fix(gateway): preserve spaced MEDIA paths

gateway/platforms/base.py

@@ -1275,6 +1275,16 @@ class BasePlatformAdapter(ABC):
     - Sending messages/responses
     - Handling media
     """
+    _MEDIA_EXTENSIONS_RE = (
+        r"png|jpe?g|gif|webp|mp4|mov|avi|mkv|webm|ogg|opus|mp3|wav|m4a|flac|"
+        r"epub|pdf|zip|rar|7z|docx?|xlsx?|pptx?|txt|csv|apk|ipa|kmz|kml|"
+        r"json|xml|html?|geojson|gpx"
+    )
+    _MEDIA_TAG_RE = re.compile(
+        r'''[`"']?MEDIA:\s*(?P<path>`[^`\n]+`|"[^"\n]+"|'[^'\n]+'|[^\n]+?\.(?:'''
+        + _MEDIA_EXTENSIONS_RE
+        + r''')(?=[\s`"',;:)\]}]|$)|\S+)[`"']?'''
+    )
 
     def __init__(self, config: PlatformConfig, platform: Platform):
         self.config = config
@@ -2136,9 +2146,7 @@ def extract_media(content: str) -> Tuple[List[Tuple[str, bool]], str]:
 
         # Extract MEDIA:<path> tags, allowing optional whitespace after the colon
         # and quoted/backticked paths for LLM-formatted outputs.
-        media_pattern = re.compile(
-            r'''[`"']?MEDIA:\s*(?P<path>`[^`\n]+`|"[^"\n]+"|'[^'\n]+'|(?:~/|/)\S+(?:[^\S\n]+\S+)*?\.(?:png|jpe?g|gif|webp|mp4|mov|avi|mkv|webm|ogg|opus|mp3|wav|m4a|flac|epub|pdf|zip|rar|7z|docx?|xlsx?|pptx?|txt|csv|apk|ipa)(?=[\s`"',;:)\]}]|$))[`"']?'''
-        )
+        media_pattern = BasePlatformAdapter._MEDIA_TAG_RE
         for match in media_pattern.finditer(content):
             path = match.group("path").strip()
             if len(path) >= 2 and path[0] == path[-1] and path[0] in "`\"'":
@@ -3150,7 +3158,7 @@ async def _stop_typing_task() -> None:
                 # Strip any remaining internal directives from message body (fixes #1561)
                 text_content = text_content.replace("[[audio_as_voice]]", "").strip()
                 text_content = text_content.replace("[[as_document]]", "").strip()
-                text_content = re.sub(r"MEDIA:\s*\S+", "", text_content).strip()
+                text_content = BasePlatformAdapter._MEDIA_TAG_RE.sub("", text_content).strip()
                 if images:
                     logger.info("[%s] extract_images found %d image(s) in response (%d chars)", self.name, len(images), len(response))
 

gateway/run.py

@@ -15778,15 +15778,7 @@ def _clarify_callback_sync(question: str, choices) -> str:
                 if _hm.get("role") in {"tool", "function"}:
                     _hc = _hm.get("content", "")
                     if "MEDIA:" in _hc:
-                        _TOOL_MEDIA_RE = re.compile(
-                            r'MEDIA:((?:/|~\/)\S+\.(?:png|jpe?g|gif|webp|'
-                            r'mp4|mov|avi|mkv|webm|ogg|opus|mp3|wav|m4a|'
-                            r'flac|epub|pdf|zip|rar|7z|docx?|xlsx?|pptx?|'
-                            r'txt|csv|apk|ipa))',
-                            re.IGNORECASE
-                        )
-                        for _match in _TOOL_MEDIA_RE.finditer(_hc):
-                            _p = _match.group(1).strip().rstrip('",}')
+                        for _p, _ in BasePlatformAdapter.extract_media(_hc)[0]:
                             if _p:
                                 _history_media_paths.add(_p)
 
@@ -16074,15 +16066,7 @@ def _approval_notify_sync(approval_data: dict) -> None:
                     if msg.get("role") in {"tool", "function"}:
                         content = msg.get("content", "")
                         if "MEDIA:" in content:
-                            _TOOL_MEDIA_RE = re.compile(
-                                r'MEDIA:((?:/|~\/)\S+\.(?:png|jpe?g|gif|webp|'
-                                r'mp4|mov|avi|mkv|webm|ogg|opus|mp3|wav|m4a|'
-                                r'flac|epub|pdf|zip|rar|7z|docx?|xlsx?|pptx?|'
-                                r'txt|csv|apk|ipa))',
-                                re.IGNORECASE
-                            )
-                            for match in _TOOL_MEDIA_RE.finditer(content):
-                                path = match.group(1).strip().rstrip('",}')
+                            for path, _ in BasePlatformAdapter.extract_media(content)[0]:
                                 if path and path not in _history_media_paths:
                                     media_tags.append(f"MEDIA:{path}")
                             if "[[audio_as_voice]]" in content:

gateway/stream_consumer.py

@@ -612,7 +612,7 @@ async def run(self) -> None:
     # Pattern to strip MEDIA:<path> tags (including optional surrounding quotes).
     # Matches the simple cleanup regex used by the non-streaming path in
     # gateway/platforms/base.py for post-processing.
-    _MEDIA_RE = re.compile(r'''[`"']?MEDIA:\s*\S+[`"']?''')
+    _MEDIA_RE = _BasePlatformAdapter._MEDIA_TAG_RE
 
