          Hjk9wVKZFxGd5QZibT0oGuccriVqDIUQu38k9Rc/Vx5BYsgEhyoZ1jucRWgEyRvHAptqwmexWOQd
+          bQnveOhuSW6w0Dm/LRKWeZdR7aGUqPooK77pAZMtEUtDUVyuC0lGuQBLdFRHNLR/DuemZ6yJjgG8
+          8vpK4ufYUv7IXxs4nfSY3N5NDHjlwTXQRCeVnG53xt0K5+2gurRu5C5anbupA0jg991kJDn7m0tP
+          6sjDfbx98TQk0Rp/AxHpW2nhi7zKrqAQYsAvyUcc2Mm9ptVsiUjVLzxREkPUloSnHtqY54eY+nak
+          w9VrKjj62JlZ7uGADT0uBur67wufnpujLXfkNPAxJSa5OJh32/6o+2jHT+K05ZHODxpB9JIKnphf
+          JNesaQQe9OqwIv6eH/vCuMwo/961mZE+XUSrTy+jhcvfOPv28R//CSS3v/TzW8sH9+vrVx/1+xk5
+          rRhyd7GpI8HZviLivBhv2Hy5K2CLyw8xqdJF6/YcPfjmvt18HC+myznaJMNQCB7Yr3TPpUkfNKgs
+          AxvfLeczLNu1k6H1yPK5VY7uIKCeWtDlGMs/7tfHWfJSuOP/H97WW3QVcviXz5yHdK1X5p2XMJP0
+          bmae1I8WLrpU8HEFBnHpFoGVeZ6Nv98j8aX5utxt+gWwnSSenAr+7dI0CxdE4rwmT+ZluFyzpg1o
+          qkUnAdXOA91gIMNY5m7YpIqdLR845EDNRgEHr/abkRKYLSxn54H13GAzmrS2BeOb98ZXbnXBGNZc
+          hfb1hi/f2htY6UcqyBv8iO2jJ7uc4vcQWnqlE6OVzy6Ht1sP6+T7nYu8ONb9c/UKWCbkPh/j8BNt
+          obGwaJaOlKjrT3ZH+96GwGviO5YZk8/IIgsiqGf1Ryz5+4q2S4NCMNs39N//q8qjCMhCDCKznZzx
+          Tj4x4D4sl318TI2iwuPhqDdXgkHxASt3OwdgxJFClJN70La4Rd5ffiT5J3Fc3pUeMjRM7jcX7Hx3
+          6UPAI9jxjezr3eXE+9mDc9EFxL1QeVhOUSLBv3jH5KvRRQfHGN7yhzofy0Gly0eWGAhP9WXeLK8D
+          n3JQfHQ5+R9fuE4b2JLc5+H3BBBW5kQYRrsfVdj9niu+KpUzsHd7YeAjldm5D7o2oo/D3EJ5P+MX
+          rrcq2vlNCP70xP5/XBpbTwMG0ofxj6VtZkQ9Zx4sjoyAbdOYo+0vnzsHnSGvyDDq5XrHFpQh/yVK
+          8ovqte3XH/hguSX+/LQH1ltAChmmwPNRvLdgGXSJlcyf0M7omGF3aS/LDA90Xfz6RSKNte9zAJNB
+          0ojbBQbY/vhSx3nGzN+Wl7acldWHZQ85ErQWO6xv+jOgUH99oie3MltMVV7QkbdVLL+EkzbufOeo
+          DL6IFaeMwPLHf0+1bGInL62IP18OLSiMKNrxoRxW0mgOSsJbTOyCDSM2l4AOjWNW+OjK2y4HkfUD
+          1Vnt50Ct/WHDthyjClWdv/8/uplnYQZBPLcEF8iq17KqEmAHe882EqiATXjgw9fZybGazDzd3rrj
+          Qy8/3kg0eV+wiNdLjvTzMpKseU4Dfb8fLTSeDtj5Q+3S8LUuaA0XgIOdf3DWd1HRUXtfZ7jrz21O
+          lwAJdn4n8e+1UkrzPIfcLW18Zj58tCU/OBU0hMtCLtVCo7kJxQIRbGOCmxEOYwKc/h9+5fVyi36O
+          Ls8QymfGHx/hO1vqWPZAywYKDhiJDAS3wQyv5lz6y84/p8o99/CQbd2uVyR384fcO96H7ULcxsNg
+          /KmbCC1O1bB8Fq/aCsXVg+SsvLF9X7JswW+uBdmTN4gerGe6RKdVh52cD1ihj7kmB2zKsHmPZ6xa
+          T3Pg3wdpFLdblWCsmPzwd38wPT1lrOvSQIlclrnkpi4k5pGNI3q4TRI06srASuobEbkgxjtmwCnJ
+          uV1VjTUybgN/fOLi3ZpsbL5JgD4oFfb18M0WxooaqPtsjXGRTe6yyhEP33k1zNwysWB0rNQCUxnJ
+          RBM3T5v1APuw8vU3ufZ7f9PGuTfgmPI/4q1FB6h2lXSAzvkFZ9KnAOP1jh3Uqc19v3+X8t0yJWB4
+          sUd8dgKdrs/8wUg56F9YXQ295s1al5B05Fsiq2EABEboclhkTYsfAvKjqRvXFv1bjySowHrguxka
+          pvDzj3f/ALZu+SRAEKX3mcaDM/BMd5vht0g5gm+/yt1ua67CcwINnNFAjNa66cM/PJ3RQVGG7dEj
+          Ha7hBogmPAVtx7tZeuHAwRlnWnRxrMCSvlUf+uxhkNxFXsMSpUenwZ7VYI0NQOgjEy8/rIbPuzua
+          RuqBxxgG83Hn6+OdrWdw4bcM7/FZ/5vf07z0//j/v/HLWJr4G7hRuprKZECf3V6TeKG8S0QoF9Dd
+          lpqEojNlxF1vueSrzoZNPCmu8Jcfrh9VOqPjOFJyUU8zRF76wM7vZ2kUHBcI84t29Fd5lbXtai0b
+          eqjkjbFwtrQlOHQ9VM5SMddW7VBeeZ1U2DE82PVxo2304MnSIedLonxrr+aU7BKgDqd05+dz/adP
+          wM7/ZsZJLFf4i+9NNq/YFnu95to7EKGEZBdHR62jRM4v6j894DxMmC3hUvPwS4qRGH7YuRs89xuw
+          X46ADRVV9UZqkoPdvyJGe3EAsQv7AWXhHRMjSmRt61i4wChuWSyr7JjR2/HlwcPM+DPabm1Esndv
+          wRczh1gzfY4uCmOxEESHzB/kxHO3yzq00CUN2v2i3KU04lp4mkbqM/vrTRW7H/TawvTzeJxpdzjs
+          PXSuIf9Pr/cosffruYmLPV/s+BX++WkYG0wLlmCjCdq/f2azJh3oKloNfNMnwG7aDkO3vekPBdft
+          ORNj6rTtfj/EsPq6L2Lv+ZM7aKmOzlcRk7MFAo0aZSAha3w78yB/O0rtS6ZDsc9v84Af8UAHMbWk
+          niQSNpTfGXBcTVvkLBaPlUPlR4J36wtQD6KEi+Ar1HTSzBT6HJvsej6KxorZIHQ5aGHrZEuA7PGD
+          vhkO8I4vLvcgpYc08XUn59mwBiKXvxx8nNX5t37WaT6WcI8vn5G0M+3FiR+hIstn4jH3ipK5BAUs
+          Xvxe0Wi1gfdMOYW7PsHOzxiHxQHvDd3u3hNH5osOZHLn4KgslYft8vFx56/lGBCEx+ZPT9WTzfM/
+          dIZCgXU9XbLZ/+ob2Pm2z+16ic5SYMBW29Q93vKM735GAdP3VGN5e29u95JnHr5+V2um5WXe35/q
+          sP19aqzIa+mu99OJBUznQCLfiaAtRpP3QKg/PrHqt5VtDap9WG/BCf/pKyo1w/Jv/k7hsx7mmMTG
+          X76cJXYxwcZDWMKkbRb8KkYHLE0hPP78lf0pfp96ySWqo/vFHQhOo4Pb7/gFKmylOL3IkNLlkZcQ
+          SduA1QtANYl+6Q8oPnPFRkuNjM3Nd472eCAXqDta372+LThwx4Kc/U2O2PfblyCvjwx+tqUeLeS2
+          lOgtkJu/fISEcnFtFzAH+4ke0eq07lYeIYCZlPmSek+z+SpfJCQXReIfgeC6mwqPBhzREBLv7jsZ
+          2f1RmAkQ7P7PYyCzFOhoAMrHl4LOyDjHcyV4P6cyVnB0dmcIGwbcP5nps7+FDLtfF0C7Kdz5wfgI
+          9EE58XCqHjN2m9nNNs+0HpA32JFk6XUC26P99CATGECsxdXoetvPIDvyPGPbyupo006pDDulrMhN
+          +ygZbxHg/30fcbQnjjY2/Y2QnyjwefEOI4qqPAR9bl+JNz5CSvG9f8DsqG7z5t6XaN7xRiIby+Ib
+          PMsZR8ZeBTt/JbY1Khr/jZcH+tMbWvwY6bjqQwh5amJ8nppJm5ZlzSHr9OquF1OXDsObB5p5eJA/
+          /cAyz98CFx1W/+6ff1SJB5P4qc7HUe6y8XW6M1D6fBwcsKk68Lt/BSmXEKKKzhRROZIciMQfwjF2
+          gTZBttvg8cEk/isy2mEhUaf/+b1YxxM/fAbZHOEvci4zy37XehlEL4cSYGKi9RXjTlyIWNH5TSx2
+          cHyJKLwUJdz5GXYenQDWc5iF8Kl+yPwUf/sJqjt2wI532HW+WbRxi+FBNckaovvda1hLkv5AerQa
+          cvpuoJ52fQrCjzNi1Tto7pCGVAI3Zc2xZ+1nvDNZjUHUiSa5v75DvaHiV0rGiUjEoVlIqT69WoEO
+          9LDji0rZpRUssPMxomcqcrdSiHjIVoNNdC0+7k8xfYpw98f3M+dFNvdH3YNgbpRZzu8lpeHruME5
+          WA7YydfWpTv+wMoWTv7hjJNo/otHZQzJv/zJQST3yJo+CjG2uzMIS3I2pMNj3IjxKS7DKn5ABdEb
+          Qx9gGLjLdrm06FtWzUxL6mRU1WH8pw+xec2vGrXbXwCm+5Hb+SahRJUbEeJPUGDsMWI9ux+PB2Gt
+          MjPplUtEPvWjgkl4jff4zfeehMGe36c3sWZ+BmPzLYK/8cfn4B1HS3R0RjiIQPCZDvjuv3y512/8
+          JdMHunU/v4BUW44k+975aHLifoHT8/Twexyv2bhd7i1MQRjiUxML2dSvigqYn3na+eKorTbP939+
+          Blbg9UtnpnvNYPfr/eP5eMz6dyCW0LaFhhimL2fC7n9DMY07Yl5dIVpEPZshfLANyRX8qxfRGizw
+          V0/AxdhTksWUh4T53LC866dtOdUy+tMjp3sw13v+c6DXJHefVx0723z5nf/lB2LaH+qSZP7JMGKm
+          H9FpCOvtxtXFHx/DYTInlK7V/hSuCGU+24WDNvdhKaGLEeTk5F25YddDMewfyncWYaQNux72IRe5
+          Fra8q+zu/HT848Nz6x6f2hL+PAYOrafjdM+nf/oPgcS0Z44rZ20N0uonmTJfEv/yCABVxRP8x7f/
+          /K8pCt0R8AR0vsT6XsQrdVxAkJxsrI2g19Z4KGMke/Hvb/xd6gszC6KzuxG1YRowE/bjQcEu7sR8
+          rB2gmZczoHk1BXY50wK7/tSlUEwuPrcwYb2Kz7iHrhZnvmTOk7sQY1ugfkiqf/lyfVRHCfzOPPLZ
+          3Y8ljX9KoPQdReKTd1kvd1uEcLuVyfxv/C8cleF6mfS5tpPGpZ9F5iFu4+tcpfxH2+slMtQupUOK
+          U7i6y/XDFWCfP+zv8bm1ghfAsPVq4qZ6oNGLep4B03v0z6+MltF2QpjZbY//8HSszLsEmOmR+tJr
+          8sEaikSFanJv5gr8LnX9h18goLy//sR3vb5ELML4pO/Pn7ldQV11a/4v/+nSuXHpX/1sr6+RS5O0
+          2lJtXg+LF6uS+OR2w1oZjAfUt+wSZb9OOniAf/VTjCOjrUmYVQnwjq4094z/BGvuXj0pdXkHm8c6
+          qelf/WCxLBNb9WZof34kOEtk+Yev43XOY2iuyQtbpLM0rhvXBkXwKxHDr1uwlW2hQztD4T9+tN3W
+          WIYfV1CwueenUc5PCRCOTIoxb68afS1pCsGYRzga7369JuYzhodSNYhnNcT9BYK6IOKbKTm5Ve7S
+          vSEv2vMtdnb+sqwkkWG1PydCexVctOTaQ4XG09r1gtmDlWRNAHpuPP0bX+FPr5Kz9ib2K5kozZ6P
+          EbpakhF5fQJtZvPFgtYmXfGZeRnatnc+RPR9ioh/u7y1zn1fdKQjj8U5X5VgjZLMgdn7s2KdPSXu
+          Ipf5BlXCRPPQ7TssGQ/HcPdjMa6XN6BFpvmQU08X7HZBC8bEPMpw10+730OHNlS7FLbJk5t5X3Gz
+          jUqwBU216eTU8CFYTNXa4OPhVFgt8y4i6c2NYQ7HZG43awQLfC0hwpV+JIl4h9k/vVCd5R6fwqdW
+          c359l+GpFO/E1ept+BmryqB4dhdyWpit3o7SRwWxejj5nCdc3NX6LjK0X5ZA/vj3WO47OO1jNM6I
+          WWq69qst/z89CsT/vaMAs25F1Htu0I1bPxX87l2wKVLFer7zkgMzWChE5Xg/+322C4985jFgbSoO
+          YNZ0g0cnuTaIgW3ZZQXHC2C5ahOx0+dMV/f+EaF1tZB/6KVbzRpMGUrFrfXm8ekydcvw7wAxZNSI
+          9xl0wMXHdYGNePPnZeh8bSXDKYd69iqx8cMvl95i1EIlyU2ShvVjGAMssrBAWuRTUWgy8s7bB6Ar
+          lrH6kqSM/r3Om4tOVCmfNGp/BBWR47HAXqlD8BudxgfjtETkFKdSRJ3nHMJZgj2xzigDi2FZKjwI
+          PEfk/X6oPb57hG3YYHM4lfV0u58ZyE7G1+dPtxpQEXUBOPm2g6Noq1w+V0oe6rb5Id66lPVqHNUN
+          4YNq+03x7KL1RGodWp1IsfeUnWiVTmiGR/N1Jafoa2r0XHAeEh0xJ9ekPGUcm0kiDIz66ddb/4wE
+          73qt/o23q0JmICPbVZCJLdeH7/WZCdwm9egGYDWTvILDtqpYl9BV/OFXEHo1d+cmGf4aQ5hRWmnZ
+          gnl/AfN7m7DFwE8myILoIylnO3z12yegRzlJ0XakL3zy5Ze7Bp+hgCl9aDMz3n/aKvepBY6R4pNM
+          t/JMsGobwmt6G4jGkLvLzi5wULxJjX/kD0omAGWK4YObNqKstZoRrRx16Mli6wuGG1LWdJABC0+O
+          cFj89ExoX14AHcNwZpEX7HqrLXWE38upIJfxpgOBJlGL7tNXxobHy0AQ8GUDWdS15O/zizizOsCR
+          QImr6B3dLksQIMfFFVG4ma+JNE8iHAxFx+Ff/IihwULG8G/EW6+Jy0lr4MDtWV2Jdn5VgFrCt0D7
+          /eIicFi6PY+pDCUlXIkLHh9tYd9Bis5WbJCXuBJAlyXfIH7rCc6+wTSwCN98KFGYkUiyHxn7+W0O
+          0rxyJK82FsHGd/kipZER/4s3yr5oCLnXdiHxXXDqRVpcGTCNkM9LPrDR8hsZC3IcvmPna13pGqpq
+          izbE68T4rG72ecW9KqH2c5y3C+rArHq3BrGXimBtOjoRr29Jgp5C8CROhd41HzwDEWW30fUPV+1Q
+          L8x51dHp2wvEeB1mSgu6Veiun2LyHAWsCcabyohYUjHDh+m53Nu7BEBondavMtK6SyP1DRxDL8N4
+          KPls456WCr9Eu/zhS8Q9o93RaB42DsB/SLuSbWV5KPtADKSTJEM6kT4IijgDRARF+gB5+lrc769Z
+          zWp41/IqJKfZe5/kHM8AbB9CCD3MPXDW9DOY8rn2UW0ZAz5/j5rGsy43IzVSZOKzhQbWF9lMRPBY
+          Ys17qYANDk2BZAhT7HhSMrAA3rc/+/Gp9PnQtVTXBj0SwZ7huec0EiHJhgfn4Pmfh9DXy7Gpr4hc
+          izsxl4+d8ZdmaRAm2TDTTWdqynYpBFbqVtjp96kO19Ga4fPYJ0QrbpdhSYBdgoKcX76oTANdmLNR
+          QN2zbWIthpYtRwWUcFilcV7Oi1WTd31Roc8d9644pj+w1+e1QmAqzvi5Pj0wSfDJQFNdVSI3z2UQ
+          5rSM96kOOsbyHFFB3BF4GDB3LN+U2pnkUxvD8ys1cDQd7WyNdKGHwiVdsKVHQrYwaB4lL8Q3bF8M
+          axCqS/lBzPNqkpyRl2H02AMPd//x165ohrV6JS2s+HHBznHS63XYziEsavOEz5Z+cQRHjW0oWt7m
+          g4uvZEtntD7U+KknZsh8nTWdrgG0FPnrV3l5zDZMrBiNXvHBbjrZYNy6zoZYeYnEUxscCbZlbrB/
+          X3zi3eNTPbFh5sOPePNJ/j0ZmmAM6QjG+50Sx5yBNmc/ZpPIljp/30f/4hF8KdSeeen9rqeF/ZpI
+          dvsBm86rdXo1f0tIGC4MwTBwqTCdW+aIwI3sJxIGh/XHnwE1VI8EmxTXayDTFl5XLGFrMepojdlY
+          R16DnL0iklNqW/IGayxORGEV3lnzr1PBhlo9zvDvqFGgfGMELXIhnsQehk2Z0ytI8uBC3HN/05rF
+          fvHHB78xc/fs3GiTZZUF+37t8SkeFtmTRFi9oYSdZ7Fmq8qPInQKF5CoAkFG36YsIeEQ19jaHurA
+          NsFLh7UadDPQTRit8/u7oESss5lhAtlhm+CuA2xGkg/uBu+MJ05uII44SvCjKzIuxnMAJ2c+z6v8
+          SYBwDY4NDBnJJtqmM8Max4kIF+2U47gSPU0okipBueeE2OO7Ids4cPKRnPgfcupxmNHiclhAcXIv
+          +NV5d617HPQR2ZdWw5589CnLwrZEj1Az5yMXynQGQV/BMfSzedFsG/DYvpSoe9g5fvTXhnKA9Vx4
+          vD0sHO7xZQLsyUXvSDWItr06Z8vyBaJg6nx8H07yIKSvjwuzaGjIY35lzoikY4uyXjfwra1WsBKM
+          rmDpPY/Yv7qM+PFdM0CNNJkYHXwN/C9/qLAYY5U4zLFyaPPSQ3T9XQfy2uPj6J+FEHrYNMl1Vipt
+          e5IlRrcTWjB+z5Kz3h8nCBicXHE46qHD02K2j9PtG+P86TDDkm7HBenx1STBoJFoTouXDue0B/NW
+          nZWozZOlAc49PWLXV/eeNM59n/tVizNDw8aZ6c3IURmmLblt0kgb49YaKPyOHnmdfxe69pqwIHCR
+          MFEPbV8v9aqKiK6eTM7OPci2w5yUgDKCjs3tYoPlUIo5KgWOzAvXFGANvnUO9vX3hRRYzqz8fjPS
+          1Uvti47+HDZ0ORbw9v342PvDE+3I25Axji3xvVHROKmfShiBl4YNOgOwaRLR4W4PxKzONNo+fOpC
+          7rVccPZBczRvS/GBTrexWJM+OmDfx6iF+fRSfFQqcrZZgZ7AO2eXBCM0RbTkrBwaax4RX43yegWx
+          rEJFmS9E/+AT5c035IEXejeiYbeO+ky0DMQEp+Z/8291LuHjAMF86P0y2rIfv6Dbpe6IFV0ujvCH
+          77av0hH/mc3O2A2JKo1BvWK18B6gu94UCLdwmPEpcvOhc8qgRNfZnLBxOMvZeFX15O998fMef+u5
+          4rQQjUK335lQFEdQpCmEvi4dfCn5NsOygksDfj6piGbLX7Ds+BnJh/hGrBcc6nUQV1Gi9cXFfld2
+          dAk+xRXGw00nCg0qsHgn1YcreMN5japwoD8vYyVyBAXxnce73r6HVYKLe3ng0/MjgE1CtQEZj/Qz
+          6pLQ2dQVJ6ChTk9Mcy2ipfieYoiVp4hdTeczEvbbBseH+tzxWFdvrz7vQcl0K4m5m+6sdbtcAZn6
+          K3F41oh4o/c+UJK259xr+JKx888J4DJ+zXkx4Tkbv14lgx3fzNLPvIHlAOUP0tB7/Pt8tHlhH0Ia
+          HC8+bO7QmdOvNcNwDYy5W8bK2fNtBT9fFOHd3pxRsN0QuK18x8/mFQFBkb4BMNYimnn+l2pjd0AV
+          JFN7xZEo6BE96I4LgyyKiLU9qqGJ3lcJFMFiE/+0PYbVKZMSJvL8wKeL1NRTzBY6PMF0Igb/kzTK
+          vkAADyfH+cenFuj2tjRcPohY6eI7HMC3BGzxHP2t/7B8vV6GO54hhsuXYMGJ8oFFbZ+I+YjLiDs1
+          TgJ7+fCaqWRJ0Tg96w+0RO43/8rpO9DmtU+p+zEWsedfCDZ68wug1W1BLmYhZxw8u+0fvsSxy8t0
+          eehVCG7ikGKPMc2May6jCwSxfOOTn1k1+/MX5thq1CTWRGu6oUlp0JlRHGK6w4lyN8NvJJd+AoI5
+          cR22eNQqiRmGAmuM9gZTao75nz1gve4aQOe0vYLtGidE2/nK+jpeeFA+3J482p+RcXe9LJB16Drs
+          PaM42vHu9uc/BIdfB5Ajl1Sgca4/crsmUvazLo8Kurc+J3s8oDufk6H2RixOc10b2J8vMhC1YT1/
+          Wu5TL44GAnivli953rqvNta3bYG3yI3wbd+/NVpSA2brLyXG4VxGVO+dDV1FDc7J5fKpt768qXBc
+          TExeyb2Nlk9CeOnpLcvMyZo/dMn5WYKT/Db8w0d5DIu8VSOExes8d/yHB1Q8KyqMYuLNNG9X5x8/
+          Bcl4mbuVSZxV+okFHAuHw+cibbTl54sQcC3ExLwIWsZupZIDUH3LuS3TkG5n5DJwj0fzcRkrbT79
+          kgpKSrBi+S1cAOVPkgEk4n6wLl02Z417PoFjw1zm4eWsdCuT0QeQ+xyxIvDvemxHxobq/BH9Pg5/
+          zvYI1Viqr+1nZrkTO8z2OBTAal4jtt9QjOZXf23R5VHxPu/X7+yPD8Kdf/jgD4/4n7wEEvE/xPkg
+          P5sQd7Wl2yl8EY/aczaPkDXQoi8GCQzN0sZvda4AOboeuWOEtW181xBeN3FXrJ+/YYDXcUaimi0Y
+          H3qPchONXfiTfjbWm+r0pw8UkIRG7q9aqNbsr1cMxDSVQZxKDxzuj1/v/BKHQdBFdKqOBtz1Cp97
+          POqBiskxB76vQ3y+WCMglt2IR9H5TcRLsQZYw5RVyBig9emlvzs07KUF3vjo5cP+otVCD7cQagc/
+          ngUtvA4j/G0B9L/uTF7bYb/BHUYujJ9ZgP2iIqA/ss5/+oR+2Waw8vMWIPaQV9gNW7JX5Ofij38S
+          d/9+Xs07Cehddt3z9X4iylfgv/xRiF2i0Xfkj8CpmA174WLQTSe7nrLjd6OvimjKsoGHpxAoRGvS
+          KdrjcwvBRcQkPH/noZskmUfH+WtgrGZsRqHjyaAUBOKLsDM0UiR9CpmmNPBTc32nTe/vFjaMfMGn
+          trpQeueX+B8es8u+ykb2HaQwHEs884tl1Ds+n4HlHg4+rAtEt2a2ciil+UzcZjNqoZh2BTbbtr98
+          V+/xr4H2w4qJUYR+tBwGJoDi3Wn++BUV3GVqYTWYAo5J/auXqnAquMczYmSPxVkO5ZIjZtYeRH1J
+          abRYri7D50eq8GlRZvqPv/TRLSVqdgX1JDqXFDLHX+izMTNmPZVLiMJTUmB3FtRMIOJUwv5M3xhn
+          8Dns/F6VxmPs+hW6DRF1b88SunDx8OurqkDIt6EBIJkvRLvoSdZnP36Dq11C7Laf2hm9k+1C7fUV
+          sf0q9Fp48aQBh5Pl4LMf+PV2QwBCbhMKrAWJ4qyfB4RwmGwOnw7fc7R0Rumj5yPu/KqiybDwGRzh
+          M7DHmarvlrZr/F6Q8gwsYgZ8qu386vqPT8nZi6839SYE6GG8MDn5zZPOkxHxyD+c4SzdD796++UX
+          GZ3jY0ocuXtla6kePxLJggcubKqCFSTIgLah29hxch1sA81TsOtZ+JEuvrbER/D52+8ZntC1Xpit
+          lGAS6ddZjKxHtnRDIMM/fn6tNdPZ+Xn+D39p8VBmZBgUFtbbJ8E3p3KixWMPLMQW88H+Acn1ohur
+          jsrHERA9t21nzzcF5BS22PXH0hES7WRLeve4YkfBhbOghJGgbIcL1vL2onWwyhcJt79u5rf8SxfW
+          kEM0/dAFv/Iqr1dpTUx0vGUWNvWmdxafPmbwDlqH7P7pbPc004F847sZpVw18MatNMD2yz7YYgU+
+          24JdMZ58XcU7/qJLsN8xu66eNAsj4oZ1u7iB9Nl7lvt2kkTT1X/FgBx9D8vxUaBrboEEtIeeYMOf
+          uXotmjSHK6gh+ePzM//iDHiVfYnI4uvpbMi8LKicYIlzL1+HVfotBbyl8o08+MM7WtRTZaJYEZKZ
+          4Z1PtPMPBr5HV8Nx7IW1YCrYheiKGYyzmKGT8FR4mCmXwV9uPwUs1+e1hJZUXshVA3vPt/XUQKu1
+          43nz7o62mNfrhh7e5YLD2Cy0dW6DBGXmacK4wEhbYi2LYQ8Fg+jbb+/k1jTyccev2JgyOLThudxv
+          VTMp1pfXjc7p8edDGoALsWRTc1gi2S2MHOtKDKGtI6q8niZMJPfkIzQhQLl7VaEzy/k+fMp91nNP
+          WQUiE93JSXpYgCW3lIWIrLxPyvzjrGiABWS8qccWe7nvfHqToB7HJsG7/26sGNio05OG/PHRnhaN
+          CZlHVOM/PXb5oFcPrr942PXZnzM/9ZsE1bkRiUHq3zAcLcYAf++zx8OIimdLheHhl/k8fPA1XY+w
+          gFbyFmfROFoRlaLnAnzh0+F4U7Jsld4nVhq9/EOiXX9fjDoxAR5bf+YuL6gRf2UbmLxVE/tHzNO5
+          Exdd+vt9Z8/vM5VbBjJLdiO+TStAxWwuIWCcZGbYoqaCoxYmKMeIJ1q0VdoEWM8HHNQy/G+9Gf2R
+          gtN378RtFAWYT3fDh93DzHGEXS1bchxI8M8fsF0GDp3T8gphr8n/9OA/PCbVYLlh3BAtWrs8ZCXj
+          K6V+ly6zM3oPMsOlx96fXgKanyKLomJWEVZ2/XY5Pp0W7PyfmAng6LqOeQJsHJ7/8Ly288sc7Pop
+          uRzKDazbRQ+gFmdghoV4cMbmwW9wy4QB29ZvrFc2zFw4VscRa0KQa89C6Vtwy5OA5Ol5dMZvhUvY
+          febbfBnWEyC7P0gpRzbiB/aVsvhSQLAt332qKUudPZ/7cHu8FqLr71u9xxfxnz7kcOcTWOTOkeFY
+          WBwxRFbR2HwqDCiGqk9sGW7OLNzHK6xZGOGiqAgdkbT2CKoAY8t6sXRT3VcDb9/G9+G6yIMgRc8N
+          GmfY+AHmTtkayKAFe33Ap6f07Izxzdkgd+xKcopPEh3r1RbhjpfxXS/CaNn04xV+HrNFFPF33fNf
+          0MOFOWi+hKWB0nf9kKWU3gX/uIyqtuiV5cMmFstZZOApoqjQU3RL1du+34szcJvUSln7yLEuw9Lp
+          Z3a4wkGcZ+yosBj+8sVfvJi/f/b6knEMc1x7/9aTDoPFS3t+IO7TKYbVHadA1NYmxJbR8MN0VvsS
+          CIm/EYMDKuWjdy5BJCIWa4M8akSoheDYivyETXf4gpkVExPWWdHs38cMM5M9TGj4S4ttwyT1Hn99
+          +D7nHrl8wk7b+1S4UCxvH+ze2bGe/VyBUDvNvV8/SAiWHb8hsSKvuRuanP7Ff/hiC3ZulWkAlFmd
+          Ba6s2hGbH/1h9OljBHu8xdZENbqlX2WE+eaL2Nv5/V/9Ce16N/HNQaDEIsUMR/0gzmtXGIOwFuMV
+          FsFmY4OVjsM2PyIfsu4i4uQrr8NLYkQfCGg4+s1yTekyCOMG0x9l8J89kUQOjL/6Azn5WVdPV3q2
+          oWl9dR8Fz/Ow50MT/vm7xXTbsDxRm0DxwCh/eJTSe9b6f3rYvMqHOqJWurpo109nBGWUbdh+lFB+
+          JxXJXLUGSz+fDJAyfkF0MlcD7d+9CcRXGxIDHm5Ze7gnG5zuoexz4mhl1BjC8R9e9QC2nX94ZPw0
+          MfaOVTdQ7t5XsHd4g7i3txeR7wJHwFxtZz5k8Fnv06tbmKWvdhZxn2c7X8ohk+4S264X7fvpw/Z1
+          k31uj/fD25RFyLUMJobDCRl9rHkLLCi6f/w+E45ykcCJvm1i7Ou53t7FDGP2reKQ4+dobUY2hrGO
+          IyyLL+RM/uHTw5Mkb1hb7oyzXu8HFex6FnafTefQ83xeIP9AeOZ2fDn96ZFaon9JfkwvdCuxXsK7
+          bAD/MyuqMyr38IP8rz//p08p5i1FjWr3RPePwbB2Qg9B/478Wey3ia4ZTyESvfD+p8/X/OkXVBAY
+          1xj7h4rX9vqsCZWr7hLrJzMa2fkSPOU6Q3b8Go3xETTgoNLgX71zPsxBCZuSkl2vX7VtFX/Xf3jQ
+          YSIt2/aZ1xDeLvbMkYyAzbXsD4CloPngW3sOd2utK7j1U4N9w9OdxWlrHZ4Fz8K2pUyA3EO5RD9/
+          qogK1IH+4Tm0PZ4LCS/VK9tKCq/gJ33tf/hs568lVMp7/Kc30j2fl9KOr4gHcO9QKz264F5t3/nY
+          /pqIzVl5/sO/REk8L9pA0JfgdwblXF0zNBBb5BPYGpa2423tP35ojGNOHmvyyGZy3keoxC6Loy87
+          D7+9vgkQZXNi4h5m44+WG8B9GP3px1qvXy8mynxAfMG5B9HfesFMiQYf7M+/0kO0/dXbiP2GSTSz
+          s5jDzFru//yHrje3B2fn/vaXImHBkrykFIpP1PpHVogjStd6hBYdMY5lx4jo58nFyKngRtKL0Q1/
+          eBPOsc9i/yMN2pYrJQv3+vsffsjIvZlV+Lce2rKZw0huKQ9FftH9b/mMHU6ZPBcOwOiIJXxLZ68X
+          SADbHUdcDrCAPNwP+/85UXD8v08URIKokrSRL9oEEL5C7gJ5n78eiUPZ+FVBE3877BjPX7SM0T2G
+          bter+PRbf9F2M1geiYVgz5w4KYAfecAD4V2b5HyptnoVxmcKK9reiB6E54i/170KPYou2AsEXVsl
+          zZPhJZGPBN/6337HgS9BUOsTDn6BTidNixPI+p6LkzoahrUvLzaqdF4khifn9aYOK0T1ZTzhwit/
+          w8psrxJcof7GTyQ9AfVefgqTJjwSvzyBiBwNrkJRG4vYbQ4kGsejyUg1f0uJcg54uujyp4CaOno+
+          +5y+w1qhbpE6PcqIdmQVbalpsyE8S968pTWg3eMOehjkfkD88xQOHHebWPhNvh1ODgPrrNuz4EEk
+          SOosoGtXr2+57JH+gbnPhd1KV9V4+sANvBafBkWKlpeeh9CLmRNxZCJH7BDHJjLO7rAryMTZonDR
+          0ZUtEl9AV6vm02TzJXROrZkatw3MFVPGiGeZjShTWTlCDz8NWqLIJFopNnTxdgBIf9oROx9W1lh+
+          veYoMuk8X5vhMExOp5eoloc3xvSTZdy5SHwEk4jD6tq/63FDQos+z2qac1XCA+ueNxWGY4WJdzod
+          tNVcUAlj01dItAJ/4F/GJYGnzgTkEVtpJshst89V2xxs1i8DrHptJ+Ao+YhYp8Jy2jG6X+HypjPx
+          hvcpYlXzWcCuO1x3+ztnbC9LMqT8TyN6rfHalq1PEy6b/iVh9NK1MXdPIYLd9YG91GMHun0vH7Dv
+          B3kyxtGZlb6p4EiQS4LwatfLr7cTqJEpIc/uVzvksD5ZaG3SEzuY7gr7Wd05tPghQcu4DnsGkgun
+          6WMQr243sEinU4hmFVnYPx2cmiedt89JFp7E7tq03vQXP0LwST3iPKOPtk244CFblBp+VcfeYTtt
+          4WFjDA8ij81tEJJstCHmZETCTC4d9tq8dHjFlYOV8qTRZeJVGzF7xfqun6/ZVovVfgZTN/zDrOJ6
+          sVc7AXctRFixMjWjkixIyA0ON2LfrM9Af+drAYVIroiMJCMizqC0qGXsASu/ydN4dV425D7uDnGf
+          yY3yhws/QvRSQ6y55hkst19VIjHe51aul3u9BFm/AXZSO6Kb6xBtB6G8olKvjtiy2cfAfoSjejTO
+          TILtVHAGerlVMjJ42cTmh7sPbKy5MSTyRyO6dPI1jtnuJTChdcCJXbuDwN0riIDHND4fH8ZhaszY
+          RpT/an/xKGtjY+/pkAhnfyvabtjqu3aFLyaxCP7gFGz7fqPg2pyx/L5cgZB39xCEw1YR13130fL7
+          OQ10bfaLbYqv2sKv1wK65CmR8yM7Oez7reToecueRDX0E2C1Vsn3CufV5/f4w7OMlUC7/cREdZp1
+          WLKbNcP6Mp/mrrtP9bY4ogFTxlN8vrpO0abO4gJdZVJnAMFroIY0tJA5tgmxbPZYkypmCmh5MCCa
+          TH8OleSDBHurLHCsDBklV+PHQNWdcuJsWu3ww2EJkZDbJ3/R7nLEZqEtwed3vu49Nk7OMvuTDa26
+          t3FeGzkYvJeRwovw0fBtYUm2no8rAw8NzLC+udkgIHTRUXf0fuTBb8Pey91U4UWxn1g5BzEgwTX+
+          wAuvd9g0CoZ2cdsZ8FN3z/mg5hHYhiw1YBex35n2aq4tn5WmaBlck8huow0rk8kyut5agG1mCRwh
+          C2CM+oN792dOuQ2UTPsdazJn5HzwHcoLPk3RLZgEv5+mNqL5GBuw+iQqCfy9QonEpw2GiJtIwJ4v
+          w8pmdY6kS3rB3qnau76Pvxzys34jUbN1zvQ4RjLqhnzGntTbYGW4IoZ16Q9+ND4qumXuPpbc6I19
+          v7+A65l1hANNBCLzjz7bhCXcwOdmC1gVpWstwKPBImGhGXYOhRJxzujEcPOKu388hR5Yfr2awmJR
+          ahyE134Yt/odo5ZTTGK0ZzXj7+XgA3T6iUQdhLGmbybLYbny0IfPhKOreYs2KIiPDzlJP93hPHWD
+          aFScZZaYzxWwi7VBlMrigVzZusrokAc9WtHR28+MqoCHauJC9nTu8PmBLUozchXRwytcolWMngnB
+          /ZmC2qI6Pp3NG92+VCxhcMQ50RFTDoJ69kX4yPBp5sLuAjgEIgbt+ZWoeeU6bLilV/jI4IZjeuc0
+          at4DEXEOb+ACSqbDnUqngfAyBiQU52O2MVe0d5HPDOy5TaYtt6k1kXvQdHyOWCfitmfMolFlFB8V
+          xs+Z49d9A0zpMdgc597ZbPulw/7A6DjmqrIWULGUaPHMBL9mldRkiq4uWpyemRc7S5xNfzEzGB/6
+          grVDPILpaKmtYJz9gSRm12djbDgmqkbZ8aF78x1BcMgMTu9qIycrekfC+4ECmHKCg/Fu3/xz7Blo
+          vmzqg2jss2W6vxv0uk05uUftoq2a/GqghAay33nA0W6PBVqkIptFtF7AkuTSCEUrN/0DFkKH6iUX
+          St43H/AjKd/aFF2OJYouRT5L4fak1PbzHNSkQETR8k1r34XdSyPz2rCy8eGw/eEtz6t7/9g5YbYJ
+          l81FL16NyEvPdI3KkWojJb7w5PHj3vUypm2CrE18ktcHS3QMkhoi21JcbHu5Fy3VcUzgyDw3rHT4
+          4awBY47I5n7VLI7SN/pNwTOBVjZZZMdP2ai1SoGe3hr5R9HMh81/ZQHojvi3K9x71/TpvoHb4R2R
+          +KOy9ZakUwMK5UAxvjwYZ+1epxHtv0fs6VjXW07eC6I/5YhfmRvSzf14OeTKV+5zuR5k600VCgmq
+          jjHTKJGG5a2WIjqLxCN+Cr6UHua7DD/DpcFW52zROt26K8xc8iP7etUsRzEPSLx3pUPfePizD3R/
+          PEYSK8Ld4ZfzxMIdD/29v0ZDVv0gV0gAue32L1zHyweqPXWxIeS7fz1+M/hlJsEu62/16Gi/BfI4
+          FLHn6NjZUjaD0Oo94rOv35kuWLZkYL6kEvv18QBmlFIbgaug7XgLZIswP3vYxEyA3QPxszHc0hje
+          +zzwbzu+ZqGcGX+/T9TLo6k3M06vUs/kPfZx2tGtFvsRTJB9+Hx0r7VuDqsK7vmUnE5qEtGT3F3h
+          rUs+RB9MjXIG9xthb4cOlu98nK3+O6vgNnjBTO/uIZsMsSmBM2omsbrapHT5GDbkWe9ETlo5ZqPP
+          oEqwons0c3XJRP/slRgZII7RVcP2lx+z2FRJqE0Hbbm55gK9fEXYc7c426giB//ym00OikZ+LeUB
+          1wrhzOSXuF79A9ThUrEhse+40pYskSvIx9YNW7s9k1R8xJDRmBfR2ivn/Pk3cIUU+OXX/mXzZyQF
+          2PEitq22HdayZD8Qm/mPaI+7Deh1fDTwjy9cT9abbt9lnaFzTQ7/+S9A5xiYQe8TzU7LPT+pDQzd
+          wwn7Vf4GTb61jTSNdUXkMZWiVSbfD/x8HjZJgdnUxHHfOlR+15TgRxgMq3c6bpAdQ/jnT9lo40GG
+          YfYa/vButkY0kCA+zxJxOfioOSor87/4K5V9OWy+tpZQmzEz0z1+L4erw0tP+vJmeohdsNxceYOK
+          0r5J+jH2O5LDwYTkbLX4vONF4uuiAQ3cvLFs4xxwiYg/wHjeU2KxolVzbeVW4A8/vQTuE23jLS9g
+          mQ6tP+ptTNe33Lb7nQl2Fl+D6nAHvlug31oL/nv+pRjWAJ6dzMbWqn8AvctqAl+KfiaKpw10O8Qm
+          A6HAHshZXfc7OTavQkEYhPkRTqNGco+rkJUecuL2PxGsHZoD+JKvKXm+56Ve3morwZ8eRvv/18OM
+          Dpcraib9gO/GY3Iofx3/4R0i21miLYX9ECGn5QHJM94C+3qx0FVzgqMbeUdLX11S9MdPIhSQaHka
+          5x451fGLFQufnS3fygaa96oneo3XaDEvRQq94uOTs/6y6bLHY4hGrfA5fl7rScGP9o9v+gCCw7AG
+          wJJg0WYPn20EibYyJ+qwLplkZna+OnuGGMNesVvsD80+5zLAJtjjlS/GQU1J+hA3eDG3ZuZJHUTr
+          0VJ7SN+uQi5KhenEcHEMhE5isD+t6sA9jpkKf8d9UNG2lmDT9CsDfK3K8MkFNhUut0oFt7Ayic8p
+          t3pdmyiEe/4jj/AdUCoHTgyO2non8i8vnRXqQQHF7/TD/v381pYzx36Q5737P3ysbTNn5pK2yAG2
+          j2qndVC59MhPU4HgkrrDtv4ePLy33WH/e6x793MqUFgnI4naK6etR9fKIbNU3Az0+qQRxdNL2B+g
+          ji9sUgzl6eWpSDTHDKuJVNX8yFMW/vEV6/xphn29e8ARMvtCm5BdAV8qBAX+QLDAfbKFkxIJMJq/
+          YVsBa7Q0i6lC5jIl5PQVTg5/+pgQMmtY+fB7MTM6IFGFN+X1IWoiqQOnaUUKN/dBsXeKW2cBnzX4
+          i5fzW5kHbdn5zR++I/JtSSg1V+LCwZsR8dW7qi0ELw28t8MBG9Fdc8afOai74uPgtEp+A/3Dj3/5
+          ToVV4UytVjNwo+YFy4BIdAOMJ0q6X3HkXA6Jsz2kRyO9UJ9jj4U1naIpC8FD2rDPSAQOlL4iiATt
+          mfiHYz3Uy3dmGyhEakXOn+VEKWakFLzwZyQhecXO4jOoBPt+YkNkZrBs6NDyz1m38a43aMS8JxJ8
+          3yvOZ6pcoct0rTZwsaoI/8X/2a9CA724sMFK5SiRMApbip6GZftx8hbrLWAPKZw4P8AyqA/1Wh6e
+          NuCb3J3L132LNhaoH/Q+eww5G49Jm2C1qqiINp1gQ39G3YH2+wkbLsSxl1CNvnNW/8fn/vD9wqRK
+          gUYFPMnZdSuHss/ygx4nPSbB335cBCuVnM0R/cUvT1SwRyT/4Qly+r78uj96sgo3N6PzLaivEeF4
+          rUQT1Me58VyBdnmk5EiuzndsdkY5rO9KloEOeNEfQ+3mzJej1UIhWjA2LwcJUCEUt3/4SG/5QZsW
+          ey2B5Ra8jxaBd8bPPuekrD6d32joOlT6i5+BAz4hvjhWF428EBrgngID+/LlM2xahUrQmvmGcTXN
+          dNuKZwrH9GPj+Gj3zhJG6wLFOzMTN+vnYZLe0wdi2Bnzsn1PNfenn8nueP2np/BXd8tRqP8wkTXn
+          q81c/EnRyfqueLcvbamXH4Tn5XYn8p4/lyCrFnjvi8D/AMo5lEk2CeVafsJ+EzQD9XP8gWhcnX96
+          3SZOEYTxeO587e4eIqqZ6VX602d8Zyszev+0DbxX7BkrJNUzbtWpiIpFq4l5oYd6XNqskmx0SLD7
+          gOdhSZhQRyY8jlgfC3b4iUM0/nt+m0UgG7062yBmo4A4bfsBrGNN6R++xjdt3PVALo4h6A/brj8M
+          dfF7H3pQHl//AwAA//+knUuTgjgUhX8QC3lJwpKXgICEl4g7QUVARB4JkF8/hT3L2c2yq9puS2/u
+          Oee7CTn88nq9FEulydeDmSKzSk755i9KedMbdL+dkoEATU/hxjew/P3oLidf2QZsfJDEF++Zc53Z
+          r+BTvtYtnzXuGl6qCG764b/vTFRPCpmaXx7H4JJQMKsj7YFxuEZE5y4FXabzKwW/fG6uO6aeivhY
+          AEXSLsQNV1tfdP2R/fHTzZ/FczwWGCxcfSHoanL5HGQDhJCZ8i3vVWCRmvEGkWIuGx9L3VE/duK2
+          g6vCu8ts1FuexfB58Xl0Uu/nf/nFLz+lvWoM9LiuHZTroUN+cQ8pvU/fEkancSXZpvezLN3/Xc+/
+          +sZm7lXw5igVCpmLP5DmrBty5MkHYiTmqK+fl9D/8Z00m2Kw+oqEwUH4vvx842tTw82rvPkPn63z
+          sZ7cr1fCs3pvyDH4+MNSBFonb/3Wfx1zLeYf7ZTAaVi4P95KNPfqwLL3VnSE18OA9580hch+fIgm
+          eKecm8dMhGeGtj7fq0a98NyVh4QrEfL6T0bX7mtjeDh+FnJMz1w9Meuzgtf+npFsTuNh5s/rCNaz
+          +cJsOzxr+vt/QyxMRFXx4G7vr4dk+vhEv3jPeLT9SwvrD8ui+7vi4tkbMx+8PxwkqH3F9XLenrH4
+          wnlIjnos15PoMBpA2C6Rh8UXmN7hrZUKDWdEyYGhCzt80aDumjqyPKUF3S3YFcC+lD0yy5vrEr+w
+          ml/e+eldTGX5agLUlhdyPxye7qpEegO5pVTQ4VY1MVt4hwAUn8tE3HjsY6p/9yOgnXFH2ev6zZe0
+          Gh7wuF/3eGGbK9j4OrPt8HCQa9Rvdzi3OQMqqVp80GItX3DUV3CxZfWXN9yN37SgLr0BKedZpBhd
+          7z04v9acmCfvQvFLSk2wHlgfZXnL5niASwE/RhAT/7Ab6l+eAzu0WCTY/NsqJHsNrPXb8pmdchhW
+          +3q14e59V8l1uLvuCrXAg9P4qshpXbKcLF1wk0UmS3zxOVQ6FdGeAcW1YX0wGtQda4pnGMxCS7Tj
+          iGLsK+sofV1XwM3XV+tl40mQZ+FKNv9Rsw0reVCftQBt7x+sosu2oFZzf7s329Hnz3T0fvpOtLMa
+          6Ng6myK8LMmEbNXG8aIrlxaulXAjh6TTdHZtO1uCXTsi/1GjeO0viIeJmXV4Zky15kay6wEA0dff
+          n9wy56TXu4VscobkUOlKjn88HTufCKnwcKZdfQ0e8o9Xb/4D4FK/NVAlY4Ce+0sQzyeDehKPzTOx
+          ymdB54/75cHx0N78XZOn+jwfJQg/cdX64sZjf/OJP/7lK4dvjuV9jOXXUxyIpk9PfbnRewb10KPo
+          9K7O+cYXbdCwo4q08DTki9Q0GSxc00cGW1cxnT8eAz2VaMQpH4d80Y6JAvOm60iw5c15xw0+bM3v
+          FRmW3riL4qIbjK9b3isf7xgzfCHBwAVXpEZXGOOK6VLw80fXxTDow/VfKUzBBRKFPxF9+vG/e/Kd
+          f/OCmO0OjQ9tX4B/+XctyHcG335SkZO9IzoFxSuBMAs5dHqYH709vQNR3vob3o32KWaLqemhEK/I
+          p6DqXLpW0Qq3eYg/qeEtX5c+62CXyhlRif91B2fkFFgEzxBZc/ii65dpeJlzzau/fHWZdsu7wDC6
+          TQzxxu2WgzrxSpB9tNYXX+9moD0cW1h8zhPa9CVekcRp8K8/PB9xTPvdcIOl52pom0e5Sx1Yjix8
+          3XDLIz3FDJ9I0HKGu8/Blzmw79PZA4vycYmPbke6LHVRgJigHx9+6xsfSQFVtxOCv3nasJc6SCy3
+          w4t7Nur3289n8OPT19cDuHOV6yxcPvMD6dwFAnqfuwTuPUnCYP8M4+Htmf3v+/QFunf1cQoHCHQn
+          TYnx9V/1cpvTRvr5faHzEhdfNKeHe889+Xu2pPpY8VMD9pInY2ab1+GN3wKVbyby84+4fMAEXHNm
+          RbocoHiFRVvB1fcNX9jy3/I83Fi4FmNPrOfRc9mnkQRQmwFEFjgy7tKZ1SzvLtgmuqHf3NG6FBoc
+          kvcNJffpXc9pM4y/evHF3/wureoCRlJsI8eV1FyYWNrAwklfPusMp0GYgvsNVIf27beVvbrrTugS
+          +DyTgmgWY7v4V29DI4V+tJzuOT0pb1buuJUjFt0P+lxlJw1Gh20G4uSi+9P/nx5vvEGpl64ySnm/
+          NixC2O/p/DgFI9jmkXhH4QJIHgMb3oJj5IteGNOFv86mbGfBBekbH1sXex+B08S/tvkGq0+xL41w
+          CewzMhLT0/k6MUp5Wz9YcqVXTiPWaeDx5Nfkx8dX3YpnMKr7O1JnGOS8grx1v82L/XqrvzmPNFE2
+          04JBxySP81lreUUO9Lr2H8d8D0Y7fGQQu8mOBK8w2eYXBILmhHi/ypVS//OLP3+79Rd3BGaAoVBF
+          HNEMpx7IlidgZLwROi5BqXPl7uzAm13bRJ3soaZVLHSAPFJAnKrR6DKI0ATKdyrRgU+Bjn/+Awvg
+          TI5+x+Tdx/2y0ld8Ssioc2/AupWvsFMM6ANuGIdvPj1XIOjPzN9b2bDdSsCWcoM0C50+QRHjrlVF
+          GHmnDBMkrH95TSp16hGkddspuyczSpz+CNDhzkTDvEt0Fn5yh/jSI96D2fafLWxOJx45y2ty62S8
+          NmCb56Ckh3aN33Su5DgsCnK1Mjf/8XR4PbA6uvN8q9PlZSWgfTTbM0dNtZ55No5k2vEOQUqaxvPu
+          ovUyt1QKOX3zD+DeI0xBOWBv40/zsGjejZVyb9qeockrw3Kj5+z/7CiQ/ntHgZR5X3Icg29NzSYo
+          5SMb35DJJPWwSoUrAcPkeOKm3U5f4D6IZGW1XP/lGWs9zp4QyXXhieRRT2/QAfAuwfoubRKzz5AK
+          sVGN2z0cEzKp7Ncsn4WjfIqmCrmPLxpoaxcsBFIuYqjoVT5esvsILXuv+FQOTvEyvOUU+OeXhXJn
+          YV3qvTQMcZ9yeMZz7fZZdMvgfnUQsj1Gz9fXzV8hTmWIlMX8DPMBXFNYe5VOXNZq6lUKjo4sqZyB
+          shP2c2ENvg8oO97iS997Q0fjBTTI1MaC9Nwkw6KJNQ9Nb3bxMn8/+Zy0E4bgJItIiSzPnVV4WuFD
+          lQ5E269lLgSRNMJTa+RILRYXsK1AIuj045af9CgenX3gy3LQXMgT0H28JHLVy/vVRuj25CW6ivoa
+          yayasSR8pMmw9B9Hka8tBCQK7uUwR9o13SMtiog3PsKav3+3icWuZ8nhYcGcNKnIgIeLEp/PnNbl
+          uBcw5XBmaqKq031Yn10vQeelvNFJSSd9Xunehly99/13MHTD2q5iB7u7V6D71N5cLrxwDzkQ+Bvy
+          ppjksxe+RPmrySNyd6eYrqt4WWGgXimx5eRJcWTuIKiO15g8zmnoLl+1hLIzkxsx6lsPuNCVsEyH
+          5IoOrSgA6lR6ILstzxP92GT10nsXBVapVeApY2g9n095BYEjeyiR68xlpcry4cLNIfG/vlfT+ni/
+          geSbBr4MTjpg96nxkE00C5i5wT5fGr03IetnFgn953dYiHMWoblDNbGM3tZZkAaNnAkVIo9Y8YaJ
+          vBVRPlCTIFQH72HpqTdDVmItcnN9Nl9oOosQOvPB5+mGSD92x8iXz5SiSLh8db7U1EC+DPqJqNGN
+          B6Mr6qW8j7KMaM+mjEkLI00e3fMV3Z3PY+C/TdnD7PI8bbc+VAP/+zyH/FMR27Y5d00lBctzG9XE
+          5v05J8VHM+FrPz7IdeKinGT+mMKrc5qJrj2++tgbWgJb4XzELb5wMX7oKi9n1pyR4M4I+rxeTQew
+          izPgAUTmgC3PFOXp9oIovOWfeIkbI5VfcWeQ/E2eusDMtJFv37kkTn8i9FtIJZSb7DwRxxBnnfhR
+          28odATJBGcvoK92LDdBvVx7vs8x32Yt3Y6BfnQJUTDGJKf9ZSzkdmhc5Zccz4HJrfshOkhTk9jQR
+          5QteEuGSajd0MfrOnXvDSaD9NgSUNztWxyN4t/LL/GCf45UunvOZPuS+43c+CCIJzBcl8OTw04Y+
+          C/XKnW0zbmWAFQlpuD0NAnD9DKwLMMnxcJ7j+ROeeviQZBcp34OgUye1K2l+SjLSjX0Z01S7FxCw
+          t5KcnskFCFPUt+AjP97+4KhffR7uXQ+21xM7TqxBcOzFlKWXZKKtnw4zM50ryCv9FbOh3dHF0ldf
+          LqK+JPZl9wSLgI8QOv2FxfQCpnxRx4SBHwdryCyBOsy6nWB5ZO8VCkxwzSlcoAPZz9PHy1bPwgXt
+          NRjgwkZ+kz3irb4luWXTOzrFmaSPKmZSwFzxE0Wz93HX1VJMueFMESnhRd3qy8ZAnqInOVqzlfM6
+          6lhI2mdP1OXxdr+//nS4awk6jeqS/9aDtJ3LRF7OO2BmXtH6q0eUNcuoz+/I9KXxtarEDr95vbLP
+          cASXTHth7usX+ci+7wqMrMRG6N6POY2H7CG/zDdG3t655YM/XVs4LgVP4v251DlpUm8ybrTI3y2V
+          RRctZEzppy+P7EXjRR+KG1QX0SAx9kt3NW6pCPdIfyA14XyXgJ273TtuWAR9wJ3O3r3p5GJUFoKo
+          NQ/r5/3y5ac0G+RwaOyav4isIqHHF6G4XrOc3/elBkVN6snxWQf62rj8DA5mR1C+6SFFIurggT/k
+          KBJlrV7VrpWghXc3fHGVMZ+dfeDBW23NyHg0sO6y9xLJucf6JITvgyswzomVtOZSEH1J0IBjO1hl
+          AaI9sjhwqLmu0xq4W5snlihO89mKawU+Xe1ELDHs3cW+B6U8yRJHVM9YB3L56r4swNMeOdnsDrzh
+          J7bsE7/G+0iwYuHmi8Wf3tz1VgX8YlIGCOmjRy4UFXeJHKyAp58YxOnlG2B14hTQadSQ6BwqgJDm
+          nw5w5fuC7EZVXT60tzOwmfdFMXIWSoMQK3DIWh65/fuUc7vgHMkU9wGyTrler+7l7u1tfwrRgxlm
+          unKRFshVMXvEe0Apng8gTGSuBv5PT+N5+A43GBeC7y+EimAMtIMP8Sc9omtLKioYTmvI+7e1R9be
+          /NDxKAgZVM/OntiDAHKyrXf4eIUEnVvxAjicSRq0X1Qh+X3fDHi1bEN2hJuDjFlAOVtKsgl//e6u
+          pCdduKlfBd6R8STZkaour2vhDQoji0jhJDHgWu1VyV7cqOQ+tZI+6c+wB3O074kjGRKdfv1UzY5v
+          5HLjlNO7Uj7ghD0Fpdv65pIUSnCy+Bzd+1wA1Ia9LfNlFpDILgJ96W7Ahkx7uPtr2yNKt34L1/MR
+          IB0fp3zd8/0KrezN+cu7svLFvzkj/KbJmzwiF4PvqHQ3OX++CEmvBAN8Ht+2rIacRkIelTVlnAMr
+          XwyGYqa8Lzp5zkILzNS5I6NANl3i7NjIyzENydHOdwDDE5nB7buW/jwG+3gm2tGH6usloGNyL8Fi
+          IcjALn2NxO4v77iT450Py6n8EJ8+GZ3y3BHCS3hZcNeoL/cLwzML2yTM8d4dkpo0s2v/vj/kb+9/
+          5BiQwLe6qOiQhirg7eUDYXiTbLzwRzknHQgLmHu878v+pNIJeSAAynk8oayYQU09qRAh/Bx1ogrs
