fix(CVE-2024-39705): bump to nltk 3.9.1; correct model download issues

CHANGELOG.md

@@ -1,11 +1,12 @@
-## 0.15.6-dev0
+## 0.15.6
 
 ### Enhancements
 
 ### Features
 
 ### Fixes
 
+* **Bump to NLTK 3.9.x** Bumps to the latest `nltk` version to resolve CVE.
 * **Update CI for `ingest-test-fixture-update-pr` to resolve NLTK model download errors.**
 
 

requirements/base.txt

@@ -69,7 +69,7 @@ mypy-extensions==1.0.0
     #   unstructured-client
 nest-asyncio==1.6.0
     # via unstructured-client
-nltk==3.8.1
+nltk==3.9.1
     # via -r ./base.in
 numpy==1.26.4
     # via -r ./base.in
@@ -110,7 +110,7 @@ sniffio==1.3.1
     # via
     #   anyio
     #   httpx
-soupsieve==2.5
+soupsieve==2.6
     # via beautifulsoup4
 tabulate==0.9.0
     # via -r ./base.in
@@ -129,7 +129,7 @@ typing-inspect==0.9.0
     # via
     #   dataclasses-json
     #   unstructured-client
-unstructured-client==0.25.4
+unstructured-client==0.25.5
     # via
     #   -c ././deps/constraints.txt
     #   -r ./base.in

requirements/deps/constraints.txt

@@ -56,3 +56,6 @@ fsspec==2024.5.0
 wrapt>=1.14.0
 
 langchain-community>=0.2.5
+
+grpcio==1.64.3
+label-studio-sdk==0.0.34

requirements/dev.txt

@@ -310,7 +310,7 @@ pyyaml==6.0.2
     #   -c ./test.txt
     #   jupyter-events
     #   pre-commit
-pyzmq==26.1.0
+pyzmq==26.1.1
     # via
     #   ipykernel
     #   jupyter-client
@@ -360,7 +360,7 @@ sniffio==1.3.1
     #   -c ./base.txt
     #   anyio
     #   httpx
-soupsieve==2.5
+soupsieve==2.6
     # via
     #   -c ./base.txt
     #   beautifulsoup4

requirements/extra-markdown.txt

@@ -6,7 +6,7 @@
 #
 importlib-metadata==8.2.0
     # via markdown
-markdown==3.6
+markdown==3.7
     # via -r ./extra-markdown.in
 zipp==3.20.0
     # via importlib-metadata

requirements/extra-paddleocr.txt

@@ -13,7 +13,7 @@ astor==0.8.1
     # via paddlepaddle
 attrdict==2.0.1
     # via unstructured-paddleocr
-cachetools==5.4.0
+cachetools==5.5.0
     # via premailer
 certifi==2024.7.4
     # via
@@ -64,13 +64,13 @@ idna==3.7
     #   anyio
     #   httpx
     #   requests
-imageio==2.34.2
+imageio==2.35.1
     # via
     #   imgaug
     #   scikit-image
 imgaug==0.4.0
     # via unstructured-paddleocr
-importlib-resources==6.4.0
+importlib-resources==6.4.3
     # via matplotlib
 kiwisolver==1.4.5
     # via matplotlib
@@ -83,7 +83,7 @@ lxml==5.3.0
     #   -c ./base.txt
     #   premailer
     #   unstructured-paddleocr
-matplotlib==3.9.1.post1
+matplotlib==3.9.2
     # via imgaug
 more-itertools==10.4.0
     # via cssutils