     @staticmethod
     def _clean_for_display(text: str) -> str:

mcp_serve.py

@@ -32,7 +32,6 @@
 import json
 import logging
 import os
-import re
 import sys
 import threading
 import time
@@ -176,9 +175,9 @@ def _extract_attachments(msg: dict) -> List[dict]:
     # MEDIA: tags in text content
     text = _extract_message_content(msg)
     if text:
-        media_pattern = re.compile(r'MEDIA:\s*(\S+)')
-        for match in media_pattern.finditer(text):
-            path = match.group(1)
+        from gateway.platforms.base import BasePlatformAdapter
+
+        for path, _ in BasePlatformAdapter.extract_media(text)[0]:
             attachments.append({"type": "media", "path": path})
 
     return attachments

tests/gateway/test_platform_base.py

@@ -329,6 +329,34 @@ def test_media_tag_supports_unquoted_flac_paths_with_spaces(self):
         assert media == [("/tmp/Jane Doe/speech.flac", False)]
         assert cleaned == ""
 
+    def test_media_tag_supports_windows_drive_paths_with_spaces(self):
+        content = r"Here is the file MEDIA:C:\Users\Confera\OneDrive\Nusa Alam Kreasindo\Project\Foo\report.pdf for you."
+        media, cleaned = BasePlatformAdapter.extract_media(content)
+        assert media == [
+            (r"C:\Users\Confera\OneDrive\Nusa Alam Kreasindo\Project\Foo\report.pdf", False)
+        ]
+        assert "MEDIA:" not in cleaned
+        assert "Alam Kreasindo" not in cleaned
+        assert "Here is the file" in cleaned
+        assert "for you." in cleaned
+
+    def test_media_tag_supports_unc_paths_with_spaces(self):
+        content = r"MEDIA:\\fileserver\shared reports\Q2\map export.kml"
+        media, cleaned = BasePlatformAdapter.extract_media(content)
+        assert media == [
+            (r"\\fileserver\shared reports\Q2\map export.kml", False)
+        ]
+        assert cleaned == ""
+
+    def test_media_tag_supports_structured_data_extensions_with_spaces(self):
+        content = "MEDIA:/home/user/My Folder/coords.kmz\nMEDIA:/home/user/My Folder/schema.geojson"
+        media, cleaned = BasePlatformAdapter.extract_media(content)
+        assert media == [
+            ("/home/user/My Folder/coords.kmz", False),
+            ("/home/user/My Folder/schema.geojson", False),
+        ]
+        assert cleaned == ""
+
     def test_as_document_directive_stripped_from_cleaned_text(self):
         """[[as_document]] is a routing directive — strip it from
         user-visible text just like [[audio_as_voice]]. Callers detect the
@@ -728,4 +756,3 @@ def test_http_proxy_falls_back_without_aiohttp_socks(self):
             sess_kw, req_kw = proxy_kwargs_for_aiohttp("http://proxy:8080")
             assert sess_kw == {}
             assert req_kw == {"proxy": "http://proxy:8080"}
-

tests/gateway/test_stream_consumer.py

@@ -76,6 +76,24 @@ def test_media_mid_sentence(self):
         assert "generated" in result
         assert "for you." in result
 
+    def test_windows_spaced_media_path_stripped_without_leaking_tail(self):
+        """Windows MEDIA paths with spaces should be stripped as one token."""
+        text = r"Generated MEDIA:C:\Users\Confera\OneDrive\Nusa Alam Kreasindo\Project\Foo\report.pdf for you."
+        result = GatewayStreamConsumer._clean_for_display(text)
+        assert "MEDIA:" not in result
+        assert "Alam Kreasindo" not in result
+        assert "Generated" in result
+        assert "for you." in result
+
+    def test_structured_data_media_path_with_spaces_stripped(self):
+        """GIS / structured-data MEDIA extensions should strip like images."""
+        text = "Map ready\nMEDIA:/home/user/My Folder/coords.kmz\nDone"
+        result = GatewayStreamConsumer._clean_for_display(text)
+        assert "MEDIA:" not in result
+        assert "My Folder" not in result
+        assert "Map ready" in result
+        assert "Done" in result
+
     def test_preserves_non_media_colons(self):
         """Normal colons and text with 'MEDIA' as a word aren't stripped."""
         text = "The media: files are stored in /tmp. Use social MEDIA carefully."
@@ -1780,4 +1798,3 @@ def test_codepoint_only_adapter_falls_back_to_len(self):
         # auto-attr mock. Verified indirectly by all the other tests in
         # this file passing — they all use MagicMock adapters.
         assert consumer is not None
-

tests/test_mcp_serve.py

@@ -343,6 +343,25 @@ def test_media_tag_in_text(self):
         assert len(att) == 1
         assert att[0] == {"type": "media", "path": "/tmp/out.png"}
 