+          MeZ66q0Qz2O+TRCnnC4cf5Pna3MjlrYDYC1RNsL2zgjILdFXJ3i0RbkNA0Iu6vMAOJc9FxArsMHS
+          5bDLV3b82tKwfGWkp2kz0MtX92CNSoLUQlUogdkXwpDKCGnvg+XSYmc1cF7hRrOOp5wK7ZsHqv2Y
+          kd0so0tz9lpJ1jN6+xK/F4dFsBIf2q9FIX6+2K4QGo8WDqBviNa/ZX3m56cPmy4jxMZRVE9C/GD2
+          5vGRkG09gBGz71bSLsxj0+tAXz5cpAEGzieSDVYQC4meajAVo564O/kBlkIqGYkxzy1eqsvbXVfx
+          uUrH5FUTJHZCPH/DcJW7tsHoZC6GPsFR4eHH0CNy2LWvepEOJgO0u8yTWOJSdy7SogF7l3zIcbjl
+          +vztayirh4eBZ82RwTxte/jhpfAxNa9pvvlTUYr4mSXmaxjrlZiEhZdMeeHvV3rrBJWOAzf99GlY
+          8jlu6N4H6dmDRP3Vg9lkFQSP5ICe7OS4PJHeGizW7ILiNDWG5TkGjNwX2zPXSCoOOHvcHcgaeYl+
+          /WexvsiBUDW/RAvldFjQrp7hvXJ8vzxHfr6mz/Mo5c+a+PM6zPF8VjwDwNtDIt7rstQrOlsM9I5S
+          QxTRu+jzqtod2PoN+tUXFRXDh5mADHyVXrlLKe+UYHjtbKLgWXeFOTgngA0PX2LNZy8W4EwKeBot
+          irnD80yJdtU8uCsqA4tTZdVTOXgBbLT0TA6v7EXXNjt0gLXIhfjXkMkpNswCjJ3RIKX+6jphUSlB
+          ZrfukcMfp3q+rbYDv8e+QOa8ejWdg3sCzwfiE+f8/saLaHUp/PnHmEOQzvDsSjByUhap5bh3l/fd
+          vsGIrjYxjpaYU8feG5DsBOCT9HEdllv8MYBxqlosc9UFkMcuxiDFYuYXydABas15BBlF1/2l9jQ6
+          6+Dkw4H5BiTWfdYdmeleidaaJsTYMUj/+33Wv1lY3vzNx6p1Fq4P7oLC1Q9cfkcTBvaDMPiMgvua
+          7Fd2lMeEDESfrlo+jd7gSO8FTch7Rw+6jDQTwS+PXTn1oP8+D3nzHwhR7Q7WIfUqqNiDQEw+e9P5
+          kXYYfOTi7S/w/dbH97j3gH+uLbwewwlQU65GaF6PR4IeNyEfhctBg5fwvBBDb5can8fJBsQxR//2
+          sZR8HaAUwAubPzG12alemPDLgEOciAi1+Qcsu0GJ4NwGNYlYGOk0HENFEq7PlJhG5NSjOyozLEQs
+          +UxR1WC1lw8DfLHl8E6wVfcvj17ltPSZenrTWak+D3gBBkK2us71HH1wBCX28/EXg9HdSRMHHvZX
+          xyFOMYhgxCjroHWQEnTguVZvO9dJ4P3tSMi4PE7ufPput4IVs4eMIHbdhX1hRiq8SCJGB8J6ShzJ
+          gZra7slpVMOYGu/QlLGk98T3A/WvPuAtyBC5sU0DpgaNLLwj84l+63+kDPFl9cGeif3xdJ0XHqEN
+          BMIDpPnaMZ7TFbbgl2evP37xoAEjr8Or9Pkr8QF/LvsMbnqHHuSr11z5VCv5styoT4in57zXbPe0
+          uviNfDG4grX8PHxwXzgdc3u3H+jtRn2ZEyLL//WnhaxZAi55miG9GI+64O+TFHZ1qyHrOR0GsvGT
+          n56R+1k349n/ghKaxZFDp6MQ1/i4c0vpsmQUKQt458ux0FmZxPORuPuPVC/758TDmot3RNWzb03u
+          V6kD74+QE9d97dzF7GMWfurliH7+YA3PiQmr5ypiWmtTjHe+ovz0HanddKzxlAa97I1MtOXJ87ZD
+          OUjkk5k2BHWqGv/qBYZSvCI1OOiu8Jx3DdxHtwzvtPEVL/fD0/zVL9K2+h45hqbwp/8CB941HbZb
+          ldahLn3IKF1Ng8MQSdpt52BWb8N6uR8uBgAv/eMDUVvowtrJDE9jxmz5faxnEx9FeUymATm5q9XL
+          8OZSeLpmETn7QZjP1+zLgn4tC5RN94Syj7QboSPdeGQoaUHptdN6iBslIg9Q4mH51c8XwR0Wt3oZ
+          ZuPWgtFWK4QI9GN2ylEAf37ZoZiPJ9WyFbDsrBLLqpnpvRgeR/hKu5k49l4Z2N/6/uVHv9L29O2y
+          94dUcUKMdyFVYzZE3fzzg+g+qmHOV3QyoZtSlhhrIepTeoap9KtPLzob+QIe7wCWvT8j/eeXDTlt
+          oeDRE1I+z3tNRcXzIFMQFzkr4+b0cxYz+OmtERmdGf/yIgMTqC7IuO6bnO7X7Uxfa90xTINl+EZM
+          7clCJXRk84fxaqgeBkPW8KRwjQddwqtmy012mXyZv3D1OqRGJTvzdPOn0jFj3nnvE9gTtyJb/Qzd
+          zY0yeDR4jFCfnl3K5nUvZz1rosP4VIHAFmIq/fLuXZwk/b0TXEXavw97dMqvfU2/WFPk22RciGf5
+          TT6X8LhC/rjrfnqbr4u02LKyMx9bv1jqadzF3U8f0cYbXTKFdxE+hiBFz1IJKD3FIADCcX0RRfpg
+          /U+vN35FrvH4dudRP87Sgc9adGMOQ77eOexD6925SD3Qks4o+KaQcypI1PuLpwsTvhg54ZUrsmvz
+          SVd4+szyr94elXYFdHd53eT6LJ2R6rHyMLujvcLN7+Fq430sq9gm6E2jQDFmXB0fqbLKydhbSA0k
+          ha70eWr+8qcxW00+ihAbYMtTSF8SUi8XuxHhrPhHoukDG+N5OrewGLUFoaX6ULpwTAYbPEWY8R0m
+          H/37vMpc+bn4HWxYigt+FeWHe0rQbasX7hGtvSwtDSDKe/cZurGIoSy+vydkBgyN6ZU93cCL6hhp
+          IEpjTjkQ7y9/KEFVx3TYex709BNLtC74xhiD/QiXbgLk6oBvvToLa0KmNA9YUJ8HyhucKskG3EXI
+          0FSb4uteseXn5+ritjjQePPfEsjXz4OgZB/Q9XUzV7DlG+SdnqmOR5szZXdyr8g4eK6Oz+JHBJu+
+          E13RV3cdkx0L50IK8QpFQmft+Mpk4Ow8f+GRMqy73RPDiM42csbuo2NWHCSArdAijvM45sIR3AsY
+          DE6Ily1v0+e8a8HmN5Fy04+gK22vhc87wWTzQ8O6vR7a7vgmDiN93fmXf42yPWMeX7h842USaBet
+          JRYzvlzCfdobuN6bLwm0QNWnVnuVf/7IaypK54B3bvB5wSnS+qGMN94ZQRAm6x+v7m6Xuw0t7qIg
+          LaUNWPV8aMBB73Xie+9zzW88FnzK9b7xvyCetfPowD1XmEiXON6dH9VQwG09Et02tJp2ZVtA4wsa
+          zAtiDhbzERoys5v3eONhOrfOugH9MxuRwsC3Ydr0WsyLWkcKWj7DUul3BhI9h+RwvYQ57uchA7+8
+          fv79fP8OBTjclYQc2HORL5o4sGDzp1seaNx54qQATrca+pMX9fS7+UdQRF1JrutQDlN41Ry497QV
+          HRclGUj/nESY2iVFp6NlD9PaKOmPH2A+OXI1fc5CAzc+g/NQbsHAfdoMKhneE88jg7ut50R648r1
+          o7dkgDnmVB7yN61HBz9YYnrlboYk892C7P5y2J6JKcFfvvVnS/H0uXCuLMyEk4HhLax08j6PEVi7
+          kvVldvWosHOuEKid9yFGKMaA04WLA8vtzPvGA/QFu1UDaQht8szYh04t4jfSz79vPKfmGONWwV99
+          WA/RostTjDEcDwqPHEKPOtXXrgRSXpbI2PmGLqi7mgc//q9JlZvzu8/Nh3eGT4nBrkU9U+bj//y9
+          L5wvrxy3irLNMwTDF7f8uM6Jm0FOiij65bW1KCwRPj7wuf29IZ7u+yIAujWeUS5qC9h4PQMch3H8
+          6quw+XyvninY/CJRTLCPl8PQVfLm54n6sWadSEqWwEk9Dsh+vy76uJP2BqRA95F1SgMw3fdJAOPY
+          XbE4CHm87nYXDIdz45Br6OX5n1+MlUdITmfdzLttngE3nosJyhx33ee7+VdPxB3OYjyWWErhyVSv
+          SDuk5W8+VYHPfI/QNs+oF9YuZmjsmZY4bVQN9CX7Bdz4IHpal2s9/wMAAP//pF3Jsqq8Fn4gBiIg
+          SYZI3wcBFWeiiIKINAmQp7+F+wz/2R3uqi0KSb5uhZUI3zqYTKFN97e5Z1PykTOw5vVEOKfniuzw
+          VoX5GxxxwGCRT/LL8tFtS0LqNv4jmRv1W/7xrRMKJzC++iiDXsW71IHipp+n2etk07ml1HbY848f
+          //IK92w8kulpRwT9fq+66rW//Pa4OCHdH987bw74bw1D9X4h8n4cwfwJhxiKp1uH7xfCVcOBRwSK
+          eN5RtcQZmzKmEHjO02w9ZabrFxjQBQ629qKmKAG25Lp6haBXQ2zG1c1jYtiuOwY+gLpDa3lUUaYa
+          5bsoxNF786mm2/kZ/eUL3uA1bKp2fIke11HFutYc+uW1oTz45R+3Y2n289WqU7BFhwc11zxpWfN4
+          GMT0RQAP1ZwJAS4l/Y5zMn0N0esjo4pA3vgl4bSd37MxfjWgONTyyi9OQtUZE1AyKcBKlt3ZD8+B
+          Z5mb9flvEkYGW5Yv13j5+eueXLtBgeO5irC51l+Wj8uFcm23UrhDS1n9+fGVrwhc///HFxB8yBnb
+          S/AAUyiqEJWdPxHpEb2Txdl4L/Dz/94BNWw6XooIrP4/HKMg8xZKuhj2x8Zd64G4oix/vGDizGU4
+          246R8z//9L0dEembTtH4tpdauOIHPUVckiwsUQXQFJFKpqv2BcNl5hV0BiamRtpsGVmGowqN5CRR
+          dd7rbM2DeKiX9fG3vry/9aUnso+1qdYrQRXc4ucPQolJHljI0YPwts218DMdh2R58vOy9ljSyXN3
+          VUG/aLEK8009Y2PnddU0fbroT1/sc4CSqbYiGy2v8PA3HuyovWwUj9krEOxABL96F4DRXKwV9W2/
+          8I/LAJ9IPVJzV9kaOUtQgWQ8PvHFNEAyj4vuo487qPTQxEsy3JGbwlVvEvawyl9exCPlSAIctODJ
+          lmqvEEiF14DXehCbzeKiA/0e5FSLJi9nqZiE0CZLQtiqZwVVUAu02xYmDZGQaNuVf/7qJwhVdrKg
+          j6n/rf/Tq+ySlrOVBj7Ck06dNT/qZpPBX30Iu8VJB7/6DfyIrhru8lYFK343UD/JEv3Vg8bnQdYh
+          9xhFHGrOs6/Utw6B6sXH9dihwhtf2h3CNU+gTjIYmoCwUv++D/tWqOe8LWcTkLkmDPPuPeZLTJYr
+          3Hz9IOTXPGNSnHsEVv0c/uYjBUwREHX1gZ53wsLo6tcg85oX/q230d6eFOiiyyuUM6Xsp1ZkV/QS
+          3IK6H0tJpnG/9nS3/RZby6Tl/LhH+u9vamk66e8kWxTwvMDPWq/dgLYeRv3/2VEA/ntHAX/86GSC
+          1QfQTG4m9J1FuL4qcc+Xo6A1YIHvLzYkv2Wzr19CWOONTDphPFZLvDm8kG/6Ij0t14p9ZencQDsT
+          ZWrYW76a9dBpYb/J+ZBXjl0yvwvNRr4QzCHfxg6YB/U+yGkS6VhBI5/3he+kgBdFBweXs6WJs1Qr
+          UIu0FvvJBD3WFEYNTc9nVH036x6pz9eG1WOjhEy57HN+xMxFdnApQiQUes/M77UFii6+KN4FqBqU
+          GyUwrMgtFLDGeZP1Nkq4GeUOe+n5qc2bNxGg4J+T8Mi0vceCeSqQsL2oWPHuSr/lG9VEfnuD+My3
+          QT8nbUWgODE3ZPOjzUW+8UvwbjqIrd307BmR2is8+DMX8qZaV0w/KwtajmSh3no66jT2VICdefSx
+          VRqHZH5PJIZmrNzpcaPRav18hhQBGjRakk2/wHLnw0jrVcJAIGjs4lcC8hu7wnvHvvXN92WH8Ks7
+          OQ4Gv/bEVypd0YgTn2I09Xl3OEACuyt/C8sv5ZIpu0YmusddTa394ZJPkWPysFJDP9zoww0I3eto
+          or1i2njPAM3ZWetLeD6mFj6fLjmgS3NfQGQnPPXFwfZmniYtejv7L71uK0ebhtc+Q0dZTaj/qt+9
+          IMOLiq6xLIcoN3dgHiecQcmIrjQjQltNcqal8NtcSuyjq6Cx8BR10L8OFj7q/NMTVPJx4fG8hIS9
+          TKkaVcM5wcNWHInUb/Y9z/D7hYCHMHWHjQV47b2XYWQfeGrBj13NLoh4ePd4mV6Dxql45UYHOJhB
+          QY9Me3qT4XAQDPt6wFd2sPMtR6oSfXB7ovFFBf3MiUiHXE52BMkPSZsMR+BQtnAptnzTYtvn8abD
+          4+2iUBvkKB9vkxRBTm1EbFzQ2aOlN5l/4/34lGkvEBRnsBWyhqaldPS2fVpx0LFbnT7ysvGYq9cy
+          gtPFpNb1rOXb5KTU8Pv4CNS+EFMb9Jvhw8e4OYR392lowrEqb3ALY4M6hz7V2P4cqwhrnU39TV9p
+          C30uNbxazwFj3Tr1c3Kya7QMrxzfshRV8yGwY7TV/DM9HIY7E69Fy6NcPE3h6+Xiitj1k0elV9dU
+          E2ebTUv2ClFnnn2qRI2d8PGoCLDoWUn9W2x4QhPdCigl+3XHTBInSzkUE7xcth72d5dXLz5VdoXj
+          lbZU5a9NMl3tggAqPxt8irw4IRJ8Rejimyecf8uzttBed9GYLHcyh9sTY/dpIQjf1q7ETrLvxVzb
+          Kyg8VG+M89L0xOl+L+DF1080Pz6Dfhs++A7sdiSiwdloElqpLxcmwxtilQo1m63ziYeyeN/g/GQy
+          RsH6TjRK5RP108T1BL9+F/DYOVtqq+kLLBLLBjgPX42GmaP1AgfTDn2DEBO441rArtrH3sGYMbIp
+          IEuWaLy9IO7KDXWY/O7n5b6f0P7ru9S7fp/9Ug6nBZ4S90TQqwz7CcfaCz0+qhnyeIrz+fnZhNCa
+          kUZ2p/AJhPtFUsHx/dSwXd0YY3FwbtC6vvHZvQT54sZ5A/253uCDtuHB3/dfzuWCjYjH/WCL6gt6
+          9uuE9WvD51M5pClqeT2glwek1ZRX7xi5LjpTdf/R13Npu1ZqHqTDxnhqwJzLlAO7qb9hI+lUNm2f
+          mwxSXzSo3ihfbc5bnQcf3J3IFu0qbQ4WV4LEFzxsX6Wln9+7jQDLvl+wMcxlv3Q96SC3O/k0m+1v
+          IoS520H5S93wIxYK46uoiuHZ93Xq3YCa8J++bJDabHbUK4wQDCwTTZiXTwtnTXIF80MOCnjYbkca
+          4reqzfaSlcg+yJTsNseRfdMHVeGtByHZ9l8/4QUGM1CCMsLX0MNAeHCjCn/4Z17SHRiE5yWGT+9d
+          4MOL+es71OMNmnQ+4OyRlwlZxwMpauUQdGN8vriOqkJu+czYqsqO0bN8kJHmYI6GHBcC0Qh1E4q3
+          sKfa8k3BzKToBNbfi1X/bjJ+/05csFzPNXUOZsCm5QkmeCkeiIbfQ+NN9DnZaKPeML2QoM/Jtz8s
+          yB7PLnbq8uDxqbC3kXTTNYwzY6imCX8l5OAPxvsPUMDWLc8ufPcHgvXlfQfjo8gW6MziMRRypezF
+          55T5sK82F8LoEoDt8aHIkCu0GFvgkeYi29cZ2KgFxllbt9pURX0Ez+TmYCX9ADB3jRxCuNUkIp93
+          94q90nUn/ZELcQSrD2vbyJ7gXQzUsDndFzaDvadDekhPNHSkDxtuh4cKoaFy2D2nYk6OVVmgnDsA
+          nCXFnQkHu7KRegl0rAXVLWl9oVRQYOqYpuI4A9reYQpU74bxXTncPYGiT4xsexloljaxNxvLd0J+
+          JbjUBVun33a7+wlO7lWneQmPjMZzxwN/3xhUsT9BLgSO0cJjKR+oEqZWL9Dn0qDKu080LRPCmPIt
+          Y7Qk2YWw8vxmS3NpVHhK7BM9jN8QLGVlhHDanyPsfIDbL6drv4Ds/L3io1kloNXEUUHPAIh/+mB5
+          bPMbFNb2ehBG32rcOIuLfvPJUF9Kvp3S0wka90+Jg/E09cws9yb48e3xhs+MvT75C814P1Ez8pZ8
+          MsV8gth9JzRzPIfxiNktsjevPXl8Sr6aotou5UI5GtiYFl6b47nMfusFY7ur8zkMDiW8R6mH3Y1w
+          15aiKgroHuY9DvH75U3nK+eDJxr2NHVbLqH781WFyAopda48qKbH9qAgnb9GPz2nsah9lb/1TjhW
+          XzXysc4LmFvlgI8Po/emOMY2WA66gk/rfBye1aOV1/lM42ND8q9Iz9FOKe1buPNQ403f6Njt1JGs
+          pwQ9uWRa7+/HRyGzkVptxaRrfvqFpmfX8r5mmb9QKJcGxdzx0PNeXSnIHo8uDrfRXhsWnLVonZ/0
+          LB8/YPr0bQ3T6hhTL2fEG46jf4PgNe0w9vszmPVymMCKD6TM71T79h+sw13ZhGQ+8TMjnmKk8A5E
+          HvvAPuYTQ+oJ5k6L8UU8LWzINqcTDB+BRxaBPPK52V0I3EHlG0o3xQS8KkYuTI/NnQZB5msCF4oC
+          vNWcRpg7HPOVPyaYVucYm4Z8qpYd3LjyV/dyuu8hn0/nGk9wxWN6zxaFiZ/JbaA5FhI1vP4DFo04
+          ypq2S2RnDDuvhWH7AisfkF3QKRWfbmKCZMMesPuQsmoxa2CCuxXaNNBx3dOlOS5A/HwtetWLbU/8
+          UVSBqlgqNerl3VN6V2pULB+XXNbrT+t2ORhnak107A3eLLqHVR9KMrViRHomuPWAEi0bQu4ovtgg
+          ZbcGVnVxwJ5rSWDpXncdbFlwoAGcJcDiuRNg61nDH//zsRBnIHl3Mw52R69nueao8Fl6aOWDlzas
+          9wc9V6D4xyfisDNfP31Nre58YENtkQbs1ewZ1unuqs0wsU2gNYuGnY2caq3uvHSgvG5bqjVfM1lq
+          zn0B8sw6ahiXT0LnRpHgb/7e2jjpRf8BCvjY5WO4i6Nrzh6a3skrHlBjmJWeD5SwBFDJfLzyARPu
+          NpbAyichpI6c09NMblAFdwdbyCeMNNHtBr+1YVHlTjEY+UbVoU7lDVU/4vfH96VMrh8F70/Y65d7
+          5TRwHS9s87VQTTPmJrB7yR/s5+7EluOwKEjHKAm/8zer2JR6MvzpweN8lhOaHdsX/LTiJ+RJ7eaT
+          8DxEsmANT6z9/NfMuwpUBM6ggdjO3nyZPy2MRS6i2tV4eyte3ADa3XhspZ3jdXouL/Bx9toQSGCT
+          TEvWhRA5lR2aWCu8OTkpDQDeBtN9ET00etfuMsw92FJTGiWPGqGvA6WeQ5xcOy8XPq9LAQ9VYYZk
+          FDeA5NstRF+pPlPdP2Zr03ve/+kT7O0rsWKW/oyhJxUqjb91nQyeaLqox5mMzU/9BIv6vjcwPdb3
+          dYdF3S8NmuO/8cHD95YMq7+B1qb7Yq8FE5ivcmLC4BEesK82T4/ZXyeDP37A5KhpxFOCE7iU7XY9
+          RQjkI70nKnyO24o6j/M3mVb8WburN1hBdujxtyEjkDubFlXrW5SwTzMr8AWuL+o+JKl/f6a7AtlH
+          rlZ+POdzfGqGPz0bZuQLvrz09qE7k5GAfta1thoeHDjecgW7qz/jnfxF4GVrGxjjp6fN231EkN+x
+          I3WfygDGpUlSOCWlhR8XfpOz4hhLcOX7sJ4ONZjPczyh5PYi4dbepv24Z34s795YIfLmIICRpk7x
+          ywuo9QJ1NQEDmbAGEU8zmhMwfKN7C6sHUujpKeXaqv8HuOoj7ERPq1r5XOWizI5/fp9NWfXs/vSH
+          fr/nCQtBIcPvcyFYuxqGJnBGE4PtVVGoHze0n+H+AYEwTRHOUTFqTP9WMiRZ/MH+tN3nP/0Dtp0V
+          kvF42fRTMh4JiOXoG0p3eunZ7OEQnsXax0q0V735LNcpOOz9PX6U34AtpRW9kJJ/PexN50Jj0/14
+          g5tR6uh5fQNmhLdLC2+cbmDn9jx5lEseGXj0rYQPW03rtx54D2gu5JDawraq2MfbnGB/8muaOdOk
+          kXSjxMhPNgM1nOsVsPeU8+BDaxVH4S4AP33+0xehQG63/CsmXQ0uLxTivctXyR+epmdZIIhfEBu3
+          +2hAVMJByEEbVENylRuYyjbAZq9Qj+UFGoATRAI9BtkrWR7D2mMjVPtwEZydN/eqzUMniAWqhLuA
+          zboumQjWsCOCNEprzyn/Jlf5KcT4GL4rFq0VyMPhgLHLcy82EJO7wb16fYaEu7QVpXe7/uUP2O2c
+          NB+F/buB9kGi9LycBrA8rXcGvSf9hOJyGtiUX5MBNmbTYjVQR20JDzsT3EMnoekPjz59WSO+01Pq
+          fE51svozAu8ph7C3+rmlN9sMPOOLTY8GF/Tzu/BsqKhPJ4SHm6axSu1sKL/1DPulxIF5TK+DdCaF
+          g82ahNUcBpcX0KJ9S9i+CpjYezdBVhCIsNXBzOOvVXaFqz+ijnaVkhHeDh3SvdNz7WlzrqZV38Gs
+          ZHccDnagsf0XufBLixTvqyzRGEutEgoizbEVRGk+bZ/iVd4O+QdbJlK8bXV0eSBOs4s96bTzFh5T
+          DtQYyQSWUsGYtxQ1NGfo0Ds6bbT17KbrH17sZwNqs/CeePTgF0C4Z9xrv3wBRs/oSP1aMBOhiFj6
+          82PUTPbvatl05h++hYBrd/1ij+u50EAG4eJOfs5Oc1NALegUfN98Bu+rQAX++SU1dp8aFSSnhN3j
+          HlLPNI1cuNh6DE28feNrsNdy3o9vE/TbAlJH3S7aJBulj3Q+i/C+C/bVjIHnw3efEIwfz13/BQnx
+          4VV8ZNS+SnE1uuXDlc8fSqjuapK2fKzHIj8bKcNpP1b5MuEhBY+z02L97FoaM873Dsah/KRhszGS
+          5ZeH7XZDRNWouOYzpYKNngJqsBl5cU5//Ofh+zNEw+zmC//MTLi9pRdsDqPTb69yboLhrlnYQ0Wg
+          1QvOOvi+v3Y4GIHA5mI31dBK/ZC6enHsWRw8avA0uYH8rjccq0WW+7tDaVjktdfdK6cG78XIsJ3f
+          JbYUEUhheG0u1C2Ir7ENeXPwJGvB2jNE7Dse+RPEyXjCl4thg+1rB0MY2pcXDVY8G5nLSuSdEo8a
+          rLUSQX/7KfjpdYcsh0oI65sPx6428Un3KjafTr4Pu+yl0Z8/HHxH82FBjB29tyGsGG2uLvD3tUH3
+          dflJqOyqIbqQU45N3dMYI/tFQXVsL1Srs13CVEObYNCmKU2nQ83Y0twnwMiW4JUPveXBvVVkz+uO
+          hiV5VFM49iew6kH64396CNQBJbMVEOT1h4qH3dGHs5txWGv00pt8oVUhp8RFyLl3oo1GWJTgSs5H
+          vEf62dua5z2E0ts3qce1l2p0ThOBi1YB0vdzrQ1V1McQhElL9es+9HgpSxuIPgkK2bcM8q3OLAny
+          1XFH3ofLkf3ySvDLt/14/80nrtM4eJSVhF7q8qA973Ngw/YeV1gZR4Mt78vBRbnTYWxfSKMtytsm
+          8O6Uz/V5VWy+zvsS/vREZExNPuXWVCNjM52xTfOQ8fnBtSGfxD0N+ECsuppTX3Dyhzs1i1sA+PvV
+          LSDMhQ/WXj0HWj9OJ/CJ9xyZTDkG49XnIrgRlZGGCi0TYX4+UzhcB0IzX43AhDKTA4ioGXnf8BmM
+          q1+RJgJf1JYPE/gO6nEAVAoCfGCAJl1YpyE0Tt0xbLBWaCQbUAcrVCdrHgzAen8DzLkEEMgfSDW+
+          vVYFlSlw1ELz0Zs/91n+6cNQuiZDzvbnqwK0Ptv+9Ic3hfVlkTkdRL+evNWUHpAEH3EpEN7yZ486
+          3yKC4KjqNBjmLmcsxS/Zk24q9ov9GZCbBRRQo0+AjdPzy6adaaZALVXyV29g0t6R4dF4n7FmRa22
+          bI0pQtEzPhIWm533/tU3yuVxC6mDu2rVyzV0725IreSVAv4eTTKU32ZGVaOVktn7qjWYCymk6drA
+          mEFH4oB2H0MyeXelEqZsK8O5O3WEtScFtMPLucJH30mhWHZitbzXngtHLeJoSLgyYc/ddwI0OzyJ
+          sGxcjw9zJUQvrxCoUZQ1+/lx6BwWK4yUdklmDfoN/NJbSo0gU3N+QJwue4f2Qv/ur0WlCuHlFVLX
+          pK9qFuk5Bmv9AwdrfYRf8yNZW/xHOE/lESwc83QgA2lP4/F9ZuzY2iryDt2FOvjCktdl+9VhiYQv
+          tvh2rEj7Iik4vy9HHNZO7C2eAAaQvvs+hP04avRx/Cxw8smdTC+xAovQRRla/S/VoXDpmZhk7l+e
+          arThApYt3gsgW2D673lGjwjCyNjk4VZ4hAk7yxcJBifvTqYN90y+N39U5bV+RbVGVzwhvzoR5AKo
+          47wFEZslsnMhH7IrEcGDT+aTv7Rw1XfY8vtQm2xnJDAWYYRjT84AYea+RPHeB3/5JdUhVsCAoxjb
+          6zsfi3rzFRDepIX88HdJqt0AuJqHdD9+CZgmMzvBX15qq6nKtvvy6SLFcgC2Dkz1fnkMdPa7bzga
+          nJhPyeGpoN96AO9H1S9zJ8sw0r4qvgh+kywo7Di4t7SE+pXTs1FsAgI4tRZXPf7Il9mkHTydmU1/
+          +maiT8kF9qbcU80djsnQREQC98uHUledgfeXD9RCq+IbjJxqS6SFh+lZEqgbepSxIHj7iA7LTFWB
+          PBIyoKqD59shpsmuqBJ2PJ3lP/zAuiX0f89zL8ANxuV3ZJMTH2zYV+hCzcGu1lNbtwUMunakChrT
+          nFW76CXfd8883BD5m5DSyl4wlV0Qzmqkalvuplx/+S2143ROJivuMghrrqM2Eex+ETPXhr6AZ/zj
+          60nLpRPEhwyFwlHyK/7LlzI4fsQb4YHz8Ob3ThTgpw63GLdtC5bjIP+rJ+rLG7H5l3c+RCPF18Yu
+          2Sw2xgDj6wmRTW3k3vTaph3iSt/F2vO5JAPAYQePV/2DrQUknpgXWwKFcrdgY82PxV71YtAnkkeP
+          xsVKpp0ZnsDqb/APTz+P6XlDa32NOoLf5DPfuCYoauFI1fdeqJYB9S1MCtEiUnVL2Bi1pY6e3qfA
+          xoG7aSJPTyb8zSd7xc8BsIsEaypk1LqHXcI4cjBRAzyTgNgo88HKAQeld2iu+ppPFs5oIjgd+XLd
+          AV6u44NuEKXSif7wnW5uKo9Ot4uM1/nvzTbo9F/evObBSr8ddr4OQzZ3dL/VtGq8TVOEVr+Fkw9b
+          2CIbXQvNWL3TELye1RLmavvLa8Ot5NuA+aenBDUn4Ci+8Jt/8zsHmxd2TuGTsccOKJDNSkODOvh6
+          RCHxekoy+BJZnXONn8zohOoTqwkPhUslnmc7haO22xCOcEryw3O4eVV3jNf6GlnzBsmk7EDxPVAq
+          +tmlNvrK8oS1D1vAbOuA/Op/RJTl0FtuFlMheBsxDa3G8oaserY/PRxK8e7c/+Wnr0k7htNajx3q
+          fsdDp/YIdU6s1SbLyXggHB+nEBZ7ESxr3gxiDuTh9HUlb1rrkzAI9lesw0OSTIeDHEFjt4ehtNYP
+          +rXeDC9H/0JzGw1JFZ6yFhpBGeODY9+qhT7lGsLCUbA1+0oyS2R20aXstqvf0fNl93QLuPqRcHvg
+          oMeSlthAfKeEaph3EnYtSgHy29dEteGc5nOj9un/s6MA/veOgsNN57BZ6mO1dJfmBVMjkilW8m8y
+          829gwuF9U2hMLFoth9upg7kAH9jDb6Gaku2lQOasM1rsfKvfVs55gezAT/T6OksJM6RDBMv2WGGD
+          vV/VMvqHBp7u+h5r9+cOTLrMBtQ9hYjI/GzmW/4dXeVvMvVYvzimNh2/a5cJ0eyw8Ul6b9LO04KM
+          a1zQ/ZIHYIqFgodmag9hflqCXjzqTxlS8xCE1avG/byjMASULA3Vjs6aID55GU2D4GErSE/9fDHf
+          LUC17lMb5es7O2w9D6h083DT1pY2f/wphovWltQGBz8hAxhMmIPcD1sLnHsq6ZUM787nRpgDPU98
+          0UcJj4MR4hPH27lwcB0XkqOmh9Ndn9cmQIYMH4cKU9M4oWQZgoIDypbzKY7FPROXb9wgKeYKmp73
+          uGey40zIldKEZtjd5ZMk31QoXGUFa6doTJY2KRvEDoNBi03G8oE6t0k6+PgbQu79yoV32KroBfiW
+          HuPerBY+33JwuVVyKG9lmEzPia8RivYlNV7RNhlFSG6wysAVm6fxmAtBpXKIWDDG97xKAO3XFldj
+          r3hYFe63floe4w0ojS5SZ2fo/bA83gXUfT2kqS+GvRDwaIKisa9p/NxmnujOboyyLxXX8W28pcoh
+          hGZ6OFPr0hCN2eFLgUWZmGRh/qQxOOgERaerjZUHMnPh2l0h0L+HHu/faq8tb/zOoBxwDg2Lm9+L
+          /F1T0UgLiXocErxlP5QZzNBS0WLeTD0tv4caZdx+RyPrck7a1rNLeKbbL320LezZBcU8eF7ue6yA
+          ZQ/4bBPX6MGlHc31bAPmT6FcYdmeK+qyKffoIncEKY49Yf9jc9U6XyIk9yqgCi27ZIjU5oZqRw7o
+          /s6++RyyUIVW03I42bctEOds7UrP3QyaLFbS8/xeuSL9m/TUVdImn8cdPEH+HEVkfque9q0vXSl3
+          QoTojc+fv/UVQ27kLqSyRbcSmLRwcGRdRPcZacFU1W8VpeOU0OszfCVkNtQaDrtJpSr83Nni0bhE
+          3vE54QNHI42/sjdBjyOCVJe4Z8Wf73cFdk8IsCHzdU71T8UjeFI+NMw9Rdui892E4niNqXqbztX2
+          wkc6qvRUpY4r5J7Ik8RF3jG+ht9ZCnpBDJ8K2vevM3Yf+tMTs/wew0+kRNhKKymZDO/pw2AMfXxw
+          48hb1MjU4VO42PicVlI+39xWQPjWvUl0ZO9k8qR3jO73zg7nFQ8EqgwDvFDnjAPHiHqxGzSC3rd+
+          GxLJ+nhCPDYLdLr8QoOGKNrUesoLHST5iq0GNRrLrnLzN788C38rcslKG83xd6S4uRqA39ZElQ9R
+          dML6QFLGW9F3ks1U2ePgWrCcKUQWoAtBT73kufeEqdIh3OkPRKDpi2zqor2J1vVJrenK9UvwKiMk
+          KvQV8hv7VU2e5DZwdvctWcDr7k2PnlfRik/4or1Jv+ia8IK/+5Ge9ScXvoIioL4n91AORkfjzQ0f
+          o2XrbULop2HOHhdTgEYhCvjMn77eErzaGF6qEw3nU/3pmRCxAmXLGYbjrU173n7OEqJ7oaRuCjtA
+          VzxFihyHhC+zDxikryfBTzttcABzI59waboyuNkXnHTSAUzCNrlBAcY8NWRez4kTpBxAbV+T2Y3e
+          OVGRVMJN0kfY2itpMsaiE6JrOlZY+Z6++ezu5BP0q8+ROpc4YeJlUMlvftCCO7TacjriDnYX36Oh
+          1TVs/b4CFaWd0+v4qnty0lIerHiI9Y2t9uv985D4cUr9MXQTQZ9bCBkLarz3LiCfn/ojhGNo6lSX
+          LbEiX9H0ZVJLF3zIpqTf3nbgBOOj8KT7Aavacru3g0zBNsOFqDYJCy4Gj24TP+I7qkHCXndXhbsb
+          OmFNnTlAbvd2PeEPuH/za1R7QUXe8hApXvFmbvfeFVIyNfQSKLd8fZl5+P2N7608eO2MDBcqOtGo
+          +pxf/dRxk4TcOvlg821V+XJ+XzPICaNDi4O1yZfN3VZ//ITDSTLy7Xk/6pBx4ElEG580JjxbCZlZ
+          QbC9zp+54A4hdOvDBxt3L2Wkv3gCqPt8oEa6tcAfn+3en5w65fPYi49sX6CLajYhbzSqJwbb0obc
+          K7rhW7X9JNPmLenoeGY+dp/R6C2WNRO0e79zunfTKxv6J5DBAR0tXDh86zH1pdTotz5qo/UZL57Z
+          hDhi6TT0lVSbwlBQ4Z4/WHgvFlW/CIMyIOMlV/g2xg/Gu5/1HY9C2eHIOg/9jM7ihH78c1I105s+
+          MhqgduoIDkb0BOLp47rw2+KMxsS55tPldaxRvbllOPsQOxe/V1bD/Ya/0MxcnpWgjO9FHrjjleL+
+          efCEoHq3UDuVGj1u2S6f78QsUMG/bXqfrEpbBlCbv/ElQp9gbcaL0qH1aFlS40nzhMp1bTjcYIMP
+          3Gjk4kY93mQ51Hl84+++NvSsGVDXXFoyi4VWibSGA5QmLSJc2a5dhHfcCW6S8o2dYVbzbbsnOoBd
+          +sWKBCRvOiSJhOrNR6OFw9vaUg/dFTVdTWmWmBVYxuKjIzJvPOqy4V0xdd0RVSl2Tw+FnFakz54u
+          /Ca7E9ZE18knmuMaWqFTh/PheWA0f0QFSB36IVvOAhXJgX1FgT6wsDSjZz5N10ECxc0iv/WfMFc7
+          C7AzvAe1tJ3NhKLy4B/eX87uPSFfoSoRr5k7auKCVYy/lSn6tMsGK/WhBYsYfhV4TB4uDU/u0ZuM
+          e6pAw614Ij1rK18CfjvBzeB+sWpwM5ua77uWL+aXC5mEXU3oLDDI/uMQU1cGVrKl+QyhJQtH6r7O
+          Uk7Vl90g9TaV9BLvrVzILHmAH+vqkfR9t5Nl5UvEq1xIbRuqYFuWuxL6jxyRuu8LMP+uF4LrDZtR
+          leYzS5UILtpOIILUfXLC+XENf3zul9y9orS+KrDZwx6fH6ROvs45NtGqH7GxpRoQbXQQkKg0Irbk
+          Fw9oYQkEXtsJU7NPqDc5ehxCbq8mVPe3FZsEs+pgtFF9bOmc088arAoQ6IRR43v38yXhUl92HOVN
+          aq3o+4UD7lU+2VJHL5fI0bbQpB24N1UVSuAw5IOSXGXZhybB7lYwAcuG8QrCeMmofoBYE/U2fEHD
+          gLdw1309NrGiNcHKN6Q9xpQt53d8hckwl/QWl7LHLlCyARUjhfqrXplX/QJrfj+Hy3MreXS7sWu4
+          lCJa9egtnxKLquD7uLUhHOMNaH980wRtS/e9+8yH71guYNWz2OSEZP18FcNpgAJe+dwTo8K8wuOn
+          e+I9N75zariNAEXl5uMsMTUgonBUgLcPCPXCwktmq3p26Mff0VUy2HLpUAvCeMpWP/Pp26Ytfeh/
+          dmey8ktViYbZgO4SenhfSF/2W09w8hqVarbYVeOJLuGf/v3Nt9nOzgV8uc0D6+445hN6XkP5mhYB
+          xf5Ygvl5/KSQRM2Z+lnzBhMrShPG5BMR+eb5lfBgc4ZYYy8Yu283markJoGdHn6w7wUqmyuxlkH+
+          6u4UN0WeUOl0InCkn4rqiXbNZ/3hTTB72w+8vn/LZnTeTECPXhD/zacr81a9fTCJcJeeFbOGIYPV
+          s3pQP77V1VRvrwXcZa8vXecjY2fXcOGq/6l9PRq9MPN3E7JTcg+Fu7Tv+fHzsMGwk8dQNI0UzLa5
+          5354R1f8ZbNZf1T4OFxrqhrcATBDAhyUndYiTDBaMG/t/PqHv/GUgH4amK5AWulfmm7Vs8bqCnNw
+          H0sJkWXwyZmN/QHm4OLjw4/PpUe4wKWsP+HMXjut1eV1B9M8G/SxTDGYTndxgBv8kEJm747JMpuY
+          g+1FutM830ls9Efb/OlJ7IW3sprSzoqhmbrDqufdZAl4tACh6ibs5vap//k3ue8RC6/TFoBlZksm
+          87x/JmLxrcDMv6NMfgRHkdpKW1bj2O5fcHammFq+p/TsCfwrCMaTGEr7+VSNPjFuUD3oFQ4mZwST
+          72yLP7zbFec0mXI9D6HwOI0EgaussUJoSnl401O4Aaqbs13zLaAnzOupYwLw5sv2fgNyfBXDnd7W
+          1Uw8vYQseI2ETzQ5YdKpGOTdje+psZ1Jvzw/TwGe1UUKa6N5ecPJ1+U/fPWO46gN1kG/odRwMUn9
+          3SVfFFdxd7LTWX9+jh1PNwILiw3U0Tua0DCrXJjZZ406TS9U3601EbT6c3pSjQmQn3+jYnWi+12r
+          s4lIegNO78HEx9VvsMm6qyjez1tqrfqFj+ZTBK2mTP785/hVygGUJqT4aF6wxvhxKpDGbi412tPe
+          E6S5h3DhpS9OiPDxis66qHD1HzRAdZ5P1ubQgPUFsfBbXDYe0z97Fzw6taWBxxkef2uyCQ42kNfr
+          Pb2/5zdhSccRWPbsh5+wNagZEsr0fq7tOAO6b4bUGeZXMuuPpgAtObrYzEmZTFN/PcH7gfOws3Wu
+          HvscmYK4V26FsqSjahmCE4cy+6iFXLdcqtk5X03YXVKBiJt317On/BygVag7jG+pnA87z3mBlzu6
+          GPOSk7OK8DyqRnPGuLm+2RhwsgwPUTXQH14yRRkHuGy1dPVvpcbEZaeAQloeOOj2Yj7rE9cA8WJG
+          Id/3HPiAgV7BojEPm6UpJXP+NSSw+QgfrE5JXi2vXdohlwo7MpyfhjZpnRnDMrdN7MobyJbf8+al
+          1wVrs5pok3zjU+h0lwvVKPhU8246SWh7/HihHC/favpuhBj+/OPxFR2TRUCg/PPLx/b01JiNP+sp
+          hbaFD0aAq1mDfQFTw8Y0A3jfiwqReRixFNObn5Kc2uFLhZ08dvjmn3Z/eRJa9fvao2BKmKtu7b/n
+          +8yIDebdVEjwcdzAcLdsJLBwQM2gYVoW9tMO5PSC1lMinp0cCvN9SiblbQxwLvwNme/Mybey8iXQ
+          AK8PDag+ekSNTBOaQ2dQy9Rcb341mwJEgUywtuoTJulOCE+yzbBiRmtPiFyCEKXnF9Yifb/6l1CQ
+          1/WG8dUO86k5RybSaLTgvaaRfL7wkQmrbKHYXvl3lh7mhF5TcqLB9Wb3DELZBsuZe9H19yfTL48o
+          1Zcayqv++8O/0stvq14PNGGnHgpgDstEOC6JwXhokkmeG1HHvsbtPf5MLQG6lIRY90XSj8KleoGG
+          mQ5OZpVpsyIZ9Z8/fa7js1BlILv3rXLJCw1cPr32LxO6cNdjtSQ+EC3u7KMEli5Vv/qFTVc2Dj9+
+          wJ5I0rybDbeGH6lNw83Dq7VJiSV/NxfHDof34zNZttepgcN7PFH3YXpAkOuZwJUPsfZ61WAaP2db
+          fr89HP7ws+pRLoBB1jKqP5KPN6KTfIXnc/DGfrHM/VCGaYrsM7DwPm6RN9fXIPzpxbAXvgwQbx9k
+          u9IUdIoD4QT4/nVq4HOG+e/+vWXNB2W8K/bhdkvXCnsy1PDOyQeq69WQLx+VuZCa7ivcXPaVtjxb
+          vYS3raLjvX875OQKjQlywBGw7Ykv7bfekWq/HeoviQsYRKkA2+X1ovh5kdlI3CePVv7Bx32sge26
+          fiEkXU/DjtQatUZLR7UjBWTpgVZtN57xgm5eivhwqNZTG92Rh1yBKbalU6Ct+v4GD1G+D9mstPl0
+          OQQ3aJiGRaOU6GAuuEu4mxtEycYfSzbZz1lGq3/AJ2mItfkRkQmWtfYki7e3qu3FfHfQcmqfXm/O
+          kxFd414A7sUI+8tB8Vh8SF30WOYMW5cm9Jb+ogl//ioOPBtspSCB6H5fGPXl75nN35jz4WY67Ym4
+          +RzzuQmuHNw9bYLxgfc1uvs0AkgGVhJ2Cl/92rFbgvZgFjhQaNUvTYleIAoWDXv5VkvIpgiuUKOu
+          RfEtvebT3ckXEIy+T1dxAGZk+hBYcQpxyDszW46L30FzaA2avh/Mmx/4eYW3oT9S/bynFVNffQwS
+          LxswJsu9GtZ8BD7HNiO7UZaSScxOBWyN0aR4/H6qCQdSu9NnZmCrVRuwvPGYgYbpDr1+Mi3Z/vTT
+          xawS6mSeXS1nkUrgHKYuNlf+nPa1kkJ2QDn1s3aqyGt366B2hVPYSpkImLr2k/x4mypcrnVf8W+z
+          63a/358cgnM/RlokwHe0kbDPvdWc5zS5BAl8uVT9+WMc5ifZr24dPn7vQzJWEm5kgl53MkuVnm9B
+          fHuB79RY2D9cwn94OTdbnUirf15Wvwd++aC16sFvsTk1MKkkE19+/GejCw/Eix5hb65ZQs9SHME6
+          bhHVwo3CRFe93iCaky+Z3/qWMZHlK96eO7r3r00/v/t7Cle+xGu+mhC3cK+yIZ4jvM8WLpnBYeZ3
+          yxm+8L6Q035a9Qu8jucPdYPx67HGCGq4wecgpL/x31uNCjM0VXRfITcRC0HjoWHihSDhoHtCcWA8
+          xE8rocpd/WjMr+Xhh+9hRlDnEVP0/T99oP7WG/e1ZLjmq9RZENHoZh/pUD0E4Z9fXRzoL9AwuBs2
+          3bOd/OUTv8//9KGQ+1Ej3znpsPKl1c/L4rZAkbMdDrmD/VtvC0RtFmKr2USAKSSz4Xv3OGBXOc3a
+          RJ10gq6lu9R04g+Y89YuwSM4i1hn6JRM2rdaT/WabthAKvDYn177Fk/C1XXl/fQ6+khdSt11Pi4I
+          jDrsw+mEz+b/AAAA//+kXcnagjgQfCAPIiJpjuyyB8EFb4CIoIiyJJCnnw//Oc5tjl4Qkk51VXXS
+          MY6IfbuPDIfwqgTTpZvRiM9hD4Oa7gNqpVXNpPLOw8I/ycJ/sqE/7xJ0otmBmPoB0GAoT1OCmdew
+          mc0nxJb6ANLD9EDMnD00qmzCQrqm/Rd762/FCMouIfhyOuFAn6J46IxHhKj7VMlNnoJsagnyfn4f
+          lv27pm13Ar9C9tfSSL4ehZh+RjeHq/mIsZHrdk3R8JghQ9ED26RX6vFrnRPpZp/NkdNKPhsWPYK6
+          rr8Rz/UrNEUWZ6LFnyAYmXc0r7xNiLw37cef/0yqYp/snoGH8T669DHFQXyWFv8pgHIl1e35th1/
+          8xOsd5KNtj/9om3FLVYMLGZsSpwCwnPJ8K0VHnG/+PEgPU0v2Gq7li16IpV+ftHpfZAz5qhRIdGe
+          c8kvf/efsXFg8Q/GVxVusp+/Iy6XvpMArz4ui6tn+MefovtbibfGrvii/fqgYl+4p2zuPlMo7Zvv
+          iigvuYxHIqsqREp5JzfqCvUg+HyLEq9KRwrbqmOrpWfky8JWQBc/vlvw95fvggs73dhwDz1dPOTm
+          Kqg++g7Nwn2yoE0VCxsP++tOqLZXoKmBgpXFbyW32BLhStzLyJ87p+a0jyz85Ud7EJOYKQeJQ4Na
+          zdju2aANv3guMj7B3hx/2STKn15g9ulM7IsjZVQXrzL89NoSD+5Sv0mkX/4zyeEZd0cGKeDkWOOL
+          55Y1E06gAvM/Cfa4TMk2nygrRe9uvX96sJtoaJjoxz+M43bu5jpLI2Ti9TEgC97TbVLkYoe1nvj8
+          4alN/FXmxOSTP4h33+Wog66SpVp2upFJR6fuH/dnIcVu2o/gRXX30wPQHDh53Hr+0vUevxNp8bOw
+          L1VhzbuKn6KD53+wruUlm95FbaIfX9gf7yrjmss8w5Y/DERFZZT135UggnWZc6xcBqme2e7Wwt7e
+          Vtjn45HN68JIpCX/EAPtnogU/FjCUp8LpssgdUQ/xz0s/AEb5uWZTffHShQXPR/sdnucTQ0MOdx8
+          VGB78ZNYFzxlacFzYn/fikuvoaxLWXEzyIL32ixnIsCE1Cygd6mJqTkqBfSHxxr74THKiPcMZfTl
+          rYgYbq4j9nB3ImyvHhr58PGpeU+WVWmpD4x1aHrxxHfnI9DBeARNsuVrsuAN+sXL/e7qLiXZvgF3
+          VGxyftmBO//8lJVxk3G4aWU26XdtRvsqOATPprLY8NCu3O6w/x4INt8brf/Vly1urP/Glw1qnqPH
+          53heml3Z8ezJsgzMDjpiLPpz/jzpiDzzE2BlSASXrvRbIKaDt3hn3tHdrD5YhLm/YmzsC1Iv+JIA
+          f78nWAvWJfrD34Zd7kTzxTJj2ONDsFdPuviBRzY1qNXhKz7XGHOwYkOxQiIIVAlJsO+nuv/5Ay9t
+          Z+K9azkdp+7dI+ysaU38rGaLLF8XaLc5G8Re6nHMnnH7f3YUSP+9o4BySUq0A1Q1e/pBD3e/XI2c
+          s/rUxF3rK8huqCf6vY+0NhaGAvhHdAvoZWg0JkqfRjpFzoEE1HfRHBQ2oL6wCJGf2cimq0R1uPNa
+          FKxdptS8Sj4p7OQ2x4rwGtxRyD4JuFitBya8XJe/7vIWudZpuYEkkl0Cb6cEV+kY9vCOuHNzoo2U
+          uopDNNPed/z8qZ+SNJsfYh/1RzeEWGtgnT4g4NterenOujUgtucM669W1fiPep6lqtc/GLtrvmOb
+          YveFL5/ThTDO8XQLpxUU13ozcuXu003x5M3oWz8CcskavaYvtjuj8z1XRpZeD9n8NncBvB/+MK7F
+          QUMbJ2o5iAbg8KU+pmgjerUq9fYlIvLqPHYtbz2eUFrAkX2TSHE/VVcBtJizg0manzH1tKoAr6tk
+          cmEJqT/X4V1KeqnEJLOQkM0vtaLQr6XvWM1GxDb3k5VK3/DdE1Vzad1X9j4SW/zVsUVohTZYEApp
+          ZaguCfTNl/VXI+QlSRUl7F4aud6OXFYClYiC999b5U7+FUqYEiPER2dld7wTGCl4u6+B3VLDHTu6
+          /REyMbexYskcmux1tJL0m0GJewyPNT2l4wifqY/JVbUfHS+HrxzSXp/Iaf3pMv5DU1N66x7GRX3X
+          3E3szkdgk1CQ60uEerqf5ERqeN0hlm4m8UbAXCQ175WB4/L0yLaPy/oLjodO48DMO6M+Tp6QwVkf
+          Nw1fuNyAExPu0TEnkSva7pC9rzK8YaxIHgRXdzIdF4C92iM5+VfJJeZJ0KX4k3zJAfZBRt7rJIeO
+          zyiWleO7m7oWJfAO8yu5v/cDokF046Djr5ToctJoY1x5qeTtWgOfntqp5hlMiVStZZdcxs9FY+f0
+          3kPy8s1xl6+xy5oVjtArDE0cn+SXuzHfqxHWaQ3EYrBF004GETo39In7GI2OVWVboRA6cVw78z2e
+          9T6x0C2ybJKsD5w7xXfZkTbpZQrWSzwPT0bP0u55irByKgw0S95ZlXzFnIkarLquV6/JEc11bWK/
+          XSGNJkAEmNTyjs8vR4m5FhmqtNs5hBjvbRlzp9CmEGnPBv/Ga3yRXAB2KfG4y6yadZvHSoTZXx2J
+          d+kdbR60DOAVGFtiBarlbpX7nYdxnX6w5axeHUuK0QLnY9+x+o0uiLd63ZHwecsTOcsgJtOnegKg
+          IsR3lynd9lTMFqrq+oyXeMhYF0Vn6ex8n+McLj0K1vm1gefr/cT6JRMyKh61Xorh6BGFd4uMHriL
+          KhH0YgE8OOLydfko0RI/wc7wdW0OtLwAv3KuQXow23pyopKXGPcYcFjENqNXI+GguD42xAxUx+VT
+          TKh4s+oAW7f6GnOe9i0gnT41Vkzb7ja0l3i4bCKTpJ0vZbO+yinq+Y9GPC2X4w36nmZp67ZdsEOf
+          fUxfV42DlVQWxEeuXvP9dpjh9h5FHMjmHI9gJ7M0he0TpwBtPKIkcaBm84fo4eXsbrfDoMJbX9fE
+          +tZuzQmzWkjqYURYf77XbLaezwK2w73B18ozs7ldHwpJR6WFr6PLZyxXaCMFOz4hXiQdOq4IRYBs
+          JcWk0DdfxExWLT1Q+hp7kpVl9GxVOtzwWsT2Z63EbAYhhUdUEey3x31GzUPfgh3wIdGuXuqOVat4
+          0E43mThO+USzflWP0uX2MIklLychok9nwuoQnbDVKU7G0XM9S68xcUiSOL7GHxeuVYuFQuyKT93x
+          9ZYCUAm3HqGgXs2vu8NKOhninpwPSyF83AsJPGWVYXlnX+LJufkUxTerDE73pqknrvIEWMu5Quz2
+          2GU0et2/v/8ne2ECbVJWnxlJAlXwPTtGaPPZijOMs7YbYcWmeK47s0DhQB/4Ngxxt8SXLA055vA+
+          CYWacner3V1XTCZqVPZo+qgXCvmTS7APz4M2dC1LJTauEJFzZ+POT55r0KodFBI2VOj6WxK1MBSJ
+          iiM3/8bs6Zs9SMKsECfrPZflseTBRmdygDYrHNObNOiAV88NCbMsz2gkG4EUid5qRPJ31bE4rXtw
+          CL/Fhn+9aYw4TSuV/eeLlcuJoumx+kQwYx5jzzebuiWFVsBaM0ISTKdVxhpWHeHKmmLcvka741hT
+          nqX7cosD1kZALK24JySdUOAwb2PG7rYeSubh7AeMHg/ZdJmVlXTJOhfjR1ihzxxfS+BNIcDHcmfX
+          fd9yiZS7+DbyG3uKWVRpiXR5WwpJobC7wV6nKzAq5xHc5/zrMq1uQljwDOeyeYx5nd/o0oI3GB/D
+          sGMpHulv/RCT5/YZdzmG3t/3ns9T1W06s/HAKzWX5LTrO+qv9yB96zrAxtrtOypll3QpEbNla1Jb
+          z0erNeF7yuxxdnMn20xCDfBy85Z4vml221d5rKS3X/HENfzKHU3kVrARZjvYtrOnTVdZSsExig/W
+          I7XXPpmhcVKzTiZ8X8ZvEi/7Fbjne4iDAtYdO30uuhS+9/YoBv2OTW/kyWBuNG3cvUKLDfKMvpCu
+          1Ri73JS4pOZZA0JQ+wteRBnFhQawdb8dyUaXi8n5khaS8/yeiGHaRs19HmUviT54RFHSg8Y0HnQw
+          0ZgRbTNI9YDuAg/LkSyM9cPSZfBEdZB29Yns0ceOuyKsAW4PrcLOwbRq2tLSlCbDHbHDios75zEf
+          oMKij1Ec3p67jfY3EWgqu+Pasdx6U4kBB5vAeONlfDtqIaOE2rdCcmH8GP/wEuxuO+Il/9W0PYlP
+          OAS8hc1jOyDWbXsKjmUIRAZ/x8bPzQuBI16AlVWeaJPr2pY4vXJleb5Vc0VXFlBjdh9XkURqWq2+
+          EVr4EQmMe9FNgnVsJN0758TDWZjxLO09ML2KYGefqzF/e+uy1DH6IOlk+xlXnpRE4iIwyL58zBkd
+          X7UF6zmCgLnLtZFx03LozlING/c5q4fNwGQI720euNHW0rpXeSwRH1sJvi14v/mo9xkt84Gv8lLx
+          P8MgQi/LHTZuiHbMenYFIlcqjKv5LHf8zT5E6C3debL/8dHYRWe0e4zaX/6dBMyFyMv8gcSN12ej
+          dO84eJkfl+ynYddRjdQJOmp0F8wXbMSbzzot4H1lMnGVrkTz+3B2/vjdvj/62XRPjC90r29OvCxw
+          47/1f2Gpi83Vu4qn7azkAGQ5A4ZPJtt0oHjwaY2ExOVJyTi91UZgw17GeBimesDaLYcHj19ETeM9
+          ouF0oFKJWw8bsyvGva5OEWztmY5sN1cZvSVpC9nX5rF91JV6YsHahF0vvLC6rmvEdt5DhlWeb4Pl
+          MqqM8rFZIf/zNrA2rd/Z/NhaDdzKMMBHmdfR1nlEAaqUMSPGa3TrWQ6HHPXRysfXwDrU7Mc/tDtb
+          E/8x1h2rrjIPVL4esBIIqjumd3qE5X2DrZjPGvXPQrj744OPWoo/16wSIElWME6tdq1Jq55mCLRA
+          xk5Bx47aJuqBj50EaxoR2OBEJQefqHyT43nSu42/DlaifkDByAeqo3H3/KEjc8eXWKktF1H62TWo
