Luke/CVE python3.12 update (#4027)

CHANGELOG.md

@@ -1,3 +1,12 @@
+## 0.18.0
+
+### Enhancements
+
+### Features
+- Upgraded Python version to 3.12
+
+### Fixes
+
 ## 0.17.11-dev3
 
 ### Enhancements

Dockerfile

@@ -1,6 +1,6 @@
 FROM quay.io/unstructured-io/base-images:wolfi-base-latest AS base
 
-ARG PYTHON=python3.11
+ARG PYTHON=python3.12
 ARG PIP="${PYTHON} -m pip"
 
 USER root

requirements/base.txt

@@ -10,7 +10,7 @@ backoff==2.2.1
     # via -r ./base.in
 beautifulsoup4==4.13.4
     # via -r ./base.in
-certifi==2025.4.26
+certifi==2025.6.15
     # via
     #   httpcore
     #   httpx
@@ -141,11 +141,11 @@ typing-inspect==0.9.0
     #   unstructured-client
 unstructured-client==0.25.9
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/deps/constraints.txt
     #   -r ./base.in
 urllib3==2.4.0
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/deps/constraints.txt
     #   requests
     #   unstructured-client
 webencodings==0.5.1

requirements/dev.txt

@@ -10,8 +10,8 @@ cfgv==3.4.0
     # via pre-commit
 click==8.2.1
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
-    #   -c ./test.txt
+    #   -c requirements/test.txt
     #   pip-tools
 distlib==0.3.9
     # via virtualenv
@@ -23,14 +23,14 @@ nodeenv==1.9.1
     # via pre-commit
 packaging==25.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
-    #   -c ./test.txt
+    #   -c requirements/test.txt
     #   build
 pip-tools==7.4.1
     # via -r ./dev.in
 platformdirs==4.3.8
     # via
-    #   -c ./test.txt
+    #   -c requirements/test.txt
     #   virtualenv
 pre-commit==4.2.0
     # via -r ./dev.in
@@ -42,7 +42,7 @@ pyyaml==6.0.2
     # via pre-commit
 tomli==2.2.1
     # via
-    #   -c ./test.txt
+    #   -c requirements/test.txt
     #   build
     #   pip-tools
 virtualenv==20.31.2

requirements/extra-csv.txt

@@ -6,19 +6,19 @@
 #
 numpy==2.2.6
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   pandas
 pandas==2.3.0
     # via -r ./extra-csv.in
 python-dateutil==2.9.0.post0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   pandas
 pytz==2025.2
     # via pandas
 six==1.17.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   python-dateutil
 tzdata==2025.2
     # via pandas

requirements/extra-docx.txt

@@ -6,11 +6,11 @@
 #
 lxml==5.4.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   python-docx
-python-docx==1.1.2
+python-docx==1.2.0
     # via -r ./extra-docx.in
 typing-extensions==4.14.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   python-docx

requirements/extra-odt.txt

@@ -6,13 +6,13 @@
 #
 lxml==5.4.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   python-docx
 pypandoc==1.15
     # via -r ./extra-odt.in
-python-docx==1.1.2
+python-docx==1.2.0
     # via -r ./extra-odt.in
 typing-extensions==4.14.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   python-docx

requirements/extra-paddleocr.txt

@@ -14,23 +14,23 @@ annotated-types==0.7.0
     # via pydantic
 anyio==4.9.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   httpx
 astor==0.8.1
     # via paddlepaddle
 beautifulsoup4==4.13.4
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   unstructured-paddleocr
-certifi==2025.4.26
+certifi==2025.6.15
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   httpcore
     #   httpx
     #   requests
 charset-normalizer==3.4.2
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   requests
 cython==3.1.2
     # via unstructured-paddleocr
@@ -38,27 +38,27 @@ decorator==5.2.1
     # via paddlepaddle
 exceptiongroup==1.3.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   anyio
 fire==0.7.0
     # via unstructured-paddleocr
-fonttools==4.58.2
+fonttools==4.58.4
     # via unstructured-paddleocr
 h11==0.16.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   httpcore
 httpcore==1.0.9
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   httpx
 httpx==0.28.1
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   paddlepaddle
 idna==3.10
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   anyio
     #   httpx
     #   requests
@@ -68,7 +68,7 @@ lazy-loader==0.4
     # via scikit-image
 lxml==5.4.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   python-docx
 networkx==3.4.2
     # via
@@ -76,7 +76,7 @@ networkx==3.4.2
     #   scikit-image
 numpy==2.2.6
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   albucore
     #   albumentations
     #   imageio
@@ -102,7 +102,7 @@ opt-einsum==3.3.0
     # via paddlepaddle
 packaging==25.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   lazy-loader
     #   scikit-image
 paddlepaddle==3.0.0
@@ -115,15 +115,15 @@ pillow==11.2.1
     #   unstructured-paddleocr
 protobuf==6.31.1
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/deps/constraints.txt
     #   paddlepaddle
 pyclipper==1.3.0.post6
     # via unstructured-paddleocr
-pydantic==2.11.5
+pydantic==2.11.7
     # via albumentations
 pydantic-core==2.33.2
     # via pydantic
-python-docx==1.1.2
+python-docx==1.2.0
     # via unstructured-paddleocr
 pyyaml==6.0.2
     # via
@@ -131,11 +131,11 @@ pyyaml==6.0.2
     #   unstructured-paddleocr
 rapidfuzz==3.13.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   unstructured-paddleocr
 requests==2.32.4
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   unstructured-paddleocr
 scikit-image==0.25.2
     # via unstructured-paddleocr
@@ -149,11 +149,11 @@ simsimd==6.4.9
     # via albucore
 sniffio==1.3.1
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   anyio
 soupsieve==2.7
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   beautifulsoup4
 stringzilla==3.12.5
     # via albucore
@@ -163,11 +163,11 @@ tifffile==2025.5.10
     # via scikit-image
 tqdm==4.67.1
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   unstructured-paddleocr
 typing-extensions==4.14.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   anyio
     #   beautifulsoup4
     #   exceptiongroup
@@ -182,6 +182,6 @@ unstructured-paddleocr==2.10.0
     # via -r ./extra-paddleocr.in
 urllib3==2.4.0
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/base.txt
-    #   -c ./base.txt
+    #   -c requirements/deps/constraints.txt
     #   requests