requirements/extra-pdf-image.txt

@@ -6,7 +6,7 @@
 #
 antlr4-python3-runtime==4.9.3
     # via omegaconf
-cachetools==5.4.0
+cachetools==5.5.0
     # via google-auth
 certifi==2024.7.4
     # via
@@ -48,7 +48,7 @@ fsspec==2024.5.0
     #   torch
 google-api-core[grpc]==2.19.1
     # via google-cloud-vision
-google-auth==2.33.0
+google-auth==2.34.0
     # via
     #   google-api-core
     #   google-cloud-vision
@@ -58,13 +58,14 @@ googleapis-common-protos==1.63.2
     # via
     #   google-api-core
     #   grpcio-status
-grpcio==1.65.4
+grpcio==1.64.3
     # via
+    #   -c ././deps/constraints.txt
     #   google-api-core
     #   grpcio-status
 grpcio-status==1.62.3
     # via google-api-core
-huggingface-hub==0.24.5
+huggingface-hub==0.24.6
     # via
     #   timm
     #   tokenizers
@@ -76,7 +77,7 @@ idna==3.7
     # via
     #   -c ./base.txt
     #   requests
-importlib-resources==6.4.0
+importlib-resources==6.4.3
     # via matplotlib
 iopath==0.1.10
     # via layoutparser
@@ -92,7 +93,7 @@ lxml==5.3.0
     #   pikepdf
 markupsafe==2.1.5
     # via jinja2
-matplotlib==3.9.1.post1
+matplotlib==3.9.2
     # via
     #   pycocotools
     #   unstructured-inference
@@ -120,7 +121,7 @@ onnx==1.16.2
     # via
     #   -r ./extra-pdf-image.in
     #   unstructured-inference
-onnxruntime==1.18.1
+onnxruntime==1.19.0
     # via unstructured-inference
 opencv-python==4.8.0.76
     # via
@@ -147,7 +148,7 @@ pdfminer-six==20231228
     # via
     #   -r ./extra-pdf-image.in
     #   pdfplumber
-pdfplumber==0.11.3
+pdfplumber==0.11.4
     # via layoutparser
 pikepdf==9.1.1
     # via -r ./extra-pdf-image.in

requirements/huggingface.txt

@@ -27,7 +27,7 @@ fsspec==2024.5.0
     #   -c ././deps/constraints.txt
     #   huggingface-hub
     #   torch
-huggingface-hub==0.24.5
+huggingface-hub==0.24.6
     # via
     #   tokenizers
     #   transformers

requirements/ingest/azure.txt

@@ -6,9 +6,9 @@
 #
 adlfs==2024.7.0
     # via -r ./ingest/azure.in
-aiohappyeyeballs==2.3.5
+aiohappyeyeballs==2.3.7
     # via aiohttp
-aiohttp==3.10.3
+aiohttp==3.10.4
     # via adlfs
 aiosignal==1.3.1
     # via aiohttp

requirements/ingest/biomed.txt

@@ -10,7 +10,7 @@ beautifulsoup4==4.12.3
     #   bs4
 bs4==0.0.2
     # via -r ./ingest/biomed.in
-soupsieve==2.5
+soupsieve==2.6
     # via
     #   -c ./ingest/../base.txt
     #   beautifulsoup4

requirements/ingest/chroma.txt

@@ -20,7 +20,7 @@ backoff==2.2.1
     #   posthog
 bcrypt==4.2.0
     # via chromadb
-cachetools==5.4.0
+cachetools==5.5.0
     # via google-auth
 certifi==2024.7.4
     # via
@@ -51,7 +51,7 @@ exceptiongroup==1.2.2
     # via
     #   -c ./ingest/../base.txt
     #   anyio
-fastapi==0.112.0
+fastapi==0.112.1
     # via chromadb
 filelock==3.15.4
     # via huggingface-hub
@@ -61,12 +61,13 @@ fsspec==2024.5.0
     # via
     #   -c ./ingest/../deps/constraints.txt
     #   huggingface-hub
-google-auth==2.33.0
+google-auth==2.34.0
     # via kubernetes
 googleapis-common-protos==1.63.2
     # via opentelemetry-exporter-otlp-proto-grpc
-grpcio==1.65.4
+grpcio==1.64.3
     # via
+    #   -c ./ingest/../deps/constraints.txt
     #   chromadb
     #   opentelemetry-exporter-otlp-proto-grpc
 h11==0.14.0
@@ -76,7 +77,7 @@ h11==0.14.0
     #   uvicorn
 httptools==0.6.1
     # via uvicorn
-huggingface-hub==0.24.5
+huggingface-hub==0.24.6
     # via tokenizers
 humanfriendly==10.0
     # via coloredlogs
@@ -88,7 +89,7 @@ idna==3.7
     #   requests
 importlib-metadata==8.2.0
     # via -r ./ingest/chroma.in
-importlib-resources==6.4.0
+importlib-resources==6.4.3
     # via chromadb
 kubernetes==30.1.0
     # via chromadb
@@ -106,7 +107,7 @@ oauthlib==3.2.2
     # via
     #   kubernetes
     #   requests-oauthlib
-onnxruntime==1.18.1
+onnxruntime==1.19.0
     # via chromadb
 opentelemetry-api==1.16.0
     # via
@@ -192,7 +193,7 @@ sniffio==1.3.1
     #   -c ./ingest/../base.txt
     #   anyio
     #   httpx
-starlette==0.37.2
+starlette==0.38.2
     # via fastapi
 sympy==1.13.2
     # via onnxruntime
@@ -231,9 +232,9 @@ urllib3==1.26.19
     #   -c ./ingest/../deps/constraints.txt
     #   kubernetes
     #   requests
-uvicorn[standard]==0.30.5
+uvicorn[standard]==0.30.6
     # via chromadb
-uvloop==0.19.0
+uvloop==0.20.0
     # via uvicorn
 watchfiles==0.23.0
     # via uvicorn