+          UfdL/lrdtdmn5hMeIErE1yobsaK7Fkj1HJv4jjV0rEmAg8uEJazfNm+NsFgr0VwMLb4H26c7pUfN
+          ++mvYH0bH/U8Is5DopHsyC267uOtE3gpUrbnntiPcV9v3JaJ4sKPR5g5JeOqrdNAEeIAa2KouNxr
+          VYRgdDcLa5rxcqfSVFQItPUx4D/nczY35a1Ew4snwS6btx1z2kmU/FVSkR9/Ix9WnEGfVxty7IVr
+          x67eo4TO3+VE615CzXAU8OBuyy3WpOWMz/4p5uA0S08hTVRc7qlSimY+b4lfv8auozcu/+EtKSZB
+          YdNl92lhWb8BvxM+bOrMMfjDn+RW72J6vCuO5CSpiz1RT9inIg9RQl0gEeVyChmrljobvtUBdl7c
+          5NKVHc0wS46MjaDhsimQEw9CUe0CRBK342KXncX50bkkePu7jHjUaKC1Vj0xz1NVM7x/CFK2+fbY
+          pa9aY/tec6Cx6BUHTr9B090vR8gLZ01UGuzZlGBuhuClokUvXNi8Hu+BuMRrMH6NBlGJ18U//qO7
+          eZtNKzE6wuPkrsZ20VPT52aESDc+OpENVWLTD3/W+egQpfkO7oSncAbsvPtAciy/ZqFQ5ehikc9v
+          PrV5isNS2uezig3tILjzdZd/xejZ+tgmfJ/RIpwBmKddicl184/vOyCd+PuPD2oDSHUjrQzZJdFc
+          Hzpq9pGFoC1Hgr8ujd8sdkvEBkPGxnjT3flz6huptaAPGhaZGjc60lfkvy8d+x3fZtRgE48OVbnG
+          0YJnvDcoZ8mthhQv/gOixlU/wvs6yTiTv6ua5PEmgGcfAlbCQ7DMDzoiW7tk2D6Op4wWdhaKalMm
+          AT09p2zqtHYF6/vlSvQ6iLPucdm20KzTCe8/+ZpNn1RWpad9STFeYcNl43WUUQZHnZyf9lqbhm0z
+          Izl4GfinFylXGSK0fHjDZ0J1NN1PViKt0LceRwH0eCNYeQMbdE1IoLVxxn7/5++pQ1z1xHW0WlUh
+          CFK/C7hyZ3ffLAlbdA7xGfsT5hDpqa/Djw65mxWJ2958ByB5+BnM61pD80oqnlCpd0LwT08+8tiU
+          tLvc4fsrvyA+590E4l7bjZQKUj3P+sGRKldRF36b1LQhZwGyiDjYaoVTTefRUkENWBKIeaYzFu1P
+          Aix6bpwrpdCGH38Ok7WOTWMK4/6KJQqHTfEYd05DY5LJJAEdVRb2mHyOJzU/NrC/Pc7EW7Ep6xW/
+          8kThtBZ/fpE2a2UlwwZlCXYQCRf9d3gCC/ntOK3yxGWIVwTpx++zKtllU9nFBbRfOcbHj1PHEzK0
+          o7TwWbI/VlU8XLOvKC7+Aw5AbhCtDoIFxkm4YN+6Xbopr/gnTGp1x3bFiy69v00V5OZsEP3DXRY+
+          PbaoGyUNq5VSuLwgIxEGtuKwocuU0doJCzit0jTg18Zj0euWA8B/GTbu1h718n3kgahHb9GvScyq
+          q8Uh95PM2Mp3r47q6hRCl7pN8D5WajyGVZJCvC0J9sxSZQt+Bzuaf9JAWvjE7/2lRT+Q06FQ0Hzf
+          bBMYTeFMMnxqGD1UUfLTn1h1cOwO8fkegZ7mjyAnUalR7SgJaNGTAb/0O6YH7i6L9vb8xPoPj6vO
+          z0Ga9Q/2kt2jpsIlB3G/dQ74sE1ONUtklCL5aG+xuRNsxi3+DPrxU0sTNZftJlQgM/1yWNU+JJti
+          PWqlvXLwia9sY0Y3oRL+8b19uZ40Et44gFOFXgGykBD/1gsKdfs97hRrHfe3attAdtv1i95fxf12
+          97WQK4t3LJ+eh4x5+XqG/Y6VWKPLmcfR2XzBeu8PI4Ttng2nTRbA/iN7WPb3VTe+zZ0noeZaEjfd
+          3zu6LeALhmKK+OfPsKZ3vnAFZIzofLWzTfFe7kFe8O9WeGrGudGH/+VLvOB93PePvIf1HAKx2/nr
+          Dl2LUvR2W/Tjyxrl9MxBi78aCLf6mnXregqB8tEJy6Yma+MkVCD1QrXCmmm/a7pXXxEcyk4fJWau
+          GUuyUpCq+nHGe3Xpkfinh8vwTG4/PrB58IIUOG2Pndd5jml+F8/IjTkFZ9yUaDyrXAH0IdFwejOE
+          mhpX7wjWs1BHbbw17l9+1Z6vG7mO7jkbxW8+o07aWyP/vrsabxzf4r/69zV+avY1ew/OWSaM359+
+          2c52AXJzNMh9U3r1cIu3ATytxgx4gD4jG3imEnLsMw7C1mADhLvi9/3Efr/mbs7jlQfW/NyTk3kS
+          axY8TA9q/SkR07tw7pxkD1F6Wk8TB7WiZ+zRWh6yt8cnMd7XRzeJ77CUQDq7AaWBzea48hJ0Mz9x
+          II6Gg2Yh+6SoF8oVKbbEY8vGzQYpwzCRJf7coW5VDwQjaIjMSVXMznieob2YE1EjdXTn6NObaOEf
+          pHgKuJ6lt1ZBMt1CvNwf1LU/Ppve+fcIx/NKm7j1+rtsGXKxv1+aOtj02oBbmi72BGjib3D+hgg5
+          7pm4o3uMOSdVKumLuQtx1ofZnY9Waf75Y1rjeRkvidcz4nROHoHzfdS/1X4GmVt21Nl+6/7p/zv0
+          30UPv+Pt4u+Loe6+iWNtRtZZ62QFh09wwe5Fk1zWv1oLuQ8wSV5FPpoul2MBWy9bY+vw/MQLn6Vo
+          /gYF8Q4R6uYUv+lfvvvprWF4fS3QjU4PpubraxvgNWFX6u6FLM/X2FypIlKNfMKhOcSIKfOrRelt
+          1eM///CkyCk0l888UpEb4jmxchM9+wiCTblm2tic6BOCaH8luOObbAq5s/z7Pmw0fKHN5lvqUZRp
+          EZYHDmnTzY4iULeHPmDdte4YeDcZfv7m3iLvbNHf8ONrxO9VqSM9NXSAg64S7LJHzd2YZUFRWxT7
+          juV3XOPoOTje7kTw1q3qyXzucvSHt7wYoTncfwWURPGTOPJzoy3+jAncw3RGZm9vbJZXLYV06uqF
+          XwqM8RtvBD1mNjaE/TeeTumrR9svdyFWRWnHfn73VNAWG09tU5N8uaWoVz8itsNKzrbnS1RI83c9
+          j5w9RdlcXxVOgs6VsGpPx5jT2+Vu8sPZH+kW7mhO78IZZsmS8ck8pfWktc4IrD3UwXZ66N1MUyGA
+          Yoo2RPfch7bogUBa8hG5F8km/hxBFGBNMwVj/9uz6c36I+y/4h7L6/qFBqO/WlD7TrjM900j93o3
+          wyMqyR+/7tV6E8ISf6N4oHlGymatQnr8GgE6Xz8ZGa+jCkq+uhEbs31N6XSIpKU+8cdve+0gtPAK
+          9lv8i+fn4u/ArhdfRAbliSZ16+jiDw/2r9GuufYkNr/1Oib5zqi3RaK1QC6iNG7WrtcNk6yNf+/v
+          JTulm27x1pOS6PAkWBtzNqOQ5vD77fjPd1b/+NjPrzO2yaGjwZ4Ku42bJ0HbCqdu3NB1hFBND/gs
+          J093cLYsktjeh3FvdDr785dHqRoXf//VTabah7A94FtAzyp1l/zPQ3+5X7D1OXPZkFerZqcGUzLy
+          nzMfs/ytqbBp3ghb51hB3MJ/pR//34g5py31rida6iHYul70bJtluYiW5xEdyausFTxBhmC8VNjA
+          OtPmh+jz21VshWSf6HXN2psbgvjZk3FbeW3GHn40wtsveXL93kptLsQrB+kN+gWvzHi8YEbhm+sP
+          bHOvl0vjOFNB9Swbm0t+oIKRnAFX5z6Y657ThsfWfopf4dwt/uCU0XFQefT4rLMAGb6qTdpMvmhH
+          HhzxJAvF80M0ONBOo/6nrxu2OfTSvcwf2Fa6itE5FT0I/APCXti+2XyIrzLyn1NLbKcJ4/ltTp4k
+          psE52LyvSsd0xwawkDjiwFFVjWuDSJSsQ1ni6HH5xnNZzzOMvKkGlJPUjCKvl1FYZhPWleUWJKfd
+          ibDU84IeZzSj5clORcPJHou/1dVUCudc7Nj8IIse0Cbh+6Cw1LuIIzW523sBTn5+PLEU9cTmC0Yz
+          0MtWIspzvXUHbr1uETJOu5Eu+qV/+qce6nHtYux7gcbdkvQL12M6472YRy6jN6748ckxU09cPW90
+          3wF3FeAFb6RlfOZAKsPrA2uX08gmzb95cFq7I/7xpcl0NJBKpbKIbuJJGzaxX8K8njC2MXv/8MDb
+          yaTtF3+VdnQ1vgLIGZmCzd16o97nOhmiO0fwUq/VNsO2oZLWRRviVBFm9H1LQ1hV2meclaPRzan3
+          LSE7vo2/eO1vvJGAeugR/gcAAP//pF3J2qo8s70gBtIoSYb00kkQEHEGiAqIKJAAufrz8O5v+M/O
+          eL9bkVSz1qpKVVoejs2y4XHwV2+wdqbqiWvAGVC4kpoGT+eqb/UJDjZGh6hpVHjT9/oSCg0cCTc5
+          33jJfnX0h5fxcb+UXjP2Uw5vPjTppf59vZm4qIfVEgr4bMComOEvnYHjKoyenvrULMi5rbB7hRP9
+          xy9Y3vjo1LKevIwqHhY50UYQqs8M38od1kki5D40E62kx2nXxP/w9u8XdOR74JaCbfqn7PtJgs1H
+          Zw1SqLg+sM/1E2vWtIBpsw90cHb6P747i12Zwo1/0mv9+3jM3oUQVfrOwJr/VZl4xWD+y2fUhKIG
+          xM1e0caXt3qdEveq5dbwYfhXbOePFAiK5Twha+ZbQK3OjSdTFC1IO1Wjf/oYK8lv/P90FAj8/24p
+          yJVrGJAzLGLy+SwiEK/OKVhQGsQ8H2EXNv1wpoE1zw377vcyeHz5Hp/8r8fEG+pdBNvAowpzzWLq
+          WNCD7vx0sbHTecaqyyWDBpfr2Mc50ueX9E7k/lpb2LSr2Vvd6vVEL/HQ0FNYvYY5cDIbgtcFY92q
+          4oH+LgYPu6Y/47Dj9h4rH70Ld+hypLH2vYP5Rf09vH1ERt1nKHnzU37I8sU+X7Car8NAKtcIkGXd
+          AVnzx4WNB3dH4ORrBTayx1rMT/kqI1kqEgK8FMTLLahl9PRCixpP8QuWfg9rcIjcE72fvKlhD9fN
+          4XteLhgnCy5+u2qOkOe0HGHGZHjzuHR7+FR8Dl95yQBiaaQp4h7Ch5o5Is06DiiXb3H3ps6zGdnc
+          oNhH5uASrJTiorMgd3nwyF9nGncn3Zvxe1TQr32rNF+eRyC8gyiFzlUKsaJ7x2advpqILgMv0wdn
+          td6ogCgBfVkhkohuMUgvfuxgyys/Gs+u4q1UOrtQOTwcfFJ3GpPMRdve9xVjfGjreHVOfQ9ftxLj
+          R30QC3YYGg2pQuNhJfJ7fZmRwiP/67/xKe/6Yamb2UIXPamwv3Ja8SlZ0aGSCD1Vdu87EARn3MMZ
+          fXlqrW008K3uuIj3gRZIZt6D+dBkK+BO05MaogsGlklzifgmzsjuWquxePXeK9K0scDp8ZgOfHXy
+          XOiATKA3Lkh1sbrvegjG7Iqt/BEM/PIZ9yDY+S96le+uLr12uxaKnqjQ9JN9wOR8vBk+1NSnZlr4
+          gFf9OESiOVF6BeMVMPcrE/D7NnvC7frjIEQwg/DyVVzqYW/RaXLsc9lP4BXbaoSKReE8C5Hhp2H1
+          iL+xCCS7hdv7oPkvq/RlbukMz6+dE/An4RuPOrlrsI4zF8f9MxnYdz/L6OXae5qqexsIKshqeJD7
+          Lw0fH11fRS5TYB49rsGr+rxixt3sDobxotBcvqn6+qrsHF17IBCxDLm4vxQ/DXbvNaOe+90uCb0v
+          Irp8NZfwVzQXUwF2vTzk0YSx8Xr9/f8Mhr0eYMeDFuPboLTB2w73NKTSOZZs+hBhNusK9oTWGVbJ
+          2j8h9zpbm/8PgJUsbqH8jk9kPzgxmD97LoH7BzpQk9JqEF+VkqOEGhgnmz+t71vzRCE8l9g5Kb9B
+          zH+Qg4u/XZrtP6TZptrw0E+4K46YBYZ50igPE9EJsNeEF08MYZBDMvRaIC6S1Yhu9a1hwDs37LxW
+          eZjpLW3Re3xZWCUSHkSpMXJEVro1QTZFwXtnykP0uXjUZpeXTi6X+Qm/xi7Hrp3e41U9wxUeH0qD
+          Qz5wPBZd3hDyzTkj83yLCxHd3ByU3/2CnduhYTzylAoh/lngOOxkRi6Xff0X72iBgOrxXKmmSDrK
+          GpHtFMXb++qQ53RcQO92B9hCd7Zs8bmCTTCrjTAYWQrks7LDeB4cfeRKJ0VvwN9xuH8rzfrrjk+4
+          S1wtEIk2A2EwwhStFrCC7fli8XbjLSQ7nIIdfT404xcsP7g8jhEuu/Wir2EfcfCO9AV7a7kDa16X
+          NYwF+0Iv01n3JDzoIzLs8UpDu6vBSmSdINGkFNvLGXjkWHQ+2jo48IPr34BeTc4AptZHOAntHMwT
+          4RMIj1NLj7ff1WPCvk/gR2Rn6t+uubcMg+Sjh5r4NB8ypZkbVARQiS2eCN6iA7466S6KjduTZvFS
+          MH5ORw1u/k9yc68OUjddXHRTy4q6F8OPxbgRjcMIJ4Pm6sv11tu5+MnfA3fHjnFcYvbbmqRRZZs0
+          5scXY3PFhXAOjBvGeC+AJTq2GrRhrZM6IVYswFH6wZNWDFjXv89CdFLVgr+9gGnqYiUeXOFEQFvU
+          J3x6FQUYL61Wbpe8BRxzvM34ZlX8w+uY2PT2c1aPyeEjAxxXXagqzT0buw+qYDbJe2pdCs9jNL3w
+          YL54GQ0qu28WIkUQRbMXYQMED30RuDOPUuO6SQr3dmBOtvCozD95IMu3TaLMWhnC1vdoGBcXIBCN
+          aDA71g15BofEE/2kmuHL9FUSsvGjryN/5yG+SSYRt3jDlP25Q7vjUASLo9z0BZrhLN+gqVIje0Rx
+          jyogy5u9UBx8bY/fH8QVpbKqY92YWm9+Vr4CHPt1CBohaJrhPr8UNEX5ErDldgLCHMIUbvaPA3U8
+          6WsXZDX8LHyOz3Y8ecTomyfyJrchwqmOm3/xfHczDvTMbtv5qbcc4ivhaDA5X2+e01GB7eUGce7m
+          eizR9M5De7DeODBfNVgv6JoD9D4ZNOb4nn2lZ9wjIbUTXN2Mtyeyh81DOMTOf/6JiVGiLZ/gCkqX
+          QVpKncDNnmnErKIRFQXIQNwN3wCIZ2FgZ2OyDu9Ph4OxMx1Pqs13DX/LfKapeOvjsdh6IcL34Uh1
+          YzL0VT51P/iXP/EoxgUbOnNE3+PxhfVtPtwqqO4PkpOa4PM+csFSHeMSXgZRxrhPlma5XfUE7DM/
+          oj43EjYL4GVBY649bPDgBKRTUnLgKcg/jNsHAO3Pl1t4KleXVo3v63wv6ivaOaymxyCu9fng7sZt
+          bKpOtf5sNwzRrgXBfdEptnZlc79U+XpQ5zbB8cM8xIxqdQbPL+QEfWfO8bg/iDM05crEuhx99IHb
+          HzV47Q8CLk6wKZgu0hBcX7wUSF1meUIqQQsG3g9SPxP28Ry/8DaGH+UEIXco2FJ6I7S11qGXL0e9
+          9YaeNkqh4wZfcKvBoi/mDL/cIaFWnRjFhmda+Ic/HX2+NcTzuASG598BW9qbZ6PRaAq0iyXA9txi
+          fZrNV4r+7M3AhqSzDe+AwKkFap2WxzD/xVc7+HDUDaDpSXM8hKBpSI7dMxYbdkuMDGKwLTLvz/aw
+          7O8vAy3PeiGfl9gWs8yfe8T9/COZg+sak9sNGlB15U8A1CApJuNAfMC8vUPL65IMrNFkCNal+WF3
+          i3+iX0oB/MtXWmmCYjwzOYVbvA8S5Z01y0G7VzB/zYQe1Tfz2FyJEbAH403T6Un1efFuCnxYZ53a
+          ryYYptzbp6CEyp4etWinL+Pl5gPuFxzJ/Pjo3j97RHWW4NOlNnX+uoYrelGxxSq5TGBNTpkFjc4b
+          gq+Xv70lucgplEo7pOkN+UxE7T2HX8hO2xqByzbGPGphtSoC/Yu3a1qFEB2upkEju755Qv5oVnjc
+          4y82Q+HYMOuUGtBo1ZZ61QENq/tRIYpx1RP50hFvfnvfDoaVaFOVe0Nvi4ccrKg8kQ2/6s3RHZ7y
+          q2VlAJ3jAEgjfQk0XL4isEaPYr5afgiO8nUOKrltvdVdEw7yR+jhXAY1mEsOrmBH9Jm6X2cuRvf1
+          rOAnQHecr1RoyFctfNgkhU6Pr6QellMASviH7wzCuua3Ch0PzRpPBL3QvVnK962HWhcqNEpIF7OF
+          Sjac9wRR/2rYjcjvfy541cMFu5J18ljxYzaMEt6kQR5HYP6yc/Av3iVH71nMXVlXSOcrhwbPIteX
+          DL04mUqjgf2ic4vZoUQBwWTk1He+00Awe3Pw2cwNTb/DLe4NEZVAAZ+Qqr4QAKb4JwM2b/6FH/ef
+          3YzFuyjh7a3oOFTgtRBAIc6Q9vlEZH9K2GKQwgezcnDp/ah38ZwFLQGH4RoFXFLrsfS7+CJk8goC
+          1nyAN4fDTYbH7lhjPQutgh/urEQn3Mn/3tfIUX6FMDIM7AtnBiZqxzJQx9Tc8CID7FNrNvj49kQd
+          K1WHZX1oP7RIH0rT9LcH7F2WIczlHSNC59zi2WljDZH3a8GOX/Vguo4vA81tSXEZ7EEx93WWwlKy
+          VOqOZVvMeckU9NXsH03D62eY34+7Al/TeA+YAOhASoBE+O2fO3p+yitj4LCk6I7UBXs/+dmwY1Vx
+          8NBGCbU+2Yex3c+pIOzSiBwm7gNmnVwUcLl248ZnbMAOu7SE+s/8ErEvvYHPbhKBx7nSCIK902z2
+          zsH6Oz9oYuRBvNTeQ4HK56qRb7TIOvGVIYWKP1ZY4dLZm8Et0ABlk0yNxcr0RfKfAXwj2NILpVWz
+          jPflBwy8XIJd7H3BPz6xn0eKc7EdilXKGwg3fkOulUqaVatRD4kLTv/w7zxJ5Qp+QTAQSbSiZiiP
+          R+sgBUgJ5N3daNZGehHoBL37h4d1YfD3M+xPTof93+cH1mkZIqjoy51wN+Ot/3bB7geOD62hpxop
+          nph2S47ur/FNDQOG3tzNFxmkcU/oLTKI9/e+wNH+BcHD5dc//42AeILHYO964jBdlbJDv9hrqLtq
+          TiMezMAHXn0ssFPYyzD7AQrhH969GF85nvo9X6Pf97XH7shZnjQNrIUbn9ye1wXM36sa+p5L9U+P
+          8NZT7vdQqTMPn/WcNcQ2leBfvFCK9VTM8+9awS3eUOUlJ83qO94M9VgZsHOQqmbu5rsMw14NCFPz
+          0FsKIPXwE+zuVLutX0ALruv+8AD2382xmPX828F1igusNRn2Vkc4PpEmPwPSvIZXvBzDi4a+b4Kx
+          Wa1GsVZiSAA56UmA9GBuJmhmM0T6sg2J+pYek00jh3JvImy8Mm9YDZhCKPNtizPBmtmEXu0Pmg4N
+          6YaP2ATHXS9/Fa2h7iDX8SrhjIN6MJe4iGt74KUlTcGDs0SqK/sgZuwXaeDQZQXWwy5nTHS+kezf
+          QIA3PKivt/fRggm1MHXNvdpIm7+AP/0kGTW+WU3Nl+FogHMgSn3q9UHSuNCunZY8jWXS15+o2Oid
+          ZztsWfmvYaaBfjDYBa9AQK8lXj7l8ycv73uPL2kxMvaIfjzEKP9hK7x+muVZuQEcD3ISLDq8FCIe
+          zCfa4h9ZP7Qb1tmOLBkd7hGOjpUQb3irh/tC16gpByjupYtWQkUdUnwig998t/gHF5J/gwM9XYaR
+          2JfxL7/iY7kqYMt/2WFVIaZY0J7NqkxKBlusd8H8OtmD+BMVF+mHm4KxG4UNg+4aQY4rLzQ4Q1BM
+          ROs0qL/dFTulf4knT8gtGFijSW01uhf0731v8Y1asfAefsDlLdk4fYNAvCEfsEK0XEh2chjso5ro
+          //CFOUUC2YmcXIz4cC3RKK0nat/8q776ypDAP/xn3NVEn7ffi6yQL3Cw+SvVRRrBTgsYVp1aZqtJ
+          bjkUDO1EhPViDbx0/+3B7cvJOBA1rZA2vgcvT7Zi63gUmzlxvzZ4zIcPzjd7mrd4B956qmIN2082
+          9795Rp2pN4HyYydGxmedgVDwK+yZ7wHMeRK16O/7mtU6gM9wf4cwEOMnVgOjHCis3iPMUICxHdXE
+          6+NDH8AeyUUAyqLWmZq8aziBt0j1+EAalqJ7K0uu4lCnQseCNXJeQSF1Exwsg8uE8ev/oLhvh02v
+          fMfPpEo6dGe2jjc8EE++FSsolNI3dkJBidfDEhH4Cl8BPh6AXQgds36QXztKzW0f3Oon1QpHzN2w
+          PQyXZmrROYd6nWg0+CasmPpp0zdNX6VXXGPwa2PUy1t+CubJ2sX0KM4KBJG6J8vvmenr25yNPz6I
+          dU8R4k9cBdk/fcLyKqKvpAsjeOzMmiD+vOqjbyU8Or2UG67k1vAknz479OePJ+9te4u5Swxk/dQn
+          +Ze/B2cR0Rb/sXkjCuM/ezGBwZJ+sX81+ma+Hy8hdJvg+89/GLfHCjw19pOGp19XkLi497I6JiaN
+          3diJ51JMUiQdPi0+Xg+feIwPvQ83PEf9Gu3if/nzJtkuTYT+xn5ga/F0vugaaAmxCiE73yuIzU7C
+          TpvVA/2p5xxOTMTUEoHriY0GZ/iqbhZ1yYmB1zvIE2gFJCb8SXCKecMXSErVHVZ9WwPzGeYrgHL3
+          JHPpmN48XAoF2sGbww7PCm8OksYGrhx/8Km8ZvHaxsIPjonf00zoo2YFcVLLq2Gl1LvE+3jE+VmB
+          wQ/GVPm2T30FktIhvgxuNPipu2K5jl/roPu/CzXoxY6X85BZ4EoVC5eDmP3DD/KGB8kF8u94/fpd
+          AInJpURYpT6ei3OvwWsQn7FuwjNYB/x6gt+ynqluYpNJC925UHOWGz661RgP1uHXAa41RawouTAw
+          /4qfsLvFC7UXBtk0DLsAHnbvJ1XM6uCtcno0pO8B3gOwcwmTvG3UxYanqSkBj7Hscc1BDHGMNWFs
+          hhXaTg0wiBp6jAQdsK6F5J//YfAsvdV5lQbov/1pw/cHb33yXS5/OZBsku+rWIZMn2F6mQW66cH/
+          6geyRMIrLYVzzNbzhc5gPOoZ1s5FVJD1BTJoXcENH/vh2jDuprRgOeM8IKu8b1ajH+q/eEnI5Gw9
+          M843hBEIcmqnh/dA89saQnN5GRRzVqsvmn8uZXgTBnyEM2gm6Vn8YIFsix4XjXpsvZ1teG7Zgery
+          Lva2/AdBc25jctvwJ1/z8ioPA/EIg3rK+Oy2GyHwhjpYtXwB//TJ5ogZYT/abmsthif8hQmjZfkj
+          3qI9DRk8H88EqygNivGPzyRLcCV710sbUvRuBsYi7rDNjaK+7KjMQYFebli/kjaeg2Rw5S2+E3RW
+          mmLxrViTt38n4nh+FXMSOy0453sX5/wFeEtrLvI/vV1bb3hYAkZnUFpXH/uH3ovFKEIBqGpaEUSG
+          cWCiaouAz40IR+p+1tfNv8D0ufREvKERLE4cJlDWJUod7TWAMXFfLiodD1K3ODyK76fs+z98G+zP
+          8TP+4wtww+8EbXxy5gXgQ5zIQbCk6AJm92Zr8MVvQ6w/6ostWzyE5/soU5wZejEftVsEz9ms47vw
+          fLDlTy+AhswH8FE89flPLwuFoAq4T6MP427SAnhewifOlPcSszgCGbw34kit+6pteq+f729GHASs
+          zqaYiT0loK1SRLg8jthcHAIZ3ruhpKfC/hWjpirRP72iNXIS/9MT4jfPBUJa+IyvjB+BxsvnCbcM
+          P7ZIfh9APZRbepTvP2/R71EKVWoLOJDOs7dM8ukJFLG/4XRto+bv85HzcT/BDrYP0J8enHswdzkk
+          /McBHu23IXH4o12wz7knb+QdQiDzZAcrpG+Gdbc241745fUW32q2+FahwSCaAPY+pgtEJaEpzLe1
+          X9vvib+C0+7/6X8wicVh7uswAf/09ZGz9HW55RzooJRi09uGrpqHNof0dkyDP/1i9YUwRF4mX4LV
+          wsdheafripD6fdDzl43e+qcnbvojgTi/63MWjEReVQ7/w9PMDm85wEqv4DR/kGbUFtwCde4SshyX
+          wZvyn7iC0xuFVIXHbc2tw3rQPTINK/JLLkh13/3glo+p981gvMi7Hw+cQ0S2862bNV2NHFj618N/
+          +saAKGnBxpex5cIxZqKqiBCbrbTVd0Yw23FuH3YDVjaWLYN1z59CaN5RhU9hpTbCWWwsNLgoxfe2
+          MTe+XrtIUb8p9ip/GUbxKmgHb8lX7Lw/c8GYMGlQOAEQ8F1JveV5+a5QGJ9f8uTSUJ/LwUnBhqeo
+          +yAXfdPPR1jU7EOooD0HloQ2D0EOD9R8l1O8HDqnhqsnoU1v0AaJfFUFbfr+P3uYp+b7zx7xDe7s
+          WOpuXwIvZvnG2JwDwI7iXoPnJXoGAphfzSwaTgd/tHcoboq998zkJUAzm62thXPXzKcHZ0P+8J3w
+          MSucWGDt/ERfRWnIglIS94ET2qgvS4T1qd1704Yf4Lq8ftivw7hYHHnTq4B4xzquJW8K37WPHP2t
+          4TtoTdby3Lb2PM92VJ36F/vzT5BySMb2X7xIXqcUUE3JsHsrdLa2ZWXIf3w7uGBczHYcubCRuP4v
+          Hhaz/Tm1/9XrMh82qyPgGiJ1eNA/fWEdjo7xx9/+6fOSpekWBKmi4Wr7vJUT0gi+vWNPg/U0DvNf
+          vW/D2/jMvbdA9p45mCefKpDN/atZnzzJ4bXhOCL1g9Ss4j19oo1/B7stvzDViZ5Q9/sL/tOXl6i4
+          ltCEnU/GGK+Myaaf7a977OOjN8neGF94G4aNdcV4kVMgmdtFzT+8uNXb/tknfH/HFMfj5VfMu8kN
+          oDb4Ei2m0+rNbY1b6C3ZSq3Nn5YKKhniD8NEWFKEw1Y/r1EN9t12ZT30/vRjgO3bJ+hbURukTW9C
+          EfBzaoytGP/js5L0uNMNz29rZAYN/H9aCoT/3VLQCrpJ/TbeJJQgqoDU1S7F9dXUefHwDeDIFxE9
+          tWkWM93wOIjt+BKIeafHEsnFDhHaE2pML9n7peANgXpNlOBa7VQmqTbYAz4Qf1gTsqvOU5qF8OG0
+          DtaLAykofHkzXPevguwvzxFML+/1hJ9IdMkuWYE+FWv5hMKnsbAlKtOw6pm9wv3Of9No/czDkiT5
+          DAez06kfq6k33xvCgeh6/mDfaOjATPnbQQoNl54y/a3P9LhYcBfZM856eYrHctp3iMH4G+Tmt9OX
+          9yNZUWtqFVa+Ahvm0tNXSL8vhyZldNKX1N1X0F2YELD1WoE1d4kMjS6JsGI3u2Fh7y8HGDx/ccK5
+          ice/JqFCLs5UGhZGEzNt2+xyGHkFO7MRxfNyZE/YojHCQdRStlYJyg+2knE0uXgPMB8F4KNq3kN6
+          //GCt1wlTQHm+BixZ974eJHoXCFCf4Sql/0VfFrraoMHQDBYBfM08NdbqqHMxS5VnW0T6J7rK+jJ
+          9y82lL3G+Lt0T2VwC/WArYZQUN3mEtCjw7JtgtMK8aa6Lpymo4UtlBsxs+YHhN/3TSEcPcXD/Alp
+          AEtcGzSYso++KuUGn/HvSLW7VgKhmsb27/dRh9pZId2lYkWKhq/BobzaBXM61MOyrSKyQwsuBCUC
+          Bty1L0j2RuvEwueyVujt7CKsV1evkBLzG8D3NB5oGj9SfeE4MMMqMkCwbt8nkdQIYPG49vTIooot
+          7/yXwev5ldLQfNBh6tt5D6/je6OEgwr4Mz4ZkB8rk57LnanP3+9h/LMvbMqNDcTjABUoez/1tDqX
+          FRCF43xw+Oxz6geF7vHJ5HDImH0Ll9dfF4u/7yOEOurv9Jy/bF2UD/MTXYUQBbIY3Yv1afUEmoIt
+          46v3Vgv+7OYESvae0qj7WIB376UM9IyE1NHNRKcYOxbcvg9rliSweSRdDT+y0/x9vseMWMvRc0q3
+          WQeGXNCH5//+/p6sKih1dq4XGY2HZE8v0qoVvRVbGQyFA4fVRB+KKb8tCYrt/ovzD582/FFgATzR
+          7kB9ZoSN9LJKAu8slLFS2XHRX8y3D+VfqeFqO+/PA0saXKr+Ru/8bo0ZSKURpr30IjOpdMAr9gYJ
+          Uf7GQU9TNi9H8IQ7Qb4SUsvPgq+sgpc996FRpyhIzAQat6C01Tu+R1UOlk/e8pATKMCO66uFUIXn
+          CmXbXgP86eyYPwoggJ737bfnY8X2e/foI3sN6efE8IQ0uQcwxrsigOHh6Unr8cDJm38F19/OG1pH
+          QD/IpUDGuP6RZjWGpUPshUOMX/bQUOEIMziyYCTCImeD9PFYBJ9BjbBNh7gRn6scgfrGVCIc1DdY
+          nfpG4En+xjTc5cdCmr7bplImW2T+1j99ri8JQeKMa9JfTk2xPG6iD5bxuCOcYX3jNamyHCgXyAV8
+          fd9KDI8hQAfHuOEkym8FY/RsQ9cifACZMQ/SqMU5ih4cj3XaZbH4tqYOWvKRUM8c72w2ay2HttW4
+          uMwLH/R67HFwPw4tNo83qVjusO8gxzcX6kydUMzPaqfAHjzf9J57wFutUuiQKqWfgIV+UJDcJXto
+          SaTG2k+e48XKzj1g12uJXfPbeeuYhj6c2uJADqfwXND1K5XQ/FlPsrNUFwiA2+xNjAxadIMdU38x
+          XBRL5Y/6OUmL1TdVDT00s6ZhEj6YoMvJKNcw1chuB+thlQ9zjfI856mb5GK87AYzl06XyaLhASf6
+          2qyQwCemDT5KvdlI3j1OZa0wAnp0Jqm5f862BVr/EgajXH2KdTnXBK1SItC4r+1GON+RBfThWGJP
+          u7SFtD+FNWoRieh1iAR9OrmqCGdVy7Cn3OVhJM8zjx4nD+DTbb80632qbLjFU3rjp3pgWVJb8L1n
+          cbC72S+20PRlIVaJOrW89QLWMFFlkF22rnZ7m5tvvSiEZbn7BXHok3g9LXn79/6oO85bl+jO9WHa
+          C9scras1zAF2e3B8tYhGtFEG8f2NchQfK53Mm7+K+uv9hCj2RrLvPzrj23Cc4W43u1Tn8jdYQOnX
+          MBQAR2RlGgsCOteGd+ROZPpGq7femyX7s+fgFx59ILkCbyHncXr/88f1gssMuvMHU2WpY/BD+mOG
+          U3A2A7S+LoX4KxQXTbfLg+wrOussVQ7RX/zF8eeQ6aP0e/soV14qdg6x0LCL+LbR3VsGWu7j5L/4
+          9fh28zb3XPP4m6cmyF0WARdyFsUi1/ktLNsywqH5wAOTntcWpl15oTdQ9WA537QRnZ3vG8ey8fCE
+          vGpG+D2yN81v+iaZRccZ+XtzwLmS12Dd22YIvCQR6f21L2Lhzz8AzjERSXwq+GNt8GDf2j496ZM7
+          COxIazjHRUs9lmjxWFi3HKbuJuFB5xqT8QE5OIXiERfJ6QUmngYjfGbZgpV7/SzmSv1WUH6cGC7G
+          JwfY/tf06PPGKeHfgcWkyy1XIM3SJjhM8bDFf1LD3wf25DWP2JN60lvwdKEWtl/PdGCHgvZw0TuF
+          VrWsxMJaXGQ0OumZyAToYLj2WQmZVmB6guPazKwYkwPU7YoGc3n1mPR8dHAkYYUzmJzYytlLiZ5c
+          0wbNofK9VX01BnTbBOP0WdzAL8I3DrrWyOPL7lEP60dFPxiTpcTHNjALqf16HPzLTwexpvpCzscM
+          yvWu2DYfesNiOpuEPx531FVyDfBEe2TQKuIp2OuXN1j33LNEwn2IiPx5Td4iHKIZ8s2832jrRx9f
+          X8+AWuf22Lr/WDE9J6oAMTcI1py+jNdsOkfI0bgOHx/REDO+qGcw4ftAsXjdD8RbMhclaTjQHI28
+          t+jCLz08wDb3uP+B5p9938xSoaG8SN5M3oaPviNdyf6qUbAQbQ6RtXQp3vAEW5cClwCZxNrmVibD
+          yNIFwsTzz1TtXgtYajiVAHUgxtr5ApsVnl45clJywkq87WmyXh8Ovtfv8Q+PD9Nd3Tat98WO7Peq
+          Ucyf8BPA0Dxc6R3evvp675sZ7oV0pRg30jDt6G7eb/ic2t6FNuPXk0LoC+RBbgk6DzN8SwTG8aTj
+          +FIszfzOZxHtf1GGre/v7q3MtnMIf3eJOgW5gnXvSz94G5cbvdzOCZvZfldBcqhrjN+FoBNFuJUH
+          7/kRqXo8bi0DztNAW7ymkdS4hfiCDwU8x+1Wfne96DQxX/4fvqbuOz2A2TaWEe3Pbw+bd9n0xHs/
+          zOBr5jfqHElbMCdItqlIl5L8ymPazOdlkcF5sXms7L+qvn6R+YNobx4C4UOLgoXZJEKHdxjFn64v
+          FveT15DW/hWXSmh5PJr1FT4PH476xE7Zwp9BBytykXHVnW+MPS/hCINRudDLkryHZUelGZIOvYLD
+          pNTFrBi6gd4OigLZnkR9WYv7Xt7dHjX1v/W2+TObeFAuYE9g27m6VEHLBefPWcGW1qwDs+YrhD+m
+          a9S7WVxByPkZAna69zTk47xgxyCvYMYf39Qo32mzXJVOBHn35QNF3OY+b/gBRp72od75Vum/Pfes
+          oGnaMb3dDkO8ZqKoAZNxCcXJYxkmuu9a+OAHF9vF8NRf5cXJ4JZvMA6QMQj7w7aZmLUfggJviddX
+          fnuCKp9tmpJ4ilf+4KfbJtqSehv2+cO7/+xpndfWYxUMXHjz5xXbh+ZY8M3jOaOoWjcNupGa6bjb
+          i/D2y05YuQiKt6TJJYBTrU1UL4W+WPuQN+Bgzd+A0+8ZYIn5CpB4Ll4068qwYOrQj/Bwud6oiRJa
+          EGtX5wj3Z4iVGSqAKYb6RIdRVPB9yzfSddJE+Otmjfq1PXlzYZ0zpErJ5y8/NgOf7jMoB+oJWxfE
+          9DWpwhxxemrTCzSFbTGvySEGTB27B0vV+QtUcmisn4h6puWC0RuLFGzPT4jaj8V2HiN0DM/Hx2uk
+          DlLy6UaIwvVJ5nutFEKS5CtogPsLxDrWG9ZaDxdeCvygp+15J9PoeegeDzkOoHMt6Hg8rXDttjnn
+          4jUblivgMyiVv5Sq0v0SM/nlpDD8/jiKWdw3M0EjB4NRu+ANDxXLWbyLMB7zEOt66A4z/XItrIJL
+          TG2a/fTVk+Fe/uOTGIeBvtQSMGQ3UDuq6ezNJnNUFPju8iPVi0MQM98ZOTCYrY4jqx1j5tWnEUaD
+          HuCTVS3eqpfMBXlsaYSbhnl4w8GFcL3RnBpnvisW03iKCAI+opf0GcTkDF4W2s6bhps9SMDuW8h4
+          KBGQc7Y33zAjMKxTn/re51Ewc1dwKErkgrq/N9KX8WxA+GJ9gK/6XvX++Rv+mCnOwpbpA1/UK+Kj
+          LiIiB4+F0PQ7A0pbG+VFbnpAVEdLYRzMQiBd34PHDsWnl3d2BDGu1CpeZCrLUGy4mixefvaIt4Qu
+          StS0opYk5DqFh7EGdvdasK+uX33Zl5qFSLd7BUitr8U6E4GHBPdHel1vdbEK7zsH9LozCLz+uoIV
+          1dxBnZN0jN2ELxYz2kH44L9u8PuOSsM/Jqahh8xSbBy/jjcLKZ/D8ydWgjpqRX1sv6cneC/lC2PV
+          PHu9aWAofxJtJpOev/Rx4+N7P8Ahtr7XeBg141TDPDY0WtzLtz6jrU0FzJ87VTu+iDd8l6O+37XY
+          jOkRCOz4ef7jwycv/7E1m24R1BJOw+aHguKnDkcRXo9DjY3f/afTQrqlcHRriXC6S+NVKecRbPj4
+          L9+x2QoSDTD/rGE/ImPz61bJhuo1VYLPNVIbobyxGW3xFmeBJeordpAIt/MIPvMra/74ECwqR8Np
+          YluN4I7+HtpHOmO1mb1Y1DN73u/rZSHoFdmFdF9fGdL09ycQWFkMpFVZACNP+eB4qup4m/H0/OPf
+          1J+7GJCHelTg/IbvQFD6s76+4FWBxrE+ExgUjceiZZ9APzMNrJJXMcxnNxrh4XljVJtXQ5cMAe/h
+          TfyeCEgvVyCmVS7+8TN6tOBZJ3s9kuGJpS8iH5fS2/QZF0hL1+Kglp/xPPFKAOu6wkFzKc7NYF75
+          GRav/Y/aRvsteoJaCMSdBvCxGzVvtbPTD0ra3OAjAFSnS9EnsKg8jaxbvp2xYvV//kJWS7qwNcMp
+          D//0pwo0rb60JJxReQbn4I/PsE6lAfR/vUmd6ME37MAl1R8fCJ5kORXC5r8gSk0V52iHhqVv93to
+          7wIOG0cYDQzZjSJn3m3G+RYvZ9fgbSgbXo611vnpS3kDM+yEgAZ8lB9icgqmGbxc7vbHr4elI7GP
+          tvwRdN9r3CyZK+9h93Mv2NfMqFkH37BhFu2zIDpxjc66x+sHD8sa4KNzbJv5ZHCdvPFTIjZC1/yo
+          1kaHzb6oXx62zenvJULXoXDw8X7MdJ7buQF03HNGfUG4euwULxq68ZEacG9ujlf6kDL0518uDaaC
+          3pOOg7er4ONjyBtAirBeQ0W/IWp7muWJcsFZoNEqEXsbfqf8G9egPj1VfHzNfbysxWWPYLo36Vlu
+          tWGJXGuE72NnEbrFm9V19px8v7sW6SRB1pkrd+WfvojVdjTj8bh7QMgZT4L1UrALwnV+B7f3Rw79
+          rxjYfE7Cv3iKo+F0A//4nwYsgfpK2HlUFU7lnz4XAHNE7D/9Mmx7HMnqMRaX9G7ATf/F/qZfrW1e
+          QMDYfMYny/sMK8+fIXwkfbPpkTvw9YGegMcTNNTpr6vOdmUsojh63nFcIS6eyPHXw8PK8QR47bGR
+          6DOM0IbnqfYaTjpj9GaDTf8Mbu7jBxbx3ebQn/Qzdk1Im/UvvowevdPT5s+C/QkNpJ7V3T/9eUWv
+          yIe3q+STHS8IA7v1kfzn79h/CXrTD6iugSWNNTXeQQdWT3SfcGpvh8AFwdWb6xxbgCVcGJAPg2D8
+          hJ//+IdRHW8D8xTFh8iSR/rHH5avn9nwmS+Iut+dq//DL1+JU6mxSArYzt8Aoa82VDk/DgVJqiyD
+          3fWT4GPL7+NRH1oCNy0uaE/VrlnHNAugu8v6YN5NIdv0yxWcCinCpnl+FUO3d3OQecVMzXPLN//i
+          r/j2bOo0dtCsRai0e8o9UmywcSqWa1P08HWYc5rpPS6W4fgMUB2QD3WPC/QI8+UZ2DfLC5B0p7r4
+          enh7MK6KRe+tTtg47df2gOna/MNvAuO0EeQ4zalVx00zC/TuQmfdnfAJT3wxqeVVlG3101LtlemD
+          FHy+ovynP5i29gEzdx0UGKAAkN2GT9jG/8BUK9Omr/QN3fRj8KdHudiLwGoYeQYVVVlwmndNPJVV
+          Z4NNXwuG3Udo1lYFAQwfi0vN4Nk2pAjtFnrMleiWf7zVWzIb/ha+pF56ubI5gxoPH33uYvVxf7Hl
+          l9MIgL1l4CA6W8VqxiEPM0drqfMTCFh3B5SBUyFE1At6aVhv1T0Erw++EXYNK51ff30CE3irAjKX
+          0vZ9oQ1n8kDUAcsas1cllLDX+IVWnnvwZst59TBRkyoAmx6wXhJrhcjaj9hMn0Gx/Onh313EMD4X
+          asPLs5PCFyHKX/2gmQ9lWMNOlybSeCFfLJ3X5SC9s2MAXK0d1ntzyMCv3Ac44tB2Zek6aLDO2oJm
+          spMXo1m7GXSHQcWmfVniWb0yGSJztIK+K+diNOTBhZa9f+FK7cd4jt37pldFOvVI1g9/+j3409eC
+          aZgbNp3PIlrOekePslU3nSdqT9iU6RfbwW030MT8+tAu7ldsv04n/eNmT4iGmyOQ2pIEwKh2EeH3
+          uLxp/k4PbMvv/OEv31iichrmlW7F6XZ0cWbdLoxUUh38i68b0o/ZM9ct2Ptji4PZOTVMMZwadn4t
+          Y6u3ZX3T0ww0jKCk7ng/euLfeSr+0Q1k+6jrfNrnGvjxuKf2Vu/5x3/1UNepdfRrNv20oAaquXsH
+          ZOFybw6w1v/Fq4BVhsqkP/1uCPttioD9HdqmLVzQp11Glcd1BOvtVssgPrUYq/6XH2i/93mwaxuI
+          sftB7Fu4QoogfxmC/aanMByDct/pwkQdQfG98bKVCCeRk7GP8Ams+LOP/s4DW5pYevNduiTyo3qm
+          RLB7ma0f5AVQkI0jPu8tE0jrd1fBrd6GN72LLd+z+5RE63rc8MCvWPXusQePPnPpLfIHvd/sDbTn
+          uMI4Uvlmla5OBNNqN2GPPz22CxxyBm9C9SGcPZYFe6l1CgXZOtLCZ4P3PujfBL602KfqO8ri5cXs
+          EWz2Gwg+pw7/6geb/oytDa8Np/iggN10crB+sXtvPhliB1KQ1zSgz7vHvo/RAu+xjgMYJ49h3vgW
+          6mspI8KmpwiXUeBAuk4XqjJ3YIs9uqX8M04qtsM29gg6yBxILpaJHVPMG7bp+/Dk9iNZrvNrmKy7
+          r0Bh/TwDyckDT5xamsA//UC7a5CxhxVFaIvPOPKeli55ih3AdHlSbPefhhGeP3MwfaYO9gXcemxO
+          gx/U1KDASlowjznG3oXNvcHBsOmRLHHjDvJRG9E8fX+b/q/e2frXkNAilAC9LsiHf78P+E+1EelD
+          yuHz9rlSvCzmwN+mgIf+SfsFS86fm6nXfOOff12geWGCFnguBEor/ulNYNMfZODvuoh8cr3XqSqY
+          FQwe72cAThbf0C/cBdBSTgl2auGtb3wOQmN9R1j9q0+p3aVEAogmit+kalaR/hRY6MaX/FRFAPNd
+          uidgqy/+4+/suqAA9pq44D//HLd4i/pn6eO46YJmDD5NhqxovtCEByFbMst+/tOXN/sFFESnFOB5
+          5waoC2qP2KJtARHxFxyfyDmeR/+r/ONXniRfiiW/HRJ4AsJh2wupxNIfXqVVRemJk0wm/OmJW/2N
+          Ri1F+rzVt8BpHx5x+PilMX1PXCL/3OuJGknUFDN7piPyz3CPU+8xxpNyfVewaM9iwCf7NJ7Fgqzw
+          HKWMWkRp2PynT298KPjTY9llkWew1RNwJmWMzVGN5IMA7zHd8Lku7vVo25smOVjhlRCsPBIJ2OwB
+          H69vz5MOXFkC/8ztA/7N3wDbnu9fPPvDK//iW/K6mDS9G89iacIqAePlfSbCVyZs3Y4FcX10odh+
+          fLZmIURgN6SA2pL5K1ZF1kUoUbXGyrObi9kQ5hkZx+eZyJt+9ldfgeZ4H6kDKpsNWRWI8v+npUD8
+          3y0FVZLL9HiOLcbX3bGUL4/QwFYMtIZPy55AhLua+tuqmal+nlyo79+A6tZjFy9JeN5uKVw7AiUr
+          YXNvyRlcboFFWvtFdRZ4XgLDYsqx+vJFfeGA3wJ4KQOqXCIfzPnYtKgv+5g68bH21ty2W6gD/0u9
+          cTU8+uVeNgQJzPBdi8yBX28wlYNpWz06I6MQf6NpwOneYnql8Z3N50mOYKi8ftjBv7z4bpsO4D3M
+          fWpX9NuwKV5qZMoyFywVwQN7fc4VGlOOYfuqPJtveg45NPvYxgoe+WLB3T4C14ul0Fv/1OL1edNt
+          cNcPRkC3Kbv0vVoyIMwNiQjCwpuzLIdQ7k8XXIWXxyAlryKFxsFcaH4V8mYxjEA5qD3oiDyqks6a
+          OxdCUq4Ee5+XWiwoHjRYtOWPXr4ntRAEtZ/RtecZPTe6WQgjekEYe8OP8OK1judA0Ttk+f6Fls12
+          61O9/HL4kmFHc7xdmRDeYN1uhX3wqU3BMAFequBN0ELsBu1eZ6l8yODDLvqAxSen+YFj0YJR9xV8
+          inLdE1e9FGH1vRvk3e+9YW0L7Yn6BuywLpx3A3PKRIaHXXbF6ktUG9Y0OkHntV3p+bOLPEGQTjxE
+          +RxjVe26QXQUz0LfyXWCv8+bd/mkwKLxYnLAu9CbnfTX/j0Pdph8iSUufXVo8W4fXPDboA85b2uY
+          HxSO5oPp6mv4TkPoCKcKa1LaeDznvURYSgUki/N7F3S5zHv0/b0KmqXpxRv9nc5D7D8B4V5Hu1kP
+          TNujJbyv9IqzrGGyKCZQfmy3rsfF9QTpsmSoE1ZIBvj22HK+XyxgHXOLmpceeqtO7Qrtg77D2WZf
+          vKvfeljs+pxe5JPPeFk8lwiwPcRYemT6fChZDgNfNfH5TFpPdGTUwYyWmF6PnFmIeE8D8Odf6fG0
+          L0gViRocPaZhdZKVmFd3+RYyRo1e9LEZyC31LUiOY4utY010qtybDjbzXJMG/Xj9u45hjgJfNwk4
+          nruYCW8wQ1cpMPngicS0dqYeekdlwJezauqiop5X1LL7gV4fz1Tn9WzPQUfAFVkoKbz1mt9EWejb
+          GTsz9or1bKgikjLVpRW+Rc2qXuoMvsz+Rc//R9qVbKvKK+EHYiCNkmJIJz0EARFnYINgg3QB8vT/
+          wn2Gd3bHey22JJX6mgpVknbpZuaD2r/3vT7qe0yUapalmrHpGo/F33lGrG2lJIxYIxb0BYloeBw1
+          nAvTBtHn/r5FW33ycbqer7lB1RvIx2Kwn9ueJlDvOEpr/gnmBt3dpbspT0nINAcrDdq4y9aIRWk9
+          v6OU3x6IPZJdA5jRfeIG3bmeP4miSnnUXbGZGJo2b85bB/osdPBB1gdtGiNSgve+74NwY5Uud5Ss
+          EN3j13ml0JM2X2zJgrd8E8bBUcaO7gb5LSVZ05BbeNx084kOBlhCkmKlM3VNqAK6jh4+DNguzQF1
+          cZRZ2y0jPoLhpIgaNV9bAyBxBKybr8ql7vSqJPbmlTjTiiKeTEFQkXPZpiTk8yNiLWIx4L2v+/HT
+          BGXNiVrpSFYVecQ20Rwv78MnBb2sCL7JlUDnZ7HJoZ1zip0TJ3aTGR4vIE/Gl8icv7jr+1uAr3xE
+          cMIdu6mQ2Ugqj/kbe0LRdpTzrxk472X6/T6NNuYpgt1Z5rFcfQKXq8q9hRL/JhB3t+CaHLaLLEYR
+          OxBtez50/e70AgCxyohO8VQ8js+HKF0bKcA67MyOezJTK63PI1HYPCgvCFkj+LFnESOpzmhW/W8A
+          6ZPdk/SbVFQoY4uH1plTci+auZs9yhrIm0537O6JR9k1P6JUCymWu23Y0agwPHhXso2Lo8+gOcLn
+          N8SepBB1PxuUa6GNYGEmA4deWrrcNQnVX3wRHOOru7wvTgBz11XYlB8SHWXlMElOo5v4aDiO9q3N
+          94JCzcpI7hkHJIigTaAlnIUvb+Mbz/YwMjBAScnhF58asS5gfIWK2M5ZQAs3uyU8v6878anPx/T+
+          8BJopsIkjsERbanhBlDNT43sH3ZVN295CGGjmyZROw46Gt5jRnofawX7F2iK5ZzqOlgnoyDKbST1
+          dFUQwPNO9sGOmadixpdrDvFu1AOW4rDgfFVIYL/7fHDgC2eX1bhNA84DHkT5vpxO0M5HXrqcFzIi
+          cUOL54r3yHlPE/bxx0A0jkJL4qyEw/jtpEjgNwdeGs3xid37ter6815VUUZ9DUcnw9Km0Mh6qev5
+          25hul4Tyj8ZIpLDOYnIgI9IG2XtUkrsfknHjG14nPI+WB3OgHn/5gU6WegDJvuQldopILTjdLi6A
+          jyZD/Hq/RQtbyYkUaDcNp86xLTjmQ1uwSHclCf6WaHGwEUpFbcf4VDWWK0Rl3Io2NAVJ4HmgLJlj
+          GXiL7bDqu6I7TRm0ML5O87of2GXX94EzHp2Vz6y3/Ic3g9b1JHjcKi7x5yFAu32o4XQjGvX06pOn
+          tOZ3LG+PT3du+Ku1ZbTXASeJqHWTdH4l0i++5WRva5w9jADZvOuJkzc8XTKF3mDzas7Y0PNLwfqz
+          nErdVsUYj7VbLA1XhpLVcgG5vx2eUsYzRunzCBqiCVe9ni7NlYXbsT8Q30mf8fwwPjqcrYM/hoNN
+          C2o/2hT93je5X6yau1wDHdCNn7FdXrSuhXJuwOvfHZaV+16jymYzwuly6rByXipEb8zkSC+6IcHC
+          90I334WOh+/iJVgW9EcxzgXygOQdPzKsGBbTLKfrV2luPnIr31neKUxAIoPFWH1+imWkh610NiQ+
+          4N6SVPziCy6qKYzSKSVodHnCwjs/hgHPf3V3pKfhhkw3Y//Of39PvTewCjkHYuKE9TKnrwXCJqgI
+          3uVGPGpfpEofsguJsuKbED4fGfrFX3BaPhoNDe4pDQmHyJW9nd1FZ9wG9Kf2IVa8DTpKvWsPvOFL
+          IwPfc9Ei7yuDOY0iVgzB1ug7FG9QTWE/fjz1WbS7gbvA7XvXcSC6C50z3w1ASS8ODl+kdKdlyUeQ
+          O/lG1t+PulyoU7QPpxabPn93aXHXGWD0bUPSZyvEc79+xS7vISG5xWxo32vPRDrv+DvRHyJxaVTG
+          DZg8W2I/qSmaA5ZOcLlvjzhjLZuyfSpFMN70Pck+XVr0oRGOQK9jS+zLZq09NQYPW+nlkPM2brX5
+          LVaitH/W14D3rXc9TNb9shPtcA42e9VH45vEFozF9fQ7v+7CcGoI8pzJ4xuFyCV1c7aAZ8MTKb4b
+          E7HzZ8NC6+d6sA39Zzfx303ww0diT8+Hu+JVBiVazJFzJzWeuFK5gcaZGvaO39QllqVHoMrve7Db
+          cXU37U4vBu72uMGBXAloIvbuBkzNeVibjhFa81uKxg9zJ4Fx3dP52xwaGJpFxj88mY1rryMwu/XK
+          6RS4tPRpDi7SzqOUBDd3ak1RR8Y7abF6K1E8J8Nq+fv3lgRV0scDiupEEp5lRAwUIm0+s/2CVv1C
+          go+od0t3s59waMZsrJxMQst9/pa7ynBOZK0ZFPPGwTxc3VQh+imzKYffU/i3Pza/Do74PW+c2R3J
+          nsfK/cPfXkoSkt2IXU95sFmQmjLyyE2fpKOKkzdQ7xMe32/G1iU0nd5//NqLXdzNLx9t4aCmAjbq
+          c1JTmbcBvR5lTMLqM2rU708LpN82Ifb1a8a8KvAlWvE2GIOH3gldtGNBFwsrYC/fvltu4v0Jix+w
+          wSIoQ3zTt4yHROdpEc9wnxph6SYSu62MSRqdv+4sviBCthJMgXRm19Gu2cRIT3rfkdP1tK876SuW
+          MM78jvgDUDTvg1KUVv0XTIl0cslg3xqAz67F+XZ/ROwwxQnSxuFBfvg3o9M0SdentSErn3O/h6oq
+          YbPdczgIj5t6CcnsSEJgTeSAsFLPp8IYf88fN9rcaQMn7PlffiPm4PjFnFvcDe3U7oj3B2Xv8l+9
+          KxHzsmp8W/d3bo/WKNqiNmA3a9/dvOpl9E5wO7Lf79btndewoBW/yN5wWncYXu7zTz9pSr3UXYZm
+          Ax1M2yCrnupmWxYYdJgPBbGja1pP0tplruWxusbvi676LRdV8lCx3pWfYiot24CXd7OCzbDboElK
+          7lskKhxL9nsuifvHSQ/h8yRl0GjTrweMpMI8huG41QpUfLioZcViXxnEG51G+4v3H3/0632G2mPi