requirements/extra-pdf-image.txt

@@ -10,17 +10,17 @@ antlr4-python3-runtime==4.9.3
     # via omegaconf
 cachetools==5.5.2
     # via google-auth
-certifi==2025.4.26
+certifi==2025.6.15
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   requests
 cffi==1.17.1
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   cryptography
 charset-normalizer==3.4.2
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   pdfminer-six
     #   requests
 coloredlogs==15.0.1
@@ -29,7 +29,7 @@ contourpy==1.3.2
     # via matplotlib
 cryptography==45.0.4
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   pdfminer-six
 cycler==0.12.1
     # via matplotlib
@@ -44,13 +44,13 @@ filelock==3.18.0
     #   transformers
 flatbuffers==25.2.10
     # via onnxruntime
-fonttools==4.58.2
+fonttools==4.58.4
     # via matplotlib
 fsspec==2025.5.1
     # via
     #   huggingface-hub
     #   torch
-google-api-core[grpc]==2.25.0
+google-api-core[grpc]==2.25.1
     # via google-cloud-vision
 google-auth==2.40.3
     # via
@@ -64,12 +64,12 @@ googleapis-common-protos==1.70.0
     #   grpcio-status
 grpcio==1.73.0
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/deps/constraints.txt
     #   google-api-core
     #   grpcio-status
 grpcio-status==1.73.0
     # via google-api-core
-hf-xet==1.1.3
+hf-xet==1.1.4
     # via huggingface-hub
 huggingface-hub==0.33.0
     # via
@@ -82,7 +82,7 @@ humanfriendly==10.0
     # via coloredlogs
 idna==3.10
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   requests
 jinja2==3.1.6
     # via torch
@@ -90,7 +90,7 @@ kiwisolver==1.4.8
     # via matplotlib
 lxml==5.4.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   pikepdf
 markupsafe==3.0.2
     # via jinja2
@@ -102,7 +102,7 @@ networkx==3.4.2
     # via torch
 numpy==2.2.6
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   accelerate
     #   contourpy
     #   matplotlib
@@ -129,7 +129,7 @@ opencv-python==4.11.0.86
     # via unstructured-inference
 packaging==25.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   accelerate
     #   huggingface-hub
     #   matplotlib
@@ -143,7 +143,7 @@ pdf2image==1.17.0
     # via -r ./extra-pdf-image.in
 pdfminer-six==20250327
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/deps/constraints.txt
     #   -r ./extra-pdf-image.in
     #   unstructured-inference
 pi-heif==0.22.0
@@ -164,7 +164,7 @@ proto-plus==1.26.1
     #   google-cloud-vision
 protobuf==6.31.1
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/deps/constraints.txt
     #   google-api-core
     #   google-cloud-vision
     #   googleapis-common-protos
@@ -174,7 +174,7 @@ protobuf==6.31.1
     #   proto-plus
 psutil==7.0.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   accelerate
 pyasn1==0.6.1
     # via
@@ -186,19 +186,19 @@ pycocotools==2.0.10
     # via effdet
 pycparser==2.22
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   cffi
 pyparsing==3.2.3
     # via matplotlib
 pypdf==5.6.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   -r ./extra-pdf-image.in
 pypdfium2==4.30.1
     # via unstructured-inference
 python-dateutil==2.9.0.post0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   matplotlib
     #   pandas
 python-multipart==0.0.20
@@ -214,15 +214,15 @@ pyyaml==6.0.2
     #   transformers
 rapidfuzz==3.13.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   unstructured-inference
 regex==2024.11.6
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   transformers
 requests==2.32.4
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   google-api-core
     #   huggingface-hub
     #   transformers
@@ -237,7 +237,7 @@ scipy==1.15.3
     # via unstructured-inference
 six==1.17.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   python-dateutil
 sympy==1.14.0
     # via
@@ -249,7 +249,7 @@ timm==1.0.15
     #   unstructured-inference
 tokenizers==0.21.1
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/deps/constraints.txt
     #   transformers
 torch==2.7.1
     # via
@@ -264,14 +264,14 @@ torchvision==0.22.1
     #   timm
 tqdm==4.67.1
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   huggingface-hub
     #   transformers
 transformers==4.52.4
     # via unstructured-inference
 typing-extensions==4.14.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   huggingface-hub
     #   onnx
     #   pypdf
@@ -284,10 +284,10 @@ unstructured-pytesseract==0.3.15
     # via -r ./extra-pdf-image.in
 urllib3==2.4.0
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/base.txt
-    #   -c ./base.txt
+    #   -c requirements/deps/constraints.txt
     #   requests
 wrapt==1.17.2
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   deprecated