+    def test_windows_spaced_media_tag_in_text(self):
+        from mcp_serve import _extract_attachments
+        msg = {
+            "content": r"Here MEDIA: C:\Users\Confera\OneDrive\Nusa Alam Kreasindo\Project\Foo\report.pdf done"
+        }
+        att = _extract_attachments(msg)
+        assert len(att) == 1
+        assert att[0] == {
+            "type": "media",
+            "path": r"C:\Users\Confera\OneDrive\Nusa Alam Kreasindo\Project\Foo\report.pdf",
+        }
+
+    def test_structured_data_media_tag_in_text(self):
+        from mcp_serve import _extract_attachments
+        msg = {"content": "Here MEDIA: /home/user/My Folder/coords.kmz done"}
+        att = _extract_attachments(msg)
+        assert len(att) == 1
+        assert att[0] == {"type": "media", "path": "/home/user/My Folder/coords.kmz"}
+
     def test_multiple_media_tags(self):
         from mcp_serve import _extract_attachments
         msg = {"content": "MEDIA: /a.png and MEDIA: /b.mp3"}

ui-tui/src/__tests__/markdown.test.ts

@@ -170,6 +170,12 @@ describe('protocol sentinels', () => {
     )
     expect('`MEDIA:/tmp/a.png`'.match(MEDIA_LINE_RE)?.[1]).toBe('/tmp/a.png')
     expect('"MEDIA:C:\\files\\a.png"'.match(MEDIA_LINE_RE)?.[1]).toBe('C:\\files\\a.png')
+    expect('MEDIA:C:\\Users\\Confera\\OneDrive\\Nusa Alam Kreasindo\\Project\\Foo\\report.pdf'.match(MEDIA_LINE_RE)?.[1]).toBe(
+      'C:\\Users\\Confera\\OneDrive\\Nusa Alam Kreasindo\\Project\\Foo\\report.pdf'
+    )
+    expect('MEDIA:/home/user/My Folder/coords.kmz'.match(MEDIA_LINE_RE)?.[1]).toBe(
+      '/home/user/My Folder/coords.kmz'
+    )
   })
 
   it('ignores MEDIA: tokens embedded in prose', () => {

ui-tui/src/components/markdown.tsx

@@ -80,7 +80,46 @@ const MATH_BLOCK_OPEN_RE = /^\s*(\$\$|\\\[)(.*)$/
 const MATH_BLOCK_CLOSE_DOLLAR_RE = /^(.*?)\$\$\s*$/
 const MATH_BLOCK_CLOSE_BRACKET_RE = /^(.*?)\\\]\s*$/
 
-export const MEDIA_LINE_RE = /^\s*[`"']?MEDIA:\s*(\S+?)[`"']?\s*$/
+// Keep in sync with BasePlatformAdapter._MEDIA_EXTENSIONS_RE in
+// gateway/platforms/base.py. TypeScript cannot import the Python constant.
+export const MEDIA_EXTENSIONS = [
+  'png',
+  'jpe?g',
+  'gif',
+  'webp',
+  'mp4',
+  'mov',
+  'avi',
+  'mkv',
+  'webm',
+  'ogg',
+  'opus',
+  'mp3',
+  'wav',
+  'm4a',
+  'flac',
+  'epub',
+  'pdf',
+  'zip',
+  'rar',
+  '7z',
+  'docx?',
+  'xlsx?',
+  'pptx?',
+  'txt',
+  'csv',
+  'apk',
+  'ipa',
+  'kmz',
+  'kml',
+  'json',
+  'xml',
+  'html?',
+  'geojson',
+  'gpx',
+]
+const MEDIA_EXTENSIONS_RE = MEDIA_EXTENSIONS.join('|')
+export const MEDIA_LINE_RE = new RegExp(String.raw`^\s*[\`"']?MEDIA:\s*(?=\S)(.+?\.(?:${MEDIA_EXTENSIONS_RE})(?=[\s\`"',;:)\]}]|$)|\S+?)[\`"']?\s*$`)
 export const AUDIO_DIRECTIVE_RE = /^\s*\[\[audio_as_voice\]\]\s*$/
 
 // Inline markdown tokens, in priority order. The outer regex picks the