+          9XAem3GknT65C58NFgiH3ZN427XEseedCzTm0ybuWWgLeqDPN5LnXCaO3bjFhBdZ/elHom0eWsFl
+          UdigMdoeyb5kezSxo8eIhhccA0hMqHuqXBkJ2HwJmMVMNGE3WG+4N98ay/c7uMTyDyMwwmeD1eNO
+          j/mkrw2pkr5SwEdn2xWEx1sG4xCW+GRjy+1bczH+6Vc7Krr+0hxZcK1UxQ61LTSzleSJRj0ANhTg
+          0PB8eDzs79coYLTW7vpWfTaIGcgLewO3xMt+kh1J8Lgzxv1+/8fnUCBcZnIThj4eVP/hwXwSCVYt
+          PBRkU55S9MMHr459l8t8zZPW/B/Q1U9ZTD3kpebD6+MsLGe0+Kp6A2CzBWtSm2gTrk4jcKZ2C0pP
+          aeP+k2sJDFa1J0aplfESTVsWeufajU0ACp2rcu+AC8kUiOeCajRmnyFk3+RLVj1RfFODN6RiV7FY
+          /QSvjvog69LuUo/EjpY+numVLGjlT4FYp4rW19dzBjug35XvvdFy/l5V+D4LDptVvMRj/TVUdLVI
+          PAqrn8Th4CACmF+y7tdFG3qKdSBx5pLg/gg1joMiBL+bHbx/73H809ew/v+A2apVsWyCMgMSJDJO
+          FThSgeB6gp/+LO7Xqp7zwBER+PRIgpVP8PXXkKVbOxXkp0/plw0iyGVPwYfEa+te2UQ8mtNXNqIH
+          +rqrq+OgAeHLuB39oluEXppQFcrzyG/EdzcP61ddm025jBu3TNx5FostGIZORk7Yi3Q67T4ZMFOa
+          YGtdz9WvWZCwPeyJNzVBTbI6HyEWzHRcslOLqFZPI6z5AHsFDuKZkSMDkGnvsG49iNu80X1CHzXS
+          A9q8bdTO7OENa/xj631bummzDXkp6c8FXv23Yu2U0og2tEWwNQKhGzXP9CQYtXCcrmlXLzv3m6CV
+          L5GAw29t2p0GgBf/vBJ5uH216ZpHqfRbP+tNrXje3tdeHgquyZ47yK7wEpIEgj6Wseblfset8QKP
+          vO/wad3v+ad3mDMs+Jc/OWdrs1D6ZBip3j3rKVS08A/PdbakLukfnPzDq3/6S5AikJZb6ZKoGDYu
+          UWfwxLt1bsYzFbl42m3MEtK6GgPplGLad6RoQINKJvjouMV0G1sGmo1Jgsfj0VG6VzYO+ul/73WR
+          0IJY5w3+Nozw4ftq62ngmgWCVvpiWXFIQR4X5g16WRKsablCf3gpIXv5jOTodMXIPHYpTNG7/cuP
+          tBTbHMSW1Qjen7RO0G5dBKztpBiXuyQmm600Ij8OLOL3eV7M0Tz0v/NL7OOr6ebvlV/+9IGXXaKY
+          HqdbiUS2vxLzrBI630M1BHEj+lg3UFrT/XObwbGWt+MsZUu8MKfbTQyO6E6s72FEvaMdGukwxwVW
+          EFY69ulYN7DcbUsyy36h3synROrSm0SwPmw1utHCBuxkU+O9W3i1kGUbFq6O3GKDGVTK1hFfgvzc
+          brExZ0onrP4LgBNeyX2Xv+MeXakO5/jGEJ/rgE7x82RIm7N7+Pk5NT07kIDZMwZxRXMf81YQvCEs
+          6YtYDey07lrqLfq+iDdWWV92kylsVHh+P3ds7HWTzlY9ByDql4qEzmR2nC1vAIqtfx2b7yFAoyq5
+          gNJC4MaHlC3FJGjHVFr932BXXupuYSsrgdMy+ER1xVNBX/CYpN65d+OhXYJ4WbSER2Wyx9iVb078
+          86PF63W+kvttqmsiPQcDRZfBCJYda3dLPcYX9Nh1eESlvqPD6sfC0z49R06s1E74Mg8HepClsVrx
+          hzbes4cperbYHO61RvPlKsL4Os7Bky2pRssJV0D2rx7r8JxpH4iMgcryecIFfZXuj68iQ6cZUb5h
+          QoVrLgRwnRqT+EW371r+Kr7h54f5q9+0OOev9cenjiSd0cA2Tgs88TwSCrpSzM5zKyJv2KpE/fa9
+          xglqc0FvbgKsUSug/amrRSjl5IKjjSS6M7wRwOFTKGQ/gRPPpuA4YNQE8Lp/7pKiMQIh9hBRPpvF
+          pSwVIhAGMwlEqRvdyWjQ6vcuMfFfsvnzu3Qk3+p0zTe29tM7aBM+MU5vV5O2u0G6iZtXew5YA6Xd
+          xOagoipU57FDyVebjzlWgRZ7GIV1Pevm5eooPH5YYt3yGC1u80zg588rDbprAztkOvSCN+LT6XKK
+          m+LW6Mjswfjzc4kqKC0ofL8n18snqoelzzI4XTt13IrohaZ4nAEkJAdEPT6Dmpc0nwVHsg5jq58s
+          xMe+8EY/P0XV1q9ut0/2Ca2f6T++2i3Conm//QvO0ZXvanQ7MWA5rxgHK3+YSSmxUGa7G8bnu6JR
+          eYIKhlLpsZI85Jitr4ccrfUT4sSB9ueHAWYMn1grPlGfX1q0N7YhvgRYRRTd7gwoXZUQU35cUTuA
+          CQByrmPtbWkxP5nV+6e/CL5NK/hEn2q7xgfOn1GgLRm84C+eMKNIxTM2wxzqrj8Rd/WLe9yLbyBB
+          KhPlmTRun/fdE2Tew/hS4DFeknlbwq5CYTC2Q4kIF1WsdLKeW6LeuIM7ibOnwsQfr+TKH5uYZnXe
+          Q6NLBVad3+DQQYzEsig9HObBuf756ZCZ9DTSHJZ6Wu7KAtUhtbB+Yaq4t4tJhotUV9gE6aj9+Tu1
+          rqx+WeW5bKv2DSDT3QVj/tJifuVT8H0NHvFuJzlms2mXoCRVxN/+dsPP7131X4D8G4em1lx0yeT5
+          Mhj3VIyno+a0IHkSS4K1XjBQVDx//hMxTxehoMkjTn/1BGI3rlePymgtP38eFz7tuuW7SdeuaseZ
+          BOHxXk/5OWp/+T2YKJCC/PzIw7X8jNKjvhfUPpk3YF5OjQP47opZ49QFtPP3RLQVT/std1ggCv3N
+          H1/lv/aWAe4Tn7Eehpk2ny9BjtpTdMPmc96i762OPfDmhidJOjLF8DCIgZpce2EVolATrh4dgVhe
+          jhMD8TXtP+pFalnUjeyi9t3i8h8eYvfbkpgRBDQuDE6QyolPYs2tVvSwv/Qioa0RcBO0xRLHvoqs
+          KvTwlZeXehxH/YYsVT5g59A7rlA88+yn10axTz06ZvBiYNVPv3xb059f6Pmt+fND6fx85yHabif0
+          F9+TUNAe5adrT1z51hbEPxsskm+PdNwV406bAzImu7k2SuKV3BtNb1e2xF98G/XoFKzIny+ASssO
+          xFrp4qXp9rl0Dx4BPnG1QoWTDa24rk/g5JLXzUZ2a3a8gSUclNerO593SyB5BnawGjYPNK/+x89f
+          GB9r/YtrUPWEH//L1/db+ZSFDlS0x53aBmhuj3Iv5dmsEW1qrZotbqUh7fNOI87zUyMqqRnArx6l
+          cN9vTPsLecKqx/HveXyyTkKynE+MPaLy7ixqpfWrrwTSYTrVf/E68afryKz+1ZzsmRy6Fk/Y3CpW
+          vMbXBTb63sSBYYUa/2FuCbxsYe36knIu+Z1PLWfGcXrt2W7+fM4AjuQcsFOnj3WK35xDHITZX/3l
+          z6+mzbbA8Zpv6Da46hCFePPjy9rk+jiD2r1cxo1fqMXKFye08oc/vjxeLocAVr8Vx6s/uUjtJ0Oh
+          wik42JibbpFakoFzEVNiTGOqsZehi0BLrTv51TMmbd/I8EF7hRTiJo7p03klcOVuKTaYbNAW3nm0
+          u5JEm794WWSeZ6XLZB3w7ZCUaPJnK4XlVrnBw7JcOqm9IkqhFPbj/OAfdent+EzabfITds7GNV7G
+          RpAhz6iG9+5UxV36aS5odxX2xFckQpeNnQNYvfUKxPPU0K8UHJ/ACxQHAhd4Ll/esqco1oWBHYu5
+          Ixq4WgKj8mYC1gzXT/q8hyph/huSdBp5l1pkYGHVw6N44sR62tzFC+gza5BEtvJu9UMdCaJ1Elqq
+          7oqxfQstIvmXx5Y3XdBiBcETNptqGb9ZpqPpetRFOMDdJZhRrvH0pY989/9cKRD+95WC1NdEsk+A
+          jWdmOAdis88vo7HYdj1rehCBlRwtYtnJEC+74RhJKqmXgLdNrqYahxcw5Mue2G1zLRbBynt435OJ
+          WCFn1KwQvFcJ3jg4OzWnmiur8xNBsjaS9K4hmo7zpEvhzmqxtW1dtKj3voTzwHgBX9xrl74O1gXk
+          cJfiK5Nu3aXk/BTSa2iSEy+rLr+rWF1aW7cRN3QDNDrVlgdUhgd81qMNGpIqWyCl4m1MhfUW+lmK
+          VUm6eRyWzedUt0z5yGHc5buR0utbo2ZYqdLk3h7jWPijO0l6FYJ1GDkS6+DU/ffuqLBJtyx2/L52
+          R8yVFfjh8T5uWpVDsz1+WIi7fYRPeRN3nNkqrQQ9tycHLiq1gbuwARxd744VTcR0OmtEBpnFD2yi
+          +7cbq0q8wX2AN7nnuEaUoUdL6vHRIVeDQ90U2r6H3vrrhNXNeVNPAtFUKbxvZXI+pDUaTmaRowF4
+          ffyGTYFYZac70oJPz3EGkXXna3TI4aRELg4G9UB5d7LL3+/Hylh03dhzgYUG95LgvTkfC7bm7wZ8
+          TzcdB3u/r6fTRp4kV7JifNtlPhK8bKqk6bb0Y7Q53+uRVeko7V8yJWlvZi7vVBMrUfc2BoCTu8ub
+          H6RKdaleRsDC16VqBD3QcxkTz1SNgnt71IO3uKg4YPyw5nL1HACvPVV89TKjEw7DbQvnOcxIIs5J
+          vciT1YPj5JTgVj0inkwuA8lgiVhvi6wjkZZbUB1zlUTlExXkOpxGMS7fBjE1ZtBoWBqeJIxeQdIO
+          dnTeiN8tdH5cYdO9GwW7m8UKSd9CJBh6KPpWuVqAxyok3iXTCv67b3Pp+CUEe+L9Xgsn/57ANWgO
+          pBhUnQqpWzbSJkjUYHmSc0yzIHQgRfMJH+L3x+U6p6lATTcKsV6p3AkXubvBhW635EjeYkcHS5yg
+          b6cLVi7vRpst77JAez7cye04GjV9WDsGKrqLsDVGN20WD4IKzovf431ozt1SZYdIujPtbdT7axtT
+          9SPocD9y5cjROetm1ZsmaTqXBg6lWau5/MEu0vWe2KTw71nMRRuPB1x8J+xW7droxuCfQNOtjPEh
+          V9z5oT9HGBLFIAVKsTYpO8+CM95+CJaxWrBtn/fiV18b4x3pROnf+thcOn4c7tEJh4obYfRjQtbn
+          adNo1w3U+svBISZBx3fS1MBYXwp8I3pd0Nw6OmBoeob3Df3WdBPv3uAwGU/cSNhr0w7iXtre3+74
+          +cpKwdunlJdS5twGLZP+bonlici9qisxKTPFU8pgBsJhyLCvpIlGHdnnYWSdGAfJeqt2OQZPQIzZ
+          EPUFlivIg2iAYGkT1rvGjHlnyGUILrcGq0mwKWaPNXhoCZaIU7pux21yvwKWEQ28t0NJo35oyMAL
+          7z1RjLcRj1/ysWChx2Yc8YOng7Ip1xIXnrDv95O7gMmPEk7PDi5Edd/N/L3u0e3BXYkrqq+OV1t/
+          hKp8mUTekAPlo32zQBhVScBFtqYJOm89YWcEFQ7T4Urf+dFjwFs/rjlPblPMZ06+rInBJJp2rRCb
+          pUsvlZ7skliap25JBVaVnv0rxKbHBO7COG8P2P05xEf+SjR6fjuqmF4jE2uPfdYJYme0qA30HHtm
+          77rkGWQAl71bENt3X4hW7uMi3bs0JvqWvovleRFuEpuRIgDvMbrTW29amOtLQy5FXcac4tgR5Etr
+          EdzlardM0qXdXfzwQQ5n8VCz4BUgeLvBIjdhZ7j09XZYMI+OjFVbTKjQoVaFu/o4BtvbvdcmPR6e
+          sDk/zGAKn3O3HHTCwH2OloB15kPMHufJgJSvZqw6z3PHnTnrAlLI2SQJGcUllWJFv/jGx6m41lR5
+          y5nEtLTF2gHceAlkZYHpc7oR++OaBbeoOY+wYKpYbxm1pq+3ykuV0zTkHmekoFk2p+CeWYucP1yv
+          jXOThzBZfBCg76OraTueE3DUeCAaXE139r5DA0wZ3YIu8su4M8NWhmB7ehH9HO4Qdcpq7ed62wTS
+          nJ4KdvOWEmjUB4ywEbyY5U9nC15HTiS2a7PubCTGBQm7c0mCXTagqd7YTygvo4ZVBN9uGbNc/+Ed
+          CY66grjOKStJ8Ksaa961Rov5QTIYgl7g/J2sXwkoUyL1paatf9cQf/GLCDp1SbEWhwml3yy14Ey3
+          Hc7qZKSThlAjJdvjZmQHE2tC9xpZiMunQS49l7u0kfsI6ssYEqtmaEd1Xn5KcR2ccfqifs0XnWVB
+          1/Qsvh3qpqaf0mRB2/c3Eu6rjTu9NMICKqMDPt/SZ8dvXceD58HdkFxJPy6NUZNJbpWJ+Biac93R
+          eDshv9O35O7jWuNyb8kgDk+7QMr7EPEXdTHgGegyMSs6d9OKr+BHlUnwXX93nb/veTDVuCTO9X7U
+          1nyXgj1uHtjP+wnN+09uwdHnl0DajDpik0kIISfzB+f1w+gEsk1LKSf0Qxws9/U0GKwlHaoiIcrm
+          7cTtbz25Umux/JmVf3g3JJqBAzI27pw/2El67XNMbq9HVtN3JDjSD6+s1q3j5YbkRHqwozMKxszE
+          SxzNF2lxDCaoLyFHhw+uRnHlG/hYh0w9ar3kgSU5xvgMorabf/yNWnyB5TJsaO8mTg7vqgFsXuSy
+          o5cSq6BUDBMwKz9ghWBsQZwrdkSXR0tHq7QT+Hw3TsDzyad7qWsj7RPdv4lzs89ocZ+khINCLtif
+          1IG+YWFDSdMMF8vL3ivY9Lt/QgRPFe8ju9boxs4yiDTTG5kWy91yPN1bdD0w1+AjcjrqQ6ZjkHKV
+          7ziwdEXji0524BV+EoI/OztearFPJT4V/eCc2EeXLZ+lB3c/GgPOjcZ6PsysJX1NXyDX5pF0U9Qf
+          dIkdg2AdlOWiOanCSXpZcoV9cVdp/fz0VfGOdwq2Jc6nTaZM6W5nsVzAEvoq5u8neYLAbGOchs1I
+          Zz1XbtK96WV84cbSnQVZdqTjrWoDZMkT6qGfUvjI/WPcrPFF47DlIblcVZI73FjTB5P1wE1DRczK
+          7inpF6GE48zzxNiZGZ1WvoqU1+cwwr0t3DnEXx06dUpx7DYy5V4P7g2EqzRinrgqbntXyaVv9bXX
+          We+FSwdrWcC9XAoSvvZdPX3M0JHeQ37Enr42WkvJ44J2wjsKJJpZdBZDa72LY29ImM+KK6Ry0KL2
+          c9KD+Q6aS9TPRgce+I44Um8UfCArk7TTnpgomkjQssh6IJqdcRtnZXsrqOEvJbz82AqO3pwWQ3WQ
+          chBlV8f2yu8J2p4SwD6bjX3mv4rlBoKKNoc2wEG4q9HMVJMK+XQvscv5KmJBOJTQvw5fsidvsSbX
+          4T6i8rOzcGG6A12i5SrDpTB35Ifn4xQ2+U9PBPzEle6Ufv0niA2jYnwf3XjY7F0LiLLpiN3ITizs
+          h5sOM+uUxDK5fc0G50cOx+9AiG0kczEMoheK3+3nM1Y7cevObw95UM3hhiRuwbs0iLgK2sW4rfho
+          xNxbzWUofMEbi8CWNGK++hyt+0ey0H2gSZvzGwiWMhGFi2RtfquRDHemuY1LONuFcCmxDObQz7gw
+          HrZGhfAVwTKcbXw7qaH7d37oxXpj/S1uYnqR6wsSQxZIqAdcPd3PRgDnnXsYJbdl4qnT9RZc2wmx
+          69+3xXyAa4Lw2kBdsc9CNw2Qb+GxYyasSM8t6jeXiwW1xNGAZSKfTg3zeoMNp4DYz/vsTurTXmDr
+          kS5Yon2E2ncnM9LSLjy56bKE+vxceWDPrEpOxnyLV62pSvF964/bJVXiib6hhGexy3D4LAzEj1lu
+          IKs9vIlRm4LbGy7aQnGwM3IothvUPztpC/GL7/G+sZ90+RLiwPA5HkiWRFBPO4NlpP3lpmD5Ig41
+          TTsvgEQhXxI8fazR04B6dCn2u3G7uQKaEbBPQDq7wVdLGOI+ZTCIE/AH7BBB7QT92Ac/fhpsX9ei
+          6Djt8v7pWaLqxVBMLOt7aMPCORgyWaz7nqgBRMIeE+Xytlxh3R/xcOarNX4O8WyfhxuE4akn8kIO
+          Gt3RnQNBdTaxt2933Swn7Q0osDesKFummCJ+iUT+4lCituZbWyp/a6A1/xPF2o/afD/KIlxfYUP0
+          rf2op/BmvlE5qZjs/UQuljQLb9AWuwWre8Kj7/VTTCBakRNsA8qipcbvCaaLcw64n14R3Vr/02fG
+          /X1FlAxcCr/f725lVnun1dxK21fZkizkjG5xn5/qh38kUB+pS8Ou9MRr8pywfxRGbZg4x5CCM3sM
+          xMH1YnYRZQ+KtGAC+miqbuX7I9qQ+wsr1j7Q+GteM9I9xv5I929c8GDyPfRDRnC0vj95Fd8WgjN/
+          /Lde1VuY4MdX5Lkw6146Pxsg+Hgids51xbJ/aRUYcFKI5pKim9Z4Elt2nHFwkzfxIl13IgiyJRAv
+          VM1iLk/bDOVnwSbmpv3UpHnIDbKLLg2InF7dqWGG985NTkfsmu6A5kRhMykwKifgX3SoF9wzKqz6
+          PqiVYo7H99FfdkLJkpEtoq7orYm//MW7c71zLikXvkcN3fXYtwS/WGh1D396BfuoNujXsIcITlwU
+          EI/P1kbNzfmGmreekDOiF3cRs0iU9Dw748stnbu53QSXH58jVmSH2hzihwHsroixD9mXNtvng4Vq
+          YGjAFH7gTgzajBDSr4D3U3HtqEYPOno1fU7u0EO8hOss8dseynHaQklpjMpMMoqvTX7rN+27+wXW
+          eMPu077RSdj1a9e5pR5h35670c2DBNlpNuGUT8yaXc8XpHR7w6lTbLSZFR8GzMy3J+qT7Ao6dntW
+          +uHdVL9rumw5gwFSrbOrL5kWT54X9ogpw9vf+SOm+F4H8+xWvGwrNAo0DCXnxe6JK99P3SwI1yes
+          +j+AVm0QTRZuCzndU2I19kz722Pqf/wVO+ljj9jR7lrwmUIj/tRH9VRcziPi6JHDmM9mSjRJjNC4
+          z3hsfJmunuVvz6MnO56wis8Qf+WkuoD2zlKsXxkFzVs/ZCV2rnKsfN8W7VZ/AP38Gltyvy5l5ZcO
+          WXqeiStlMpqeKS2lefQU4hO/Rf0zReWfvnaIUNXTmOistCibbdBsZVObk35mJDi0ypqPD9osCMf3
+          Ty8S+fqMCpZBQg+r34ex7W4QXf0UxKamhVc9Rxf9tPDw03+nGRg6T5yqgzL0J7xf+RO9v94ybENN
+          GCkvgjuWDp4QQtcj1pnIR/whry2p2tlbElSjGc/v0yUD/mJR7Fe+o00yf2kgmwUN+3N6iqfMO5eQ
+          7lydeG0/1ITLbmtJzdCD9iWP8ZIOEQ/J4Ij4p2enlY/BV5EOOPNkjdLUbVpY6KkZDzQMYlqlTgv9
+          K/5iS7879QKGJUtj+EJYH+YYvcRQTn/+ETbSVi6EOKxY6OkpIHvj6iJOfVAHIBU/WD4EaUGP1s5B
+          1c7dBqIlyHEnhEMIa3wQhQmkbh7zpoeLWlujsL8mtHvI5gXqz+4YSGu+m8pn4/35Kwp3FilBA2/B
+          BW9Z4ja+qnGTo5egLhxDVnzqZkdUWOm5YyTskUdYsCvflKIvqYjKFlP3JUv0/Okbou73iza/2saA
+          rXnd46gOD7Vw2l62YNfT9eev1fwhVAASLlICFAkvd8FcWUr0u2GIblHNXepCniBJ1kFlghnTaUG8
+          KBmOdMDHYL+howFSBvjgPImpRCmdq5diSI6TURJ9r49ichjTQtmtFIitPUp3sTh3BH2LvtgirUPH
+          XFbfkH0JEC2fA5dTHCWUHNydR6pdVTr1RPWQZEWnERkP2+UYNW4l3Xzsxtcab6OoZRVq3kZC7DhZ
+          XKpZcSnRQ+EHYsmFxVgX1gLKFl1+fIPSZ7ZY0Mr9SGIWBtrHQR9AoomHYHPXjXrc+McejZZmj+2c
+          tjFNFknc/fgWFmuMaPOwWlj5JtE+Zqt9kypcYP9SKcb+7kOXpVMW4EUvGyd/lmOeQzYD/ouPxo1Y
+          E1T//AQ/PN2x16oWopJRitCfLgdyvreBJvhIe4N6N0xsGm1dr/gjQqIMX/zjT3372JXwDAyZBJqb
+          05//C8dvqP/8k2J5mpwOY30rxpxJO42tvkYGxn7syM8v/PMjdf6Q4vycCO6iCWkGmL2RcZtEl3qQ
+          TjcV6u3Lx3jCr65/enkKj7R5jRSuH5fyY+4AujB7bELidYK1SDoUO0jJyj/i6Tw+36h1byYJKr2q
+          adHMJXoKSbvm8083RAmO0K7nt4HIqiWdof2ug34OhKiHYtSmw93KwXhnNf7h0fglxELRxjuSY54o
+          LpXrcAtrPiL7U+PUf/lps9AZa98wrRdUOBkSu+cR23zCdsOLsyP0i2dt5Q+LJlsAVfkxiV6sg5nk
+          Rx7C6lcTo4Mz5T/V7QLu51tivaJ6x740wiOJEYdRcDilXrZcwIj1pQ9Jtm07tBiHDYvS5uqN/K6t
+          6Hz45IBW/kvc/H5e/TymByfavDDW8zKm/Bg50i4JTWzfUr3mswvfAIPu1Sh+15K8zCctbBN+CbjV
+          b1q6dR5ss8vuuHi7U0Hd7cWA8MXfiZwOVzQ7pGnAuMUX7NiPvSbALnMQ07gnouE9T4k0DS2g+zvD
+          uNE/9d95xmnhBFP/ftLJ468XkT9XJVGzqKr7n97z5NudqKZZd9N87lh4l+oX68+oj398TSL4dAom
+          NthSEr8GD2K+eo4SzzGo/yYND4Gx4QhGwi6m3SyzP/9x3K31Csq/SIs+Wspia2jO8Xz4RAyEdeAR
+          xxLKeNoZwMAoO+HIjdHBnfBbNNBPb4crvs/q6QwgQZiveOcg9npXPDSr04SVYruh9OfXaoIREJ3Q
+          fcy9OCX86ccRVv3aZ49Y3K3xiYPZx2iydiVItelf8YqnlD+qx+jHr8ed5Norf550KG2uJnK2bxGh
+          8bSA7Wxua1dDoo3upJRICgUbG7coWfEsySUjRNPIrfG5fE5SBLZ1dEmweeTudLjLGaz+G7ajdigW
+          Kc5G8GnXYuz6c0Hqga1gOV06YtwiNi75x5LClHZv7Oi9WnDtURmldnNziPaiZb2MWWRA0pYfbBIa
+          FVR6exfURItN3Phho6UZpstvvXA4F5+a+kWpguNrBvbxqNS8PG8yuJ6wFnwfNofoJp7ff/yqsBOx
+          mCKlKEH+yhein8whJkHElUBv0464gS25s3u5MEg/LQZ28/vOHX/8L2aKayAG6QfNNYgArXsxA3gJ
+          DqLHK6pgWPRu3K3v/7L62YF9kiJiSFxJl5z7pmg6FDIxP01SjNzUTLtf/p45eP72V9/FiDpBvubn
+          KSHWG1Y9MCLKcaifhLCH05kHHLB15y4rvklrvhjn4z6gfM49UmnNN398clKyzwJ1WzfBmq+Lljni
+          HH761n1zR3dePFB/+h6v9Zr/AAAA//+knUvTuj4Phj+QCzlJ2yXIGYSCHISdoCA8IgLSQj/9f/S3
+          fXfvupnpIpn0zpVMM3HTocuhArMXPoXz7F1p2OSQW5sryS3+2TGfjjnI7b+C+PLh6G3GtU2Q+nxG
+          C8Uk6Hirbu5IPUbmIq1MLD+u7HDQkoAQFHamaf946hTaGr6lYqAJ4Nn1iBVtjFUWLiVTpDKEe7Gf
+          iVeLT7ZC18/g4r6rQE7Fh7bUdmQixX+52PrwukbJaCugQWKPMcJmzJt7XUFff5Drc4y19bKFEFmT
+          fie6vHIT5WqXgz9/6nSwSzG9gUYujV38j+fObaIrsGDyhFWhd6aPsbIE7bd1XeSwTkpqekyCcXcq
+          iJk7g/fjNUDeyc2vPo0HjjNOv/oTV25dxUt6vIXAHuMea+eeetuFugn86kVymr00/uWzn74kxTdf
+          itfzLgFfHkicely0lbcOAujqrMN6YwXxPz5rV/eAfN/PaeSN7oRkxdHx5dv/ISp99/AYLeUyfHn0
+          V++0aBC2E7G/kyXUFL58dNyEZT3JlG16WymQ85sMl4xP2bcfR8GPz5tHx4yF2VNz6PtqjE23rD16
+          y7JZ/sWPcuiNSejkv0SuU7HBFnH8iS57N4cWm0KshX/3jpl5vKDKWhdSSgrnLeMLVVBcmxexnmOv
+          /bt/Cl/LIkfsxt4XuLnoEXnbt/+OytnzLjJ8zc8bORWPAIhL8zdAxZgj4sLdvWQWNXpYWWwhgaT/
+          xZTjjOD/Wnwg/e+RgmmIKXE+5cg2AysQpVCJsHd34nh92N0OjjvutQxtrDDOaeIWibqt4GvPPMY+
+          k3BFz2WxiOk3D28e1D+IEpzNWKPhMebnyGngw1o7fCzcSuOircuhtKMlUY6y7XF3e5pB0SVnrKXt
+          4G21tlfAWAZkkT+km4i++hmILraAU/dDptWTPj1cljAgvn99gVUz9gksn6JFMK4/8XY5KmckKuqC
+          7d3+UTLffR7ROWt9HIBT3vHFbuiRUSk+iToh8D7ZsGYoS5cX1tOojF8/+9iRpUBKYAfE06ycYA8e
+          gETl7RiPN0Ma4ZkGHVaj3Y5ti80H8OEIc1ABZ56YXkgJ/BhBiw3o/HlcogQuPPeRQkr1YXpsbw0L
+          XMX1FezePOjGILcq8ED+Bbs1sOIt2yYZSuMhIwnKtmmtj4WCBDm4kUoe5HL72RdFTojhBSoT+JjP
+          0PMI3yQfvrvJ/IWdIaDjsAxzXwNBPLIzcv4mSgLsJRNRddVESbY/YNxtrUY1vUhggq0aGyqLJv6a
+          8iOE1wRh5dEAsB5f5xbKHl8R4y3Z4HN5wQbWF33Eapv2YFUtukP+6YRIsH+LHd1H0xm+bqlICkiD
+          Unjd4wpJT9387m5VgZiPdQ9B/zGwHhQCoAd/k+D+5Dvkd76FOILoIKsNNle917i3zS3weuc6nMYv
+          fxLl8T0ACpuW3B9Np9F92FdQLNkpAFIGSnboIhny2+6zMBq2cc/DIYRb/DBJEhQZoLRSc1hZxfej
+          XtfvVvt80FHltTbJxfYSz5I5uSANGwUnBvwPAAD//6R9Sc+CzNrm/vyKL++WnMgkVXw7JpFJiknF
+          pNMBRARUZKqCSvq/d/A56UWnd700xoHirmuqm6orEHbHboRC8337sqku7ksVrgQKu9yY2YCF+foW
+          elMOljuHdPPUNsLJmRQ4ucEOB5rU6TRQpBE+uQuDTmdKm2WnewpIAZegzH7fXBaNQQE/EoiwoTae
+          vviXZyGPJ7jg2HJdnTZx/gafsHTwdr9cXB1ECSZ2gPB9aGBOPtK2xA6xiczO6dyJFmoPSKiEKNPR
+          G6yPYXLg4TYJMyzyaHtOcZ+CD3AecxWPj2aW0rqSb6dvgLbvi5dwTGf5I+0jnPKHJuZPNyOCYFxq
+          hA5G52JZ9C5A/2rMzDO57gqZptRy5cw1dkn5BOtdHjUAk6DCfi0L7no856Lkx8WMfvdr0X1dlBlP
+          WOcJ3CJX4Kctsl1xin94QeL0yoDGuFsoqA9qzjWpEsB2d2yQGaZSM4+N0kMtNcWZVTSDckpKNbna
+          9oFRVkEEkyjqrFynkPf3g/LK8UtNZrkq6hJ7mJwB685UA1pqiPjklIGLU4fXZM1KDuiyff9a3+YI
+          BrRi0Jk7eTF9tvcOqvB6w3b/iF3+yHIluAjxBZ2y5EDZyYwseO79M/IqaACBUQIGHo1PtR00Ql0i
+          lCwDqiQVsW0KEl0an5Hgr148nb7cBcxhDW+3DOOjysvNlL9CRvaqeUQuroKc7LmnJtukK9D5obQx
+          jmdK4GmiIz45uzrnnUF3ZBTxB/zDN96c61pmjlrt7yGCOrlcJA/GmvZAVzs+DoIIuxpmKTFRzH59
+          l3o5NiFf5BxWsRsO28mtFUwOroydbtq5y3xNStkoeh4dYjPP6fGTK5J1/3ro/vT6nPLtrodByCDk
+          p/UCSDaMPTwGbw/r38QZqFsMImy6ozIzz0p32SLrW3g58Sd8mptPTofXUEJB0Wd0KO7iQDXdiyBG
+          vIyPcX9rhMPTMuH00Uqk88NnIN65zeDJSFps4PoS8+HEj6DpLRun9zMaZi9TTZjyiuFX0L/oPLI0
+          Xy5DSUemq1rxMiMkQjl8z7M0BD0g75xksgrPN58+9pnONxWvwY0fUPCsciA4ySMRS1qs+O4D5K6y
+          OXTgDRqwtdgcGqL6bgcn8XNCh1dug1powh6eysicyb3Tco7CSpEt4VojD1lWs+64b/13/Wr0WPQf
+          HkMWjBCHIVNSsp+SEkbGHPrAF3WddS9fD+4qE2G9ysZhOT+fLbTFVcOnjynns5aCAPzwQgsCxqXx
+          5cnIG/76fMoZAx+gG4R9LUB/aeNTvNVHJR+GhvrMd++DJclCHgL9ePKx9q0HMh/uFynXTsjvz5Xm
+          /vhP7pJXhw+7bReIp/LI4N3iL/4qiLw7rCKxZN5Yv8gM5Uqfla+cwGtEu/mD1soVrHfpw64rJhR0
+          RdysviIVIEruBjrXcNGXp3JNZZpOk8+k6q5ZM9RDKGimgC1OL8G68/g3NNXtLOosiIblUDQZ2C2S
+          i261eYx51pta6NuNixQh2bqo34sGBXsXYFs6W4B9gxDKvvtOUPheX2DD2xK++PcTFxejynlFO3ry
+          vg8mFEufJCfvXMygidwBHytVpXy2Hd3FWkmIDE9EA1nr5AKPZv/d8OWQs3WZQXhnImemtzKhw1mO
+          AzCJrxNWILZichufpWyQCCNlqwfWAEiEsVwTpO9eh4EyD6nek1ALUVL1kcuFrWfJXjnUGMVwjBe3
+          ioncuqOI/ZesuV1cjTU8SnqCfvpnTVypBc+YIHSbRQbMx3Msyulzv+LYPBnDsrSCI39fBOATy3iU
+          5CtM4W+8xYPpu1TNSwO2+X3w+4t80VcOdhFAbMqilCmR/mUMhZV7dhDn+jb4+XgmFgtbdxaRRnNB
+          Xx332YJYUx6oCD/PfHlIXgbtSGCRxR+afA22gxEI8hPsyd0hngP/7Eg+p97xCRVqTldRdMA2/7By
+          Gfth0UxoQY6AFtk3z45nlgsqmX9f07/7L/DtroOfJjKx+64f+tJe/A6ibC6RfzrNOQmz2xsOJlaR
+          ppdmTJ17NYMHoQq22+dI6U0SZ0j56owvZvvNp3f7dmRZTjyfDc/+wAMlHaHajifk7qKVrrLZdPKd
+          PeT4+gUs+PElDPAgzhRjDJbTTAPokfsDu+rzrVO0lAlIbrqGLFhZgOpfuQW/+vvhPSkkw5K28UDG
+          5yDEpPUSE9yjavwPngqJlMHu0i1IK71qINOb1DKUGc6X+K86CF7+MWFSwxe+/fA9uhVEuuoKh2NZ
+          u9Ht8ylkjkqNk3MIcmIpCwMTdn1iJelDlzvcOw/WKKr+ro8uoaPBijmesEOc7xZ5VaW88esf/s/1
+          U+JhLnx1/IdfVRxf5GLQZ3QUgyYf9Khr4cAGBY5y79zMxicissEfdjgJbypgj4Hgw8GcVH8ZkEFp
+          b6qiTLrvHRfsfMy5oQ+sHz7g7UnxnBRZ/wYzMV185BZnID++aBdnQuaZ2zedkc8XuCv1A1LpqANy
+          KraNls/NOsvVILrLGKoVdLJji/y4o8Nawo//41d0PBFz4M6q64O92XH4tOmn17GhFtxeo/PnFuh/
+          fuJxd4/YOPMK4LfrB6TcHf1pJxn5DEu5gO5QG/Nq7126mr7ngI2vsXtdt5bIwz2Bu8pAsyjqNhXM
+          Xcvu52En+pUbUzBlpeNB6T6kM6+3Zb7UR2yB3/jd2Gs9rE7NvGE83D/Y0ZFJOfwmK1BwlKJjX3x0
+          bPtGBi84hXjTp7RVp5cPgxAifLcKvlnA22CA3GQ9dtnMH3gOVoF88rZdMB7vrsGSys/wh0e6Ho50
+          +vmz8cy5SL9Wc7zxUwXuY+HjIOuEeN3bpr//6Rnvc3V0wQ/hCpXphNDV4K2B3YVNBLrqnSBNElWX
+          k+G2C9WqNQjh42uYLt0+gd/deEdxS8V8yQ4kg6EeLtjb/NKy6Z2/+jqNxkxp9Gk1mS9nF1+ePsm/
+          ouiy8OdPLVvLKPl2ZQS5Qx7NdW1I+uynqICqwlmb3jUBt+EV3PgF+1nr5vxdHhX4ocoe5+IFxDOy
+          NE9u7zWLbCnkmrUonRIypQhwFuq6zqH74suPhx/4rPHQ6UjY3Ic2fx9ncrhVlJhrlkIw0tpnVuoB
+          YoCjCGg735CNlhQsQnAfgTIh5GOhWnICDeUtXxN2wceLHeUjt1cy+fRhCqzb3cMlHq8RuSLO1WeP
+          94dOSGFnP36c5XJaBtoJ/QWE5+OE/UE468Su8hp+L4qF7cJ6/vCvl9rTqqKTXov5rx5hLFcEeW9e
+          AOuRlQvgtOoRHwT0oiSPIxZu+nUWVi6itHplJTgvQo1PpbeL55DXI2lfHRnsXkOfLr5fiLC5xAY6
+          jPEcr717JdBXuxUZVk6GFYcqkSVd1mdwRly+tOW2kfmxpdhK8mVYyAswMBbKGOtO3wzzNYtM2TDW
+          O/YNVhs2feBJOfeR5o9yvuS0CDoNAtvukWmVRTNcWq+XUy31Zlbc02aMu70Jpo9S+vSyK9y1ivNE
+          GvGzQmVRMg3Vo+otd9oy+s/g6QD+zW9ns2//Nz099hT/9K7A3Wt0U/bssKLH1YP34N5su/b1+ZIb
+          oIXWdNv5ey0DYCyy+i1bOiNgV07GhjJKCn9+CimIRFT4+ek4pJEvL5riztpyMmCcQxVnbULjTV+m
+          QAMjQUgxedAtoGfAdSYx0peJzyfIPQN5yy/8yZI7Sg7adYbnNLthb5mLnEzMqYJ71gvQgZApngNB
+          MeHM7XOfq6Rr02//X0qt8jtL3OeuL6rwIFB+np/I6t0s/tUH2D7vv1z/CbpjsPOkpnTv+HB+Je66
+          PJ6RjFTFRhf5xgwrX9jin7/Qv0nfkOFa+PDI+gayXIONyXz6ilC+5gzyxtCKSZVaGvTz89Fnqz7S
+          e+Nh+tB6VhF+mC+c4+uQZLDpHRsplX6k34h2BZQ5/+0vjzOgS34/pPKqPmRkNoFNp0QinvTDc+5i
+          vvMVk+YC9rV9RIW6uA039KkF70zgoMhLlpic30IpXT5Vg22wFmD9tu9U3vQ9tsUYxWv6/PKw/OB4
+          BlU36qQAFoSZ1GrYk7os77i9ku43vTnzm/9rt/EB3Aon7MTqN6dreiN//sZw7o77599/ekXl3oJO
+          +3spwn7SBuTexiTn56Vy5LcSCNjf6nG8COkbqvVNQG7hntzxuO+d/TsyKTK86z3e5mcprW4/Ir8p
+          7Xze8h3pl1f4r7QY6GlwU7jpw1l8xBwgRXTRQDGoM1KSftn4eWXhc+i7mRcebk4jlHny+CATvoeX
+          ZFgOScnuxWNcYqsNPzG1kO1BQTMEZFevSl/EtI/gdPBqrG3j2edCaEL30aTYwZMNeNFxFaia4IUO
+          ZkAG8gYhs2f6LkMpffdDv+EFxKp0RKndTPrq6HcNSqBQ0PG163Sad3sWOkmU+E/HPOR/8wc/bxNy
+          hT10iSVOCSyvTIUNKw8aLHgTCzY8QT9/tG7+Dj4YzKPToik6/34EEVQjKcQWetJ4kserBL+7+e5T
+          r3N0dgE9hJFoQBTJ8XsgGV182YSTiQq5e8WLs68zeXpYBnKAqTTsiTsx8HEIZnSIP1ND/Whp5R+f
+          KoisdOHtAwP1p/PyBegRly5LlYHUQgXWakNy6dVdWdjHlwvy9sfY7eOXvfHFY0S+uVpgLUqtkHlP
+          QrOEPWGg0DZLmN/h9Je3Uvr4VvCXT6iqd4wFm1srOECLQ9a+kik2biSRaXv30OGhGDFfnpQULmZx
+          QHaAnXgN/LMlk2ox0I9fOH2uI1kzdYwVjpJhqbtDJQf+0faH3cXUObXSVjCJbIZulylo2IP7NOR+
+          Ugaks5Mac+pRhLBmx/PGN5y72GvUyRwwYnzfxp+uwcmXcNhLM1PtAkov8FLKGx9t/mV1yfOqe1Ao
+          tBu2hfOdLlXFrVDHhxybw/fcLJ9RmuHt69Y+t/n9hZhRBje++MvH5muWmfBUanufHXULsO0pEOWz
+          P+o4Kj2lWVX7AiVFcFafSd8vd5vvmTxnLpzjvMh1wu2tFA5K12JFjL45MZpVlLUrN/iQnM767Ohn
+          DbQ6+5hbtmCbfsc9K6CTffrnv4vA4SLYJAj4t8tEmtGbGA/amZxhL9XelBJbCsBpx3LYFOGu6d70
+          lAHp/Fyx76IjWOOC7cB33yk+vximvj5XKMEqyURfIPQWr3fuWgDl/NaRdhv8eFWcnEheykfIHpRD
+          vIiizsNzUCvYMJVBXzk2niGM0NWX5+YY92607Zqx6VG+t/ycfjpTgafkeMVmuPfAXJw5CaYNA/7m
+          B30a0IBTBVX8uOQvnZsfXwWi6C1i1PdmM51QOEP50bRIv+2OrjB6BQvZ717B9pafrVs9g6Kxz8iJ
+          VTteT8+UB2cvn+bjvDfdNfDvlkSOxoy88j66K/oMBI6g8mZp01PC60lNaOGHhdHjrOvLfpe30j4t
+          FnQTWXVYZ1vSAIfeKvLqm0gneXz85WPbfPhs87XP4DG47PxF0i8uGYS1hfu0XLBp1BiM6iX1fnnI
+          vOuOEsUvtRhhGYo6eoTSCiaN+VRwFpgWHZWbPYybH4c0xRP2qusjJvxjT6DX9z7WZiZ1qSwaCdjw
+          DB8N3mp+elI2+73jCw7JB/6VWjXo1izY9NIjnqY0DAAaqwxfd8N9WH/8fHjNjc+ZS03H1yD30Noh
+          MMuLVrlU4AICR9NIUVLMr5jsPd6Am1/05e371rFROmjrF39moNO75H6bVpjK4n2WvENIydeCs1Q+
+          +D1yw9gauOwgZjC8Zjr2wdTTrwz3Cry2r3ZrUTIpd7usJUhq5uXvXW50+yHCFfjNN/9yHHN8J2or
+          Hf1Gm8G7frjrqPURJMhLcOKJuKFqU/fylvfNcJlhPsdGz/7he3TDz2bTK6Pshu96q8c5XvCbEKha
+          zIwNV2L01V0jCRgOK+H8Ro4NL2gggF1XTsg48xVYpJLz9i9SG1seNVNi2p8Smrl1xcmod3R53csM
+          7rTwgw/XRooxyuwE/vzM1SGgIQBQCL+arc/Pjng5++TbFfb9rvbFTOr1haG4k7Kqq+aFdkFDXSMy
+          4bFDKbYawoLNf0DpcAQAoWIV8mXbdkH2BfM9MxWQG2pfAgamwfG3XjA1pIHCKB0lNfHxPbYHfpuf
+          e0dmvrMgUMPthmviQ8Mgd3SrXopLt/qXn3LZYt8On83v/oLidlqxUY/DMNRJmu63/MinW166cP3X
+          gYTtbeTf9dbF08Bk4OUaCJ+eXh8vm7+CbzrUyA/7U84Zl6CUf3rx9DHvMXmDGwOlK2uh2zE9uTR7
+          1gY07UuOT+NL1Ve/61ppD40veqS806wS3lXgO32Tud/mN6vJaiUr96JA5hPXwzpEnxoOaS1j/Vxb
+          7rjkXgs3PsM2M7PN1D4CBb5s1cT6L09+K1cJLmmgIlM5BHS5MdUb9jf7PbcRzwH6OPMWtM+O40us
+          eaaz2D8zKBTKbd7yl2b5DJwEv4bnYW2RMKVh17Fgy0/xATnzMH0GWQSmneToOApeM/bpTZE+2KZ+
+          IXVZTDAXaL/xx6c0GuOVzgqRHY7TfdmovmDNhdCQJX2n+z2ebMqpu/ECT7lMULKWvs4/9J0mXUfs
+          +PRa+dt8tHrokccDKczyBOOZKLz8W4+6k+cE6LlgxZ//R/rn0rr9kIsS1CIlx1mZVfrKzCGBgWLF
+          2L6/zu5UqYn3W//Ah3hQKAenYoaP/Mj4/c8vbNcDL8/S9NdBG/Jvyyxv+HgqJ5SVmaILQrJm8tF/
+          auieL3uwXIlqyEy9qvh0pnEzPW6z9svvf/xJF2JmGfDEq4j8kXs3P/yF1mM+I+1QeJRlvpwh2yLR
+          cOF/Ljlp20mERpedfTjEPFjvB7GDGtfHyPlG7bCqqyrClCMCci7yxcXc/SPCjQ/QaScZMe+ukSh/
+          Do8SKy85o5ufU+CWr/okQu6wdq/Gg7883jo9bnT6DJwIys8Uo9jp9YG8kqqT5duczJyF7Zw7386K
+          nPrhHqvTyY25IME8pGO74rR3pXiVmYcCtjwam73lx+xNZ+r9GWohMsLbk9JcdUfgHiUPG/JlbaYf
+          nm/6HZWbvhm/39iEu9f+gfS9U8XLzDMaVOzPFevSh81fXU46WXrW5Zb3IbA0wtZSnUIeqXfxA+Zn
+          VFbgWsArjm7DHJOD9hjh5/QyfBoebPe3PvpX/7vkNgxLAO8QelKuIo3PNZfYj8v4y5+wuYNOvP7W
+          c6Agpji1tQwMv/Vn3X00/hCxmNIp+abQSJ7B33rs/PNv2/ye6VZ/ZCwbBsbNx59pXgCXsPZhhLzo
+          JrO4HCOXdqm1SiIvajhkDxjgYSLK/1dLwf7/3VIwHt/bkubrFZPDmUvl+/ogyLW4gdIiUlZY9gFE
+          x1FbKa77oIKLrN7nN74o20bp8Vue2uyALQVM+oQ0Ku1dGonz6plRLOTBroWBmZxR7n1CymvnM5QD
+          9nVAVj89By7hdQVa1gf6TNd6Lilz2YTHFc++2E/PZllWMYPHI7tH11PouOzlqLOgbqg8t/zh466H
+          u8pvjZ8uDph4yEmSdzXs4EX1dzHe5WQanhm8tp2OvKB2YhbvmlXGF6/EDtMzA3Ftr4V+dhCQ/iIN
+          WO9u1csH5qZij/twOXlfRAcuqSZg7ZAeXGqfMQPLsnqgYD+8m0WUq0L+dGvq89QyXd786AoUL3sJ
+          +U1S54ISQgke33I4f7ioyMn+23VAEL/OnMKdRGnD3kbo39cDOob7mTbaxypA+LjvcSpeLy6NOl2U
+          40/B4muEp2G5Pt8JGOr4Ns+W3DY0M/lMZpj3BTvpMdAHzREkKATRF3m7tc654fYoYchILT56uZtT
+          9lU5cjJIB6RYZ1Nf0j3y4TEpGOyq+gGwDD9r8FrtOKQ9cA3I2w5KOZA5f27PnUJZTolG2DCHAnn7
+          W57T/evEw+3gN//+uNmUzLpzAY95POF7+6UDNYU2gYgUR+Q1sgJ4CpNZXj2bmxnnq7nCtVsJdGkg
+          4qQjKV35h6LIqd1/EToIoc7P9ukN6z1TItePUc6nWEqgO6VfHMIBNNuuGg58UkVC2nS75kKtLiVk
+          D1qKHZg/9ekUrwas72D2mZ3dUAyNqJMb891gNw20hue56yifly/Gp4PoAI5k/gWysAnRqRKmYRGJ
+          QuR1KTDSjuMnJzt6rKC61CY+3acxXi7XlyPfuJZFl2RXx1w6khl6CWiw0VRYx7hdIrhKo4iMFxsO
+          hJakhfvooKF7gSzAkc9bg3P61HGRwrNLb1C29o5aPvD1FPY6USfyhrTbdb64PyWASPTLihoDTlgb
+          m3Sgl70IIYJURHog2A0/3PtI+s03hbXKeP6rZzYM5lU3tIbmj6oEz2t0R76H+wa/pmKEjdk2KFKI
+          EnPTU7/IMTIjfBuFK+XMnUXgVo/IzTN9WPB95CU+567Ir28MmPEBm/CxeE9841o1JqdALOE3mRvs
+          uAoY5lefmzC8nAiKP4Uac8D3Uvh51D46xeTmCrW6L+Fw09EsHN1nQ+iYRjBkxBaVH9PKWcUSNUAm
+          R0MPUScNnTPRAOWVfLEuioCOU6LUMs+e9tgNET8Q3GmifFrwBSlP/uCy+2/VQzNIFFys98Tlee4x
+          wnd7Pc67fpNgugo0cLU+MTrdwDv+XhOzh9f88Ebhld12dWArS1aelxRri3vPeRybK3TLKJxX4fFo
+          XvuH1spbPSJ3/Si5YACNh6+9vcPhsUnAIvHbknkQfJHBHz76Vr+XPzxyG89uqJozBiz7CGLntWNi
+          3L2vo1Qxn85n5LuZsy1r8DIOLoHPWs9xGH/Xd+QPX2xG79zlw65a5Vq+3LDz2HOU7LtrCUPlXCMj
+          N9t4ZWOuh8UyTuhG1JsrRNnow6DwGpTe3qeY353jTPaPn9anIEtd2hLgwcggFKN3cs2/p+dtlcU7
+          o/uS2F5yEn92GngMSYRKbb3mxP7UpVzLyQ3dJ0YFQiUdKnlypwBfa8fIl+PlOwMpttK5epc2JWEM
+          O+haqY7Qdj/7MXNXWPdPii685LlcMx0iyKVvFZciKSgbf3bKPvouN3xK7hplu16qoPZVdshihjTm
+          VbdY4RnQCvvjJXJXnnvM0Py2Njper1Es2G3Ky+O9qLBy20n5unRRIuMFAuTeTDUX4LgfgWtlOjaY
+          R0jp9DFb6J4VC6UnuRh41POm3Hjlgk6ngqdzoFe+LJ0CG2UqWXUc6noClY/3nmmiOoAm9qzB/KEM
+          /sKGBFD7slhy8rIQKpn45XY0TCXoZ6mDs7U8Afb7KTQAGpX+fh+s8Csav/FCdtxGDaujbIQJk6UY
+          Pcd1oBJ8j7As68c2HjXF9hlDeHE9F5fx3h0ECc4z9EfxjorvyRnYtJtEGMV7F7mu7evccLsWsg81
+          DjkCXPK1acT0j08r6XnI+TjxDMhJ3ugLG9/yCydpklpOD2xbnTuwn2p/AdKSW8hFS0RH7hUH8Nvw
+          AB2zeN8su+i2gsNFqXDAWky+8Gtp/sYDFexyaYTLeIv2QiFOqIxOvrt+JQbCtq5mVB6zQl+S0Clg
+          RiOE7fLaNrO3ABM6dw8g9ONzF7Q8XPjwg7L0SPSlAzEP57NroRPgA3f6WKCF3uyOWEuru8sZPsfK
+          e8+jKOMyzuWf/shD1bww+HIUT4AzWzWV5ZfzmIXaMeIZHz6m7IeOiqN25w0LyEcDho1jIVXUL4DL
+          FODAtfNe2KlPaU5v+F7AZ1V98CXZaTF7jCtDfsy55DMLyzW0y2wfdifCIDUtUL64i0/Aqu3HWebI
+          B5DgGlvy7JE3cgvxs/FzZ8mibOb42JYm5Ta+l58GcZAvkgJQ9iJE8o9PvDJLB/ZgnWdo7HcEGaXY
+          50tc1pb8RpcKx+Fk5bS/9LMcWdaMVSFfwFpajgXFFfY/PotfupCI8sbX6FiMCmUzKgbgfmwBujsV
+          ariDvGvhKTAHdKrPi0t8btEg7sYFOR9SNevdeZSAUCVE5c5uwCp8bR+O97JCKhZeYN2dt12Tpt7H
+          LlF2OYbhmsh4NEZsXauWElg3PURv0PvLx4H5CpUjC61J5P70IdEOYSWHjWXhwziEOeVHjYX9Q9Fw
+          EOCAkvd0ZGHHvV2Eol/XuMU78HxZJWxLnqmvj0hP4LDky1w9aq0ZvttBGkvSQXxbz3m8vOiHkW/X
+          i4Gs4NIAVvKfgVxFnIc3/UNXpcjg7/N/+LOKqavAq7vzsfF9CwPhnViEL9oayDvgNF+YM4lksJbl
+          LHy6DyWcEs3SOrQVtoMsbAjuHAlwn/mG0CMNdCKp9hvks62jvOpPseBpRfnTD/N7q8/Vung8vIRG
+          5m/8OCz+EDHShr84frwuDasJ+xUs0/6A3PxcbvWxC+BUnhq88X++4HvLwtf9Vvvne5AN3NBZF+Af
+          X+121vFEySe+9ZAjsMHuDSouW4wgAGbxXbDX63ZO2G6/wok49Qzqi0P5Xnz2cj17KroV+6dOdJUq
+          f/ylmocwx+CcsmDTS/gxrBldbuH+Ai2pkvxy08OLdRgq2MXDCZuBJA9LKgWpLIiD4/c+eg7kpgaa
+          3GvHYV6DWXApulaKvBalhY7+0QLUv5Q8lESH2fR8nxMeQgtKRfr04W7V8vWpnhyA9dFFl0C6D1Qo
+          G0e+QJ7311I/5CSM2R70XeDhNAh3+pyV0ALfdRcjH1exvh6btoInxyvwXU5isODg5EiwPdrIuGcm
+          WN8qq0mz1nM+OF47nX539gpAo9P5U4Iy3/6/Asn3ziItbDlKI25owRslFTpUD2Xg4tFmwW53vmFf
+          73h39by1huV1/WIbrd7AArlR9ormK8i5PlydN6oqgUaJYp9fo0+8vO5pAMvcanHC3uxh2fAc6hXv
+          zQz3zhuSXfEI5aj3Nv+1xlN44QNYfwoX3T6fWR/J1Zilo//GyJSVpOHld9jKE7FqVAhAaKioHCIQ
+          XWMTWcphe+ozexMgp1WGz00LKeFecSTbondGhjuchpWN5Q4cw6OOXA3rOb+ucSWLrn33hRd08lFS
+          7Rbe8a6YmTS+DDSBfgC+09GYmWMGXayErAT061vDevrFwzgN3wwq6n7CacJYzfir782P/PBcX0LA
+          8/LwoAd07IKRErwbCDjhk4biqzuApWKdAE6fVUGIw77LtrdDIb+rnYdPmtmCmezTCgYn08VufmYo
+          deZMg2AM4YbnYkzrgvAQcPzNJ2svDeOmpwDj2IcZ7m8gXzmR9/780c8v8MfF8OD9gmOktodKx9dO
+          WiE5CjbSmGoY1pI8PWggA8yr7bg6mUt7hsxo6liJ9wtYmyU1YUlcF6O6DuPxUy0XyGB3wGg667lw
+          KfdvKKd1hs0Tc4lX1DMG2LbWxVYzcs3y6G0TZvxZQP6soWZpHq92PxreFaXMIOa/+y8D2CJ83pWN
+          vpbnJwvfLnyhc/+yKCe4ig/5s2z7O3aPY7KjqNoP0iDhA0w0d5mebgLms21tfg/kXV6vBtjqx9+b
+          deMuEt8UcomyIzpwQAKr6hYEbk9r+cJS8fmif2UfvrWxwO6bwQ0BaV9C91PlOPBzP6eyH7PwFkMb
+          37PlpdPt2T54Kmbsr7e8+uljUybZtcD65h/IZHY8/OkHhYmHeERrVv/57UwS04Zj9p9Oyu67MzKk
+          aQB0OIwVjGLgImvzT4uNxhF8GxagMBCbfLFi3Yf6VAbIzO4HQA7N2YThmIVI9R+FPsrhVML0UQt4
+          0/fx93h5jtCo8AcpNxs08+i1pny4Zfmf/sBH9dWCza9hz42qhopdm4Dbl78ird2NzaI2USUl0fX7
+          q19A6BgEsvkCYJbN+dusVWHU8pYHzG9QEzAhrnDA5u/9fRnt6OqgIw/TxW187iUb7m88gaT0B//L
+          fmW61Oc2A7vd9Tbzt89Hn9JRnMH77GFsXXpTp0PViPCGxxfyK0fN197c1svcu7bN/+tA5lId5bMi
+          jhu+fBvyBCovX1Z8QPpU22CFRZDKDp58hPK341LlUbzB9hp7/b7TJ4UNZijVEZn330NIF+50r6Xq
+          dX1s/FsPtD1TFsaz1/ukSNWGuLbRyvKTtOhcmrcYn81jDSVb9PFZCwVAH8cpgiaXvNEdWUm8KGw6