requirements/ingest/clarifai.txt

@@ -15,14 +15,16 @@ charset-normalizer==3.3.2
     #   requests
 clarifai==10.7.0
     # via -r ./ingest/clarifai.in
-clarifai-grpc==10.7.1
+clarifai-grpc==10.7.2
     # via clarifai
 contextlib2==21.6.0
     # via schema
 googleapis-common-protos==1.63.2
     # via clarifai-grpc
-grpcio==1.65.4
-    # via clarifai-grpc
+grpcio==1.64.3
+    # via
+    #   -c ./ingest/../deps/constraints.txt
+    #   clarifai-grpc
 idna==3.7
     # via
     #   -c ./ingest/../base.txt

requirements/ingest/confluence.txt

@@ -42,7 +42,7 @@ six==1.16.0
     # via
     #   -c ./ingest/../base.txt
     #   atlassian-python-api
-soupsieve==2.5
+soupsieve==2.6
     # via
     #   -c ./ingest/../base.txt
     #   beautifulsoup4

requirements/ingest/databricks-volumes.txt

@@ -4,7 +4,7 @@
 #
 #    pip-compile ./ingest/databricks-volumes.in
 #
-cachetools==5.4.0
+cachetools==5.5.0
     # via google-auth
 certifi==2024.7.4
     # via
@@ -15,9 +15,9 @@ charset-normalizer==3.3.2
     # via
     #   -c ./ingest/../base.txt
     #   requests
-databricks-sdk==0.29.0
+databricks-sdk==0.30.0
     # via -r ./ingest/databricks-volumes.in
-google-auth==2.33.0
+google-auth==2.34.0
     # via databricks-sdk
 idna==3.7
     # via

requirements/ingest/delta-table.txt

@@ -4,7 +4,7 @@
 #
 #    pip-compile ./ingest/delta-table.in
 #
-deltalake==0.18.2
+deltalake==0.19.0
     # via -r ./ingest/delta-table.in
 fsspec==2024.5.0
     # via
@@ -16,5 +16,3 @@ numpy==1.26.4
     #   pyarrow
 pyarrow==17.0.0
     # via deltalake
-pyarrow-hotfix==0.6
-    # via deltalake

requirements/ingest/discord.txt

@@ -4,9 +4,9 @@
 #
 #    pip-compile ./ingest/discord.in
 #
-aiohappyeyeballs==2.3.5
+aiohappyeyeballs==2.3.7
     # via aiohttp
-aiohttp==3.10.3
+aiohttp==3.10.4
     # via discord-py
 aiosignal==1.3.1
     # via aiohttp

requirements/ingest/elasticsearch.txt

@@ -4,9 +4,9 @@
 #
 #    pip-compile ./ingest/elasticsearch.in
 #
-aiohappyeyeballs==2.3.5
+aiohappyeyeballs==2.3.7
     # via aiohttp
-aiohttp==3.10.3
+aiohttp==3.10.4
     # via elasticsearch
 aiosignal==1.3.1
     # via aiohttp
@@ -21,7 +21,7 @@ certifi==2024.7.4
     #   elastic-transport
 elastic-transport==8.15.0
     # via elasticsearch
-elasticsearch[async]==8.14.0
+elasticsearch[async]==8.15.0
     # via -r ./ingest/elasticsearch.in
 frozenlist==1.4.1
     # via