+          QqMXUr//+kbMzfHQS4RPPRwoZR6T8+s5ymj6VtiCuwx8n+7gQ4WUBG34su3adE/A6/hM8OEkWc1q
+          WS8HsGp/99mTvZ3l+qw8+cd/xrlTAItaiYUhqL++sOUn5LZiA276H9uveKB00nxF2h3ZClk+zXM+
+          xm4JrvnxjVW/rehyeJczlE3l6C/HhqUEnW6MyHDJCeXsodXndUd9+Ik+AtK1KGxIkleVvPveXayu
+          dHJn+CWmtAO7EulscHW5+jymsL2QELkEaEBAsHvD4bEc/OXb6A1/ZyxRznxWxpfpdadzswSmnCxr
+          gpWn8o47tJM8sOUHyFS2g8f8LryALe/D/uQFdHkybgnSSzDj+4xzQB4vJoDnU0VQZCVevIqvfv7T
+          e6mkx8NURpUvG5/7HZ073WvY1TRrACvs+8J6ObtER9kMzUCfZhiPbk4RVkRgUWqgTf+BFfriCIPC
+          b7b7JQGsM98E3uCl95lv1lMSzFCB5MjZ6DA9iU6XOxttD8bjGfZ7yyVV0zHQccZuzob63XDX55z8
+          6cerxLXN3FzSGWz1tG3c/ohJ5oa+LOCbMtdpJev07tcRRJTXkMs1x+Eb2pdWXs9z9NNvYJQfB+eX
+          FyLP2N8GPilPEbiPrye2v4AM5P4t1r98Ryv1Vzxv9x+Ga1Ahe+PbbstT4emWsMi3ji/wN395xQTo
+          4FhaQ7XXLgBu2J3wrdirOvurr/E99ihZ7lpMo2z0wJZf+AxriTpRjm8ees+mm6Wgad05utatzBGm
+          QQYPDoD1s6WXN3zD504f//JCue55EzupPNN1yyeBZ9sm8lQ5irk3IyWwKGYTefE4xOvtmEcgsw5H
+          X/Qee5cix0rA5OIAe3p+BsRwoQ/uxzfwsxudmmVsn7xcf0oX+e/gO3ypKbFAYqmOUf7u9eU4ay24
+          hGaGXckd82HVWAlemveEdLWIKJ0zYsoee/GRllay24tUy+TVc7mZex0PufDMPx1sho890854xz++
+          BB6B8U8/u8JHfTgQtgd7y0/KHJdbC2FgXs5I2Q/vgbwvxJI3PkIIybctsg9SuWmFeP7A2Ip/flt6
+          p+8eW9J+zZf2OxOgajsdHc2A0SfmLAbwq3kBPp2KC10Bb7bwh6dneT0MC3c61/DLX33sgPm66dum
+          BnkrSkjF6tGln0DRZIZpL395K0uu3gg+oxPPx4ZXAeFeeSTdTssHHaeZ5Ou65jVUDmHn7yo+phxQ
+          xhGw12pAysaXONuXPVx8MqGf/1jyF51hvYclOu2UJcbj/ZTBPO8J9m5bx7D+5Xypnn0VH136jjnT
+          6yrIe/MOm3P1zKdPI0Hg2a7pV4M5D+svH9uNgvmXF5P1/skkQXVCfz2kL52+v4z/0xs+mIzUnV1t
+          vx2sOX02/FqGbX4G0Ag/aGZOr29Mntb3DTZ9jvXB/8Q9Fzwcqb/u9sh6K6m+bvwJz3G680F+LikJ
+          Rc6DYsNHPlc9qmbtnxaBJ5N/zpKk3ehSvesWvtO2R+bV34H1gHpr30vrA+mkDN2uGGkgb3kIipcE
+          0SUeVRaKR1ojcym64S9PzRquxRt+UvLV8gKWnuZh73vqBxoGjQa4kHeRpx5ps6xOmMjb9c9im2SD
+          MPYvE2758bxXfIXyInVSgBVdxnbDWe7qDGkK2tOJxYb19Aas+pcMxD2P0e99wRTaC7TyrPb5XXKh
+          QhgMCtz838y14btZvGu65U+ahYzDdchX6JPxlx9gFz5m2p+D6Q0S/RT6+/LaDhRIYQZr1ghwvOWh
+          01YPIFyjCisDnOKVNe8M3PIG7NtUj+e2ukmw3TkrssXKGtbOuozQkCJ9Fgnt9edp3hewe9sVDm01
+          HbY8wZHRIdghVNdLPmu13IFcDnN0kxlPX022HmXyJceZPzF8TLb8A9TJ8MK6zctDd8G9BOJPyWLb
+          4F/DXFqaJctkOwj4nQj5mPgWCz3vqc7vTd8u1bt///I0/Njwgy5DDkGCAwVrfuc3hFyNEX41P0Cq
+          LQjxn989F1/Lb7b6W779vYen0/uKXAG9dHLdj8Zf3ht+S4Yuj+KZwTSS6M/Pg6WL2BWq7WdGysE1
+          AOnuESvz9tWYoUfa3/dHEJa+5At9i358mUD5EQT4ZM7fYRXybwXlOet84Zm8BnKWtbdk6Zo68yfr
+          2uAfPr0cgLD3uuzoNn49WKPL1ed8coy3vFCBmx/C1oa363rKGFgfnidkONxMt7yDh6fbhcWalYz5
+          fNZP4q/ekCeXissrRcbAQXfVmd3WTyjnVEQqrT5EqnGL6ebfCdyT64qP79XW53O37UqKbwrKgsbQ
+          uYMsvCVGnC1fOrsdXWzUjvLXLW9Yu3hS02dhUsrvkGexqnwmujaXYIYERhn2rOHRjLG36/fbfJ1X
+          R7EpOUgchAdhcNEp/+iUJMrXgedvfPJXozZz4W1wHQhkwZ9FZhBjyidfBioTLLb8qaLE5/YaRH5j
+          YZ+JXzoxCEvAnlUU7FTPDNCWySS4MiZFinhq9PUlSET6dCTFj7BSKR9d6zf0R+nuc3akU5olFwMC
+          fZ+jI4y7nKggCH73C3nWsGto8X32csEZHUrnx7Hhxt2WJBzZ6s9/zM4QZPLUpgf0CKsnJXL4KuUb
+          THoUv593QHXPNiAx7j46TcyTEva0ZvC1d3dY/ThFjLf1Njglx9Vns+DVTHVjGPKWl2K72D9d2nHW
+          G+q++sKmlvUx/XrPSt7yV+yj8A2WqzlDeHkzHTolaamzxSN/w/HYRihRX68cs6lV/OH1CTFUX8L5
+          4cD5uZywi5aVjnXUr5DXCx6p4Vzrm78PoJeGJTJIzVDy/RTbI9zdERfIeA/rrY46ebreJ2R4cTSQ
+          yrMM+MvL/f1Q5bMtvCVoKvszcjb9sZzXsYLtqTqi6KentI9SyL/1t219qlkDfdPTIzTQ6XY4DqN1
+          fwbwUZv2DHZfTNcYDg48MecM3dTPo/nqUZFsLWH53F6FMv/lS/IzFmu85Yt0/a13/PhWm25CTPNH
+          V0rBkABcRP6RsnktmdATZhtpzXOr5340oHaTPli13D3Am/+RpVc3oJA8ekA7rbr8/AVW/QfUCT65
+          NeQGI/n5iXxm3MiTsUmpv1qJl09nxdvylNJFx9pp82mCzggepnqby/uS6evtGAfw4JEL1hRS5bNR
+          dReAQr3wx18e/wFk/OEVUls+05dx9/Vg59x07L3ydlhYvbjABGh37D4rT181ck5gbPsGUg9obRbE
+          FRa8SlWINc3X6fouQPZbH5j3DOsAfstfgXq+X5B3b29g3ZZd4OYv/eHHr7/1k0PUvfAN31h3ltVQ
+          gf9pKfjXf/3X/9gaBP55d/fytTUGTOUy/fv/tAr8W/j3+M5er19jwT/zmFXlP//9nxaEf75D9/5O
+          /3Pq2vIzbr0GsiRyf+0G/0zdlL3+r7f+tf3g//rX/wYAAP//AwBtWE9GugUCAA==
       headers:
         CF-RAY:
-          - 96665ce339ae15ca-SJC
+          - 96a9b5586b97eb34-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -2942,14 +2944,14 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:36 GMT
+          - Tue, 05 Aug 2025 22:25:06 GMT
         Server:
           - cloudflare
         Set-Cookie:
-          - __cf_bm=UcbpJtgMpUaG.Rib..BOZBsOPT0SXZZ85Jts36QHT1Y-1753726536-1.0.1.1-cGUj49faUJP4hz3Xy2fAVRF9haI91DkQJseN3aHmIrkDYwbjrPkrh2d0SU2U_kw4J5Ul23qyvPXNp8cLbgrQ.6I0TiPvYimdY3f9Y8SUec4;
-            path=/; expires=Mon, 28-Jul-25 18:45:36 GMT; domain=.api.openai.com; HttpOnly;
+          - __cf_bm=G4YeGSZWqxPbFv6GvdcCHzuPYepwD6TqJ8wi6BfygIc-1754432706-1.0.1.1-jQd0DIz1sKP7nBY8_8DnOD.vIyZUBZlhmmljSGe9bTdRUNQLqfyHhaVxf3pItONIa5v.YcwBCzxWMWGDftBHm8S3FP_64dDL6tpAzQZRDr8;
+            path=/; expires=Tue, 05-Aug-25 22:55:06 GMT; domain=.api.openai.com; HttpOnly;
             Secure; SameSite=None
-          - _cfuvid=_jNn2qGdBfM4a6pMe.Bu.RC7.CjnKDDPoTXaLz1sZQI-1753726536577-0.0.1.1-604800000;
+          - _cfuvid=aeQt45A4HT3v386R9a07QeEJ6maS3ZnOfwZDCeUuDoc-1754432706038-0.0.1.1-604800000;
             path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
         Transfer-Encoding:
           - chunked
@@ -2968,7 +2970,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "297"
+          - "513"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -2976,23 +2978,23 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         via:
-          - envoy-router-f55cd75d4-kqpb7
+          - envoy-router-7c6b8c8c54-9hw7j
         x-envoy-upstream-service-time:
-          - "306"
+          - "517"
         x-ratelimit-limit-requests:
           - "200000"
         x-ratelimit-limit-tokens:
           - "200000000"
         x-ratelimit-remaining-requests:
-          - "199998"
+          - "199999"
         x-ratelimit-remaining-tokens:
-          - "199979817"
+          - "199979820"
         x-ratelimit-reset-requests:
           - 0s
         x-ratelimit-reset-tokens:
           - 6ms
         x-request-id:
-          - req_8fe3b8b26f1694dbc67d25981a335e9e
+          - req_326c08fb380a1e0ea8d076aa237e34ea
       status:
         code: 200
         message: OK
@@ -3016,7 +3018,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3026,7 +3028,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3153,7 +3155,7 @@ interactions:
           98//e/xf61/917/+FwAAAP//AwBe8oQ14CAAAA==
       headers:
         CF-RAY:
-          - 96665ce60d3715ca-SJC
+          - 96a9b55ddeb7eb34-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3161,7 +3163,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:36 GMT
+          - Tue, 05 Aug 2025 22:25:06 GMT
         Server:
           - cloudflare
         Transfer-Encoding:
@@ -3181,7 +3183,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "180"
+          - "67"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -3189,23 +3191,23 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         via:
-          - envoy-router-dbb4fbdcc-pbpnl
+          - envoy-router-896675b9b-cs9dk
         x-envoy-upstream-service-time:
-          - "196"
+          - "69"
         x-ratelimit-limit-requests:
           - "200000"
         x-ratelimit-limit-tokens:
           - "200000000"
         x-ratelimit-remaining-requests:
-          - "199998"
+          - "199999"
         x-ratelimit-remaining-tokens:
-          - "199999930"
+          - "199999933"
         x-ratelimit-reset-requests:
           - 0s
         x-ratelimit-reset-tokens:
           - 0s
         x-request-id:
-          - req_8960eb8fa29d2dfa78ee7cdc30460c2c
+          - req_0ff41cbb2670f2ced2b1753cd00d5819
       status:
         code: 200
         message: OK
@@ -3225,7 +3227,7 @@ interactions:
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3235,7 +3237,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3251,118 +3253,118 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA1SaWw+ySrel779fsbJu7R05KFW17jiLnAoBETudjiAioKJAFVTt7P/e0ffL7u4b
-          E5FEpeac4xmj6j//9ddff/dFW5XT3//89fejGae//8f32vUyXf7+56//+a+//vrrr//8vf5/d1bP
-          orpem1f9u/33YfO6Vsvf//wl/PeV/3vTP3/9nTgHgRabpEjmps8uMKBdgJNdMwF2aJYYTY+rj3dd
-          6vGFunkOl8U5UHVrNca8uj4YPKjl8/feG9XVXlLQfnOhscbkdoLBoiDZvuwIvRrnhGGZlHBAZx8b
-          kDoDxUYUo2ZjnIigfSIuZHeQgdoqylAcDnJLkehuIL7xHpvdhYHFSN82enuqjs1PSgcGIBagFbg1
-          dl7ePIx9GTko1VOHYgGnBo/me4NuF7nGjlqEYAphp6Boqq9YF/daKz0lsYOPi2JQt7g17TIsQoz0
-          WGFEIfyU0LKPa/SMWEo4iSkYL45dIWOjvqlf6wfOANxJsN6nhB4O+VDMumL5oEsXHWe3dg9k+bDN
-          ITocrtQ93rct28Q8hEfHz3Fivx7t6J2SHkJTd/EtC1ND5PftBQ7vcaC43VSJtBz0El0/5xT70u2T
-          TH7jZrDdRStsPmsRfOpkNMGnfX3I0LhaMV/5OUTO7oLpfnUePLkSDB2dk8+IddE6JvK7tBWAzo5G
-          xMN6yxcB6wR040UMWTM6XPaVewztDBo09VLg8QYuK0SP2xxfs8vLG8/F7COjziVcNr7mLfFRU5G5
-          025YbQQ8yHOlEDibe5HG+nQDMtaP4UbTSEB3TuAAcaydD8Js1dDA0G2DkVPwhIeCKdgcQ39Ytuls
-          IsFsHgT1PGnZseAm/K33QQ5MIFubTQPjR3elaeJrhizPeQePTpjT8HCNEyFb6gqNPLli5xn1xczF
-          9wZKL6emR/WhG/NGKUMoHpYNxTdNT4S1K5tKfrxgmjh3vZWnzJj/1McZVTpnkRmb8KZYO+xysm4f
-          m0zqoHL7DGSx6ZhQ+2EzOPuMYYvYJV/E6a7AvmwEaoqw9Zbf/YN9Z9Qe7QdgYi5l4GTAbaicXrBl
-          Yr5KYXbgLvVVTysEoQEh/OS1jbOqOxayJYslNJujQCN7Ggc+dUiCeB2pNNmXLp8L6cKg2mAxRGlZ
-          GaLSNg6cJjknG1QanvRQlBBqa/gmr0/Yt6zahBcYyL5GU3m1BSwArYlK6Mi4MtbEmDZSJKH0EUX0
-          5nd7LouWJaD48bxibbu7g0X69CkMovRFT5/Qaen7qkEkPQHFXg+UZFkydQNjveiwPernQZJNzYUq
-          YSbeG3EyyO1GzOC1uFQ4lBK5YOZkM7jryh096bfYkB55dYHvbXTD1z2/ecMW2CpSvYeHD0nsePJw
-          cS7wdYdbfN1glc9THF3Qt37pPhJMj52k4wfeP/ck3H7Xb4mcawV8tJ9C6XUyBqmEng6dmHR0Jy1d
-          woXX/QPzx74iIOoXb2l4YEJDy9Vv/a7BzHbHDt0T90AxrRcwYyL0cDkoHzLrW8YnNB5sKHarhaxt
-          KfGWtRw28EBHnx62O43La5VXiG6+88c39ILXuB/RzagqupuMAMwWxyo8N0ND1uYcGex6aSIUTq8t
-          kRKQJS8/tyqUCVNMNSjjRLjvZR+xE1OoV1wzT0bmxoYHSnwcsyjy5vrar2AyVFO4OA/kLWg8mHB+
-          ZSRME0cAlH0GH4ZnkFO9JjfA3o2To0yABrX20Qk0atYL6Dg4Nc0vo2bIyJxtoDaBiFXx0CfLx2xV
-          FFWRRDN/B/molXcTga5lWL1YUzuvrhOD06LGOHBZ0srOLMdglH2KC8vcc7ERVAdOlbnHWR/7nqSJ
-          jgqNo3nHfi/WHutGx4VgczPp2fYsQxxv1hOBjrXYlV6LwcRN9ETX3c2ndk46MGWlsIFCMvv00C7y
-          MCJ13yPi9Ba1rLLypnwT1xDVrUlxP4x8rq81RBk8WXgnTRtvhgcxhq949e/5PxfiicDTabhQP63c
-          QbxWUw9Dd9NQZ96+CsZWCYHrq4FCIDZNMt/PYwY3huJhLQ3mltWmosMrMxwytrt3MfbeawUvwVqi
-          rhbOxSikmosOJ7GnyU08eUtXV7mSMHmm1lfv+DqqO3TzXlvs3I+qQeZ260AUPAENDdMCwguKAjpL
-          TwE7RBVBt2yGaisQmVL1YgXDUhWdDcuro1EHla035/C4gfftrOAimfqBRebF/PUbvmnABXL1iFbI
-          NLJVKFplZTCAdAWenkCjpv44eyKLzRgNj+WMU/1tJeTOjBjpmWDiveVHg2xnKYHGui9xtNWOXC7P
-          dQk1Png0GF9BIaLxbMLR7zk9WalcTIvQm2i3ZkW4soUkmYT1sYbZ5szJg1qHRPzNl8IbB+o7ipHI
-          93DqoSc3IS6quChmS9BtJK1YQuD38/7BnMsfnlBXqpuQy70vUa0RQLUHrTkn8c4F8l6UqebetoBG
-          sZ6iZ7/PaAY+PmB3ZkTovXxG7C592S6aM3doH8Q3HMqxx8WEPRyYOPUDV2JOPfKeWYmKD4yx4Vyz
-          Qpyl5xNW48ugR+dIB/oQAYO35+VJMbmjohfV2wXO6jWl6fu8K8RwYhJk7Q5T8zbMwzJo7xKCYw2x
-          163tlr3sZkbMmU18S84hn0G6EmRrzHcEwr0AmOCaMVRJm2C7vaGBTdt3DtWmfFJvc9q1VC9PEng0
-          B5umgm4n830vh3AQDZXApRb5OLKVCQdU+FQTnCkZX2hWUKeZd1yUkTnIiec78MSqHUFiRpPFvkYq
-          2r8eHxrEzyCZj3HoQ1Xa76i7rfWWU7Vo4NaS91hP0hf4o5f5XtxTwz1xgzF67yE4lyFpJPPNR0W6
-          PYGP8oog2ufDzM1URVt4Y1gd2qVlJzkf4ZmsVey3u3fCbqd6RN2Yi+GyIzVnspCsUCFFL3qErztg
-          t2fMAH/7Ec217GSIsVUoyrMUNRzcLAhaX3lHULzHEd05+7LlzryOoHlZLfTXX7x83SpojLeWmtpn
-          BhwbeQS6SJypJSrSQO5N5cB8XA7UXL1yvhgmgPDa6jHeS49DwTaGbcJE4jp2iWy0Mkl0Hx7U6kmk
-          j98DNr2FGCZOIoTQ1hZALZs+lV+/mlNwN2bptnyQ67KWGrK28OUuayNqK2XBWEjqZD4M3EZq4q/D
-          sYzNhDvqWMLR/3CKyWPfzg/m5OCkSAnVCvvO5xnWJiRl9gyZPq35clO8CHzrieaLQNtn6gcQDk/V
-          pclaUA0xUO8dfMVTjLNdfhk46G8ZRIcmDJtmdABbtGsGvQE54UqOPUAzsZBgzYctmcWgNKR1d7Zh
-          c7Qw9d4K4TNeveff/KeumvQeq3agg7uvBrrbMksoMXMH6lvm4zDNDM52L6Pfxoe3h28IfQa+BaEK
-          tXOq0rJfOGfKdJ0hS+0zPTvakb/EVIHQxMWArfbuenTxfRVoJZ/JtqrCghnh6wlhcfapEwuFxx/m
-          YwMri75DqNnvgcPHlcHy7FOs2pM/0MYbTWhZCaReQNcD9c9qDU/Ko8GBbhzaheknF95J9aC6JrWc
-          X8pGhQXpOT5/eYbFWlSj99KPdF/mXiIv6LAB1f4DaFCvt3wWtUsJXTG7YV0y34BARcvQ5tSN5PHV
-          W76PDh1K9d2Jqmf77A2qvsng4zLEFH/7n8GtHcNPGb5ClDw0IG1Vv4FGOabUo/O64Ftg6/Ds7J5Y
-          e99IMXrO+6l82scHY7jKvMf11o9A/zQPbH8c2LJYy2uo4sNIlnoeC+Grd2AKHwt1rnc7EZzJ24D0
-          WgxEWjRiLNSNLqgctxH+1qu3tPdDiU6n9yUEtb8p+rHzVThV9j6cx+aZzPH+FgMbnQB2Eo6/8+me
-          Q/zqA5q4p8QYvZfK4M8/6N3zAX6/H6T3dIVv+CEa5GVRAq303pKnexi9pX6VnSIeqUu6XRe0S9IC
-          E55WtfDTk2QxWeDDcr92sDtPaULv1YWBPV1LofRxyoFQcWTwnVlP+uUjg7sTIhCbo4Wjr18hhjo/
-          oejmItX67mmMjyaJkXjuy28/aZydeWoiwXdyeqX9pl1W508O4sdppH/+n7jJO3hv9rtQ6uPRoGU+
-          5X/4YP2tZ1ZttRL99MuFxx3/8d2PX6n+qZv28fo4+fY370+PRE1YG/Sr3+/FTrx5FfwD7i7Qb2H0
-          R594us1LWAVdQGP38QDzba2NMBYeSShUI/QW7McmdFxxjXf34pKMlg5DcLGEAgfsegVLqYYlMD9+
-          ThjatENHh7UCch1n4aZ3rURKA02H0kfUsDWbtielfrCCl3n//PrZN180Z9OB4ysIiVT192FCQ6WD
-          +B2tsBeT1CCyuXcV+XhB1GjfEZDwLumBjboR+zDQDOYHFYM3ZbcjL6I3xawDJ1KaYPTppYqKYdm/
-          nRqwRO7JnB0SsLw2SoV+PBmft0o7Z+tTBVY39UADfXXmQtNXOUzn5wPv1DM12HLQK9jGrwTrNRsN
-          ZhWFC7/+hBr7fNWOGyX14WC3jO6mVeZxsDFLxMNPSoMsu4BFHJ8SdLCu04N/zrjgnYoPbKxUxZWh
-          ngdysnYzkm4HgMOpExMCt2EEy1pqqf8BrTHXY6HCyd1GP37zhAg9RyVYH0sapOXK4/726cOfnwnV
-          lQ7GU7WdIXbyJ759/f6Sv+sKSerBprr7sIB4cu3w19/U8Y2mmN/7e4l+9SV0lxiwYwFscMbQpS6K
-          Zo8jA6mgJW6Ond61CnG8Bc9tbX2Cf/uHtbu2IWEoov7hsfVGX1QkqIuSgrVcc5N5ivMchBd3H8pl
-          zYx+7EwVJJFzxjusB3yxqNeDKngGWEeCXLz0Ff+gFWYf7GVNwOXhJV9gfmUF3asHJ2Fy95Rgotxa
-          jIvAaaXvegBVMc/UeXnRwNe3Sw89tRfwV2/b0XpEOjpKlUg1evaM5ZPmDtx11Y46VUnaUVymGEwu
-          iKjWii1n42XnQlKmT7rf3EhB+QnWcDbzJ91nX164a5sK0szZh+ud+Gr57UBDONhzi308lB5vioOP
-          1hGKKMbve8KGabOBJ9V4YQuHKph3K7lGzsd64tASV2D5NNiF7ns5h8rT0Vq2TocVhEyTqKaKtB0B
-          r3TYim1B4pNzH5blGEjgq2f4Ir0OBhPAxVFGXoc4YnnGmTmFDHz1GofjO/AmUpo6bLpMwsEOkYL7
-          uVXC4279CllNbny+u0oEhuqShZuE02HxkjyCF0sqqP3xez4H6cX+U2/7EewMcacILpwWPcaOTMaE
-          faicoy9v01+etKyA50KLLy5Vb6ZuyHowMThKaEXDS6QO02PSIAwzX8ZaPfuF0BmCjaKpuZJFFfEg
-          oLksob6dfRw5LUrm4XiA8KAs81ffWz53hmBCcxdFNEtU4rHAy2O4UWoBJ+e13Up+eBCQfwkeP/4B
-          vW92ECreysWnaFMN7zgN4K+ecNiJSSF9nx/0satRB6+PnHXFbQW3QnqiiX+YPcHIuxIcV7L2nQfY
-          e0f7qIZdfmuxP42lt1jRWP78xlcPZuMP7zz5bQgHsWkKucZf3hTUN64+dDbYoucX9OMtNTrJHsf6
-          0f/xesglHRXzp1c/CGhSEY5L1xjLIXc7aPWViL2Ae3wS1tcabk5HEL4iTR+EJx67P/74clhGg/zq
-          F8z2Btu/+WoHHwUc0v2OWv7+nZB3o+YoDB8GgdfAH1hxbSW42N0R7490LNjXrwP/UmVY23dLS3O8
-          KqGm3UqyKjyHC3RkEHnApxTvadcyj6wb8M27sHlEikHpcr/AMY1eOPJ6AfBNIMwQvz4Bxq/ZTJYA
-          Jx9kdN88tUKSQcopWqFcKCysme/9bz3cX3/h8svrvFBCBwYveKHHw/rMeWZpOgickVN36eGwfFo3
-          hOJmiGloKlLLloNbwvzafrDnruxWjuZ3AxPqhdh6ZrRYxqhoQCvOPk3s9pzIz7RQ4fqqoZCJlpjM
-          u7AjiHBfo0XWTIAM06zAWP8cyaIJpicbQLGBq4M13mXSq51KDJ7g5j221JLEFSfq46KAwiMDdvyV
-          2y67soSKdGtGGma5MnBWLA7Urvqdep0aG9KrsxXI6GUK+RkoCZ2EZgUlVomkFdRny5abUINHBpef
-          Xy7m3aI2KN9kp6+/iDh/rxoChxCsiBwa54Jfj62EvvpCNTnowFQ0nwjWEV3wj/9mLhABMnW/w+XS
-          6R4felYCRfFf9LSL1UQQUs0BO8uvqc+fOZ9B167gvKIz3fPxBpi0cwm8PEwd68YoJLJXujaEq6eF
-          v3zzzdPvNpzu84DxTWsSnly2MfzyBuH00BVc1ecMPZDkhZLnC8WMV/cZlsfkioM36w1eGZsQNOWk
-          Y1sQd5wl5D6DL5+R9llJ3rSW7ebn30nttNfk846LGAT0GWBTfz+KJU6D1Z980xXuoSH+/GYSuWea
-          0qTz2G89f/rjr0O/GIXENpH70V1aDaQbFn+GOdz4xs9vOYB96PoCEybO2Fs+MJk2p+1HGZ1VjMMs
-          v/zJp0D4KCuCNlgFgnJcFDg9bn44e+PdG3HtdrBgp2MIkqlvf3k0+PolfNoMWkF//PX2dJ0a6rsc
-          +AE7TPlTfy/rPZAPPpNf/oLVT/9O+LYkApTtfUYEs9UTCeX1Ex7UQMQ7ojcJz29DB39+tpSKlTFt
-          E6QqX39PJH8HAZOOkIHv/PrzfKSf3/jxeOjxF/8zT+WA7/FOv8XeM17aFL5QEmAXy9QYu+Qe/tFv
-          I7X2XNIX4wL9x12hpnn7rk+flnDnbDXCvvOwP4Deh258OZGvX2pHLr4VqOL6ijV6XHFmf4IQrhZ4
-          IcqPN9wg1396SR6nrTTQX/9uz+0Hm+9jYfTey5nhXRR7uiO6XrAt3vbw0okHejSeU9JvlDSExWcV
-          U916HhKuXy8KxGCHwuZTSsn0SSMHVfI2x4Z/mI3ZOyUfoJPApnvLn9sPzgICqfswqLNLa4PCx3UG
-          fmO72Cujbpjuxxqil7jxaOZ7e4MNoyiAr1/56ocIlm696X9+F/vqTjQILxmE2V1fCHvfQ4OhvO8g
-          BhnBtnvwPaZUA1G+eS01KT7xMWBLjJYpXJFtIuJCIo91D8Y0fpEFVyT55hkEBJYa48Ma2kAuZf8J
-          A/9jUev7vG78BBv4zd+odnRyzsWuttF6DmUCaj9P5uW9XOBLVDzq6MdhWCrkS0qXkzSU9rHccq/U
-          bTTKISXC6pWDeex8HXogpCE4n7hHLP9uQnrHE1ldorqdshIqwErnlO5Ehrk0TPNG+eVzxvLdf2An
-          l4DidvaxnSqeIQ9nK4NdbZ3C9UPaDbLKDRemmf7G6inqEyb2bQnIOiaEPliR/PFnHfiUIfCF07/z
-          ox+PWwdzaZc6/zQg4BGl5v0xtct3Pwji0tpRNbwqxvz1X4CGSRcC5eq2y13ej/BOygc9bsmc0FVT
-          hso3r8d7MyiNkY5sBUbgauFq6zuAzGwIwRidjj9+9ASVpAQ2Z/2O949dUIj4LfnI1x8Mm35C+cTv
-          2xwYJUlxwMu4YC3VRjiv4J0m65mDaS+zpxLD9S0UVsPDG09XaYTYJFa41ZO1x1aST4C9VbxQtIt4
-          EOJ+qYGNrYoW7D1zIu10Al8hX3/5A3rLn/27WDBpprbHYeyLow3DrbOmppcCYyykeEaBQzj1425q
-          R19kXz4qM+qJVWaIra7E0CZHC4fBOx/4abn58EXiUyi5h9EYhXTvbvcq0qi1JVExJ7tohsQcOf7p
-          F+uK0wquTvuJSPfzpeWp4z6V3HZ7HDokKvrvfgf85RH61y9/efECDyuqh+hYsGEZFhhBdTdsyUa7
-          tMkfv7S2PoBatykqxH0flBCaqhuiM/aLUUZHBp147MJ5jjnnjTfa4GHEUcg4NluunKdY+e1frms5
-          8RarPj/BN7+g5nf+0/d1v/r1B3l++1ca/cj5s1/lfPO+5UjbJyrPIQ1Zsb4OBJ8cBp/yBWHr+poS
-          pj4uG1jvM0JN98wAE4ZJB8zHQrjNPunA30EpQSuSdUKkkzEIm6FxwHtaud/5rCfSEu3yXx4TNkm6
-          4/J8lXu0wasN9RtHAH0qNDr68ZS6Vl4et2RUKvJhLHG19XswH43UgV8/h4MzHpMRpCsJ1uv1iTqX
-          XuCL0RYuCN33go3Z9zmX5+iJiLrc6eHyrdfbcPaB4GAbG1p/MBY3OwgIdHOLMa0PXFh7zAGLIDEc
-          0NEEIt1fM7TdzwW2Czp7s+e8O+j5MKYH8/1OvvPbh4HJ7/THd+wpb3r49+9UwH/966+//tfvhMGz
-          v1aP78GAqVqm//jvowL/If/H+Lw8Hn+OIZDxUld///PvEwh/v4f++Z7+99R31Wv8+5+/RPHPWYO/
-          p366PP7f6//6ftV//ev/AAAA//8DAGq57+XgIAAA
+          H4sIAAAAAAAAA1R6WROyOrfm/fcrdu1b+pRMkmTfIfMkQVDErq4uQERxYEyAnDr/vUvfr05333iB
+          FGKy1jOt/Oe//vrr77ZoqnL6+5+//n49xunv//G9ds2n/O9//vqf//rrr7/++s/f5/93Z/Uuquv1
+          8al/t/++fHyu1fL3P3/x/33l/970z19/p2eBp8lrLJJZT9MK+rfjHhcJnMCcdkuMtoIZYPUc+c16
+          c7MM+iWMqPdWH/qqBuYKG9V5Uw3tHv5YPV1O4T9tThOrkxhRjncFqe/QJrRBF31WSFhCab8LsNE5
+          DhhDrY6R5w5nsjkZEeMvmZ+DWtfLkK8GqZns0ZOhlZw77ETHFcyPpLPQhx91bJ0wHZbgQXm4idY7
+          Nt/tXBAxiRwELNWh3sFP2Ao/9wey+luNVXEbFmOvGxxapuCKXfjaNeK0N1t4tohObRs8GpbMxxi5
+          UcWI+CTnhJzPcY2kZ3Uk0jjM/vhdHxTQtqPuCA+MWetZhI9OJTR9sqFYROcUgdW6aDj3Kxfw0s7N
+          YDCikrpxvm2WwzYJYVXIGU63zauZTEtvodqIHi7T9qjzQ73NoWGXA91LS5VIM15LtJ3QEYfw0jc0
+          z5QUpqPM4WBzFYqe06AKnlLSEVrbu4J1qxui9WJhasf84EvZyjRU6+mI9111SoRlFUWgM3VHNrfz
+          ljHlqq3gAi0hZBvbYbzn3lNotaNOy34C/pKChUMv4Z7hbMt/ijF/zAH6rKqIo9Dd+UwoFhVxx9cN
+          uyLBg+hPPYGrPgk0Nd83wIvDh8jqodpTh28dwG8/To+2df+g9puzdKbvpzcsb7mC1bgPitUtZgMx
+          znsRkY1Js9iPxoAgnzUcr50BpLskPyB/iq70LDk7nd/T9gmjE5dRbcrihFfmukJbS7piTxnaYhae
+          nQxjUa5pEe00ff3IMITR6yRTVeS0RNhv7EjRbQ/T7Ia0RIyPbP5TH5W21djsNA8VJtfJxr6xRc0n
+          PVlPyLp8IPKrH5PJAOkKA+exYt+9lWxRjUWBqhzzNOilxl+il/WEkWKudM/fX2C+pGIO5Gsrh0Cq
+          YTPP5+oI3/fCo/7xsyv4/D6E8KlEFr5cuVPBl9aphlIj8DRtPuMwv5NJhK+bqtKMiB6bF8Fb4apT
+          IYTlqdLFrI8daFyajPBI1n1xu+tjKCZ8Sybits0ciiSHBxnu6FnMtmAGWuKgcOAlfDxyEyNsikQU
+          mmpEk6B2mTjqLx7F2/iKjZN+B/Nz2x7htVY/NDVnm5HncQeRbjUUm9pRSdbrOMvw8No8ceiDS8H7
+          yc6DIfcwsLcJkkHc3cwc7i9ehU1lkgrmDekK1Xdg06S4xLp4s7gc2p16w9V9rYohW94qgsj1cZlO
+          ji+ldyeHlt5ucaLt1WYpD1GO+t4SqD3Ohr/ceKGHBneIwzXMNcZe5FqBRNyO4XJj+iBlK9Dg9ZE/
+          qbvun8maW0sPd+VUkdVoFp8t98mAJ3LeULN4bcCqnl5PdNGUA1UHsAAWX48tvF77jsyf09qMKTtY
+          cOf1CxHKKCnYeCIPSJtsT48nfcckf9Pk6JzdamzlteazIWxHRF5eRW3xugfLpd0Y8EaSO5ntEbNl
+          6OIIOXt/S9bjWuvP9v2qkFddDtSXF5zwsmMH6GZXCg0UKfWF2HAseMBVgPPSjYoFH1sOthdvClej
+          Qv5aNgcDmpzChdG14cF0vQzBrz6oJX1uA0vvTobcaNb/PO8OTi2Plpyvaam1O11ItzdOoUUnYPsa
+          tMnS6omK9AmKNEUbyMZHcjeQGtgrdsPb1DClQCs0R/6AsTcmjeSEUgqmt0xxLtxcJrRAdeB9w7s4
+          rffB7/0NuK3bB9aOa+2zaXA8mGa+QU8P29R57vp6o8WpGrzLH4s+72n9RllnBzT8OE8wqfpThkJS
+          BzR9EqmgV277RCDLTLr30isgvqTVsOOwTk3hPTLmPiOIqL8xsSnasj9v5lcI5QfRqZ8Uj4YxdiYw
+          fyU59fitN0jPJ2phZrQP6nzxj9WTTmBlNVyITtUjWfZqmUKF9T52DXtuFr/oNZj4Z4/cBaMrpmkr
+          KTA1bxK1om4uptDfeag/X1qa+Pjsr8c3yRT0Ps808HupWQp7fiLpZm/xfn1tWOsnrgM/1xhQZ6Oa
+          QMqAyaPbGPLYMk1++Az7gmwPwY1S09rsB9YrR+1Pf6tr1vgL/zhBGF5KBWd824JFSjwDHo6BjUsl
+          9YBojRGHyr2GQt49VTrTwKrAi/bZ0cCqLj5/d/gYzexy+fKV2VC1ZTE6vJ8GVic7GiTuyBN4uh+v
+          +FyAUyP56lzC/YB96h8f+4JP2cWAr+DIaLUhkk93rDXQVsjLUPnyMw0noYbp+CJkWDeHRPgs7RHu
+          btlAfVnUE0EQrk/ofOIQ53VfDGvjPTxU1X1COPum6V15yPI/esLccC4btUwu0bTPFeoiv2brQf14
+          IDWvElVBvh1onq1HVJ+XlJ7sVwBmILMIzfd0xKq+K/+9/ifC3bBHJ59JkSoE8CUHL5y8FlpM/W0t
+          UaSMMVazc1oIpiH2UCmxTgt5oQX1xmGFuLfe1LtVEAy8dcvh3jCP9Goe7ULoglWEXs0wVXVrHtjV
+          uZTw+QogdrjcamZ0imekyk8Dx+EpZDOQ3pGsk8giSNV5MH/3A25CO8EWztCwNHKXQcY5L+oIoaNP
+          cm4rAO22Fk2KytLX9SyFsJs/O6IIucBGuAsdKO31gKqyOenjGc4Kwhxf44sRGIN03EIDrnNok9mI
+          abJcn5GKypvZ//YzmV8qCSDfTTZ1vny58hu/gp7QuNj94A+YNyBWYXe5ulRrV6YvAbm3UIqNMyEp
+          17GvWHsA8nIqwjn3DKycxqt/8EPT0dKsVzEboUI+KjbTpkvWNq1HdDZUMURVX7N5B3QOQR1+aDnq
+          d7Cq+3gFYwsjmgjRWRdm3heVj3PfYQttIGiSpYugii1Mdze9TNZIwtEPr3/4zqht3h7wFegNxSdj
+          BgtanQjQ+jJTc5OIA906nAOL4HKg+6rOGDvqBYQ8tx4w5m6HYsbu24Aaf9Pwnkv0RjLuawDzD/ch
+          svVpwZy2zxj67oaFrLisPjHTTaqo1Y1SNXTv+orf9x7VedVQQyoWxpiwjGh6PlbsvGidsN2aWEjL
+          0T4kAzH09U3LDB6rlFFMVffLf1kGfO+dUH3Q7mwGeu3A4cR9QuH7/5jaggjMXm3Q6BwNzRN6E4TT
+          fvTo5TyrukDw4Qm39SnGSR/nwyxxNIXKEofhGNsOmK8YpXA83N1wex4Cnw46UGDR4C3ZXF6lLt2K
+          iwUlr8PUugiEzbepm6Ek8ge6V0nrL4U1PKF68BC1JuXEJg60BtTdR4B3N6Y3C/ok5TZKDj5Ou0s/
+          zNglBjxfDJXGbcDYKjVohp4gXuiNux2SzzbpIbTUzYjNA/J82iulCkKjmMn81feMh9ID2sAMqP4a
+          i2K9NyceCq+kC+eL1g3f560wV/kZu4kZAHrLAwMqBoPUMt8bQLparWEeug9sbatDw47yx4JfPUqD
+          96Fhq8ZiFer1kcf5V8/MylyXCJ+OI/XmxE8kFNxl4GxTQI1su2WrLHsZfJfcDetB3YFJ2e5SBEae
+          Evpyz2D5rj+6UppS43K7+C1QnBi+vSSmJuX/9H/8q6dwc7vvgLAXYQVX43iimG83AwsOqQX1aXjj
+          /eVCCoKUzlP04tBhTMBpeOHKmcH3/bGT72CzrE5Ww0NwpURw/LEQjf35AYTQXag2Xa2EP4NBBp9x
+          M5AVQqL/0Vff+sfGpbv786k+lOgBzCLk9rbsD0I+qpAKqxcC9fpOWIVvMYBQAtg1JgzY1TmUMDxn
+          e3oqoyQZg7e6wp9/+OnhhWWiB3yqIpxsNJ4RQ98QaLVmQ15iMPqL30Begcj3yfPI7dliPwYDhlzA
+          0+owScm8macAUvfjYLPoE0akl7eCjZTIIUJNWdCSD1b4crs3vaI2SlZ/vBJYPksTn8ch8umYqQ94
+          gYZA3Wn3Yn/4Tdofy2937dhaPgwDff0aTTRdbhiYvQwkIhipOYR4WO1N9oTFa2uFzCIDI7vbPoPN
+          e0Dhet32A7PtpUafIgyxxRa7mdsljuCHtjL13e7J3rdYDrY/vM9eVE3Y0Wg5eCy6N1Yf/KdYi8sS
+          g9LjIhy87mWz8mpWw3J/3NOrV7/ADy+gKB/iEIQm9FfF17769bTBeh/nzaR3owc2r7rAhpReAXtw
+          3BFkw3whrDhW/kvOsQLqM0vDuX2biaDzEQd/+GlItQl46E0clOn0woe56vU1ubVPUFlCSOAXf6d4
+          RyzgvCGH/fZzTMghaDkFdRqi3q2JgBC9kyfgPtGIg6+eZCefrPCe6ibp9tyjmOcmc5Q7KgN6Ebti
+          WKuNXIPLtmgJWvsErA+pr9AzmmN8XlalYbZ5fgCemw8U384XJvlclcF2jl94f2RUX6fzWsEmxsmX
+          vwc2zw/fgz99tDsduIZ2Mx9AE+KFmgc59Ze9yJcoifITDec5B/O2sESYh6tGr884ZeLmBHrY2oaK
+          zw66FNOkn2c0hgLAaomEZgQSiSCvPhrq3M6Nvp4UoELM3aOvnlV9oQDiqEwyKmlQnjifnVsrgD8/
+          E2pbDZDXtZshCdU3vrSHqpl//nW/QV/9WptAejhpCF/u8KbmeHsUv35CsL9xIW/wMWBfPwN+eI5P
+          /ewzs9+rAG7EDLv62/R/+7ethHz/b/8wLTcNas4r+uLN1p+qsRchFSoFq1npJT8/AG726oXzABY2
+          3HLDAHDDZzhcilBn0jS0oNyne+xZTBxeC2p65ErvHuNbuWfim9g5HHFVUMwHTsLEhyhCq7UbvFcd
+          h4li7Hugu8wX6uRWNLDn22vh4R3xOIRHsRnzMdJQYHkC9aujr89Abx340EKbGseMNPQtoRhoziei
+          Bv7isVWfPRis6odqEBKfbqqy/uk1atp9zVbHkSsonGQ3FNnyaZiANyGs87LBmmuUxUIfhwD9+MuF
+          r3szn6Esw3XzeeNd76tg2U52jYRP98Z6TjiwhA71ILc9XcJt+Nw1C04KDu43nUC9mtLm1w9QbXBP
+          TrpxH5ZquIrgVz/XrXbQlxFeZKXXoxCXyiFliwGqFdSZ9f09KwDT3D81aKWaiJ1GJsXMLKGEqvD5
+          hGh93di8WZQA8L12DlG0p2Cx8iyC9fIuaBDb7Tev6DVlULdXqrWSrQvK7hjCZ7TGWB/dsVnfxM6Q
+          7yJG7XPeJ3/6wzQvHtWwqum8Ca8EFtWVoyqc1WKsnjsIVUGWsAqHoBA/2tFCm8SriFJTPHz5qITr
+          UAY4OxUoYcWwQKiQw4z3Udww9q0fiFIY0crOib8clDaEsxfw+OoqViOk5oFHfbe8qLZwy9CNwJAh
+          AsTDl+V69dufXoGoKvAumpNCOK25A8NZ0ai52Z7Ycr9Q7qcvaXL3Z1/i7WcGksdNw+aYh0X/43NB
+          0RuML6/SX7JzUMJCQozeTv2ss82bplC+nttwOlWPQhrCekQzGztchN6ss/suy5HV5owGQST5s50J
+          AVTYEYTyqUDFyotzj2yQl+HHuT/0GVXKE37rF+sW8RkJBFTDzXMLwvaGtEJiXFDDH57lX3wiX38C
+          y/sq4/Crj+eC70Ug3RebeoHRJTS9qxly4r1BJNAFBROfugg3SnT67sdYMPGV8cCTwxTrt82S0B8e
+          ss4uyC/PkwiLIeKOMqXBfXg27Ke/v3kXdoGg6GOt3iv41dO4eO4EnzH+OcLwnO/xHo9GMmuh3v/J
+          U3fzRdSnLz+g1dBN7Fa6W0iR5oXwKckuPvh6lyytTBz4y3/LnXxh60e/W0BQjowGrzsc1l2nhNDg
+          kpha3VFs1umslNC/2wM21m897M7bB9yuFGOcR7RYVtt/AN2tA5r35SUR68JXf/wWiupNSJbd3iDo
+          m2fSX149vU6qAlcjPRHgYcPnO0WxwHYtNl/+/TRjFg5vkAF3Sw2DcoycjFwBD5AP2OouXrPccwiV
+          4BqP1F5WZVjNx+LAoSf3b/5zYFL9SBV4Hrwp3P7WbxkeHJxiTyBdW7wbtqmONTjI7YKPnmkWczOr
+          D5Ra1plqRnPQf37llx8QToCXYoHHRETWYKhUW7snGNslj6Ak4gX/0X9E4Xj4ZouN029+tMrk8QTb
+          Qv7Q0g7VRETK3QHGDGtq2nXGFpwkHCy980ytOq59djEVAsNZ1rD2efMJzxeeBeUotTD+6sl5YosG
+          bVz3OBiuj2QtLtsYfvMushT7ZzEDRY2Rfyd+KDKLL5Zvv8JzL12xMZFW/+U14N5eNByWs83Y4bTj
+          wXMLRfKeFAGMz+n9gM/Ew+Rj+0XSWRc/Bl9/gTH3fBVsG185qAxb9sXvUJeaY6SiL3/RWMJPf/76
+          U/jjn2B+7n3iXVIDhZfVo4XUP4c/fkj4DG+yYW8HLI2IczjKpxnvUQMTAiqXKPeMxNi2l7xhurFf
+          gXc3bkRsRnWQ1vGgwK1gByGj7d0naqo8YbvZHEPh7LYNK7sbVKgwe/i0mLuB/PQXyFeNfvOCgZ3O
+          7Vv51d/x+egGyqUugTci7/Du2w8r6zkeMiScyBZsNF20N9EbRp9JwKZWPRKWpsUT5h/4obeEcfp0
+          P1wdxU42JyK1KgTzbSxX8OHthtpOGeqSJEclqsd7hL/+Sl/dOIthBm4uVq/NEbw+c3KEsbzZY1uP
+          aDLevV0Mioa/YL+p3UZMPiyHib9TvvgpgOmrf+AzumoEuM+ZdQelDWDxDlMifP332BsXDn79OsZ4
+          z7Gv3g5hlDwLwr54O7+CVoOPZPMkXbaKYBSl7gERsHvsumGhD94kz9Bn1/arX7RiPqbbFl7564Ge
+          xevE2rEzQqgGJKb2cDskiz16CnQTRw035CQmVPYjB9m7V4YNHFDGTgFbgXIRLOqKhDbtoZwI9L2D
+          Tnf9UuvjO5l48JunuEbwHMagrCHqzdan5+zt6msavEagF4r7wzN/5UW5/fldbH3xZOKbB4THmiyE
+          ++qdn96FM+EIDsHeB+yb7yiliB80EIZzM/7mY4IqbggIZ1wIwYhbAHXu8/WjJFnj1FmBxs8xLnTJ
+          AkIzwTd0w9ykQZnH/gO+ywe8HDmOumiX/SyChdaNIhLUWJnOuvaeQ2fb+9Tpo2GYj6exV8y+OobL
+          LpCalfSrhcSAzGTpdxlY2WPUYJArNAQTW8Evn4U3g02Eafu6ocgvFdCROqHqMmHGbz9qq/zyuT/z
+          B3TKV/C+7QLsqIKvS1g6pXBbvM7hxlvsQZo65sHaJB3WPkObrMmHZeDRaYTUxr5I/viz7pyXIbch
+          Z32ty0hFpzKJsW9wS8NC6j3A7qoSai7q1Kybs9NDjU421aCk6OtxSUIwX8AzVCzZaxgTtuOf+eJR
+          /szN6DtlqhzXpfrxuT5dG10Dj5Tbhfx3nkjcgx+Db/9851++/yef3FvrHduuti+kr/5Ga2qu2Jf9
+          WSfvucsAD/Mj9tJzXPzxU1413mm2dXl/SoJYUdxHU4er0b58OgbWCPs+N0PRCzb+Ij8hAULyCELk
+          CvEg4nlXgtvc3Wjx7Scavh7kN0/4vi/02W9+p9etQXP/cgKTdBAsqFv85ptXgGRcBG3+ozd+efpY
+          8A8eZZyTUveapfqfec7XH2B732TDfLdpBLVay0JlN476OPaXcFutk0r3qI0KdgrVGbpRybCrgQNb
+          fnl0mC4TASjOm5k5XqUYF7HFWgxCvztWZg2DhZ6p5rpRsjSinUNanbVwOaTrMKftGP3Jf6QpbZKZ
+          KKII+1MK6D75RAXfvqcSpvTph4IYBD6l4LXC3dN5h+s3z1meXmkBeteicJ4Co2FLto+VsPN7wl0P
+          ib/O9uUNbrbbUydfGjDdS5f75UdkeFquLv3wYMKzhIMgOuvzniRvlKvi/Jt/+eTbb/DrP7EN9Emf
+          o2Muw299UrxNVrAqLrLAx5T4UAlex2ENCyjCMQEWebi8PkiRpgXgyy94V2+0ROKtc/bLY8Lug20m
+          xaXdog/tZWoWPQO9xmINJYKH8Y4TP/58KKdR8ZqyxMfOacEvP4DURyYOxc/YjNsiFOF58zlTzO78
+          H38KJHG/YF1zgubPfONYX+407+/7QnhqbgR+8wor0g/6fNwfeLQT6wb/9IggJZoBSvGx4vDjGECo
+          bZSiyq0LrEne/FvPJ1SDMab5V08yo54CqK/FnRryaDR/8O/v36mA//rXX3/9r98Jg3d7rV7fgwFT
+          tUz/8d9HBf5D+o/xnb9ef44hkDGvq7//+fcJhL+7oX130/+e2mf1Gf/+5y9B+HPW4O+pnfLX/3v9
+          X9+f+q9//R8AAAD//wMAMsX+eeAgAAA=
       headers:
         CF-RAY:
-          - 96665ce7ff9d15ca-SJC
+          - 96a9b5600e2ceb34-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3370,7 +3372,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:37 GMT
+          - Tue, 05 Aug 2025 22:25:06 GMT
         Server:
           - cloudflare
         Transfer-Encoding:
@@ -3390,7 +3392,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "308"
+          - "203"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -3398,15 +3400,15 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         via:
-          - envoy-router-86d465658-qm2rk
+          - envoy-router-6bd4b5cc9f-9jg9p
         x-envoy-upstream-service-time:
-          - "404"
+          - "271"
         x-ratelimit-limit-requests:
           - "200000"
         x-ratelimit-limit-tokens:
           - "200000000"
         x-ratelimit-remaining-requests:
-          - "199998"
+          - "199999"
         x-ratelimit-remaining-tokens:
           - "199999990"
         x-ratelimit-reset-requests:
@@ -3414,7 +3416,195 @@ interactions:
         x-ratelimit-reset-tokens:
           - 0s
         x-request-id:
-          - req_15c42738bb030b0a6ef2b2572298b27b
+          - req_d608fdb8d4b7a16b81cdcf1b280859cb
+      status:
+        code: 200
+        message: OK
+  - request:
+      body:
+        "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
+        the relevant information that could help answer the question based on the excerpt.
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from wellawatteUnknownyearaperspectiveon pages 9-12: Geemi P. Wellawatte, Heta
+        A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
+        of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n------------\\n\\nthat
+        gives subgraph importance for small molecule activity prediction. On the\\n\\nother
+        hand, similarity maps compare model predictions for two or more molecules based
+        on\\n\\ntheir chemical fingerprints.83 Similarity maps provide atomic weights
+        or predicted probabil-\\n\\n\\n                                       9ity difference
+        between the molecules by removing one atom at a time. These weights can\\n\\nthen
+        be used to color the molecular graph and give a visual presentation. ChemInformatics\\n\\nModel
+        Explorer (CIME) is an interactive web based toolkit which allows visualization
+        and\\n\\ncomparison of different explanation methods for molecular property
+        prediction models.84\\n\\n\\nSurrogate models\\n\\n\\nOne approach to explain
+        black box predictions is to fit a self-explaining or interpretable\\n\\nmodel
+        to the black box model, in the vicinity of one or a few specific examples. These
+        are\\n\\nknown as surrogate models. Generally, one model per explanation is
+        trained. However, if we\\n\\ncould find one surrogate model that explained the
+        whole DL model, then we would simply\\n\\nhave a globally accurate interpretable
+        model. This means that the black-box model is no\\n\\nlonger needed.79 In the
+        work by White 79, a weighted least squares linear model is used as\\n\\nthe
+        surrogate model. This model provides natural language based descriptor explanations
+        by\\n\\nreplacing input features with chemically interpretable descriptors.
+        This approach is similar\\n\\nto the concept-based explanations approach used
+        by McGrath et al. 85, where human under-\\n\\nstandable concepts were used in
+        place of input features in acquisition of chess knowledge in\\n\\nAlphaZero.
+        Any of the self-explaining models detailed in the Self-explaining models section\\n\\ncan
+        be used as a surrogate model.\\n\\n   The most commonly used surrogate model
+        based method is Locally Interpretable Model\\n\\nExplanations (LIME).35 LIME
+        creates perturbations around the example of interest and fits\\n\\nan interpretable
+        model to these local perturbations. Ribeiro et al. 35 mathematically define\\n\\nan
+        explanation \u03BE for an example \u20D7x using Equation 4.\\n\\n\\n\\n                                   \u03BE(\u20D7x)
+        = arg min L(f, g, \u03C0x) + \u2126(g)                           (4)\\n                                      g\u2208G\\n\\n
+        \  Here f is the black box model and g \u2208G is the interpretable explanation
+        model. G is\\n\\na class of potential interpretable models (e.g.:  linear models).
+        \u03C0x is a similarity measure\\n\\n\\n\\n                                       10between
+        original input \u20D7x and it\u2019s perturbed input \u20D7x\u2032. In context
+        of molecular data, this can\\n\\nbe a chemical similarity metric like Tanimoto86
+        similarity between fingerprints. The goal for\\n\\nLIME is to minimize the loss,
+        L, such that f is closely approximated by g. \u2126is a parameter\\n\\nthat
+        controls the complexity (sparsity) of g. Ribeiro et al. 35 termed the agreement
+        (how low\\n\\nthe loss is) between f and g as the \u201Cfidelity\u201D.\\n\\n
+        \  GraphLIME87 and LIMEtree88 are modifications to LIME as applicable to graph
+        neural\\n\\nnetworks and regression trees, respectively. LIME has been used
+        in chemistry previously,\\n\\nsuch as Whitmore et al. 89 who used LIME to explain
+        octane number predictions of molecules\\n\\nfrom a random forest classifier.
+        Mehdi and Tiwary 90 used LIME to explain thermodynamic\\n\\ncontributions of
+        features. Gandhi and White 10 use an approach similar to GraphLIME,\\n\\nbut
+        use chemistry specific fragmentation and descriptors to explain molecular property
+        pre-\\n\\ndiction. Some examples are highlighted in the Applications section.
+        \ In recent work by\\n\\nMehdi and Tiwary 90, a thermodynamic-based surrogate
+        model approach was used to inter-\\n\\npret black-box models. The authors define
+        an \u201Cinterpretation free energy\u201D which can be\\n\\nachieved by minimizing
+        the surrogate model\u2019s uncertainty and maximizing simplicity.\\n\\n\\nCounterfactual
+        explanations\\n\\n\\nCounterfactual explanations can be found in many fields
+        such as statistics, mathematics and\\n\\nphilosophy.91\u201394 According to
+        Woodward and Hitchcock 92, a counterfactual is an example\\n\\nwith minimum
+        deviation from the initial instance but with a contrasting outcome. They\\n\\ncan
+        be used to answer the question, \u201Cwhich smallest change could alter the
+        outcome of an\\n\\ninstance of interest?\u201D While the difference between
+        the two instances is based on the exis-\\n\\ntence of similar worlds in philosophy,95
+        a distance metric based on molecular similarity is\\n\\nemployed in XAI for
+        chemistry. For example, in the work by Wellawatte et al. 9 distance\\n\\nbetween
+        two molecules is defined as the Tanimoto distance96 between ECFP4 fingerprints.97\\n\\nAdditionally,
+        Mohapatra et al. 98 introduced a chemistry-informed graph representation for\\n\\ncomputing
+        macromolecular similarity. Contrastive explanations are peripheral to counterfac-\\n\\n\\n
+        \                                      11tual explanations. Unlike the counterfactual
+        approach, contrastive approach employ a dual\\n\\noptimization method, which
+        works by generating a similar and a dissimilar (counterfactuals)\\n\\nexample.
+        Contrastive explanations can interpret the model by identifying contribution
+        of\\n\\npresence \\n\\n------------\\n\\nQuestion: Are counterfactuals actionable?
+        [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+      headers:
+        accept:
+          - application/json
+        accept-encoding:
+          - gzip, deflate
+        connection:
+          - keep-alive
+        content-length:
+          - "6353"
+        content-type:
+          - application/json
+        host:
+          - api.openai.com
+        user-agent:
+          - AsyncOpenAI/Python 1.99.0
+        x-stainless-arch:
+          - arm64
+        x-stainless-async:
+          - async:asyncio
+        x-stainless-lang:
+          - python
+        x-stainless-os:
+          - MacOS
+        x-stainless-package-version:
+          - 1.99.0
+        x-stainless-raw-response:
+          - "true"
+        x-stainless-read-timeout:
+          - "60.0"
+        x-stainless-retry-count:
+          - "0"
+        x-stainless-runtime:
+          - CPython
+        x-stainless-runtime-version:
+          - 3.13.5
+      method: POST
+      uri: https://api.openai.com/v1/chat/completions
+    response:
+      body:
+        string: !!binary |
+          H4sIAAAAAAAAA4xUTW8bOQy9+1cQOtuB7bgJkFu3RRfFYj8ORS7rwqAlzgxbfUxFyokR5L8vJDvx
+          dNsF9jLA6JGPfI+inmYAhp25A2MHVBtGv3i3+u2Prmz+vH8/vv3rlw/9l2/Lw9vu9/tfPy7X92Ze
+          M9L+C1l9ybqyKYyelFM8wTYTKlXW1e2bzeZ6fbu8bUBIjnxN60ddbNJivVxvFqvVYr08Jw6JLYm5
+          g79nAABP7VtbjI4ezR0s5y8ngUSwJ3P3GgRgcvL1xKAIi2JUM7+ANkWl2Lp+2kaArZESAubj1tzB
+          1rxLJSrlDq0W9ECPo8eIVZQAZgJHYjPvyQEK0CNWyQIPrAMEjhzQg6MDtwzocgqgA0HK3HNEDxxr
+          Q5ZgX/SUhVA7yijKsYdU1KZAV/BpoCMgB9AEGOWBciP6VkgadergYUAFCeg9iYIdMPYENhXvAL2e
+          E86ENR7jpXzqgKtOEr2Cj7GFNmcetWIhebLFY4Yxk2PbSraxybyWmFgkUIQAwfGZOpBmtnOQYodq
+          UqX+hJFD0nSJ2pM+EMVJpY5jT3nMHFXmVbajjiOBcGCPmfXYXBH6cSjYGsS9p3PBI4w5HdhRVcz9
+          oFLlJpCRLHdsz24JRCJHrplsB6ZDU0LCmdyLdXMI+LXORgcKVWxXPHQpQ4mOcpXjKorRVYe4O9a/
+          5tXEPLnamvnpumXydKge7MSmTKdrt1q+4kXI7ThgT1KxDr3QNj5P73CmrgjWFYrF+wmAMSY9GVO3
+          5/MZeX7dF5/6Mae9/CvVdBxZhl0mlBTrboim0TT0eQbwue1l+W7VzJhTGHWn6Su1cqvr5c2J0Fye
+          ggm83pxRTYp+Amyuzwv9PeXOkSJ7mSy3sWgHcpfcy0uAxXGaALOJ8B/7+Rn3STzH/v/QXwBraVRy
+          u8uwfxaWqb6V/xX2anRr2AjlA1vaKVOuw3DUYfGnZ8zIUZTCbrIuNaQbd+vVDd282diNM7Pn2T8A
+          AAD//wMAKpmyltQFAAA=
+      headers:
+        CF-RAY:
+          - 96a9b5629caefab6-SJC
+        Connection:
+          - keep-alive
+        Content-Encoding:
+          - gzip
+        Content-Type:
+          - application/json
+        Date:
+          - Tue, 05 Aug 2025 22:25:08 GMT
+        Server:
+          - cloudflare
+        Transfer-Encoding:
+          - chunked
+        X-Content-Type-Options:
+          - nosniff
+        access-control-expose-headers:
+          - X-Request-ID
+        alt-svc:
+          - h3=":443"; ma=86400
+        cf-cache-status:
+          - DYNAMIC
+        openai-organization:
+          - future-house-xr4tdh
+        openai-processing-ms:
+          - "1492"
+        openai-project:
+          - proj_RpeV6PrPclPHBb5GlExPXSBj
+        openai-version:
+          - "2020-10-01"
+        strict-transport-security:
+          - max-age=31536000; includeSubDomains; preload
+        x-envoy-upstream-service-time:
+          - "1494"
+        x-ratelimit-limit-requests:
+          - "10000"
+        x-ratelimit-limit-tokens:
+          - "30000000"
+        x-ratelimit-remaining-requests:
+          - "9999"
+        x-ratelimit-remaining-tokens:
+          - "29998478"
+        x-ratelimit-reset-requests:
+          - 6ms
+        x-ratelimit-reset-tokens:
+          - 3ms
+        x-request-id:
+          - req_9f6f47acdb5fd61041428fa2f4d55a2a
       status:
         code: 200
         message: OK
@@ -3422,14 +3612,17 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
         from wellawatteUnknownyearaperspectiveon pages 12-14: Geemi P. Wellawatte, Heta
         A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
         of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n----\\n\\nnterfactual
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n------------\\n\\nnterfactual
         approach, contrastive approach employ a dual\\n\\noptimization method, which
         works by generating a similar and a dissimilar (counterfactuals)\\n\\nexample.
         Contrastive explanations can interpret the model by identifying contribution
@@ -3492,8 +3685,8 @@ interactions:
         although both are derived from the same optimization problem.100 Grabocka\\n\\net
         al. 111 have developed a method named Adversarial Training on EXplanations (ATEX)\\n\\nwhich
         improves model robustness via exposure to adversarial examples.  While there
-        are\\n\\nconceptual disparities, we note that\\n\\n----\\n\\nQuestion: Are counterfactuals
-        actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        are\\n\\nconceptual disparities, we note that\\n\\n------------\\n\\nQuestion:
+        Are counterfactuals actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -3502,13 +3695,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "6066"
+          - "6342"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3518,7 +3711,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3534,21 +3727,22 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFRNbxpBDL3zK6w5A+IjgZRjW6W9RKqqfqgqEZhZ7+6ks+PpeAaCovz3
-          anYJkDSVekGreX7287PNQw9AmUItQOkao268Hbzd35T+ZvRp90M+fv/6Y7Sd8LfPzfvZh2uc/Vb9
-          zODNHen4xBpqbrylaNh1sA6EkXLW8fxyOp/MLqfzFmi4IJtplY+DCx5MRpOLwXg8mIwOxJqNJlEL
-          +NkDAHhof7NEV9C9WsCo//TSkAhWpBbHIAAV2OYXhSJGIrqo+idQs4vkWtXr9fpO2C3dw9IBLJWk
-          psGwX6oFLNU7Ti5SKFHHhBbo3lt0mLsTwECAOn/jxhKgQKxpDz7w1hQEljXaPhiXi2saWNrSiwyx
-          xgiSqookwq42uoaSMKZAAhodbAjQRgpUQGTQNbqKchHgFDU3NIRrDkD3mC3PpUDX1BiJYd/vwo2r
-          AKHeF4F9zRujoUyu02yhCpx8ZiE0bEknS7nOMd5Yow9BWY1xeZZCIGzTxlgT90P4Uhs5utC+AVrL
-          O4EkFCTnS66gkE0ooOYdiCdtSqMP/UgW8KxrS9j2W5iypEAugg9UGN15xuGpexkuVb+bWSBL2+zy
-          SjQHyrMbj5bucenW6/X53AOVSTCvnUvWngHoHMduLHnjbg/I43HHLFc+8EZeUFVpnJF6lZ1hl/dJ
-          InvVoo89gNt2l9Oz9VQ+cOPjKvIvasuNJ+OLLqE6nc8JfjM9gJEj2jPadDTvv5JxVVBEY+XsHpRG
-          XVNx4p6OB1Nh+AzonfX9t5zXcne9G1f9T/oToDX5SMXqNN3XwgLlv5d/hR19bgUrobA1mlbRUMiz
-          KKjEZLvLV7KXSM2qNK6i4IPpzr/0Kz2bz2dX5exqpHqPvT8AAAD//wMAEWKifwcFAAA=
+          H4sIAAAAAAAAAwAAAP//jFRNixtHEL3rVxR9ikFaJFlebXRbZBaMIeQQQiAyo1J3zUzZPV1Dd7Wy
+          Ytn/HnpmV9LaDuQyDP3q49Wrj6cJgGFnNmBsi2q73s+2i8+/1dv1rf34cXv/8OfDr8u7B9tuf1/J
+          Xc1mWjzk8JWsvnrdWOl6T8oSRthGQqUSdbH+sFq9X67n6wHoxJEvbk2vs5XMlvPlarZYzJbzF8dW
+          2FIyG/h7AgDwNHwLxeDo0WxgPn196SglbMhszkYAJoovLwZT4qQY1EwvoJWgFAbW+/3+a5KwC0+7
+          ALAzKXcdxtPObGBntpKDUqzRakYP9Nh7DFiqS4CRAG35x4MnwATa0gn6KEd2BF4s+ilwKMktzTwd
+          6bsI2qJCyk1DSeGflm0LNaHmSAksBjgQoFeK5EAFbIuhoZIEJKuVjm7gQSLQIxbJSyqwLXWcNJ6m
+          ozmHBhDak4vSt3JgC3UOI2cPTZTcFy+ETjzZ7KnkOduzZ/tiVNhwKL1MBEl8PrBnPd3AHy2nswrD
+          G3T4rfB/o1wChJyozh5UxJekgxI8anf/CX756/7TO6glXtfwqqnUNUXgoJmVjwQ5OIpFVzcUGBzk
+          YOVIEVKfI0tOEMmPKrfcp5JPI3Io5g4Vb3ZmOrY7kqdjaVCVrEQa276Yn/GcyFXcYUOpYDX6RLvw
+          vAv7/f56oiLVOWEZ6JC9vwIwBNGRSpnlLy/I83l6vTR9lEP6ztXUHDi1VdFcQpnUpNKbAX2eAHwZ
+          tiS/GXzTR+l6rVS+0ZBusVwvx4DmsphX8OL2BVVR9FfA+7u76U9CVo4U2aerVTMWbUvu4nvZS8yO
+          5QqYXBX+I5+fxR6L59D8n/AXwFrqlVzVR3Js39Z8MYtULtd/mZ2FHgibRPHIlipliqUZjmrMfjwq
+          Jp2SUlfVHBqKfeTxstR9tVzc0u2HlV05M3me/AsAAP//AwBHs/jNYgUAAA==
       headers:
         CF-RAY:
-          - 96665ceb39a8fa86-SJC
+          - 96a9b5625b77fc54-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3556,9 +3750,11 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:38 GMT
+          - Tue, 05 Aug 2025 22:25:08 GMT
         Server:
           - cloudflare
+        Strict-Transport-Security:
+          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -3572,15 +3768,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "982"
+          - "1307"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "985"
+          - "1312"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -3588,13 +3782,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998548"
+          - "29998481"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
-          - 2ms
+          - 3ms
         x-request-id:
-          - req_c6a5a5a859f388af82eb3aaeb2041b8f
+          - req_348812526ed94625b6e64207fb61123a
       status:
         code: 200
         message: OK
@@ -3602,198 +3796,17 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from wellawatteUnknownyearaperspectiveon pages 16-20: Geemi P. Wellawatte, Heta
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from wellawatteUnknownyearaperspectiveon pages 20-22: Geemi P. Wellawatte, Heta
         A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
         of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n----\\n\\nssion
-        challenge and is\\n\\nimportant for chemical process design, drug design and
-        crystallization.133\u2013136 In our previous\\n\\nworks,9,10 we implemented
-        and trained an RNN model in Keras to predict solubilities (log\\n\\nmolarity)
-        of small molecules.127 The AqSolDB curated database137 was used to train the\\n\\nRNN
-        model.\\n\\n   In this task, counterfactuals are based on equation 6. Figure
-        3 illustrates the generated\\n\\nlocal chemical space and the top four counterfactuals.
-        Based on the counterfactuals, we ob-\\n\\nserve that the modifications to the
-        ester group and other heteroatoms play an important role\\n\\nin solubility.
-        These findings align with known experimental and basic chemical intuition.134\\n\\nFigure
-        4 shows a quantitative measurement of how substructures are contributing to
-        the pre-\\n\\n\\n\\n                                       16Figure 2: Descriptor
-        explanations along with natural language explanation obtained for BBB\\npermeability
-        of Alprozolam molecule. The green and red bars show descriptors that influ-\\nence
-        predictions positively and negatively, respectively. Dotted yellow lines show
-        significance\\nthreshold (\u03B1 = 0.05) for the t-statistic. Molecular descriptors
-        show molecule-level proper-\\nties that are important for the prediction. ECFP
-        and MACCS descriptors indicate which\\nsubstructures influence model predictions.
-        MACCS explanations lead to text explanations\\nas shown. Republished from Ref.10
-        with permission from authors. SMARTS annotations for\\nMACCS descriptors were
-        created using SMARTSviewer (smartsview.zbh.uni-hamburg.de,\\nCopyright: ZBH,
-        Center for Bioinformatics Hamburg) developed by Schomburg et al. 132.\\n\\n\\n\\n\\n\\n
-        \                                      17diction. For example, we see that adding
-        acidic and basic groups as well as hydrogen bond\\n\\nacceptors, increases solubility.
-        Substructure importance from ECFP97 and MACCS138 de-\\n\\nscriptors indicate
-        that adding heteroatoms increases solubility, while adding rings structures\\n\\nmakes
-        the molecule less soluble. Although these are established hypotheses, it is
-        interesting\\n\\nto see they can be derived purely from the data via DL and
-        XAI.\\n\\n\\n\\n\\n\\nFigure 3: Generated chemical space for solubility prediction
-        using the RNN model. The\\nchemical space is a 2D projection of the pairwise
-        Tanimoto similarities of the local coun-\\nterfactuals. Each data point is colored
-        by solubility. Top 4 counterfactuals are shown here.\\nRepublished from Ref.9
-        with permission from the Royal Society of Chemistry.\\n\\n\\n\\nGeneralizing
-        XAI \u2013 interpreting scent-structure relationships\\n\\n\\nIn this example,
-        we show how non-local structure-property relationships can be learned with\\n\\nXAI
-        across multiple molecules. Molecular scent prediction is a multi-label classification
-        task\\n\\nbecause a molecule can be described by more than one scent. For example,
-        the molecule\\n\\njasmone can be described as having \u2018jasmine,\u2019 \u2018woody,\u2019
-        \u2018floral,\u2019 and \u2019herbal\u2019 scents.139 The\\n\\nscent-structure
-        relationship is not very well understood,140 although some relationships are\\n\\nknown.
-        \ For example, molecules with an ester functional group are often associated
-        with\\n\\n\\n                                       18Figure 4: Descriptor explanations
-        for solubility prediction model. The green and red bars\\nshow descriptors that
-        influence predictions positively and negatively, respectively. Dotted\\nyellow
-        lines show significance threshold (\u03B1 = 0.05) for the t-statistic. The MACCS
-        and\\nECFP descriptors indicate which substructures influence model predictions.
-        MACCS sub-\\nstructures may either be present in the molecule as is or may represent
-        a modification. ECFP\\nfingerprints are substructures in the molecule that affect
-        the prediction. MACCS descriptor\\nare used to obtain text explanations as shown.
-        Republished from Ref.10 with permission from\\nauthors. SMARTS annotations for
-        MACCS descriptors were created using SMARTSviewer\\n(smartsview.zbh.uni-hamburg.de,
-        Copyright: ZBH, Center for Bioinformatics Hamburg) de-\\nveloped by Schomburg
-        et al. 132.\\n\\n\\n\\n\\n\\n                                       19the \u2018fruity\u2019
-        scent. There are some exceptions though, like tert-amyl acetate which has a\\n\\n\u2018camphoraceous\u2019
-        rather than \u2018fruity\u2019 scent.140,141\\n\\n   In Seshadri et al. 31,
-        we trained a GNN model to predict the scent of molecules and utilized\\n\\ncounterfactuals9
-        and descriptor explanations10 to quantify scent-structure relationships. The\\n\\nMMACE
-        method was modified to account for the multi-label aspect of scent prediction.
-        This\\n\\nmodification defines molecules that differed from the instance molecule
-        by only the selected\\n\\nscent as counterfactuals. For instance, counterfactuals
-        of the jasmone molecule would be false\\n\\nfor the \u2018jasmine\u2019 scent
-        but would still be positive for \u2018woody,\u2019 \u2018floral\u2019 and \u2018herbal\u2019
-        scents.\\n\\n\\n\\n\\n\\nFigure 5:  Counterfactual for the 2,4 decadienal molecule.
-        \ The counterfactual indicates\\nstructural changes to ethyl benzoate that would
-        result in the model predicting the molecule\\nto not contain the \u2018fruity\u2019
-        scent. The Tanimoto96 similarity between the counterfactual and\\n2,4 decadienal
-        is also\\n\\n----\\n\\nQuestion: Are counterfactuals actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
-      headers:
-        accept:
-          - application/json
-        accept-encoding:
-          - gzip, deflate
-        connection:
-          - keep-alive
-        content-length:
-          - "6080"
-        content-type:
-          - application/json
-        host:
-          - api.openai.com
-        user-agent:
-          - AsyncOpenAI/Python 1.97.1
-        x-stainless-arch:
-          - arm64
-        x-stainless-async:
-          - async:asyncio
-        x-stainless-lang:
-          - python
-        x-stainless-os:
-          - MacOS
-        x-stainless-package-version:
-          - 1.97.1
-        x-stainless-raw-response:
-          - "true"
-        x-stainless-read-timeout:
-          - "60.0"
-        x-stainless-retry-count:
-          - "0"
-        x-stainless-runtime:
-          - CPython
-        x-stainless-runtime-version:
-          - 3.13.5
-      method: POST
-      uri: https://api.openai.com/v1/chat/completions
-    response:
-      body:
-        string: !!binary |
-          H4sIAAAAAAAAA4xUwW4bOQy9+ysIncdB4qZ24tvuBr0U7aHtntaFLWs4M2w1oiBScbxB/n2hmSae
-          ulnsXgYDPT3y8VHk4wzAUG3WYFxn1fXRz38/fmji4f2fnz8d2uNHf7P8FFd8R/63v+8Od6YqDN5/
-          Q6fPrAvHffSoxGGEXUKrWKJerd6+WS2Wb9+sBqDnGn2htVHn1zxfXC6u51dX88XlD2LH5FDMGv6a
-          AQA8Dt8iMdT4YNZwWT2f9ChiWzTrl0sAJrEvJ8aKkKgNaqoT6DgohkH1brf7Jhw24XETADZGct/b
-          dNyYNWzMlw4BHxymqFCTuCyCAtohZEHgBhznoJga6zRbL0ABevbosrcJYsKaXPEChmqlgmiT0oD6
-          IzScQNjnPXnSI9hQgzgMOiFewB9nGWxCaDFgKraCMuBD9JwQPDvrwXXYU/mRaB0OManGoNQcQTRl
-          pzlZX/RQQ86WHKUgq0Ch8RmDQ4iJIyYlFPD0HV/VeAHvOAE+2NLu6jweA4pigjZxjjKwOlRMbJV7
-          gQMmhIZzGAqgPlqnkywVWE9toNDCgbQ71URBM422fKaeBherX3owRM8yuiOa6+Moef5cP0JCP0rt
-          KEoF0vGhZOv4MDXJdTa0KOBsAOtLPeftkQv40qEgNBRqCq2A5LZF0dHSc2Ux8T3VCHYg271HoCDU
-          diqwP740qkiRiK44OnlO/9Gzk6aNqca3nNDjvQ0Ot+I4YXnTt5vwtAm73W46DgmbLLZMY8jeTwAb
-          AuuYsAzi1x/I08voeW5j4r2cUU1DgaTbJrTCoYyZKEczoE8zgK/DiOefptbExH3UrfJ3HNJdLW4X
-          Y0Bz2ioTeFwkAEZZrZ8A14vb6pWQ2xrVkpfJnjDOug7rE/e0VGyuiSfAbFL4r3peiz0WT6H9P+FP
-          gHMYFevtqaOvXUtY1u6/XXsxehBsBNM9OdwqYSrNqLGx2Y8b0chRFPttQ6HFFBONa7GJW7dcrZY3
-          zfLm0syeZv8AAAD//wMADF3tyx8GAAA=
-      headers:
-        CF-RAY:
-          - 96665ceb6c45f555-SJC
-        Connection:
-          - keep-alive
-        Content-Encoding:
-          - gzip
-        Content-Type:
-          - application/json
-        Date:
-          - Mon, 28 Jul 2025 18:15:39 GMT
-        Server:
-          - cloudflare
-        Strict-Transport-Security:
-          - max-age=31536000; includeSubDomains; preload
-        Transfer-Encoding:
-          - chunked
-        X-Content-Type-Options:
-          - nosniff
-        access-control-expose-headers:
-          - X-Request-ID
-        alt-svc:
-          - h3=":443"; ma=86400
-        cf-cache-status:
-          - DYNAMIC
-        openai-organization:
-          - future-house-xr4tdh
-        openai-processing-ms:
-          - "1587"
-        openai-project:
-          - proj_RpeV6PrPclPHBb5GlExPXSBj
-        openai-version:
-          - "2020-10-01"
-        x-envoy-upstream-service-time:
-          - "1591"
-        x-ratelimit-limit-requests:
-          - "10000"
-        x-ratelimit-limit-tokens:
-          - "30000000"
-        x-ratelimit-remaining-requests:
-          - "9999"
-        x-ratelimit-remaining-tokens:
-          - "29998541"
-        x-ratelimit-reset-requests:
-          - 6ms
-        x-ratelimit-reset-tokens:
-          - 2ms
-        x-request-id:
-          - 27351c35-5230-407e-a392-137197c64966
-      status:
-        code: 200
-        message: OK
-  - request:
-      body:
-        "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
-        the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from wellawatteUnknownyearaperspectiveon pages 20-22: Geemi P. Wellawatte, Heta
-        A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
-        of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n----\\n\\nnal
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n------------\\n\\nnal
         molecule.  The counterfactual indicates\\nstructural changes to ethyl benzoate
         that would result in the model predicting the molecule\\nto not contain the
         \u2018fruity\u2019 scent. The Tanimoto96 similarity between the counterfactual
@@ -3858,7 +3871,7 @@ interactions:
         the input to models informs the XAI method.\\n\\n\\nConclusion and outlook\\n\\n\\nWe
         should seek to explain molecular property prediction models because users are
         more\\n\\nlikely to trust explained predictions, and explanations can help assess
-        if the model is learning\\n\\nt\\n\\n----\\n\\nQuestion: Are counterfactuals
+        if the model is learning\\n\\nt\\n\\n------------\\n\\nQuestion: Are counterfactuals
         actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
@@ -3868,13 +3881,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "6100"
+          - "6376"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -3884,7 +3897,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -3900,22 +3913,22 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFTLbhtJDLzrK4g+S4YkO+tEt2QPC2ywyCVYbBAZAqebo2HcLzQ5jrSG
-          /z3oGVmaJA6QiwBNkcUqPvpxBmDYmQ0Y26HakP3i3fGfNvu3n/z7T+HDf3F10/4f+r8e3r3/+9/0
-          xcxrRmq+kNXnrCubQvaknOII20KoVFlXt6+ub9d/vLq+HYCQHPmats+6uEmL9XJ9s1itFuvlKbFL
-          bEnMBj7PAAAeh98qMTo6mA0s589fAongnszmHARgSvL1i0ERFsWoZn4BbYpKcVD9uI0AWyN9CFiO
-          W7OBrfnYEdDBUskKjsX2IiSgHUEvBKmFkDzZ3mMBm/qoVFq02qMX4Ah0yB45ctxD49HeL5p0gMGu
-          QJvKJDmXlKnoEXIhx7Z2Ta7gzx8osRA4Elu4IQcogEMkNp6gIYtVk3Z0HAIL5UJCUcdQ21Fgix5E
-          S2+1LyTQYmDPWEATuBRwlExFZT5QSMYiNAeMrip8YEfTkhyF953KFXx8Luqo5TgWPJkjga+sHQSO
-          HNDXLipGS9CWFAChQaFzLDS9juF1LgVFa+/O0k9dYhpLgtJBoeN95wcd1TsX6JU967EOoI+OSq3n
-          Kk+1EZLj9lj//dR8JplDwPsKakdVXC7Vbi2tKflhZt/36Wpr5uPaFPL0UI3txKZCdX1Wy218mi5b
-          obYXrLsee+8nAMaYFIep1zW/OyFP58X2aZ9LauSHVNNyZOl2hVBSrEssmrIZ0KcZwN1wQP13N2Fy
-          SSHrTtM9DeVW69evR0JzudkJvLw5oZoU/QS4frOev0C5c6TIXiZXaCzajtwl93Ky2DtOE2A2Mf6z
-          npe4R/Mc979DfwGspazkdpeLeymsUH3UfhV2bvQg2AiVB7a0U6ZSh+Goxd6P742RoyiFXctxTyUX
-          Hh6dOszZ0+wbAAAA//8DAH8mwjByBQAA
+          H4sIAAAAAAAAAwAAAP//jFRNjxs3DL37VxA62wuP4/UWvi0C9FI0lwQ5tA4MWuLMsNXHRKQSO4v9
+          74VmvLbT3QK9DKB55ON7lMinGYBhZ7ZgbI9qw+AX75vfPrT3zaN+/vH7rx+Pm2PzuPnjq3v8wbJZ
+          mnnNSIe/yOpL1p1NYfCknOIE20yoVFmbh/v1+t3qYfkwAiE58jWtG3SxTovVcrVeNM1idea1fWJL
+          Yrbw5wwA4Gn8VonR0dFsYTl/+RNIBDsy20sQgMnJ1z8GRVgUo5r5FbQpKsVR9dMuAuyMlBAwn3Zm
+          CzvzPpWolFu0WtALYCZwJDbzgRygANrqEA+egCNoTzAyHhVSCyF5ssVjhiGngbKeYMjkeMyB0bfc
+          waeeTmfiluNEe80UzcVqySTwnbWHwJEDenCjF0vQ5hQA4YBCL2kEh6JTeFWTUZRjB7anwBb9ixqm
+          qboQ0HHwGLHqmkxmGjIJRZ30XFKvcubwvWfbj9EtBvaMGTSBSwE5VkbKKnPA6MYYGTALTcdL08b6
+          UIRquz5h5JA0gXBgj5n1BG3KV6uBUEqmQFFHntruIdX7Y/RjaFuqttFOyqOdSpy0pwyBtE9usode
+          ElQeTpHcHfx8z6/b0XPXe+76czuKUFumiiU6ylWfqy2uqkJy3J7q6dX9147vzHx6aJk8fau+9mJT
+          punBNcsLXoTcngN2JBVr0Qvt4vPt683UFsE6PLF4fwNgjEkn/XVuvpyR58uk+NQNOR3kX6mm5cjS
+          7zOhpFinQjQNZkSfZwBfxoksPw2ZGXIKg+41/U1juebdejMRmusSuIGbX86oJkV/A6w36/kblHtH
+          iuzlZqyNRduTu+ZedwAWx+kGmN0Yf63nLe7JPMfu/9BfAWtpUHL764i/FZapbsn/Crs0ehRshPI3
+          trRXplwvw1GLxU8LzMhJlMK+5dhRHjJPW6wd9qtmQ5v7tV07M3ue/QMAAP//AwCK+MpYzgUAAA==
       headers:
         CF-RAY:
-          - 96665ceb6b2beb35-SJC
+          - 96a9b56299f215d4-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -3923,7 +3936,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:39 GMT
+          - Tue, 05 Aug 2025 22:25:08 GMT
         Server:
           - cloudflare
         Strict-Transport-Security:
@@ -3941,13 +3954,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "1754"
+          - "1712"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
         x-envoy-upstream-service-time:
-          - "1765"
+          - "1720"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -3955,13 +3968,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998540"
+          - "29998471"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
-          - 2ms
+          - 3ms
         x-request-id:
-          - 802b946f-ebac-45d3-92cc-6c065a461c68
+          - req_f6c874a7a5da45838d5a7896a9a6813e
       status:
         code: 200
         message: OK
@@ -3969,80 +3982,83 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from wellawatteUnknownyearaperspectiveon pages 9-12: Geemi P. Wellawatte, Heta
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from wellawatteUnknownyearaperspectiveon pages 16-20: Geemi P. Wellawatte, Heta
         A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
         of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n----\\n\\nthat
-        gives subgraph importance for small molecule activity prediction. On the\\n\\nother
-        hand, similarity maps compare model predictions for two or more molecules based
-        on\\n\\ntheir chemical fingerprints.83 Similarity maps provide atomic weights
-        or predicted probabil-\\n\\n\\n                                       9ity difference
-        between the molecules by removing one atom at a time. These weights can\\n\\nthen
-        be used to color the molecular graph and give a visual presentation. ChemInformatics\\n\\nModel
-        Explorer (CIME) is an interactive web based toolkit which allows visualization
-        and\\n\\ncomparison of different explanation methods for molecular property
-        prediction models.84\\n\\n\\nSurrogate models\\n\\n\\nOne approach to explain
-        black box predictions is to fit a self-explaining or interpretable\\n\\nmodel
-        to the black box model, in the vicinity of one or a few specific examples. These
-        are\\n\\nknown as surrogate models. Generally, one model per explanation is
-        trained. However, if we\\n\\ncould find one surrogate model that explained the
-        whole DL model, then we would simply\\n\\nhave a globally accurate interpretable
-        model. This means that the black-box model is no\\n\\nlonger needed.79 In the
-        work by White 79, a weighted least squares linear model is used as\\n\\nthe
-        surrogate model. This model provides natural language based descriptor explanations
-        by\\n\\nreplacing input features with chemically interpretable descriptors.
-        This approach is similar\\n\\nto the concept-based explanations approach used
-        by McGrath et al. 85, where human under-\\n\\nstandable concepts were used in
-        place of input features in acquisition of chess knowledge in\\n\\nAlphaZero.
-        Any of the self-explaining models detailed in the Self-explaining models section\\n\\ncan
-        be used as a surrogate model.\\n\\n   The most commonly used surrogate model
-        based method is Locally Interpretable Model\\n\\nExplanations (LIME).35 LIME
-        creates perturbations around the example of interest and fits\\n\\nan interpretable
-        model to these local perturbations. Ribeiro et al. 35 mathematically define\\n\\nan
-        explanation \u03BE for an example \u20D7x using Equation 4.\\n\\n\\n\\n                                   \u03BE(\u20D7x)
-        = arg min L(f, g, \u03C0x) + \u2126(g)                           (4)\\n                                      g\u2208G\\n\\n
-        \  Here f is the black box model and g \u2208G is the interpretable explanation
-        model. G is\\n\\na class of potential interpretable models (e.g.:  linear models).
-        \u03C0x is a similarity measure\\n\\n\\n\\n                                       10between
-        original input \u20D7x and it\u2019s perturbed input \u20D7x\u2032. In context
-        of molecular data, this can\\n\\nbe a chemical similarity metric like Tanimoto86
-        similarity between fingerprints. The goal for\\n\\nLIME is to minimize the loss,
-        L, such that f is closely approximated by g. \u2126is a parameter\\n\\nthat
-        controls the complexity (sparsity) of g. Ribeiro et al. 35 termed the agreement
-        (how low\\n\\nthe loss is) between f and g as the \u201Cfidelity\u201D.\\n\\n
-        \  GraphLIME87 and LIMEtree88 are modifications to LIME as applicable to graph
-        neural\\n\\nnetworks and regression trees, respectively. LIME has been used
-        in chemistry previously,\\n\\nsuch as Whitmore et al. 89 who used LIME to explain
-        octane number predictions of molecules\\n\\nfrom a random forest classifier.
-        Mehdi and Tiwary 90 used LIME to explain thermodynamic\\n\\ncontributions of
-        features. Gandhi and White 10 use an approach similar to GraphLIME,\\n\\nbut
-        use chemistry specific fragmentation and descriptors to explain molecular property
-        pre-\\n\\ndiction. Some examples are highlighted in the Applications section.
-        \ In recent work by\\n\\nMehdi and Tiwary 90, a thermodynamic-based surrogate
-        model approach was used to inter-\\n\\npret black-box models. The authors define
-        an \u201Cinterpretation free energy\u201D which can be\\n\\nachieved by minimizing
-        the surrogate model\u2019s uncertainty and maximizing simplicity.\\n\\n\\nCounterfactual
-        explanations\\n\\n\\nCounterfactual explanations can be found in many fields
-        such as statistics, mathematics and\\n\\nphilosophy.91\u201394 According to
-        Woodward and Hitchcock 92, a counterfactual is an example\\n\\nwith minimum
-        deviation from the initial instance but with a contrasting outcome. They\\n\\ncan
-        be used to answer the question, \u201Cwhich smallest change could alter the
-        outcome of an\\n\\ninstance of interest?\u201D While the difference between
-        the two instances is based on the exis-\\n\\ntence of similar worlds in philosophy,95
-        a distance metric based on molecular similarity is\\n\\nemployed in XAI for
-        chemistry. For example, in the work by Wellawatte et al. 9 distance\\n\\nbetween
-        two molecules is defined as the Tanimoto distance96 between ECFP4 fingerprints.97\\n\\nAdditionally,
-        Mohapatra et al. 98 introduced a chemistry-informed graph representation for\\n\\ncomputing
-        macromolecular similarity. Contrastive explanations are peripheral to counterfac-\\n\\n\\n
-        \                                      11tual explanations. Unlike the counterfactual
-        approach, contrastive approach employ a dual\\n\\noptimization method, which
-        works by generating a similar and a dissimilar (counterfactuals)\\n\\nexample.
-        Contrastive explanations can interpret the model by identifying contribution
-        of\\n\\npresence \\n\\n----\\n\\nQuestion: Are counterfactuals actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n------------\\n\\nssion
+        challenge and is\\n\\nimportant for chemical process design, drug design and
+        crystallization.133\u2013136 In our previous\\n\\nworks,9,10 we implemented
+        and trained an RNN model in Keras to predict solubilities (log\\n\\nmolarity)
+        of small molecules.127 The AqSolDB curated database137 was used to train the\\n\\nRNN
+        model.\\n\\n   In this task, counterfactuals are based on equation 6. Figure
+        3 illustrates the generated\\n\\nlocal chemical space and the top four counterfactuals.
+        Based on the counterfactuals, we ob-\\n\\nserve that the modifications to the
+        ester group and other heteroatoms play an important role\\n\\nin solubility.
+        These findings align with known experimental and basic chemical intuition.134\\n\\nFigure
+        4 shows a quantitative measurement of how substructures are contributing to
+        the pre-\\n\\n\\n\\n                                       16Figure 2: Descriptor
+        explanations along with natural language explanation obtained for BBB\\npermeability
+        of Alprozolam molecule. The green and red bars show descriptors that influ-\\nence
+        predictions positively and negatively, respectively. Dotted yellow lines show
+        significance\\nthreshold (\u03B1 = 0.05) for the t-statistic. Molecular descriptors
+        show molecule-level proper-\\nties that are important for the prediction. ECFP
+        and MACCS descriptors indicate which\\nsubstructures influence model predictions.
+        MACCS explanations lead to text explanations\\nas shown. Republished from Ref.10
+        with permission from authors. SMARTS annotations for\\nMACCS descriptors were
+        created using SMARTSviewer (smartsview.zbh.uni-hamburg.de,\\nCopyright: ZBH,
+        Center for Bioinformatics Hamburg) developed by Schomburg et al. 132.\\n\\n\\n\\n\\n\\n
+        \                                      17diction. For example, we see that adding
+        acidic and basic groups as well as hydrogen bond\\n\\nacceptors, increases solubility.
+        Substructure importance from ECFP97 and MACCS138 de-\\n\\nscriptors indicate
+        that adding heteroatoms increases solubility, while adding rings structures\\n\\nmakes
+        the molecule less soluble. Although these are established hypotheses, it is
+        interesting\\n\\nto see they can be derived purely from the data via DL and
+        XAI.\\n\\n\\n\\n\\n\\nFigure 3: Generated chemical space for solubility prediction
+        using the RNN model. The\\nchemical space is a 2D projection of the pairwise
+        Tanimoto similarities of the local coun-\\nterfactuals. Each data point is colored
+        by solubility. Top 4 counterfactuals are shown here.\\nRepublished from Ref.9
+        with permission from the Royal Society of Chemistry.\\n\\n\\n\\nGeneralizing
+        XAI \u2013 interpreting scent-structure relationships\\n\\n\\nIn this example,
+        we show how non-local structure-property relationships can be learned with\\n\\nXAI
+        across multiple molecules. Molecular scent prediction is a multi-label classification
+        task\\n\\nbecause a molecule can be described by more than one scent. For example,
+        the molecule\\n\\njasmone can be described as having \u2018jasmine,\u2019 \u2018woody,\u2019
+        \u2018floral,\u2019 and \u2019herbal\u2019 scents.139 The\\n\\nscent-structure
+        relationship is not very well understood,140 although some relationships are\\n\\nknown.
+        \ For example, molecules with an ester functional group are often associated
+        with\\n\\n\\n                                       18Figure 4: Descriptor explanations
+        for solubility prediction model. The green and red bars\\nshow descriptors that
+        influence predictions positively and negatively, respectively. Dotted\\nyellow
+        lines show significance threshold (\u03B1 = 0.05) for the t-statistic. The MACCS
+        and\\nECFP descriptors indicate which substructures influence model predictions.
+        MACCS sub-\\nstructures may either be present in the molecule as is or may represent
+        a modification. ECFP\\nfingerprints are substructures in the molecule that affect
+        the prediction. MACCS descriptor\\nare used to obtain text explanations as shown.
+        Republished from Ref.10 with permission from\\nauthors. SMARTS annotations for
+        MACCS descriptors were created using SMARTSviewer\\n(smartsview.zbh.uni-hamburg.de,
+        Copyright: ZBH, Center for Bioinformatics Hamburg) de-\\nveloped by Schomburg
+        et al. 132.\\n\\n\\n\\n\\n\\n                                       19the \u2018fruity\u2019
+        scent. There are some exceptions though, like tert-amyl acetate which has a\\n\\n\u2018camphoraceous\u2019
+        rather than \u2018fruity\u2019 scent.140,141\\n\\n   In Seshadri et al. 31,
+        we trained a GNN model to predict the scent of molecules and utilized\\n\\ncounterfactuals9
+        and descriptor explanations10 to quantify scent-structure relationships. The\\n\\nMMACE
+        method was modified to account for the multi-label aspect of scent prediction.
+        This\\n\\nmodification defines molecules that differed from the instance molecule
+        by only the selected\\n\\nscent as counterfactuals. For instance, counterfactuals
+        of the jasmone molecule would be false\\n\\nfor the \u2018jasmine\u2019 scent
+        but would still be positive for \u2018woody,\u2019 \u2018floral\u2019 and \u2018herbal\u2019
+        scents.\\n\\n\\n\\n\\n\\nFigure 5:  Counterfactual for the 2,4 decadienal molecule.
+        \ The counterfactual indicates\\nstructural changes to ethyl benzoate that would
+        result in the model predicting the molecule\\nto not contain the \u2018fruity\u2019
+        scent. The Tanimoto96 similarity between the counterfactual and\\n2,4 decadienal
+        is also\\n\\n------------\\n\\nQuestion: Are counterfactuals actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4051,13 +4067,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "6077"
+          - "6356"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4067,7 +4083,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4083,22 +4099,23 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFRNb9tIDL37VxBzloPYcZLGx+0lXWC3e0hbFHVh0DOUxWQ+1CGVxAjy
-          34uRlUjdZoG96DCPfOR7JPU0AzDszBqMbVBtaP38j8Nfdfvx3fn1l7+vPq0+frq+lX/Szef0Ve/+
-          /GGqkpF2t2T1JevEptB6Uk7xCNtMqFRYF5fnZ5fLi/Ozyx4IyZEvaftW56s0X54uV/PFYr48HRKb
-          xJbErOHbDADgqf+WFqOjR7OG0+rlJZAI7smsX4MATE6+vBgUYVGMaqoRtCkqxb7rp00E2BjpQsB8
-          2Jg1bMz71EWlXKPVDj3QY+sxYhElgJnAkdjMO3KAAvSIRbLAA2sDgSMH9ODonvsMqHMKoA1Byrzn
-          iB44loYswa7TYxZC6SijKMc9pE5tCnQCNw0dADmAJsAoD5R7oh8dSU+danhoUEECek+iYBuMewKb
-          Ou8AvQ4JA2GJxziWTzVw0UmiJ/Ah9qG9M49asJA82c5jhjaTY9uX7McmVSkxsUigEwIExwN1IM1s
-          K5DONsWkQn2DkUPSNEbtSB+I4qRSzXFPuc0cVaoi25FSDhwJhAN7zKyH3hih3+eCfY+48zTUPECb
-          0z07AmnJcs12sKigqINRntD1FoPjuqZMUV8sqyDgXZmJNhSKyLrzUKcMXXSUiwxXUIwO2lRWitH7
-          Q3GJ60NBRuvkZGOq47Jl8nRfHNiKTZnK0l1t4vN0QzPVnWA5kNh5PwEwxqRHzeU2vg/I8+s1+LRv
-          c9rJv1JNzZGl2WZCSbFsvmhqTY8+zwC+91fX/XJIps0ptLrVdEd9ucVy9e5IaMZDn8CLiwHVpOgn
-          wNnFqnqDcutIkb1MTtdYtA25MXe8c+wcpwkwmwj/vZ+3uI/iOe7/D/0IWEutktuOw3wrLFP5E/5X
-          2KvRfcNGKN+zpa0y5TIMRzV2/viTMnIQpbCdHMMwzNnz7CcAAAD//wMA983HQacFAAA=
+          H4sIAAAAAAAAA4xUwW4bOQy9+ysInceG7TpN61tbYC/BFli0QA91YdAazgxbjaQVKSdGkH8vNGPH
+          jpsF9mKM9UTyvSeSjxMAw7VZg7Edqu2jm35a3H1uPq/mP7+9+2Dv/sHdl93Hv+/2rtV/vTVViQi7
+          n2T1FDWzoY+OlIMfYZsIlUrWxe3NavVmeTu/HYA+1ORKWBt1ugrT5Xy5mi4W0+X8GNgFtiRmDd8n
+          AACPw2+h6Gt6MGuYV6eTnkSwJbN+vgRgUnDlxKAIi6JXU51BG7ySH1g/bjzAxkjue0yHjVnDxnzt
+          COjBUooKNYvNIiSgHUEWgtCADdkrpQatZnQC7KEPjmx2mCAmqtkWC2AQKRVETMoD6g7QhAQSXN6x
+          Yz0A+hrEkteLwBl8uqqAiaAlT6m4CRqAHqILicAFiw5sRz2XD4loacjJNXnl5gCiKVvNCV3hww1b
+          LDWKIFRg37hM3hLEFCIlZRJw/Ite5TiDv16wr65TBiBRStCmkKMMgR0ppYAaeoF7SgRNyH7QsCMQ
+          bv0Q77UCdOWfb+GetSsKKXFPXtEBe808WjMwuDKs+uNJhkpZRrNe88J26Fs6uoCukD5lPflQgWTb
+          Acqo8lCoHd+ZBq1jjtIZfIq2rnTcyZEZfO1ICBr2NftWQHLbkuhY9ZpzTGHPNQEOmnDnCNgLt52W
+          FtMAXbi/aLSTGBKw6Iubg4xRMtqOaU9Qk3A5OquabUw1Nn0iR3v0lrZiQ6Kx+d8/w8W9LffYkhSo
+          QSe08U+Xg5SoyYJljn127gJA74OOTVFG+McReXoeWhfamMJOrkJNw56l2yZCCb4MqGiIZkCfJgA/
+          huWQX8y7iSn0UbcaftFQbvHm5rgdzHkfXcCrd0dUg6K7BN6fkBcptzUpspOLDWMs2o7qc+x5HWGu
+          OVwAkwvhf/J5Lfconn37f9KfAWspKtXb81i8di1RWdj/de3Z6IGwEUp7trRVplQeo6YGsxt3qZGD
+          KPXbhn1LKSYeF2oTt8vFW3p7s7Kr2kyeJr8BAAD//wMA/BcJjFkGAAA=
       headers:
         CF-RAY:
-          - 96665ceb69a4cefd-SJC
+          - 96a9b5629adccf45-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4106,9 +4123,11 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:39 GMT
+          - Tue, 05 Aug 2025 22:25:08 GMT
         Server:
           - cloudflare
+        Strict-Transport-Security:
+          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -4122,15 +4141,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "1869"
+          - "1910"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "1874"
+          - "1914"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4138,13 +4155,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998546"
+          - "29998474"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
-          - 2ms
+          - 3ms
         x-request-id:
-          - req_c31bdf1c31ce474267efebbfc009fcde
+          - req_1a776dc1da4f49d59f67a73ba32ec37a
       status:
         code: 200
         message: OK
@@ -4152,14 +4169,17 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
         from wellawatteUnknownyearaperspectiveon pages 14-16: Geemi P. Wellawatte, Heta
         A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
         of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n----\\n\\nsame
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n------------\\n\\nsame
         optimization problem.100 Grabocka\\n\\net al. 111 have developed a method named
         Adversarial Training on EXplanations (ATEX)\\n\\nwhich improves model robustness
         via exposure to adversarial examples.  While there are\\n\\nconceptual disparities,
@@ -4223,7 +4243,7 @@ interactions:
         solubility prediction is a classic cheminformatics regression challenge and
         is\\n\\nimportant for chemical process design, drug design and crystallization.133\u2013136
         In our previous\\n\\nworks,9,10 we implemented and trained an RNN model in Keras
-        to predict solubilities (log\\n\\nmolarity) of small molecules.127 The AqS\\n\\n----\\n\\nQuestion:
+        to predict solubilities (log\\n\\nmolarity) of small molecules.127 The AqS\\n\\n------------\\n\\nQuestion:
         Are counterfactuals actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
@@ -4233,13 +4253,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "6062"
+          - "6338"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4249,7 +4269,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4265,23 +4285,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFRNiyM3EL37VxQ6JdA2Y89OZvHRA4Ec8kVCQogXUy1Vd9dGrVKkktdm
-          mP++qO0Ze3YnkIsN/aqe3tOr0uMMwLAzazB2QLVj9PPN8cfu379+/emj/Lba93/yLxFx/0dfHh5+
-          uP/ZNLVD2o9k9blrYWWMnpQlnGCbCJUq6/L+7vZ+9d3d7fsJGMWRr2191Pk7ma9uVu/my+V8dXNu
-          HIQtZbOGv2cAAI/Tb5UYHB3MGm6a5y8j5Yw9mfVLEYBJ4usXgzlzVgxqmgtoJSiFSfXjNgBsTS7j
-          iOm4NWvYmgcpQSl1aLWgz4CJAG01ha0nwAw60BFikj07ghzJcscWRnH1H2tlBhUYxZMtHhNkTcVq
-          SVR7UcFiAE/oapWjzIkc2AFDTxk4VOpISZnyAr6XBHTAeq9NxXQgmAwcFKSD1ou4eZuQA7SYElOC
-          bzabzbcQKY00iYGYyPHkoAH7yh3QIXoMZ8259D1lJfe1l+lYTK0cjp4toGUHfZISqwh8tkoNUL0l
-          Dj2w1r6zCpoINpvNAn4fKL+6UA6Z+0EzoOc+wCfWocqixCMFRQ8dB8ehzw04GiVkTaj1hEE+wXB0
-          SeIgLVvgauxEnAGDg1yqT6oJInDofKFgJxlnXS171uMCvoxch5JBuo4SxInRooe+sMPa30m6yrbm
-          1wdop4mIkquy5yxfwkY7MO2vhkWKWhkpL7amOc1gIk/7yr/LVhLVWVzebMPT9eQm6krGujiheH8F
-          YAiip7Tqznw4I08vW+Klj0na/EWr6ThwHnaJMEuoG5FVopnQpxnAh2kby6sFMzHJGHWn8g9Nxy1X
-          d3cnQnN5AK7h2zOqouivgNv7980blDtHiuzz1Uobi3Ygd+m97D8Wx3IFzK6Mf63nLe6TeQ79/6G/
-          ANZSVHK7y3a9VZaovpD/VfZy0ZNgkynt2dJOmVINw1GHxZ8eL5OPWWncdRx6SjHx9ILVMGdPs88A
-          AAD//wMA2J8Vr78FAAA=
+          H4sIAAAAAAAAAwAAAP//jFRNjxs3DL37VxA6tYC98NcmqY8OUCRNUaBob3Vg0BJnhq1GEkhp18Zi
+          /3uhGa/tTbZALzro6ZHvUSSfJgCGndmAsR1m2yc/+7j48lu7Kr9+Wuvvn/ORHv+QlfvpSxOWnz/9
+          YqaVEQ9/k80vrDsb++QpcwwjbIUwU426eH+/Xq+W7+cfBqCPjnyltSnP1nG2nC/Xs8VitpyfiV1k
+          S2o28NcEAOBpOKvE4OhoNjCfvtz0pIotmc3lEYCR6OuNQVXWjCGb6RW0MWQKg+qnXQDYGS19j3La
+          mQ3szMdYQiZp0OaCXgGFAG01hQdPgAq5oxMkiQ/sCLS0LWmFFZoo0EfHDVscb3KEPnqyxaOAZik2
+          F6HhHm3H9EDgSFnI1YCJJDPpHfwcBeiItZpT4FAzwiD7mCE2cPAxutlBkAMcUIRJ4IftdvsjJJKe
+          htyQhBwPuqdgX3kCOiaP4SyRHYXMDZP7XvuQF+UQjyfPFtCyg1ZiSZA7zEBDSRzgi0mqpLMGGtjb
+          7fYO/uxIXxWRg3LbZQX03AZ45NxVUSTcU8jooeHgOLQ6BUd9DJoFM4cWuvj4jRkFiwHaMvzFucDo
+          wXYY2teF1kS22oNYso096RS02K7+KIUOgx0SnJzE1MUDW+CaZlStEAW01KRUOwKrKVbAlCSi7YZ2
+          IIWi1BRfv8xJacHRA/mYqinA4CDmjqSS/KXKjx3JVTnNzm1wAiE/Puk4jV1ohTNb9Hc7Mx0bV8jT
+          AwZLe7VRaGzgxfyCFyW35x5b0oo16JV24fl2GoSaoliHMRTvbwAMIeZRQZ3Dr2fk+TJ5PrZJ4kG/
+          oZqGA2u3F0KNoU6Z5pjMgD5PAL4OE15eDa1JEvuU9zn+Q0O6xWqxGgOa61K5he/PaI4Z/Q2wXn+Y
+          vhFy7ygje71ZE8ai7chdudedgsVxvAEmN8a/1/NW7NE8h/b/hL8C1lLK5PbX2X3rmVDduv/17FLo
+          QbBRkge2tM9MUj/DUYPFjwvR6Ekz9fuGQ0uShMet2KT9cvGO3t2v7dqZyfPkXwAAAP//AwD+VLwd
+          HgYAAA==
       headers:
         CF-RAY:
-          - 96665cf25822fa86-SJC
+          - 96a9b56d1c6dfab6-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4289,7 +4310,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:40 GMT
+          - Tue, 05 Aug 2025 22:25:10 GMT
         Server:
           - cloudflare
         Transfer-Encoding:
@@ -4305,7 +4326,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "2029"
+          - "1676"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -4313,7 +4334,7 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "2038"
+          - "1703"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4321,13 +4342,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998552"
+          - "29998483"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
-          - 2ms
+          - 3ms
         x-request-id:
-          - req_ba4bbca77fa7440662f0f273117ecbbf
+          - req_d04450d28c8eafed7ccad747a940dc7a
       status:
         code: 200
         message: OK
@@ -4335,14 +4356,17 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
         from wellawatteUnknownyearaperspectiveon pages 3-5: Geemi P. Wellawatte, Heta
         A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
         of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n----\\n\\n
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n------------\\n\\n
         a passive characteristic of a model, whereas explainability\\n\\nis an active
         characteristic which is used to clarify the internal decision-making process.\\n\\nNamely,
         an explanation is extra information that gives the context and a cause for one
@@ -4407,7 +4431,8 @@ interactions:
         \ We present an example evaluation of the SHAP explanation method based on the
         above\\n\\nattributes.44 Shapley values were proposed as a local explanation
         method based on feature\\n\\nattribution, as they offer a complete explanation
-        - each feature i\\n\\n----\\n\\nQuestion: Are counterfactuals actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        - each feature i\\n\\n------------\\n\\nQuestion: Are counterfactuals actionable?
+        [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4416,13 +4441,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "6081"
+          - "6357"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4432,7 +4457,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4448,23 +4473,23 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFRNbxs3EL3rVwx4SQtIhiXHda2bE/TgtjkUKNKkVSCNuLO7k3JJgjOU
-          tDX83wvuKpLcpkAvC+y8mcc3n08TAMOVWYKxLartopu96d/V6ffHd98ffr57/KV/89Purbu9f998
-          9P7HazMtEWH7max+ibqyoYuOlIMfYZsIlQrr/O725m7x3e3N/QB0oSJXwpqos9dhtrhevJ7N57PF
-          kde2gS2JWcIfEwCAp+FbJPqKDmYJ19Mvlo5EsCGzPDkBmBRcsRgUYVH0aqZn0Aav5AfVm83mswS/
-          8k8rD7AykrsOU78yS1iZtyF7pVSj1YxOABOBRaUmJP6LKkABFyw64OIWEymW1AXYww+H6JA9bh3B
-          wyN88+Hh8VvYksUsBNpSDzR6gESyXLMF9kWpJbmCX1sCpYNCxWKzCAmgauJtVhKoQwLaocuo7JuR
-          yI9PT4G9dbkq9lcPttiKhOkr2LdsW8DCVej2LWlLCViBBawjTNCGPbCPWaEm1JxIwKKHLYFt0TdU
-          gQboQsV1X1KAkDVmvYLfWnYE9ivV8kEHeWxZXQ8Ot+TGwuFJW6lWYaODpRR1Wn44QcwpBiloS4l8
-          CWe/C25HAlyRV677kuUoTYq0QbuAtqiwD9lV4AgH0RXX9cBSNNvQkUwBHTe+MOxZ20GBDd5SVAj1
-          SR471v5qZabjeCRytCs9WosNicqY3K/888pvNpvLCUtUZ8Ey4D47dwGg9+E4JWW2Px2R59M0u9DE
-          FLbyj1BTs2dp14lQgi+TKxqiGdDnCcCnYWvyi0UwMYUu6lrDnzQ8N19c34+E5ryoF/B8fkQ1KLoL
-          4GZxXLeXlOuKFNnJxeoZi7al6hx73lPMFYcLYHKR+L/1fI17TJ5983/oz4AtTaVqHRNVbF/mfHZL
-          VC7Zf7mdCj0INkJpx5bWypRKMyqqMbvxyBjpRalb1+ybchV4uDSlmZPnyd8AAAD//wMA0YvWlGcF
-          AAA=
+          H4sIAAAAAAAAAwAAAP//jFTbjhs3DH33VxB6SQvYC9vxXuq3bfrQNEUvSIsGqQOb1nBmuNVIgkit
+          11nsvxea8dpOuwX6MsDwkEc8vD2OAAxXZgnGtqi2i27yZvbup+aX2/fff+TFd9ff3v3IP3e/Tt/9
+          /sPlDqdmXCLC9o6sPkdd2NBFR8rBD7BNhEqFdXZ9uVi8nl9Pb3qgCxW5EtZEnSzCZD6dLyaz2WR+
+          4LVtYEtilvDnCADgsf+WFH1FD2YJ0/GzpSMRbMgsj04AJgVXLAZFWBS9mvEJtMEr+T7rzWZzJ8Gv
+          /OPKA6yM5K7DtF+ZJazMm5C9UqrRakYngInAolITEn+mClDABYsOuLjFRIpFugB7oIfokD1uHcHt
+          W/jqw+3br2FLFrMQaEv7Zw+QSJZrtsC+ZGpJLuC3lkDpQaFisVmEBFA18TYrCdQhAd2jy6jsm4HI
+          D0+Pgb11uSr2V2iLraQwfgW7lm0LiWpKAhpg15K2lIAVWMA6wgRt2AH7mBVqQs2JBCx62BLYFn1D
+          VQnsQsX1vmiAkDVmvYA/WnYE9oVy+aB9fmxZ3R4cbskNlTslV8pV2OjBUoo6Lj+cIOYUgxCg48YL
+          7Fjb3u0s8FgTEPb2UNeYwj1XhVW4abV0Q0MvbRDRt2dQaUN2FTjCXljFdU2JvBZdNnQkFyszHuYi
+          kaP70py12JBomI9vjnAWqtbcYUNSoBqd0Mo/rfxmszmfvER1FiyD77NzZwB6Hw7TU2b+0wF5Ok65
+          C01MYSv/CDU1e5Z2nQgl+DLRoiGaHn0aAXzqtyl/sSAmptBFXWv4i/rnZvOr64HQnBb4DJ7dHFAN
+          iu4MeH1zOX6Bcl2RIjs5W0lj0bZUnWJP+4u54nAGjM6E/zufl7gH8eyb/0N/AqylqFStY6KK7Zea
+          T26JyoX7L7djofuEjVC6Z0trZUqlGRXVmN1wfIzsRalb1+ybci14uEB1XM9nV3R1ubCLyoyeRn8D
+          AAD//wMATYPDP4oFAAA=
       headers:
         CF-RAY:
-          - 96665cf75e95eb35-SJC
+          - 96a9b56e892415d4-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4472,7 +4497,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:41 GMT
+          - Tue, 05 Aug 2025 22:25:10 GMT
         Server:
           - cloudflare
         Transfer-Encoding:
@@ -4488,7 +4513,7 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "2341"
+          - "1587"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
@@ -4496,7 +4521,7 @@ interactions:
         strict-transport-security:
           - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "2371"
+          - "1593"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4504,13 +4529,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998546"
+          - "29998479"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
-          - 2ms
+          - 3ms
         x-request-id:
-          - req_99bd26d935838897a0f3e188c1a097ec
+          - req_581c8c76eac198eaa38b0194dd1be5c0
       status:
         code: 200
         message: OK
@@ -4518,80 +4543,84 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from wellawatteUnknownyearaperspectiveon pages 33-35: Geemi P. Wellawatte, Heta
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from wellawatteUnknownyearaperspectiveon pages 25-28: Geemi P. Wellawatte, Heta
         A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
         of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n----\\n\\n13,\\n\\n
-        \    1\u201320.\\n\\n\\n(78) Mastropietro, A.; Pasculli, G.; Feldmann, C.; Rodr\xB4\u0131guez-P\xB4erez,
-        R.; Bajorath, J. Edge-\\n\\n    SHAPer: Bond-Centric Shapley Value-Based Explanation
-        Method for Graph Neural\\n\\n     Networks. iScience 2022, 25, 105043.\\n\\n\\n(79)
-        White, A. D. Deep learning for molecules and materials. Living Journal of Computa-\\n\\n
-        \     tional Molecular Science 2022, 3.\\n\\n(80) \u02D8Strumbelj, E.; Kononenko,
-        I. Explaining prediction models and individual predictions\\n\\n     with feature
-        contributions. Knowledge and Information Systems 2014, 41, 647\u2013665.\\n\\n\\n(81)
-        Erhan, D.; Bengio, Y.; Courville, A.; Vincent, P. Visualizing Higher-Layer Features
-        of\\n\\n     a Deep Network. Technical Report, Univerist\xB4e de Montr\xB4eal
-        2009,\\n\\n\\n(82) Weber, J. K.; Morrone, J. A.; Bagchi, S.; Pabon, J. D.; gu
-        Kang, S.; Zhang, L.;\\n\\n     Cornell, W. D. Simplified, interpretable graph
-        convolutional neural networks for small\\n\\n     molecule activity prediction.
-        Journal of Computer-Aided Molecular Design 2022, 36,\\n\\n     391\u2013404.\\n\\n\\n(83)
-        Riniker, S.; Landrum, G. A. Similarity maps - A visualization strategy for molecular\\n\\n
-        \    fingerprints and machine-learning methods. Journal of Cheminformatics 2013,
-        5, 1\u20137.\\n\\n\\n(84) Humer, C.; Heberle, H.; Montanari, F.; Wolf, T.; Huber,
-        F.; Henderson, R.; Hein-\\n\\n      rich, J.; Streit, M. ChemInformatics Model
-        Explorer (CIME): exploratory analysis of\\n\\n     chemical model explanations.
-        Journal of Cheminformatics 2022, 14, 1\u201314.\\n\\n\\n(85) McGrath, T.; Kapishnikov,
-        A.; Toma\u02C7sev, N.; Pearce, A.; Wattenberg, M.; Hass-\\n\\n      abis, D.;
-        Kim, B.; Paquet, U.; Kramnik, V. Acquisition of chess knowledge in Al-\\n\\n
-        \    phaZero. Proceedings of the National Academy of Sciences 2022, 119, e2206625119.\\n\\n\\n\\n\\n
-        \                                     33(86) Bajusz, D.; R\xB4acz, A.; H\xB4eberger,
-        K. Why is Tanimoto index an appropriate choice for\\n\\n     fingerprint-based
-        similarity calculations? Journal of Cheminformatics 2015, 7, 1\u201313.\\n\\n\\n(87)
-        Huang, Q.; Yamada, M.; Tian, Y.; Singh, D.; Yin, D.; Chang, Y. GraphLIME:\\n\\n
-        \    Local Interpretable Model Explanations for Graph Neural Networks. CoRR
-        2020,\\n\\n     abs/2001.06216.\\n\\n\\n(88) Sokol, K.; Flach, P. A. LIMEtree:
-        Interactively Customisable Explanations Based on\\n\\n     Local Surrogate Multi-output
-        Regression Trees. CoRR 2020, abs/2005.01427.\\n\\n\\n(89) Whitmore, L. S.; George,
-        A.; Hudson, C. M. Mapping chemical performance on molec-\\n\\n     ular structures
-        using locally interpretable explanations. 2016; https://arxiv.org/\\n\\n    abs/1611.07443.\\n\\n\\n(90)
-        Mehdi, S.; Tiwary, P. Thermodynamics of Interpretation. 2022,\\n\\n\\n(91) H\xA8ofler,
-        M. Causal inference based on counterfactuals. BMC Medical Research Method-\\n\\n
-        \    ology 2005, 5, 1\u201312.\\n\\n\\n(92) Woodward, J.; Hitchcock, C. Explanatory
-        Generalizations, Part I: A Counterfactual\\n\\n     Account. No\u02C6us 2003,
-        37, 1\u201324.\\n\\n\\n(93) Frisch, M. F. Theories, models, and explanation;
-        University of California, Berkeley,\\n\\n     1998.\\n\\n\\n(94) Reutlinger,
-        A. Is There A Monist Theory of Causal and Non-Causal Explanations?\\n\\n    The
-        Counterfactual Theory of Scientific Explanation. Philosophy of Science 2016,
-        83,\\n\\n     733\u2013745.\\n\\n\\n(95) Lewis, D. Causation. The journal of
-        philosophy 1974, 70, 556\u2013567.\\n\\n\\n(96) Tanimoto, T. T. Elementary mathematical
-        theory of classification and prediction.\\n\\n     Internal IBM Technical Report
-        1958,\\n\\n\\n                                      34 (97) Rogers, D.; Hahn,
-        M. Extended-Connectivity Fingerprints. Journal of Chemical In-\\n\\n      formation
-        and Modeling 2010, 50, 742\u2013754, PMID: 20426451.\\n\\n\\n (98) Mohapatra,
-        S.; An, J.; G\xB4omez-Bombarelli, R. Chemistry-informed macromolecule\\n\\n
-        \     graph representation for similarity computation, unsupervised and supervised
-        learn-\\n\\n       ing. Machine Learning: Science and Technology 2022, 3, 015028.\\n\\n\\n
-        (99) Doshi-Velez, F.; Kortz, M.; Budish, R.; Bavitz, C.; Gershman, S.; O\u2019Brien,
-        D.;\\n\\n       Scott, K.; Schieber, S.; Waldo, J.; Weinberger, D.; Weller,
-        A.; Wood, A. Account-\\n\\n       ability of AI Under the Law: The Role of Explanation.
-        SSRN Electronic Journal\\n\\n     2017,\\n\\n\\n(100) Wachter, S.; Mittelstadt,
-        B.; Russell, C. Counterfactual explanations without opening\\n\\n      the black
-        box: Automated decisions and the GDPR. Harv. JL & Tech. 2017, 31, 841.\\n\\n\\n(101)
-        Jim\xB4enez-Luna, J.; Grisoni, F.; Schneider, G. Drug discovery with explainable
-        artificial\\n\\n       intelligence. Nature Machine Intelligence 2020 2:10 2020,
-        2, 573\u2013584.\\n\\n\\n(102) Fu, T.; Gao, W.; Xiao, C.; Yasonik, J.; Coley,
-        C. W.; Sun, J. Differentiable Scaffold-\\n\\n      ing Tree for Molecule Optimization.
-        International Conference on Learning Represen-\\n\\n       tations. 2022.\\n\\n\\n(103)
-        Shen, C.; Krenn, M.; Eppel, S.; Aspuru-Guzik, A. Deep molecular dreaming: inverse\\n\\n
-        \    machine learning for de-novo molecular design and interpretability with
-        surjective\\n\\n      representations. Machine Learning: Science and Technology
-        2021, 2, 03LT02.\\n\\n\\n(104) Lucic,  A.;   ter  Hoeve,  M.;   Tolomei,  G.;
-        \  Rijke,  M.;   Silvestri,  F.  CF-\\n\\n     GNNExplainer:  Counterfactual
-        Explanations for Graph Neural Networks. arXiv\\n\\n----\\n\\nQuestion: Are counterfactuals
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n------------\\n\\n2021,
+        25, 1315\u20131360.\\n\\n\\n (9) Wellawatte, G. P.; Seshadri, A.; White, A.
+        D. Model agnostic generation of counter-\\n\\n     factual explanations for
+        molecules. Chemical Science 2022, 13, 3697\u20133705.\\n\\n\\n(10) Gandhi, H.
+        A.; White, A. D. Explaining structure-activity relationships using locally\\n\\n
+        \     faithful surrogate models. chemrxiv 2022,\\n\\n\\n(11) Gormley, A. J.;
+        Webb, M. A. Machine learning in combinatorial polymer chemistry.\\n\\n     Nature
+        Reviews Materials 2021,\\n\\n\\n(12) Gomes, C. P.; Fink, D.; Dover, R. B. V.;
+        Gregoire, J. M. Computational sustainability\\n\\n     meets materials science.
+        Nature Reviews Materials 2021,\\n\\n\\n(13) On scientific understanding with
+        artificial intelligence. Nature Reviews Physics 2022\\n\\n     4:12 2022, 4,
+        761\u2013769.\\n\\n\\n(14) Arrieta, A. B.; D\xB4\u0131az-Rodr\xB4\u0131guez,
+        N.; Ser, J. D.; Bennetot, A.; Tabik, S.; Barbado, A.;\\n\\n     Garcia, S.;
+        Gil-Lopez, S.; Molina, D.; Benjamins, R.; Chatila, R.; Herrera, F. Explain-\\n\\n
+        \    able Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities
+        and Chal-\\n\\n     lenges toward Responsible AI. Information Fusion 2019, 58,
+        82\u2013115.\\n\\n\\n(15) Murdoch, W. J.; Singh, C.; Kumbier, K.; Abbasi-Asl,
+        R.; Yu, B. Interpretable machine\\n\\n     learning: definitions, methods, and
+        applications. ArXiv 2019, abs/1901.04592.\\n\\n\\n                                      25(16)
+        Boobier, S.; Osbourn, A.; Mitchell, J. B. Can human experts predict solubility
+        better\\n\\n     than computers? Journal of cheminformatics 2017, 9, 1\u201314.\\n\\n\\n(17)
+        Lee, J. D.; See, K. A. Trust in automation: Designing for appropriate reliance.
+        Human\\n\\n     Factors 2004, 46, 50\u201380.\\n\\n\\n(18) Bolukbasi, T.; Chang,
+        K.-W.; Zou, J. Y.; Saligrama, V.; Kalai, A. T. Man is to com-\\n\\n     puter
+        programmer as woman is to homemaker? debiasing word embeddings. Advances\\n\\n
+        \    in neural information processing systems 2016, 29.\\n\\n\\n(19) Buolamwini,
+        J.; Gebru, T. Gender Shades:  Intersectional Accuracy Disparities in\\n\\n    Commercial
+        Gender Classification. Proceedings of the 1st Conference on Fairness,\\n\\n
+        \    Accountability and Transparency. 2018; pp 77\u201391.\\n\\n\\n(20) Lapuschkin,
+        S.; W\xA8aldchen, S.; Binder, A.; Montavon, G.; Samek, W.; M\xA8uller, K.-R.\\n\\n
+        \   Unmasking Clever Hans predictors and assessing what machines really learn.
+        Nature\\n\\n     communications 2019, 10, 1\u20138.\\n\\n\\n(21) DeGrave, A.
+        J.; Janizek, J. D.; Lee, S.-I. AI for radiographic COVID-19 detection\\n\\n
+        \     selects shortcuts over signal. Nature Machine Intelligence 2021, 3, 610\u2013619.\\n\\n\\n(22)
+        Goodman, B.; Flaxman, S. European Union regulations on algorithmic decision-\\n\\n
+        \   making and a \u201Cright to explanation\u201D. AI Magazine 2017, 38, 50\u201357.\\n\\n\\n(23)
+        ACT, A. I. European Commission. On Artificial Intelligence: A European Approach\\n\\n
+        \    to Excellence and Trust. 2021, COM/2021/206.\\n\\n\\n(24) Blueprint for
+        an AI Bill of Rights, The White House. 2022; https://www.whitehouse.\\n\\n    gov/ostp/ai-bill-of-rights/.\\n\\n\\n(25)
+        Miller, T. Explanation in artificial intelligence: Insights from the social
+        sciences. Ar-\\n\\n       tificial intelligence 2019, 267, 1\u201338.\\n\\n\\n\\n
+        \                                     26(26) Murdoch, W. J.; Singh, C.; Kumbier,
+        K.; Abbasi-Asl, R.; Yu, B. Definitions, meth-\\n\\n     ods, and applications
+        in interpretable machine learning. Proceedings of the National\\n\\n    Academy
+        of Sciences of the United States of America 2019, 116, 22071\u201322080.\\n\\n\\n(27)
+        Gunning, D.; Aha, D. DARPA\u2019s Explainable Artificial Intelligence (XAI)
+        Program.\\n\\n    AI Magazine 2019, 40, 44\u201358.\\n\\n\\n(28) Biran, O.;
+        Cotton, C. Explanation and justification in machine learning: A survey.\\n\\n
+        \    IJCAI-17 workshop on explainable AI (XAI). 2017; pp 8\u201313.\\n\\n\\n(29)
+        Palacio, S.; Lucieri, A.; Munir, M.; Ahmed, S.; Hees, J.; Dengel, A. Xai handbook:\\n\\n
+        \   Towards a unified framework for explainable ai. Proceedings of the IEEE/CVF
+        Inter-\\n\\n     national Conference on Computer Vision. 2021; pp 3766\u20133775.\\n\\n\\n(30)
+        Kuhn, D. R.; Kacker, R. N.; Lei, Y.; Simos, D. E. Combinatorial Methods for
+        Ex-\\n\\n     plainable AI. 2020 IEEE International Conference on Software Testing,
+        Verification\\n\\n    and Validation Workshops (ICSTW) 2020, 167\u2013170.\\n\\n\\n(31)
+        Seshadri, A.; Gandhi, H. A.; Wellawatte, G. P.; White, A. D. Why does that molecule\\n\\n
+        \    smell? ChemRxiv 2022,\\n\\n\\n(32) Das, A.; Rad, P. Opportunities and challenges
+        in explainable artificial intelligence\\n\\n      (xai): A survey. arXiv preprint
+        arXiv:2006.11371 2020,\\n\\n\\n(33) Machlev, R.; Heistrene, L.; Perl, M.; Levy,
+        K. Y.; Belikov, J.; Mannor, S.; Levron, Y.\\n\\n     Explainable Artificial
+        Intelligence (XAI) techniques for energy and power systems:\\n\\n     Review,
+        challenges and opportunities. Energy and AI 2022, 9, 100169.\\n\\n\\n(34) Koh,
+        P. W.; Liang, P. Understanding black-box predictions via influence functions.\\n\\n
+        \    International Conference on Machine Learning. 2017; pp 1885\u20131894.\\n\\n\\n(35)
+        Ribeiro, M. T.; Singh, S.; Guestrin, C. \u201D Why should i trust you?\u201D
+        Explaining the\\n\\n     predictions of any classifier. Proceedings of the 22nd
+        ACM SIGKDD international\\n\\n\\n                                      27     conference
+        on knowledge discovery and data \\n\\n------------\\n\\nQuestion: Are counterfactuals
         actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
@@ -4601,13 +4630,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "6114"
+          - "6398"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4617,7 +4646,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4633,23 +4662,23 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA4xUTW8bNxC961cMeAhqQBIk2ZYc3RKkSHIIELQpUqAbSCNydndiLknwQ7Zi+L8X
-          5MrSOmmDXPawb+bxzZuPhxGAYCXWIGSLUXZOT14fPtT+KnV/3aXf44rfz25oT9dsW/z77VKMc4bd
-          fSUZn7Km0nZOU2Rrelh6wkiZdb66vlwtlteXLwvQWUU6pzUuTq7sZDFbXE3m88lidkxsLUsKYg3/
-          jAAAHso3SzSK7sUaZuOnPx2FgA2J9SkIQHir8x+BIXCIaKIYn0FpTSRTVG+326/Bmso8VAagEiF1
-          HfpDJdZQiU8tAd1L8i6Cp5o8GUkBuqQjO01wZ/1tAMVBphDYNCBtMpF8jTIm1ED3TqPBbEcYAxup
-          k8phsSX2kCUCGwiSyUSuWQ4TAI2CV+8BZSHFHWuOhyn86UjmWNT6MIZgk8+SNN8SfEbZRvJAEVBP
-          4bfFbL66eJL3M21ZRWwJii/3EWwNmKLtcutAkeRQorIiTQ1qqD121FdfXn775uMfYwipaSjEc4Up
-          FtGZ3nm751I8yvwm7krxgZs2Brjj2NqUTd4T6iMBcNZrUBef2TQhCyuDE6bwSikuRMWHXsvuAO+q
-          qhK21uRz/bPri6L6s7XqDr2CF/COo2yllbcFv7woRlhP3/lTPJGYAmpgc2x94XoyzvoDNGTIo+Zv
-          5x4rljiwwJOmPeZcNpCMIp+nsTfCKHA2DyLnIooxpoHk7OllT7pnbtmFaSXG/ZCeSDdBWk/9sL6s
-          RGUeK7Pdboez7qnOVGINJmk9ANAYG3v2vGVfjsjjaa+0bZy3u/BdqqjZcGg3njBYk3coROtEQR9H
-          AF/K/qZnKymct52Lm2hvqTw3X82OCyzOJ2MAXy6OaLQR9QC4OSHPKDeKIrIOgyMgJMqW1Dn3fDEw
-          KbYDYDQo/Ec9/8XdF8+m+RX6MyAluUhq4zwpls9rPod5yjf1/8JORhfBIpDfs6RNZPK5GYpqTLo/
-          dyIcQqRuU7NpyDvP/c2r3UYuV6vlTb28mYnR4+hfAAAA//8DADU17Yv8BQAA
+          H4sIAAAAAAAAAwAAAP//jFRNb9tGEL3rVwz2kgSQBElmrFi3IKfA6AfaAHFbBcRoOSSnXu6yO7Oy
+          BcP/vVjSlpTEDXrhYd+8xzez8/ZhAmC4MhswtkW1Xe9mH5bXP7d/Lq//ueHreFj/9iv9si+634tI
+          y8s/zDQzwu5vsvrMmtvQ9Y6Ugx9hGwmVsupy/bYoLlbrxdUAdKEil2lNr7MizFaLVTFbLmerxROx
+          DWxJzAb+mgAAPAzfbNFXdG82sJg+n3Qkgg2ZzbEIwMTg8olBERZFr2Z6Am3wSn5w/bD1AFsjqesw
+          HrZmA1vzqSWge0uxV4hUUyRvSaBLTrl3BHch3gpEcrk10AB03ztkjztHgFG5ZsvogL2Sc9xkOry+
+          ef/xDaCvwIbklWKNVhO6kewxz0ymwN66VLFvAEF6slkLOvIZhlDDq5/y4AAbH0TZQkOeIj6jP5CG
+          OkTogiObHMkr2B3gMzmHd6hKQAro5vB6tVit3szhww90MBIg3NIBNPRsgT3cvP84hVAreUgyjqSP
+          Yc8VAdrMGibDXrhpVfKvpQ13uck23IFt0TckmcW+Twp7jJwZAhY9oFOK0EeqeNCSOXxqWUBS05Co
+          gLao33Q+uqxIuPGjn91XVnYH8Kgp0hQwC9ABkLtc16TsOglFgeAHfxoAbcu0HxUjVRCS2tDR0EoX
+          Kq4PuZnjfWUXIcp8a6bjekVytEdvqRQbIo1rdnWE89BK7rAhyVCNTmjrH89XNlKdBHNifHLuDEDv
+          g45Xk8Py5Ql5PMbDhaaPYSffUE3NnqUtI6EEn6MgGnozoI8TgC9DDNNXyTJ9DF2vpYZbGn63XF9c
+          jYLmlPwz+KJ4QjUoujPg3fpi+oJkWZEiOznLsrFoW6pO3FPwMVUczoDJWePf+3lJe2yeffN/5E+A
+          tdQrVeVpKV8qi5Sfxv8qOw56MGyE4p4tlcoU82VUVGNy46tl5CBKXVmzbyj2kcenq+7L1fKSLt8W
+          tqjM5HHyLwAAAP//AwB0BIFAwwUAAA==
       headers:
         CF-RAY:
-          - 96665cf63f01f555-SJC
+          - 96a9b56f7fe6cf45-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4657,7 +4686,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:42 GMT
+          - Tue, 05 Aug 2025 22:25:11 GMT
         Server:
           - cloudflare
         Strict-Transport-Security:
@@ -4675,13 +4704,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "3114"
+          - "2023"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
         x-envoy-upstream-service-time:
-          - "3116"
+          - "2036"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4689,13 +4718,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998546"
+          - "29998475"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
-          - 2ms
+          - 3ms
         x-request-id:
-          - b852a502-afe3-4058-88ad-292c2aa4aa0b
+          - req_897d416094f04025bc8b9b1ff6b5a9f5
       status:
         code: 200
         message: OK
@@ -4703,82 +4732,83 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from wellawatteUnknownyearaperspectiveon pages 25-28: Geemi P. Wellawatte, Heta
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from wellawatteUnknownyearaperspectiveon pages 5-8: Geemi P. Wellawatte, Heta
         A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
         of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n----\\n\\n2021,
-        25, 1315\u20131360.\\n\\n\\n (9) Wellawatte, G. P.; Seshadri, A.; White, A.
-        D. Model agnostic generation of counter-\\n\\n     factual explanations for
-        molecules. Chemical Science 2022, 13, 3697\u20133705.\\n\\n\\n(10) Gandhi, H.
-        A.; White, A. D. Explaining structure-activity relationships using locally\\n\\n
-        \     faithful surrogate models. chemrxiv 2022,\\n\\n\\n(11) Gormley, A. J.;
-        Webb, M. A. Machine learning in combinatorial polymer chemistry.\\n\\n     Nature
-        Reviews Materials 2021,\\n\\n\\n(12) Gomes, C. P.; Fink, D.; Dover, R. B. V.;
-        Gregoire, J. M. Computational sustainability\\n\\n     meets materials science.
-        Nature Reviews Materials 2021,\\n\\n\\n(13) On scientific understanding with
-        artificial intelligence. Nature Reviews Physics 2022\\n\\n     4:12 2022, 4,
-        761\u2013769.\\n\\n\\n(14) Arrieta, A. B.; D\xB4\u0131az-Rodr\xB4\u0131guez,
-        N.; Ser, J. D.; Bennetot, A.; Tabik, S.; Barbado, A.;\\n\\n     Garcia, S.;
-        Gil-Lopez, S.; Molina, D.; Benjamins, R.; Chatila, R.; Herrera, F. Explain-\\n\\n
-        \    able Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities
-        and Chal-\\n\\n     lenges toward Responsible AI. Information Fusion 2019, 58,
-        82\u2013115.\\n\\n\\n(15) Murdoch, W. J.; Singh, C.; Kumbier, K.; Abbasi-Asl,
-        R.; Yu, B. Interpretable machine\\n\\n     learning: definitions, methods, and
-        applications. ArXiv 2019, abs/1901.04592.\\n\\n\\n                                      25(16)
-        Boobier, S.; Osbourn, A.; Mitchell, J. B. Can human experts predict solubility
-        better\\n\\n     than computers? Journal of cheminformatics 2017, 9, 1\u201314.\\n\\n\\n(17)
-        Lee, J. D.; See, K. A. Trust in automation: Designing for appropriate reliance.
-        Human\\n\\n     Factors 2004, 46, 50\u201380.\\n\\n\\n(18) Bolukbasi, T.; Chang,
-        K.-W.; Zou, J. Y.; Saligrama, V.; Kalai, A. T. Man is to com-\\n\\n     puter
-        programmer as woman is to homemaker? debiasing word embeddings. Advances\\n\\n
-        \    in neural information processing systems 2016, 29.\\n\\n\\n(19) Buolamwini,
-        J.; Gebru, T. Gender Shades:  Intersectional Accuracy Disparities in\\n\\n    Commercial
-        Gender Classification. Proceedings of the 1st Conference on Fairness,\\n\\n
-        \    Accountability and Transparency. 2018; pp 77\u201391.\\n\\n\\n(20) Lapuschkin,
-        S.; W\xA8aldchen, S.; Binder, A.; Montavon, G.; Samek, W.; M\xA8uller, K.-R.\\n\\n
-        \   Unmasking Clever Hans predictors and assessing what machines really learn.
-        Nature\\n\\n     communications 2019, 10, 1\u20138.\\n\\n\\n(21) DeGrave, A.
-        J.; Janizek, J. D.; Lee, S.-I. AI for radiographic COVID-19 detection\\n\\n
-        \     selects shortcuts over signal. Nature Machine Intelligence 2021, 3, 610\u2013619.\\n\\n\\n(22)
-        Goodman, B.; Flaxman, S. European Union regulations on algorithmic decision-\\n\\n
-        \   making and a \u201Cright to explanation\u201D. AI Magazine 2017, 38, 50\u201357.\\n\\n\\n(23)
-        ACT, A. I. European Commission. On Artificial Intelligence: A European Approach\\n\\n
-        \    to Excellence and Trust. 2021, COM/2021/206.\\n\\n\\n(24) Blueprint for
-        an AI Bill of Rights, The White House. 2022; https://www.whitehouse.\\n\\n    gov/ostp/ai-bill-of-rights/.\\n\\n\\n(25)
-        Miller, T. Explanation in artificial intelligence: Insights from the social
-        sciences. Ar-\\n\\n       tificial intelligence 2019, 267, 1\u201338.\\n\\n\\n\\n
-        \                                     26(26) Murdoch, W. J.; Singh, C.; Kumbier,
-        K.; Abbasi-Asl, R.; Yu, B. Definitions, meth-\\n\\n     ods, and applications
-        in interpretable machine learning. Proceedings of the National\\n\\n    Academy
-        of Sciences of the United States of America 2019, 116, 22071\u201322080.\\n\\n\\n(27)
-        Gunning, D.; Aha, D. DARPA\u2019s Explainable Artificial Intelligence (XAI)
-        Program.\\n\\n    AI Magazine 2019, 40, 44\u201358.\\n\\n\\n(28) Biran, O.;
-        Cotton, C. Explanation and justification in machine learning: A survey.\\n\\n
-        \    IJCAI-17 workshop on explainable AI (XAI). 2017; pp 8\u201313.\\n\\n\\n(29)
-        Palacio, S.; Lucieri, A.; Munir, M.; Ahmed, S.; Hees, J.; Dengel, A. Xai handbook:\\n\\n
-        \   Towards a unified framework for explainable ai. Proceedings of the IEEE/CVF
-        Inter-\\n\\n     national Conference on Computer Vision. 2021; pp 3766\u20133775.\\n\\n\\n(30)
-        Kuhn, D. R.; Kacker, R. N.; Lei, Y.; Simos, D. E. Combinatorial Methods for
-        Ex-\\n\\n     plainable AI. 2020 IEEE International Conference on Software Testing,
-        Verification\\n\\n    and Validation Workshops (ICSTW) 2020, 167\u2013170.\\n\\n\\n(31)
-        Seshadri, A.; Gandhi, H. A.; Wellawatte, G. P.; White, A. D. Why does that molecule\\n\\n
-        \    smell? ChemRxiv 2022,\\n\\n\\n(32) Das, A.; Rad, P. Opportunities and challenges
-        in explainable artificial intelligence\\n\\n      (xai): A survey. arXiv preprint
-        arXiv:2006.11371 2020,\\n\\n\\n(33) Machlev, R.; Heistrene, L.; Perl, M.; Levy,
-        K. Y.; Belikov, J.; Mannor, S.; Levron, Y.\\n\\n     Explainable Artificial
-        Intelligence (XAI) techniques for energy and power systems:\\n\\n     Review,
-        challenges and opportunities. Energy and AI 2022, 9, 100169.\\n\\n\\n(34) Koh,
-        P. W.; Liang, P. Understanding black-box predictions via influence functions.\\n\\n
-        \    International Conference on Machine Learning. 2017; pp 1885\u20131894.\\n\\n\\n(35)
-        Ribeiro, M. T.; Singh, S.; Guestrin, C. \u201D Why should i trust you?\u201D
-        Explaining the\\n\\n     predictions of any classifier. Proceedings of the 22nd
-        ACM SIGKDD international\\n\\n\\n                                      27     conference
-        on knowledge discovery and data \\n\\n----\\n\\nQuestion: Are counterfactuals
-        actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n------------\\n\\nnct?\\n\\n\\n
+        \ We present an example evaluation of the SHAP explanation method based on the
+        above\\n\\nattributes.44 Shapley values were proposed as a local explanation
+        method based on feature\\n\\nattribution, as they offer a complete explanation
+        - each feature is assigned a fraction of\\n\\nthe prediction value.44,45 Completeness
+        is a clearly measurable and well-defined metric, but\\n\\nyields explanations
+        with many components. Yet Shapley values are not actionable nor sparse.\\n\\nThey
+        are non-sparse as every feature has a non-zero attribution and not-actionable
+        because\\n\\nthey do not provide a set of features which changes the outcome.46
+        Ribeiro et al. 35 proposed\\n\\na surrogate model method that aims to provide
+        sparse/succinct explanations that have high\\n\\n\\n                                       5fidelity
+        to the original model. In Wellawatte et al. 9 we argue that counterfactuals
+        are \u201Cbet-\\n\\nter\u201D explanations because they are actionable and sparse.
+        We highlight that, evaluation of\\n\\nexplanations is a difficult task because
+        explanations are fundamentally for and by humans.\\n\\nTherefore, these evaluations
+        are subjective, as they depend on \u201Ccomplex human factors and\\n\\napplication
+        scenarios.\u201D37\\n\\n\\nSelf-explaining models\\n\\nA self-explanatory model
+        is one that is intrinsically interpretable to an expert.47 Two com-\\n\\nmon
+        examples found in the literature are linear regression models and decision trees
+        (DT).\\n\\nIntrinsic models can be found in other XAI applications acting as
+        surrogate models (proxy\\n\\nmodels) due to their transparent nature.48,49 A
+        linear model is described by the equation\\n\\n1 where, W\u2019s are the weight
+        parameters and x\u2019s are the input features associated with the\\n\\nprediction
+        \u02C6y. Therefore, we observe that the weights can be used to derive a complete
+        expla-\\n\\nnation of the model - trained weights quantify the importance of
+        each feature.47 DT models\\n\\nare another type of self-explaining models which
+        have been used in classification and high-\\n\\nthroughput screening tasks.
+        \ Gajewicz et al. 50 used DT models to classify nanomaterials\\n\\nthat identify
+        structural features responsible for surface activity. In another study by Han\\n\\net
+        al. 51, a DT model was developed to filter compounds by their bioactivity based
+        on the\\n\\nchemical fingerprints.\\n\\n\\n\\n                                                              \u02C6y
+        = \u03A3iWixi                                      (1)\\n\\n\\n   Regularization
+        techniques such as EXPO52 and RRR53 are designed to enhance the black-\\n\\nbox
+        model interpretability.54 Although one can argue that \u201Csimplicity\u201D
+        of models are posi-\\n\\ntively correlated with interpretability, this is based
+        on how the interpretability is evaluated.\\n\\nFor example, Lipton 55 argue
+        that, from the notion of \u201Csimulatability\u201D (the degree to which a\\n\\nhuman
+        can predict the outcome based on inputs), self-explanatory linear models, rule-based\\n\\n\\n\\n
+        \                                      6systems, and DT\u2019s can be claimed
+        uninterpretable. A human can predict the outcome given\\n\\nthe inputs only
+        if the input features are interpretable. Therefore, a linear model which takes\\n\\nin
+        non-descriptive inputs may not be as transparent. On the other hand, a linear
+        model\\n\\nis not innately accurate as they fail to capture non-linear relationships
+        in data, limiting is\\n\\napplicability. Similarly, a DT is a rule-based model
+        and lacks physics informed knowledge.\\n\\nTherefore, an existing drawback is
+        the trade-offbetween the degree of understandability and\\n\\nthe accuracy of
+        a model. For example, an intrinsic model (linear regression or decision trees)\\n\\ncan
+        be described through the trainable parameters, but it may fail to \u201Ccorrectly\u201D
+        capture\\n\\nnon-linear relations in the data.\\n\\n\\nAttribution methods\\n\\n\\nFeature
+        attribution methods explain black box predictions by assigning each input feature\\n\\na
+        numerical value, which indicates its importance or contribution to the prediction.
+        Feature\\n\\nattributions provide local explanations, but can be averaged or
+        combined to explain multi-\\n\\nple instances. Atom-based numerical assignments
+        are commonly referred to as heatmaps.56\\n\\nSheridan 57 describes an atom-wise
+        attribution method for interpreting QSAR models. Re-\\n\\ncently, Rasmussen
+        et al. 58 showed that Crippen logP models serve as a benchmark for\\n\\nheatmap
+        approaches. Other most widely used feature attribution approaches in the litera-\\n\\nture
+        are gradient based methods,59,60 Shapley Additive exPlanations (SHAP),44 and
+        layer-\\n\\nwise relevance prorogation.61\\n\\n   Gradient based approaches
+        are based on the hypothesis that gradients for neural net-\\n\\nworks are analogous
+        to coefficients for regression models.62 Class activation maps (CAM),63\\n\\ngradCAM,64
+        smoothGrad,,65 and integrated gradients62 are examples of this method. The\\n\\nmain
+        idea behind feature attributions with gradients can be represented with equation
+        \ 2.\\n\\n                                          \u2206\u02C6f(\u20D7x) \u2248\u2202\u02C6f(\u20D7x)
+        \                                   (2)\\n                               \u2206xi
+        \     \u2202xi\\n\\n\\n\\n                                       7                                  \\n\\n------------\\n\\nQuestion:
+        Are counterfactuals actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4787,13 +4817,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "6122"
+          - "6361"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4803,7 +4833,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -4819,23 +4849,21 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFRNbxtHDL3rVxBzaQtIhqU48cfNbYPCh55qtGmrQKBnubuMZ4eDIVex
-          Yvi/FzMrW0qTBr3sYR8f+Ybk4+MMwHHjrsD5Hs0PKSx+3P3a5l/wfMd//Sm3v7e/vf05XZxd3Ke3
-          zacPbl4YcveBvD2zTrwMKZCxxAn2mdCoZF2ev351vnrz+tVlBQZpKBRal2xxJovV6epssVwuVqd7
-          Yi/sSd0V/D0DAHis3yIxNvTgruB0/vxnIFXsyF29BAG4LKH8cajKahjNzQ+gl2gUq+rHdQRYOx2H
-          AfNu7a5g7W57AnrwlJNBppYyRU8KwxiMUyD4KPleQSLQQwrIEe8CAWbjlj1jAI5GIXBXaPD9u+ub
-          HwBjA17GaJRb9DZimMgRS690Dhx9GBuOHfxBIeBHNCMgAwwn3ymsTlerWrZUra1bYBdFjT10FCnX
-          NCDtt4pAKxkGCeTHQHoCP30jFDMBwj3twEv0lAw4wrvrmzmkLFuuQjEY5ULYEqiniJlFweS5LZNQ
-          SJka9jXtCdz2pPRlKWmNYiml3FCmBrASpsYqWE870LHrSA18j7GjWgh9z7QtShtSLjwZzctAcxjw
-          vmi0ngYYldqxjAUa8qwscbGHUxZPqjQpA89GzX68NKQelT9RyQE8JMmGZaDS1gHnlMnwjgPbro73
-          SDJH5a43LSWvb6Y+6BxS2RE/BsxhVyD1TLFuTU2QqRsDmuTadKMH05O1m08LminQttTfqJdM06Je
-          rt06Ph1vdqZ2VCzGimMIRwDGKDZ1vHjq/R55enFRkC5ludN/UV3LkbXfZEKVWByjJslV9GkG8L66
-          dfzMgC5lGZJtTO6pllu+uVhOCd3hQBzBq8s9amIYjoCL5d7mn6fcNGTIQY8s7zz6npoD93AfcGxY
-          joDZ0cO/1PO13NPjOXb/J/0B8MU41GwODvhaWKZyQf8r7KXRVbBTylv2tDGmXIbRUItjmI6b050a
-          DZuWY1fWk+uFK8OcPc3+AQAA//8DACeg17/fBQAA
+          H4sIAAAAAAAAAwAAAP//jFRNTxsxEL3nV4x84ZJEJCQgcuvHoRVS1aoIoTZoNbFnd128tuUZB1KU
+          /17tbkhCoVIvPszze55nv/HTAEBZoxagdI2im+hGHyZXX+zt5src/qjml+/nH2/0Q/79rbr99PWG
+          1bBlhNUv0vLMGuvQREdig+9hnQiFWtXJxXw2O5teTE47oAmGXEurooxmYTQ9nc5Gk8loeroj1sFq
+          YrWAnwMAgKdubVv0hh7VAjqZrtIQM1akFvtNACoF11YUMlsW9KKGB1AHL+S7rp+WHmCpODcNps1S
+          LWCprmsCetSUooCxrDMzMeiQvVAqUUtGx4AMCLKJBKEEeowOPbbGwXpogiOdHSaIiYzVXb2zzGP4
+          LMCCQgxSo7zWTQQ6eLaGEhk4WZEIpZPjIxhWpDEzgdS06RirIDVgdxCuHAF6AxwxMY3h3aH8QiSm
+          sLaGAIFJWhcloeS0bww96Bp91R0DIYsODQ2hwXvrq7bWQGYqs4MyJDCkLdvgRz0+huvatrfmJSEL
+          w4OVGr7XGB1tYI0uEw/hoba63lnukkOwygI+yLGZkJ69LNWwf7BEjtboNRWsQ6L+4SanezwzmcI2
+          WBG3WImOaem3xylIVGbGNoQ+O3cEoPdB+jtq83e3Q7b7xLlQxRRW/BdVldZbrotEyMG36WIJUXXo
+          dgBw1yU7vwiriik0UQoJ99QdN5lezHtBdRimA3x5uQMlCLoj2tnFbPiGYmFI0Do+mg6lUddkDtzD
+          KGE2NhwBgyPfr9t5S7v3bn31P/IHQGuKQqY4jMtb2xK1n82/tu3vuWtYMaW11VSIpdS+haESs+v/
+          AcUbFmqK0vqKUky2/wzKWEwn53Q+n+mZUYPt4A8AAAD//wMAOWJx7RUFAAA=
       headers:
         CF-RAY:
-          - 96665cf81b01cefd-SJC
+          - 96a9b578cd58fab6-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -4843,7 +4871,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:42 GMT
+          - Tue, 05 Aug 2025 22:25:11 GMT
         Server:
           - cloudflare
         Strict-Transport-Security:
@@ -4861,13 +4889,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "3345"
+          - "1186"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
         x-envoy-upstream-service-time:
-          - "3359"
+          - "1194"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -4875,13 +4903,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998542"
+          - "29998477"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
-          - 2ms
+          - 3ms
         x-request-id:
-          - de9b0e57-8a4d-455b-be71-7af502c53792
+          - req_d274f9f4676b442a832a9ca4b3bf070f
       status:
         code: 200
         message: OK
@@ -4889,14 +4917,17 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
         from wellawatteUnknownyearaperspectiveon pages 1-3: Geemi P. Wellawatte, Heta
         A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
         of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n----\\n\\n
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n------------\\n\\n
         A Perspective on Explanations of Molecular\\n\\n              Prediction Models\\n\\n\\nGeemi
         P. Wellawatte,\u2020   Heta A. Gandhi,\u2021   Aditi Seshadri,\u2021 and  Andrew\\n\\n
         \                          D. White\u2217,\u2021\\n\\n\\n     \u2020Department
@@ -4963,7 +4994,8 @@ interactions:
         a passive characteristic of a model, whereas explainability\\n\\nis an active
         characteristic which is used to clarify the internal decision-making process.\\n\\nNamely,
         an explanation is extra information that gives the context and a cause for one
-        or\\n\\nmore \\n\\n----\\n\\nQuestion: Are counterfactuals actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
+        or\\n\\nmore \\n\\n------------\\n\\nQuestion: Are counterfactuals actionable?
+        [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
           - application/json
@@ -4972,13 +5004,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "6106"
+          - "6382"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -4988,7 +5020,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -5004,23 +5036,24 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFRNbxNJEL37V5T6Akh2FJsAi28EFikIbkiAdpDd7q6ZKdzT3aqqcWxF
-          +e+oZ0LsQJD2Mod59d6rj666mQAY8mYJxrVWXZfD7PLwqe6/fPjq9PJbPd99O3zBy2779v2ir/sr
-          My2MtPmBTn+xzlzqckClFEfYMVrFojp/9eL5q8XLFxeLAeiSx1BoTdbZRZotzhcXs/l8tji/I7aJ
-          HIpZwn8TAICb4VtSjB73Zgnn019/OhSxDZrlfRCA4RTKH2NFSNRGNdMj6FJUjEPW6/X6h6RYxZsq
-          AlRG+q6zfKjMEirzuUXAvUPOCp7E9SIo8O8+B0vRbgLCG1aqyZENcBUVQ6AGo0N4+vXN1TOgCNoi
-          DHZ7hVSDR8wQ0HKk2MDTdx+fwdAJgToxuBY7EuXDGVwptNS0gZpWZVChLidWW9RTDRQVOTOq3VAg
-          PYCNHrBkFm1pvxTzPnrkUr0vbk82wbrtbJP2T+5Mz6BUaHttEwt0GAsTOtQ2eYFAWxxTcjaAS32x
-          rK3T3gYZ/DyKY8qa+IH1FK5bci1kTjvyCBRlLIOiJhDl3mnPOMucMrIegDGMzJayHCuhCO8+Qmb0
-          5Ab4DN7+ngVjCRCMih6sgIVrewBN9wondNgcQNp0XZphgyKXfHcI4jBapiQwzMDGBkvLrZaigy/z
-          8kXSU10jY1RIvbrUoUyhs9sipy12YAeb4V380fxSFGNNw9yH7j8orDLT8f0xBtyVIa/EJcbxHb6u
-          TBVvq7her0+fMWPdiy1bFPsQTgAbY9Kxo2WBvt8ht/crE1KTOW3kN6op+Um7YrSSYlkP0ZTNgN5O
-          AL4Pq9k/2DaTOXVZV5q2ONjNF4uXo6A5XoMH8B2qSW04AZ5f/DN9RHLlUS0FOdlv46xr0R+5x2Ng
-          e0/pBJicFP5nPo9pj8VTbP6P/BFwDrOiXx2n+lgYYzmXfwu7b/SQsBHkHTlcKSGXYXisbR/GS2bk
-          IIrdqqbYlENAwzkrw5zcTn4CAAD//wMAVWAS/MwFAAA=
+          H4sIAAAAAAAAAwAAAP//jFRNbxtHDL3rVxBzSgDJsGTFaXSzkxYxGrSXokhQBcJohrvLenZmQHJs
+          q4b/ezG7sqQkLtDLAjuPfCQfPx4nAIa8WYFxnVXX5zB7P//1N/oTry/85XX82Df317c34ePn69+/
+          /HL1xUyrR9r+jU6fvc5c6nNApRRH2DFaxco6f/tmubxYvJ2fD0CfPIbq1madLdNscb5Yzubz2eJ8
+          79glcihmBX9NAAAeh29NMXp8MCsYaIaXHkVsi2Z1MAIwnEJ9MVaERG1UMz2CLkXFOGT9uI4AayOl
+          7y3v1mYFa/NHh4APDjkreBJXRFBAO4QiCKmBnx9ysBTtNiBcsVJDjmyAm6gYArUYHcKrz1c3r4Ei
+          uA57EuXdFLJlJVeC5bCrEEVFzoxKsQWPmCGg5Vj/Xn349BoGkeQMbhQ6artAbacCPWqXvECgWxzZ
+          nQ3gUqlsjXVabBCw0YNHcUxZEwPWlKOtnZEp3HfkOsic7sgjUJSRmaImEOXitDDOMqeMrDtgDKNn
+          R3lkxlEB+PAJMqMnN8Bn8P77LBirgWBU9GAFLNzbHWg6MJy4w3YH0qX7Wr8NilzzvUMQh9EyJYHE
+          4Dob26EdVmvRwVfRfKX01DTIGBVSUZd6lCn09rbSaYc92CHM0DWKUKJHrqPhh3jRA2NDg/iD7t8W
+          VmdC8UEB+9xZoX/2E0F9Tqy2djw1x4baLQXS3alWz08UYVsoDFGVi+hoFKXwPtOqNz2bp6aKvJ+E
+          tZmO88oY8K5G3YhLjOPcvjvARdBvqLctSoUaGwTX8el0BxibIrauYCwhnAA2xqRju+v2fd0jT4d9
+          C6nNnLbynaup4km3YbSSYt0t0ZTNgD5NAL4Oe12+WVWTOfVZN5pucQg3X/y0HAnN8ZScwBeLPapJ
+          bTgBlvPL6QuUG49qKcjJcTDOug790fd4SWzxlE6AyUnhP+bzEvdYPMX2/9AfAecwK/rNceReMmOs
+          t/a/zA5CDwkbQb4jhxsl5NoMj40tYTyDRnai2G8aim2dVxpvYZM3i/klXr5ZuqU3k6fJvwAAAP//
+          AwA0kytcFAYAAA==
       headers:
         CF-RAY:
-          - 96665d070f3eeb35-SJC
+          - 96a9b5795e8e15d4-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -5028,9 +5061,11 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:44 GMT
+          - Tue, 05 Aug 2025 22:25:12 GMT
         Server:
           - cloudflare
+        Strict-Transport-Security:
+          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -5044,15 +5079,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "2141"
+          - "1499"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
-        strict-transport-security:
-          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "2145"
+          - "1502"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -5060,13 +5093,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998540"
+          - "29998471"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
-          - 2ms
+          - 3ms
         x-request-id:
-          - req_3416fe1b58726521ab94a5b97a03cb7b
+          - req_3b5bcd2fbe8c4e58bd95883f7d4243f7
       status:
         code: 200
         message: OK
@@ -5074,79 +5107,83 @@ interactions:
       body:
         "{\"messages\":[{\"role\":\"system\",\"content\":\"Provide a summary of
         the relevant information that could help answer the question based on the excerpt.
-        Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\": \\\"...\\\",\\n
-        \ \\\"relevance_score\\\": \\\"...\\\"\\n}\\n\\nwhere `summary` is relevant
-        information from the text - about 100 words words. `relevance_score` is an integer
-        1-10 for the relevance of `summary` to the question.\"},{\"role\":\"user\",\"content\":\"Excerpt
-        from wellawatteUnknownyearaperspectiveon pages 5-8: Geemi P. Wellawatte, Heta
+        Your summary, combined with many others, will be given to the model to generate
+        an answer. Respond with the following JSON format:\\n\\n{\\n  \\\"summary\\\":
+        \\\"...\\\",\\n  \\\"relevance_score\\\": \\\"...\\\"\\n  \\\"used_images\\\"\\n}\\n\\nwhere
+        `summary` is relevant information from the text - about 100 words words. `relevance_score`
+        is an integer 1-10 for the relevance of `summary` to the question. `used_images`
+        is a boolean flag indicating if any images present in a multimodal message were
+        used, and if no images were present it should be false.\"},{\"role\":\"user\",\"content\":\"Excerpt
+        from wellawatteUnknownyearaperspectiveon pages 33-35: Geemi P. Wellawatte, Heta
         A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
         of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n----\\n\\nnct?\\n\\n\\n
-        \ We present an example evaluation of the SHAP explanation method based on the
-        above\\n\\nattributes.44 Shapley values were proposed as a local explanation
-        method based on feature\\n\\nattribution, as they offer a complete explanation
-        - each feature is assigned a fraction of\\n\\nthe prediction value.44,45 Completeness
-        is a clearly measurable and well-defined metric, but\\n\\nyields explanations
-        with many components. Yet Shapley values are not actionable nor sparse.\\n\\nThey
-        are non-sparse as every feature has a non-zero attribution and not-actionable
-        because\\n\\nthey do not provide a set of features which changes the outcome.46
-        Ribeiro et al. 35 proposed\\n\\na surrogate model method that aims to provide
-        sparse/succinct explanations that have high\\n\\n\\n                                       5fidelity
-        to the original model. In Wellawatte et al. 9 we argue that counterfactuals
-        are \u201Cbet-\\n\\nter\u201D explanations because they are actionable and sparse.
-        We highlight that, evaluation of\\n\\nexplanations is a difficult task because
-        explanations are fundamentally for and by humans.\\n\\nTherefore, these evaluations
-        are subjective, as they depend on \u201Ccomplex human factors and\\n\\napplication
-        scenarios.\u201D37\\n\\n\\nSelf-explaining models\\n\\nA self-explanatory model
-        is one that is intrinsically interpretable to an expert.47 Two com-\\n\\nmon
-        examples found in the literature are linear regression models and decision trees
-        (DT).\\n\\nIntrinsic models can be found in other XAI applications acting as
-        surrogate models (proxy\\n\\nmodels) due to their transparent nature.48,49 A
-        linear model is described by the equation\\n\\n1 where, W\u2019s are the weight
-        parameters and x\u2019s are the input features associated with the\\n\\nprediction
-        \u02C6y. Therefore, we observe that the weights can be used to derive a complete
-        expla-\\n\\nnation of the model - trained weights quantify the importance of
-        each feature.47 DT models\\n\\nare another type of self-explaining models which
-        have been used in classification and high-\\n\\nthroughput screening tasks.
-        \ Gajewicz et al. 50 used DT models to classify nanomaterials\\n\\nthat identify
-        structural features responsible for surface activity. In another study by Han\\n\\net
-        al. 51, a DT model was developed to filter compounds by their bioactivity based
-        on the\\n\\nchemical fingerprints.\\n\\n\\n\\n                                                              \u02C6y
-        = \u03A3iWixi                                      (1)\\n\\n\\n   Regularization
-        techniques such as EXPO52 and RRR53 are designed to enhance the black-\\n\\nbox
-        model interpretability.54 Although one can argue that \u201Csimplicity\u201D
-        of models are posi-\\n\\ntively correlated with interpretability, this is based
-        on how the interpretability is evaluated.\\n\\nFor example, Lipton 55 argue
-        that, from the notion of \u201Csimulatability\u201D (the degree to which a\\n\\nhuman
-        can predict the outcome based on inputs), self-explanatory linear models, rule-based\\n\\n\\n\\n
-        \                                      6systems, and DT\u2019s can be claimed
-        uninterpretable. A human can predict the outcome given\\n\\nthe inputs only
-        if the input features are interpretable. Therefore, a linear model which takes\\n\\nin
-        non-descriptive inputs may not be as transparent. On the other hand, a linear
-        model\\n\\nis not innately accurate as they fail to capture non-linear relationships
-        in data, limiting is\\n\\napplicability. Similarly, a DT is a rule-based model
-        and lacks physics informed knowledge.\\n\\nTherefore, an existing drawback is
-        the trade-offbetween the degree of understandability and\\n\\nthe accuracy of
-        a model. For example, an intrinsic model (linear regression or decision trees)\\n\\ncan
-        be described through the trainable parameters, but it may fail to \u201Ccorrectly\u201D
-        capture\\n\\nnon-linear relations in the data.\\n\\n\\nAttribution methods\\n\\n\\nFeature
-        attribution methods explain black box predictions by assigning each input feature\\n\\na
-        numerical value, which indicates its importance or contribution to the prediction.
-        Feature\\n\\nattributions provide local explanations, but can be averaged or
-        combined to explain multi-\\n\\nple instances. Atom-based numerical assignments
-        are commonly referred to as heatmaps.56\\n\\nSheridan 57 describes an atom-wise
-        attribution method for interpreting QSAR models. Re-\\n\\ncently, Rasmussen
-        et al. 58 showed that Crippen logP models serve as a benchmark for\\n\\nheatmap
-        approaches. Other most widely used feature attribution approaches in the litera-\\n\\nture
-        are gradient based methods,59,60 Shapley Additive exPlanations (SHAP),44 and
-        layer-\\n\\nwise relevance prorogation.61\\n\\n   Gradient based approaches
-        are based on the hypothesis that gradients for neural net-\\n\\nworks are analogous
-        to coefficients for regression models.62 Class activation maps (CAM),63\\n\\ngradCAM,64
-        smoothGrad,,65 and integrated gradients62 are examples of this method. The\\n\\nmain
-        idea behind feature attributions with gradients can be represented with equation
-        \ 2.\\n\\n                                          \u2206\u02C6f(\u20D7x) \u2248\u2202\u02C6f(\u20D7x)
-        \                                   (2)\\n                               \u2206xi
-        \     \u2202xi\\n\\n\\n\\n                                       7                                  \\n\\n----\\n\\nQuestion:
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\\n\\n------------\\n\\n13,\\n\\n
+        \    1\u201320.\\n\\n\\n(78) Mastropietro, A.; Pasculli, G.; Feldmann, C.; Rodr\xB4\u0131guez-P\xB4erez,
+        R.; Bajorath, J. Edge-\\n\\n    SHAPer: Bond-Centric Shapley Value-Based Explanation
+        Method for Graph Neural\\n\\n     Networks. iScience 2022, 25, 105043.\\n\\n\\n(79)
+        White, A. D. Deep learning for molecules and materials. Living Journal of Computa-\\n\\n
+        \     tional Molecular Science 2022, 3.\\n\\n(80) \u02D8Strumbelj, E.; Kononenko,
+        I. Explaining prediction models and individual predictions\\n\\n     with feature
+        contributions. Knowledge and Information Systems 2014, 41, 647\u2013665.\\n\\n\\n(81)
+        Erhan, D.; Bengio, Y.; Courville, A.; Vincent, P. Visualizing Higher-Layer Features
+        of\\n\\n     a Deep Network. Technical Report, Univerist\xB4e de Montr\xB4eal
+        2009,\\n\\n\\n(82) Weber, J. K.; Morrone, J. A.; Bagchi, S.; Pabon, J. D.; gu
+        Kang, S.; Zhang, L.;\\n\\n     Cornell, W. D. Simplified, interpretable graph
+        convolutional neural networks for small\\n\\n     molecule activity prediction.
+        Journal of Computer-Aided Molecular Design 2022, 36,\\n\\n     391\u2013404.\\n\\n\\n(83)
+        Riniker, S.; Landrum, G. A. Similarity maps - A visualization strategy for molecular\\n\\n
+        \    fingerprints and machine-learning methods. Journal of Cheminformatics 2013,
+        5, 1\u20137.\\n\\n\\n(84) Humer, C.; Heberle, H.; Montanari, F.; Wolf, T.; Huber,
+        F.; Henderson, R.; Hein-\\n\\n      rich, J.; Streit, M. ChemInformatics Model
+        Explorer (CIME): exploratory analysis of\\n\\n     chemical model explanations.
+        Journal of Cheminformatics 2022, 14, 1\u201314.\\n\\n\\n(85) McGrath, T.; Kapishnikov,
+        A.; Toma\u02C7sev, N.; Pearce, A.; Wattenberg, M.; Hass-\\n\\n      abis, D.;
+        Kim, B.; Paquet, U.; Kramnik, V. Acquisition of chess knowledge in Al-\\n\\n
+        \    phaZero. Proceedings of the National Academy of Sciences 2022, 119, e2206625119.\\n\\n\\n\\n\\n
+        \                                     33(86) Bajusz, D.; R\xB4acz, A.; H\xB4eberger,
+        K. Why is Tanimoto index an appropriate choice for\\n\\n     fingerprint-based
+        similarity calculations? Journal of Cheminformatics 2015, 7, 1\u201313.\\n\\n\\n(87)
+        Huang, Q.; Yamada, M.; Tian, Y.; Singh, D.; Yin, D.; Chang, Y. GraphLIME:\\n\\n
+        \    Local Interpretable Model Explanations for Graph Neural Networks. CoRR
+        2020,\\n\\n     abs/2001.06216.\\n\\n\\n(88) Sokol, K.; Flach, P. A. LIMEtree:
+        Interactively Customisable Explanations Based on\\n\\n     Local Surrogate Multi-output
+        Regression Trees. CoRR 2020, abs/2005.01427.\\n\\n\\n(89) Whitmore, L. S.; George,
+        A.; Hudson, C. M. Mapping chemical performance on molec-\\n\\n     ular structures
+        using locally interpretable explanations. 2016; https://arxiv.org/\\n\\n    abs/1611.07443.\\n\\n\\n(90)
+        Mehdi, S.; Tiwary, P. Thermodynamics of Interpretation. 2022,\\n\\n\\n(91) H\xA8ofler,
+        M. Causal inference based on counterfactuals. BMC Medical Research Method-\\n\\n
+        \    ology 2005, 5, 1\u201312.\\n\\n\\n(92) Woodward, J.; Hitchcock, C. Explanatory
+        Generalizations, Part I: A Counterfactual\\n\\n     Account. No\u02C6us 2003,
+        37, 1\u201324.\\n\\n\\n(93) Frisch, M. F. Theories, models, and explanation;
+        University of California, Berkeley,\\n\\n     1998.\\n\\n\\n(94) Reutlinger,
+        A. Is There A Monist Theory of Causal and Non-Causal Explanations?\\n\\n    The
+        Counterfactual Theory of Scientific Explanation. Philosophy of Science 2016,
+        83,\\n\\n     733\u2013745.\\n\\n\\n(95) Lewis, D. Causation. The journal of
+        philosophy 1974, 70, 556\u2013567.\\n\\n\\n(96) Tanimoto, T. T. Elementary mathematical
+        theory of classification and prediction.\\n\\n     Internal IBM Technical Report
+        1958,\\n\\n\\n                                      34 (97) Rogers, D.; Hahn,
+        M. Extended-Connectivity Fingerprints. Journal of Chemical In-\\n\\n      formation
+        and Modeling 2010, 50, 742\u2013754, PMID: 20426451.\\n\\n\\n (98) Mohapatra,
+        S.; An, J.; G\xB4omez-Bombarelli, R. Chemistry-informed macromolecule\\n\\n
+        \     graph representation for similarity computation, unsupervised and supervised
+        learn-\\n\\n       ing. Machine Learning: Science and Technology 2022, 3, 015028.\\n\\n\\n
+        (99) Doshi-Velez, F.; Kortz, M.; Budish, R.; Bavitz, C.; Gershman, S.; O\u2019Brien,
+        D.;\\n\\n       Scott, K.; Schieber, S.; Waldo, J.; Weinberger, D.; Weller,
+        A.; Wood, A. Account-\\n\\n       ability of AI Under the Law: The Role of Explanation.
+        SSRN Electronic Journal\\n\\n     2017,\\n\\n\\n(100) Wachter, S.; Mittelstadt,
+        B.; Russell, C. Counterfactual explanations without opening\\n\\n      the black
+        box: Automated decisions and the GDPR. Harv. JL & Tech. 2017, 31, 841.\\n\\n\\n(101)
+        Jim\xB4enez-Luna, J.; Grisoni, F.; Schneider, G. Drug discovery with explainable
+        artificial\\n\\n       intelligence. Nature Machine Intelligence 2020 2:10 2020,
+        2, 573\u2013584.\\n\\n\\n(102) Fu, T.; Gao, W.; Xiao, C.; Yasonik, J.; Coley,
+        C. W.; Sun, J. Differentiable Scaffold-\\n\\n      ing Tree for Molecule Optimization.
+        International Conference on Learning Represen-\\n\\n       tations. 2022.\\n\\n\\n(103)
+        Shen, C.; Krenn, M.; Eppel, S.; Aspuru-Guzik, A. Deep molecular dreaming: inverse\\n\\n
+        \    machine learning for de-novo molecular design and interpretability with
+        surjective\\n\\n      representations. Machine Learning: Science and Technology
+        2021, 2, 03LT02.\\n\\n\\n(104) Lucic,  A.;   ter  Hoeve,  M.;   Tolomei,  G.;
+        \  Rijke,  M.;   Silvestri,  F.  CF-\\n\\n     GNNExplainer:  Counterfactual
+        Explanations for Graph Neural Networks. arXiv\\n\\n------------\\n\\nQuestion:
         Are counterfactuals actionable? [yes/no]\\n\\n\"}],\"model\":\"gpt-4o-2024-11-20\",\"n\":1,\"temperature\":0.0}"
       headers:
         accept:
@@ -5156,13 +5193,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "6085"
+          - "6390"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -5172,7 +5209,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -5188,22 +5225,23 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAAwAAAP//jFRNbxoxEL3zK0a+pJUAZQltUm60aiQUpZfkUKlEMNizrBOvvfXMEmjE
-          f69skkA+KvXiwzzPm/c8M37oAChr1AiUrlB03bje181lKRfn3tCPyZ8v366+n51eTIbLy/H4vP2t
-          uikjLG5Jy1NWX4e6cSQ2+B2sI6FQYi1OP52cDj5/GhYZqIMhl9KWjfSGoTc4Hgx7RdEbHD8mVsFq
-          YjWCXx0AgId8Jone0FqN4Lj7FKmJGZekRs+XAFQMLkUUMlsW9KK6e1AHL+Sz6vl8fsvBT/3D1ANM
-          Fbd1jXEzVSOYquuKgNaaYiNgLOuWmRh0aL1QLFFLiw5o3Tj0mDwzWA9SEeQCa4FQ7mDrceEIxhP4
-          8HM8+diHiQALCjFIhfKKkgFj5mBrKJKBowWJUDx6WWtBGlumVHCTMxZBKkCd0FwOvQFuMDL1YbwP
-          vyBpYlhZQ4DAlPWWhNLGZ2HoQVfol7kMhFZ0qKkLNd5Zv0yxGlqmsnVQhgiGtGUbfG+H9+G6spxf
-          IyILw72VCq4qbBxtYIWuJe7CfWV1lQ0YYh3tggwggw++d2DmhVsTwIckt6QICNoRxrcGAqATill5
-          E8nYzPZkoj9V3V3PIzlaodc0Yx0ipd4Xx1O/nfr5fH44N5HKljGNrW+dOwDQ+yC7F00Te/OIbJ9n
-          1IVlE8OCX6Wq0nrL1SwScvBpHllCozK67QDc5F1oX4y3amKoG5lJuKNcrhgUpztCtV+/A7goHlEJ
-          gu4AOBmcdd+hnBkStI4PFkpp1BWZfe5++7A1NhwAnQPjb/W8x70zb/3yf+j3gNbUCJnZvrPvXYuU
-          /qd/XXt+6CxYMcWV1TQTSzE1w1CJrdt9HYo3LFTPSuuXFJto8/+RmtnZdv4CAAD//wMAGgP9KD0F
-          AAA=
+          H4sIAAAAAAAAAwAAAP//jFTfaxtHEH7XXzHsQ0lAEpYs24n7ZFxo3UIbkkACUZBGs3N3G+/tLjtz
+          loXx/172zpVkNyl9Obj95pv55ufDCMA4ay7BUINKbfKT69kff7rt5vrm9/pU79urqpn/db37vL34
+          kK6SGRdG3Hxj0n9YU4pt8qwuhgGmzKhcvM4uzhaL0/nF7KQH2mjZF1qddLKIk/nJfDGZzSbzkydi
+          Ex2xmEv4MgIAeOi/RWKwfG8uoXfTv7QsgjWby70RgMnRlxeDIk4Ug5rxAaQYlEOver1ef5MYluFh
+          GQCWRrq2xbxbmktYmo8NA98T56SQueLMgVig7by65Bm2Md8KWCfUibhQA8UuKOcKSTv0wPfJY8BS
+          DhmDC+Q7W8y0YZehSAQXQMhxUFc5OiYABgtXN4DUO8WN8053U/iQmIoter8bg1NoCzkGeRlcG47Z
+          sUCsfhTi1Xvu1LtQcx7D/GR2/jN8itFuMVv4CX5zSg1Fui3YyenrXtEgHVPyjgYnLsDVzRikowZQ
+          /qsEUMUM2Glsy0yAZXLSv3fBcoZff3n3Hl59QmqUM7AC+mkv6+L1FD42LAwSu1w6IF1dsyhog/oi
+          ogBmhk7YgkZIOd45y89lbHZAMYiznEs30Cvngt0xCHHA7KKMYds4aoAwwIYBqVBxM3Ss11umqu9m
+          KYtr+0ih3mc1afG2/KcciUVYpkszHmYss+c7DMQroZh5mLW3e7hoX7kWa5YCVeiFl+FxGdbr9fEU
+          Z646wbJEofP+CMAQog65lv35+oQ87jfGxzrluJEXVFO54KRZZUaJoWyHaEymRx9HAF/7zeyeLZtJ
+          ObZJVxpvuQ83uzh7Mzg0h2NwBJ8unlCNiv4IePN2Pv6Oy5VlReflaL0NITVsD9zDLcDOungEjI4S
+          /7ee7/kekneh/j/uDwARJ2W7Spmto+c5H8wyl2v5I7N9oXvBRjjfOeKVOs6lGZYr7PxwyIzsRLld
+          Vf3ypuyGa1al1Xx2zudnC1pYM3oc/Q0AAP//AwC4qaal1gUAAA==
       headers:
         CF-RAY:
-          - 96665cfffe8dfa86-SJC
+          - 96a9b56d19d0fc54-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -5211,11 +5249,9 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:44 GMT
+          - Tue, 05 Aug 2025 22:25:12 GMT
         Server:
           - cloudflare
-        Strict-Transport-Security:
-          - max-age=31536000; includeSubDomains; preload
         Transfer-Encoding:
           - chunked
         X-Content-Type-Options:
@@ -5229,13 +5265,15 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "3301"
+          - "1771"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
+        strict-transport-security:
+          - max-age=31536000; includeSubDomains; preload
         x-envoy-upstream-service-time:
-          - "3307"
+          - "3761"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -5243,13 +5281,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998545"
+          - "29998478"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
-          - 2ms
+          - 3ms
         x-request-id:
-          - 87d657dc-16e4-4ead-80ee-55454fe55ee8
+          - req_279faf78e070ac830abeb0b8a90f2c90
       status:
         code: 200
         message: OK
@@ -5258,59 +5296,60 @@ interactions:
         '{"messages":[{"role":"system","content":"Answer in a direct and concise
         tone. Your audience is an expert, so be highly specific. If there are ambiguous
         terms or acronyms, first define them."},{"role":"user","content":"Answer the
-        question below with the context.\n\nContext:\n\npqac-e186cea0: Counterfactual
+        question below with the context.\n\nContext:\n\npqac-12c335ad: Counterfactual
         explanations are actionable as they provide local, instance-level explanations
         that suggest which features can be altered to change the outcome. For example,
         in chemistry, changing a hydrophobic functional group in a molecule to a hydrophilic
-        group can increase solubility. This actionability allows users to understand
-        how specific changes in features can lead to different predictions or outcomes.\nFrom
-        Geemi P. Wellawatte, Heta A. Gandhi, Aditi Seshadri, and Andrew D. White. A
-        perspective on explanations of molecular prediction models. ChemRxiv, Unknown
-        year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02, doi:10.26434/chemrxiv-2022-qfv02.
-        This article has 1 citations.\n\npqac-c65669a0: Counterfactuals are actionable
-        as they provide specific modifications to molecular structures that can lead
-        to desired changes in properties. For example, in the context of blood-brain
-        barrier (BBB) permeation prediction, counterfactual explanations suggested modifications
-        to the carboxylic acid group of a molecule, enabling it to permeate the BBB.
-        These actionable insights align with experimental findings, demonstrating how
-        hydrophobic interactions and surface area influence BBB permeability. Counterfactuals
-        thus offer practical guidance for molecular design by proposing changes that
-        can achieve specific outcomes.\nFrom Geemi P. Wellawatte, Heta A. Gandhi, Aditi
-        Seshadri, and Andrew D. White. A perspective on explanations of molecular prediction
-        models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\n\npqac-135a48bc:
-        The excerpt discusses the use of molecular counterfactuals in explaining black-box
-        models for molecular property predictions. Counterfactuals are described as
-        actionable because they are represented as chemical structures familiar to domain
-        experts, are sparse, and provide actionable insights. They are defined as molecules
-        with minimal distance from a base molecule but with contrasting chemical properties.
-        The text highlights their utility in understanding and modifying molecular properties,
-        making them a practical tool for domain experts.\nFrom Geemi P. Wellawatte,
+        group can increase solubility. This actionability makes counterfactuals a useful
+        tool in explainable AI (XAI) for chemistry, as they offer intuitive understanding
+        and uncover spurious relationships in training data.\nFrom Geemi P. Wellawatte,
         Heta A. Gandhi, Aditi Seshadri, and Andrew D. White. A perspective on explanations
         of molecular prediction models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
-        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\n\npqac-931e8443:
-        The excerpt discusses counterfactual explanations in the context of explainable
-        AI (XAI). It states that counterfactuals are considered ''better'' explanations
-        because they are both actionable and sparse. Actionable explanations provide
-        a set of features that can change the outcome, making them useful for decision-making.
-        This contrasts with Shapley values, which are described as non-actionable because
-        they do not offer a clear set of features to alter the prediction outcome.\nFrom
-        Geemi P. Wellawatte, Heta A. Gandhi, Aditi Seshadri, and Andrew D. White. A
-        perspective on explanations of molecular prediction models. ChemRxiv, Unknown
-        year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02, doi:10.26434/chemrxiv-2022-qfv02.
-        This article has 1 citations.\n\npqac-462019b7: The excerpt discusses Explainable
-        Artificial Intelligence (XAI) in the context of deep learning (DL) models for
-        chemistry. It highlights the importance of interpretability and explanations
-        in understanding ''black-box'' models. The authors mention methods like chemical
-        counterfactuals and descriptor explanations, which provide insights into structure-property
-        relationships and explain DL predictions. Counterfactuals are presented as a
-        way to explain predictions by showing alternative scenarios or changes that
-        could lead to different outcomes, making them actionable in understanding and
-        refining model predictions.\nFrom Geemi P. Wellawatte, Heta A. Gandhi, Aditi
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\n\npqac-1bc58f72:
+        Counterfactuals are actionable as they provide suggestions for modifications
+        to molecular structures to achieve desired properties. For example, in the context
+        of blood-brain barrier (BBB) permeation prediction, counterfactual explanations
+        identified modifications to the carboxylic acid group that enabled a molecule
+        to permeate the BBB. These actionable insights align with experimental findings,
+        demonstrating how counterfactuals can guide structural changes to achieve specific
+        outcomes, such as enhancing hydrophobic interactions or surface area. This approach
+        proves useful in drug development and other applications where structure-property
+        relationships are critical.\nFrom Geemi P. Wellawatte, Heta A. Gandhi, Aditi
+        Seshadri, and Andrew D. White. A perspective on explanations of molecular prediction
+        models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\n\npqac-32de2cc6:
+        Counterfactuals are described as actionable in the context of molecular property
+        prediction models. They are defined as molecular structures with minimal distance
+        from a base molecule but with contrasting chemical properties. These explanations
+        are represented as chemical structures, which are familiar to domain experts,
+        and are sparse and actionable. The use of Tanimoto similarity for distance measurement
+        and the potential for future exploration of other methods are also mentioned.
+        Counterfactual explanations are highlighted as useful for understanding and
+        modifying molecular properties.\nFrom Geemi P. Wellawatte, Heta A. Gandhi, Aditi
+        Seshadri, and Andrew D. White. A perspective on explanations of molecular prediction
+        models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
+        doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\n\npqac-c566bbf3:
+        The excerpt discusses counterfactuals as a type of explanation in molecular
+        prediction models. It states that counterfactuals are considered ''better''
+        explanations because they are both actionable and sparse. Actionable explanations
+        provide a set of features that can change the outcome, making them useful for
+        decision-making. This contrasts with Shapley values, which are complete but
+        not actionable or sparse.\nFrom Geemi P. Wellawatte, Heta A. Gandhi, Aditi Seshadri,
+        and Andrew D. White. A perspective on explanations of molecular prediction models.
+        ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02, doi:10.26434/chemrxiv-2022-qfv02.
+        This article has 1 citations.\n\npqac-d7a22feb: Counterfactual explanations
+        are described as examples with minimal deviation from the original instance
+        but with a contrasting outcome. They aim to answer the question of what smallest
+        change could alter the outcome of an instance of interest. In the context of
+        molecular prediction models, counterfactuals use a distance metric, such as
+        the Tanimoto distance between molecular fingerprints, to define similarity.
+        These explanations are actionable as they provide insights into specific changes
+        needed to achieve a desired outcome, making them useful for understanding and
+        modifying model predictions.\nFrom Geemi P. Wellawatte, Heta A. Gandhi, Aditi
         Seshadri, and Andrew D. White. A perspective on explanations of molecular prediction
         models. ChemRxiv, Unknown year. URL: https://doi.org/10.26434/chemrxiv-2022-qfv02,
         doi:10.26434/chemrxiv-2022-qfv02. This article has 1 citations.\n\nValid Keys:
-        pqac-e186cea0, pqac-c65669a0, pqac-135a48bc, pqac-931e8443, pqac-462019b7\n\n----\n\nQuestion:
+        pqac-12c335ad, pqac-1bc58f72, pqac-32de2cc6, pqac-c566bbf3, pqac-d7a22feb\n\n------------\n\nQuestion:
         Are counterfactuals actionable? [yes/no]\n\nWrite an answer based on the context.
         If the context provides insufficient information reply \"I cannot answer.\"
         For each part of your answer, indicate which sources most support it via citation
@@ -5332,13 +5371,13 @@ interactions:
         connection:
           - keep-alive
         content-length:
-          - "5629"
+          - "5715"
         content-type:
           - application/json
         host:
           - api.openai.com
         user-agent:
-          - AsyncOpenAI/Python 1.97.1
+          - AsyncOpenAI/Python 1.99.0
         x-stainless-arch:
           - arm64
         x-stainless-async:
@@ -5348,7 +5387,7 @@ interactions:
         x-stainless-os:
           - MacOS
         x-stainless-package-version:
-          - 1.97.1
+          - 1.99.0
         x-stainless-raw-response:
           - "true"
         x-stainless-read-timeout:
@@ -5364,28 +5403,29 @@ interactions:
     response:
       body:
         string: !!binary |
-          H4sIAAAAAAAAA4xWTW8cNwy9+1cQc2qAWWN3vV5/3NIGbYM2h6BNi7YJFhyJM8NaI6kSZXsQ+L8X
-          0uyXkzXaywIjihQf3yO5n88AKtbVLVSqR1GDN7Nvx3ftw4/vf3j95/L9m99/+k3pn9OHVde8W7+x
-          Y1VnD9f8TUp2XufKDd6QsLOTWQVCoRx1cXV5cbVcX65WxTA4TSa7dV5mKzdbzper2WIxW863jr1j
-          RbG6hb/OAAA+l9+cotX0WN3CvN6dDBQjdlTd7i8BVMGZfFJhjBwFrVT1waicFbIl6z8o1qBcskKh
-          RSUJTQQMBKgyCmwMncOvPY3gg7tnTWCcQlMD2xxW0czQPRmgR2/QYvaJ0IwQU9dRFLYdRE+KW1bQ
-          EkoKBIPT+Xt7WXoUUGhB9Wg7AukJXBLlBqphwLscQnoaIEVqk4HWBdCkOLKzs639G/8PqhktrteK
-          cF5D+by5WND1anXx6hzeWhicIZUMBvCBNBd0UFj4zwJAQwpTLKmVOngXCQa2PKCBKCEpSQHNFkEE
-          cbvnaIsvUExGgC1oihxIlzAUZAQ0QmGqRQ0xqR4wTiUaMzSEftTB+d41rKALLvkcf3/MhhU4S/mQ
-          bdZbJIjOpIYNywguTC9MsRSGxj2O2QcV60M8sn1mExrjnJ41AdlCgyEwBfAUBsJtvJOlVuvL9foG
-          56+KVuKXFKPhzsIDS591QoEHsoIGWraabRdr0DQ4GyXXYaKbM0+ZAoUGkkxv8zGNuZCd3eZzSOCj
-          /Wi/O0Wnie5FTvOF6DFEArQaAvlAkayQzmSonoaSx45qitDiwIYx5NppN+RyFWgSa6D8QsahDGEo
-          ETnn4wNJebtLrEu1s5YPVB9LtIiDKW7xLS4ucXXdqFfn8MEaviMYSHqn414xv/ToDY1wjyblnn7o
-          WfVgUN0do2YbuevlhORzA7NiMSOwJivcjruGPerRhiY1kS4iVD3TPYHmtqVAVnaNG593rscgXJCZ
-          KcFtLtPU4Cm112+3WI/69vsUpKcwuED1vvvKFNoByZV1B2Jm+7YKZCb19exjDS5nmBM6qOoZC4Fa
-          tpN9Nx1iYS5ZTSGPuqxU0EQeMqt2IiyPjwykSCRK2PXHar2cL26aq9IQ6aUJA949UMhDTZybJtsR
-          V1/r5tmQ5RdmWnabeuP8eOYHalPEvHJsMubIgNY6mWLmbfNpa3na7xfjOh9cE79wrXLBYr/JE8fZ
-          vEuiOF8V69MZwKeyx9Kz1VT54AYvG3F3VJ5bXKyWU8DqsDoP5uX1zdYqTtAc+a0vFvWJkBtNgmzi
-          0TKsFKqe9MH3sDkxaXZHhrMj4F/ncyr2BJ5t93/CHwxKkRfSmwNtp64Fyv8tXrq2L3RJuIoU7lnR
-          RphCJkNTi8lMa7+KYxQaNi3bLsuJp93f+o1aX12tr9v19bw6ezr7FwAA//8DALWSShAECQAA
+          H4sIAAAAAAAAAwAAAP//jFZLbxw3DL77VxBzioFZI7vrR+pbHSCHFu0luQRNsOBInBnWGlHVw/Ei
+          yH8vqNnHpHGAXgx4KJLfgxL36wVAw7a5h8aMmM0U3Ort+vc/H0d+tGWfGX9z40jjH+++2I8f0y/v
+          mlYzpPubTD5mXRmZgqPM4uewiYSZtOr67ub6eru5W29qYBJLTtOGkFfXstq83lyv1uvV5vUhcRQ2
+          lJp7+OsCAOBr/asQvaXn5h5et8cvE6WEAzX3p0MATRSnXxpMiVNGn5v2HDTiM/mK+iOlFowUnyn2
+          aHJBlwAjARplgZ2jK/gw0h5ClCe2BE4MuhbYa1lDK0dP5ICeg0OPmpOg20Mqw0Apsx8gBTLcswEz
+          oh8oQRboCXOJlEAiTOLIFIcRUo7FzN/ziBkMekCXKUIeCaRkIxOB9MAKl1K+gncSgZ5RVVdMYEaa
+          OOW4byFScGgUAcK4t1HCKB0b6IufuTkYopQAXziP5zPs2IB4qt3Zq4OJIIkrHTvO+xYsTeJTjljp
+          5ZE4Qsk1qBCGwrZ2PSkIk1hV4CDPq/APmtV6Y7bbG7SXV/CeJ3YY3b5SUK7VoeesXDsnYlddRPbQ
+          YYxMEV49PDxcQqA4Ua0JIZLl2u9HN9mSz9wz2ZPA6I5mtJCKGQHTjHE/y2UwdvK8VyXQsJ11atU4
+          mgk9PDws2reAjgevuVVLeg4UeSKf0UHPXuVIgN7CyMPoeBgXyoWoQhl0gCE4NtidlLSxDGB1viRo
+          taNynbl5099tLq/gk//k3740vS4JWOrZk53JvTBjFevEnid02oZnLfsoEyB0avshj6AreT6vzkSc
+          JztECRQzq4wTPh44TZACxkSV8A/XSNFFCpES+TyDqzOrApyxtfBlZPUlkg574cxPBL1EsDLpJFSJ
+          c2prk6JIKUc2CRw/EnxAz5NkgTRPluqZBSbCVCIt1LA83+IW+hLzSBFSCUHiwp8jg5MrxVuKmjZP
+          ubeL0TlXPmtzcG27sbQx5raF+q+9w82mp+5g4q/W8nwp9Ra89CAZ8YktRR3jovMlUTlJRb14flSJ
+          Uexist+PGBzt4QldUWk7MqiS5aMfnegDcL6uymn2sD08e8rNOMJYb7cqNptBhhOLXx3cn5mam9vb
+          ruu3l2p5+dn7WuHUdlnE1XILCOyTXpOkgi9FPV5zqAskXS2f9Uh9SahbxRfnFgH0XvL8+OhC+XyI
+          fDutECdDiNKl/6Q2PXtO406fQPG6LlKW0NTotwuAz3VVle+2TxOiTCHvsjxSbbfe3q7ngs15O57D
+          mzc3h2iWjG6Rd3t9275QcmcpI7u02HeNQTOSPeeelyMWy7IIXCyI/4jnpdozefbD/yl/DhhDIZPd
+          nS176Vgk/fnws2MnoSvgJlF8YkO7zBTVDEs9Fjdv9ibtU6Zp17MfKIbI83rvw+6m67dbsyZz3Vx8
+          u/gXAAD//wMAkJHWaOcIAAA=
       headers:
         CF-RAY:
-          - 96665d15ac10fa86-SJC
+          - 96a9b5858e7bfc54-SJC
         Connection:
           - keep-alive
         Content-Encoding:
@@ -5393,7 +5433,7 @@ interactions:
         Content-Type:
           - application/json
         Date:
-          - Mon, 28 Jul 2025 18:15:48 GMT
+          - Tue, 05 Aug 2025 22:25:15 GMT
         Server:
           - cloudflare
         Strict-Transport-Security:
@@ -5411,13 +5451,13 @@ interactions:
         openai-organization:
           - future-house-xr4tdh
         openai-processing-ms:
-          - "4298"
+          - "2812"
         openai-project:
           - proj_RpeV6PrPclPHBb5GlExPXSBj
         openai-version:
           - "2020-10-01"
         x-envoy-upstream-service-time:
-          - "4301"
+          - "2839"
         x-ratelimit-limit-requests:
           - "10000"
         x-ratelimit-limit-tokens:
@@ -5425,13 +5465,13 @@ interactions:
         x-ratelimit-remaining-requests:
           - "9999"
         x-ratelimit-remaining-tokens:
-          - "29998630"
+          - "29998609"
         x-ratelimit-reset-requests:
           - 6ms
         x-ratelimit-reset-tokens:
           - 2ms
         x-request-id:
-          - 15f7d53f-6f1c-4e3b-a97f-8599930c37aa
+          - req_71d82903e7fc44fb9267de86729f2794
       status:
         code: 200
         message: OK
diff --git a/tests/stub_data/sf_districts.png b/tests/stub_data/sf_districts.png
new file mode 100644
index 000000000..70dccf614
Binary files /dev/null and b/tests/stub_data/sf_districts.png differ
diff --git a/tests/test_paperqa.py b/tests/test_paperqa.py
index c83368f72..9ffba17f2 100644
--- a/tests/test_paperqa.py
+++ b/tests/test_paperqa.py
@@ -1,5 +1,8 @@
+import asyncio
+import base64
 import contextlib
 import csv
+import io
 import os
 import pathlib
 import pickle
@@ -11,9 +14,11 @@
 from io import BytesIO
 from pathlib import Path
 from typing import cast
+from unittest.mock import MagicMock, call
 from uuid import UUID
 
 import httpx
+import litellm
 import numpy as np
 import pytest
 import pytest_asyncio
@@ -51,10 +56,11 @@
 from paperqa.core import llm_parse_json
 from paperqa.prompts import CANNOT_ANSWER_PHRASE
 from paperqa.prompts import qa_prompt as default_qa_prompt
-from paperqa.readers import PDFParserFn, read_doc
+from paperqa.readers import PDFParserFn, parse_image, read_doc
 from paperqa.settings import AsyncContextSerializer
-from paperqa.types import ChunkMetadata, Context
+from paperqa.types import ChunkMetadata, Context, ParsedMedia
 from paperqa.utils import (
+    bytes_to_string,
     clean_possessives,
     encode_id,
     extract_score,
@@ -62,8 +68,10 @@
     maybe_is_html,
     maybe_is_text,
     name_in_text,
+    string_to_bytes,
     strings_similarity,
     strip_citations,
+    validate_image,
 )
 
 THIS_MODULE = pathlib.Path(__file__)
@@ -1228,6 +1236,7 @@ async def test_parser_only_reader(pdf_parser: PDFParserFn, stub_data_dir: Path)
     assert parsed_text.metadata.total_parsed_text_length == sum(
         len(t) for t in parsed_text.content.values()  # type: ignore[misc,union-attr]
     )
+    assert not parsed_text.metadata.count_parsed_media
 
 
 @pytest.mark.asyncio
@@ -1274,6 +1283,10 @@ async def test_chunk_metadata_reader(
     assert (
         int(last_page) - int(stlast_page) <= 2
     ), "Incorrect page range if last chunk is a partial chunk"
+    assert not metadata.count_parsed_media
+    assert (
+        sum(len(t.media) for t in chunk_text) == metadata.count_parsed_media
+    ), "Expected chunks' media to match parsed media"
 
     chunk_text, metadata = await read_doc(
         stub_data_dir / "flag_day.html",
@@ -1321,6 +1334,127 @@ async def test_chunk_metadata_reader(
         )
 
 
+@pytest.mark.asyncio
+async def test_image_aggregation(stub_data_dir: Path) -> None:
+    png_path = stub_data_dir / "sf_districts.png"
+
+    # Test how self-comparisons work
+    ((_, (parsed_image,)),) = cast(dict, (await parse_image(png_path)).content).values()
+    assert parsed_image == parsed_image, "Expected equality"  # noqa: PLR0124
+    assert parsed_image.to_id() == parsed_image.to_id(), "Expected same ID"
+    assert len({parsed_image, parsed_image}) == 1, "Expected shared hash"
+    assert not parsed_image.text, "Expected no text for later assertions to make sense"
+    assert (
+        parsed_image.info.get("type") != "table"
+    ), "Expected no table for later assertions to make sense"
+
+    # Test how self-comparisons work
+    ((_, (parsed_image2,)),) = cast(
+        dict, (await parse_image(png_path)).content
+    ).values()
+    assert parsed_image == parsed_image2, "Expected equality to persist across reads"
+    assert (
+        parsed_image.to_id() == parsed_image2.to_id()
+    ), "Expected ID to persist across reads"
+    assert (
+        len({parsed_image, parsed_image2}) == 1
+    ), "Expected hash to persist across reads"
+
+    # Test different read details
+    ((_, (parsed_image3,)),) = cast(
+        dict, (await parse_image(png_path)).content
+    ).values()
+    parsed_image3.text = "Golden Gate"
+    parsed_image3.info["type"] = "table"
+    assert parsed_image != parsed_image3, "Expected tables to be differentiable"
+    assert (
+        parsed_image.to_id() != parsed_image3.to_id()
+    ), "Expected ID to mismatch between tables and images"
+    assert (
+        len({parsed_image, parsed_image3}) == 2
+    ), "Expected tables to be hashed differently"
+
+
+@pytest.mark.asyncio
+async def test_read_doc_images_metadata(stub_data_dir: Path) -> None:
+    png_path = stub_data_dir / "sf_districts.png"
+    doc = Doc(docname="stub", citation="stub", dockey="stub")
+
+    # Test parsing only
+    parsed_text = await read_doc(png_path, doc, parsed_text_only=True)
+    assert isinstance(parsed_text.content, dict)
+    assert "1" in parsed_text.content
+    page_content = parsed_text.content["1"]
+    assert isinstance(page_content, tuple)
+    text_content, (parsed_image,) = page_content
+    assert not text_content, "Expected no text content for an image"
+    assert isinstance(parsed_image, ParsedMedia)
+    assert parsed_image.index == 0
+    assert isinstance(parsed_image.data, bytes)
+    assert len(parsed_image.data) > 0
+    assert not parsed_image.text, "Expected no text content for a standalone image"
+    assert parsed_image.info["suffix"] == ".png"
+    image_id = parsed_image.to_id()
+    assert image_id.version == 4, "Expected a uuid4-compatible ID"
+    assert image_id == UUID("f6426bc3-382a-45a4-8677-08744044864f")
+    assert parsed_text.metadata.parse_type == "image"
+    assert parsed_text.metadata.count_parsed_media == 1
+    assert parsed_text.metadata.total_parsed_text_length == 0
+    assert parsed_text.metadata.chunk_metadata is None
+
+    # Test parsing + 'chunking'
+    (text,) = await read_doc(png_path, doc)
+    assert isinstance(text, Text)
+    assert text.doc == doc
+    (image,) = text.media
+    assert image == parsed_image
+
+    # Test including metadata
+    texts_with_metadata = await read_doc(png_path, doc, include_metadata=True)
+    assert isinstance(texts_with_metadata, tuple)
+    texts, metadata = texts_with_metadata
+    assert len(texts) == 1
+    assert texts[0] == text
+    assert metadata.parse_type == "image"
+    assert metadata.count_parsed_media == 1
+    assert metadata.total_parsed_text_length == 0
+    assert metadata.chunk_metadata is not None
+    assert not metadata.chunk_metadata.chunk_chars
+    assert not metadata.chunk_metadata.overlap
+    assert metadata.chunk_metadata.chunk_type == "no_chunk"
+
+
+@pytest.mark.asyncio
+async def test_read_doc_images_concurrency(stub_data_dir: Path) -> None:
+    png_path = stub_data_dir / "sf_districts.png"
+    doc = Doc(docname="stub", citation="stub", dockey="stub")
+    validation_mock = MagicMock()
+
+    async def validate(data: bytes) -> None:  # noqa: RUF029
+        validate_image(io.BytesIO(data))
+        validation_mock(data)
+
+    # Check we can concurrently read in the same image many times
+    concurrent_call_count = 10
+    seen_media = set()
+    bulk_texts = await asyncio.gather(
+        *(
+            read_doc(png_path, doc, validator=validate)
+            for _ in range(concurrent_call_count)
+        )
+    )
+    for (text,) in bulk_texts:
+        assert text.doc == doc
+        assert len(text.media) == 1
+        seen_media.add(text.media[0])
+    assert (
+        len(seen_media) == 1
+    ), "Expected the concurrent reads to all have the same parsed result"
+    validation_mock.assert_has_calls(
+        [call(next(iter(seen_media)).data)] * concurrent_call_count
+    )
+
+
 @pytest.mark.asyncio
 async def test_code() -> None:
     settings = Settings.from_name("fast")
@@ -1334,6 +1468,98 @@ async def test_code() -> None:
     assert "test_paperqa.py" in session.answer
 
 
+@pytest.mark.asyncio
+async def test_images(stub_data_dir: Path) -> None:
+    settings = Settings.from_name("fast")
+    # Let's use default prompting set up, so we can get JSON summary-support
+    settings.prompts = type(settings.prompts)()
+    # We don't support image embeddings yet, so disable embedding
+    settings.answer.evidence_retrieval = False
+    settings.parsing.defer_embedding = True
+
+    docs = Docs()
+    districts_docname = await docs.aadd(
+        stub_data_dir / "sf_districts.png",
+        citation=(
+            '"File:San francisco districts.png." Wikimedia Commons.'
+            " 7 Sep 2023, 07:38 UTC."
+            " <https://commons.wikimedia.org/w/index.php?title=File:San_francisco_districts.png&oldid=799209398>"
+            " July 2025."
+        ),
+        settings=settings,
+    )
+    assert districts_docname, "Expected successful image addition"
+    (districts_doc,) = (d for d in docs.docs.values() if d.docname == districts_docname)
+    session = await docs.aquery(
+        "What districts neighbor the Western Addition?", settings=settings
+    )
+    assert (
+        sum(
+            district in session.answer
+            for district in ("The Avenues", "Golden Gate", "Civic Center", "Haight")
+        )
+        >= 2
+    ), "Expected at least two neighbors to be matched"
+    assert session.cost > 0
+    contexts_used = [
+        c
+        for c in session.contexts
+        if c.id in session.used_contexts and c.text.doc == districts_doc
+    ]
+    assert contexts_used
+    assert all(c.used_images for c in contexts_used)  # type: ignore[attr-defined]
+
+
+@pytest.mark.asyncio
+async def test_images_corrupt(stub_data_dir: Path) -> None:
+    settings = Settings.from_name("fast")
+    # Let's use default prompting set up, so we can get JSON summary-support
+    settings.prompts = type(settings.prompts)()
+    # We don't support image embeddings yet, so disable embedding
+    settings.answer.evidence_retrieval = False
+    settings.parsing.defer_embedding = True
+
+    docs = Docs()
+    districts_docname = await docs.aadd(
+        stub_data_dir / "sf_districts.png",
+        citation=(
+            '"File:San francisco districts.png." Wikimedia Commons.'
+            " 7 Sep 2023, 07:38 UTC."
+            " <https://commons.wikimedia.org/w/index.php?title=File:San_francisco_districts.png&oldid=799209398>"
+            " July 2025."
+        ),
+        settings=settings,
+    )
+    assert districts_docname, "Expected successful image addition"
+    (districts_doc,) = (d for d in docs.docs.values() if d.docname == districts_docname)
+    for media in (t.media for t in docs.texts if t.doc == districts_doc and t.media):
+        for m in media:
+            # Validate the image, then chop the image in half (breaking it), and
+            # confirm it's no longer valid (and that we can detect it's no longer valid)
+            validate_image(io.BytesIO(m.data))
+            m.data = m.data[: len(m.data) // 2]
+            with pytest.raises(OSError, match="truncated"):
+                validate_image(io.BytesIO(m.data))
+    with pytest.raises(litellm.BadRequestError, match="unsupported image"):
+        await docs.aquery(
+            "What districts neighbor the Western Addition?", settings=settings
+        )
+    settings.answer.evidence_text_only_fallback = True
+    # The answer will be garbage, but let's make sure we didn't claim to use images
+    session = await docs.aquery(
+        "What districts neighbor the Western Addition?", settings=settings
+    )
+    assert session.used_contexts
+    assert session.cost > 0
+    contexts_used = [
+        c
+        for c in session.contexts
+        if c.id in session.used_contexts and c.text.doc == districts_doc
+    ]
+    assert contexts_used
+    assert all(not c.used_images for c in contexts_used)  # type: ignore[attr-defined]
+
+
 def test_zotero() -> None:
     from paperqa.contrib import ZoteroDB
 
@@ -2152,3 +2378,23 @@ def test_maybe_get_date():
 )
 def test_clean_possessives(raw_text: str, cleaned_text: str) -> None:
     assert clean_possessives(raw_text) == cleaned_text
+
+
+@pytest.mark.parametrize(
+    "value",
+    [
+        pytest.param(b"Hello, World!", id="simple-text"),
+        pytest.param(b"", id="empty-bytes"),
+        pytest.param(bytes([0, 1, 2, 255, 128, 64]), id="binary-data"),
+        pytest.param(b"Test data for base64 encoding", id="base64-validation"),
+        pytest.param("Hello 世界 🌍".encode(), id="utf8-text"),
+    ],
+)
+def test_str_bytes_conversions(value: bytes) -> None:
+    # Test round-trip conversion
+    encoded_string = bytes_to_string(value)
+    decoded_bytes = string_to_bytes(encoded_string)
+    assert decoded_bytes == value
+
+    # Validate that encoded string is valid base64
+    assert base64.b64decode(encoded_string) == value
diff --git a/uv.lock b/uv.lock
index 2af3428d3..e3d2958a0 100644
--- a/uv.lock
+++ b/uv.lock
@@ -1,5 +1,5 @@
 version = 1
-revision = 3
+revision = 2
 requires-python = ">=3.11"
 resolution-markers = [
     "python_full_version >= '3.13' and platform_python_implementation == 'PyPy'",
@@ -1993,6 +1993,7 @@ dev = [
     { name = "mypy" },
     { name = "paper-qa-pymupdf" },
     { name = "paper-qa-pypdf" },
+    { name = "pillow" },
     { name = "pre-commit" },
     { name = "pydantic" },
     { name = "pylint-pydantic" },
@@ -2015,6 +2016,9 @@ dev = [
     { name = "types-setuptools" },
     { name = "vcrpy" },
 ]
+image = [
+    { name = "pillow" },
+]
 ldp = [
     { name = "ldp" },
 ]
@@ -2064,11 +2068,12 @@ requires-dist = [
     { name = "mypy", marker = "extra == 'dev'", specifier = ">=1.8" },
     { name = "numpy" },
     { name = "openreview-py", marker = "extra == 'openreview'" },
-    { name = "paper-qa", extras = ["ldp", "local", "pymupdf", "pypdf", "qdrant", "typing", "zotero"], marker = "extra == 'dev'", editable = "." },
+    { name = "paper-qa", extras = ["image", "ldp", "local", "pymupdf", "pypdf", "qdrant", "typing", "zotero"], marker = "extra == 'dev'", editable = "." },
     { name = "paper-qa-pymupdf", marker = "extra == 'pymupdf'", editable = "packages/paper-qa-pymupdf" },
     { name = "paper-qa-pymupdf", marker = "extra == 'zotero'", editable = "packages/paper-qa-pymupdf" },
     { name = "paper-qa-pypdf", editable = "packages/paper-qa-pypdf" },
     { name = "paper-qa-pypdf", marker = "extra == 'pypdf'", editable = "packages/paper-qa-pypdf" },
+    { name = "pillow", marker = "extra == 'image'", specifier = ">=10.3.0" },
     { name = "pre-commit", marker = "extra == 'dev'", specifier = ">=3.4" },
     { name = "pybtex" },
     { name = "pydantic", specifier = "~=2.0,>=2.10.1" },
@@ -2099,7 +2104,7 @@ requires-dist = [
     { name = "types-setuptools", marker = "extra == 'typing'" },
     { name = "vcrpy", marker = "extra == 'dev'", specifier = ">=6" },
 ]
-provides-extras = ["dev", "ldp", "local", "openreview", "pymupdf", "pypdf", "qdrant", "typing", "zotero"]
+provides-extras = ["dev", "image", "ldp", "local", "openreview", "pymupdf", "pypdf", "qdrant", "typing", "zotero"]
 
 [package.metadata.requires-dev]
 dev = [{ name = "paper-qa", extras = ["dev"], editable = "." }]