requirements/extra-xlsx.txt

@@ -10,7 +10,7 @@ networkx==3.4.2
     # via -r ./extra-xlsx.in
 numpy==2.2.6
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   pandas
 openpyxl==3.1.5
     # via -r ./extra-xlsx.in
@@ -18,15 +18,15 @@ pandas==2.3.0
     # via -r ./extra-xlsx.in
 python-dateutil==2.9.0.post0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   pandas
 pytz==2025.2
     # via pandas
 six==1.17.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   python-dateutil
 tzdata==2025.2
     # via pandas
-xlrd==2.0.1
+xlrd==2.0.2
     # via -r ./extra-xlsx.in

requirements/huggingface.txt

@@ -4,17 +4,17 @@
 #
 #    pip-compile ./huggingface.in
 #
-certifi==2025.4.26
+certifi==2025.6.15
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   requests
 charset-normalizer==3.4.2
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   requests
 click==8.2.1
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   sacremoses
 filelock==3.18.0
     # via
@@ -25,7 +25,7 @@ fsspec==2025.5.1
     # via
     #   huggingface-hub
     #   torch
-hf-xet==1.1.3
+hf-xet==1.1.4
     # via huggingface-hub
 huggingface-hub==0.33.0
     # via
@@ -33,17 +33,17 @@ huggingface-hub==0.33.0
     #   transformers
 idna==3.10
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   requests
 jinja2==3.1.6
     # via torch
 joblib==1.5.1
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   sacremoses
 langdetect==1.0.9
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   -r ./huggingface.in
 markupsafe==3.0.2
     # via jinja2
@@ -53,11 +53,11 @@ networkx==3.4.2
     # via torch
 numpy==2.2.6
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   transformers
 packaging==25.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   huggingface-hub
     #   transformers
 pyyaml==6.0.2
@@ -66,12 +66,12 @@ pyyaml==6.0.2
     #   transformers
 regex==2024.11.6
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   sacremoses
     #   transformers
 requests==2.32.4
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   huggingface-hub
     #   transformers
 sacremoses==0.1.1
@@ -82,19 +82,19 @@ sentencepiece==0.2.0
     # via -r ./huggingface.in
 six==1.17.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   langdetect
 sympy==1.14.0
     # via torch
 tokenizers==0.21.1
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/deps/constraints.txt
     #   transformers
 torch==2.7.1
     # via -r ./huggingface.in
 tqdm==4.67.1
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   huggingface-hub
     #   sacremoses
     #   transformers
@@ -102,11 +102,11 @@ transformers==4.52.4
     # via -r ./huggingface.in
 typing-extensions==4.14.0
     # via
-    #   -c ./base.txt
+    #   -c requirements/base.txt
     #   huggingface-hub
     #   torch
 urllib3==2.4.0
     # via
-    #   -c ././deps/constraints.txt
+    #   -c requirements/base.txt
-    #   -c ./base.txt
+    #   -c requirements/deps/constraints.txt
     #   requests

requirements/test.txt

@@ -2,21 +2,21 @@
 # This file is autogenerated by pip-compile with Python 3.10
 # by the following command:
 #
-#    pip-compile requirements/test.in
+#    pip-compile ./test.in
 #
 annotated-types==0.7.0
     # via pydantic
 autoflake==2.3.1
-    # via -r requirements/test.in
+    # via -r ./test.in
 black==25.1.0
-    # via -r requirements/test.in
+    # via -r ./test.in
 click==8.2.1
     # via
     #   -c requirements/base.txt
     #   black
-coverage[toml]==7.9.0
+coverage[toml]==7.9.1
     # via
-    #   -r requirements/test.in
+    #   -r ./test.in
     #   pytest-cov
 exceptiongroup==1.3.0
     # via
@@ -26,24 +26,24 @@ execnet==2.1.1
     # via pytest-xdist
 flake8==7.2.0
     # via
-    #   -r requirements/test.in
+    #   -r ./test.in
     #   flake8-print
 flake8-print==5.0.0
-    # via -r requirements/test.in
+    # via -r ./test.in
 freezegun==1.5.2
-    # via -r requirements/test.in
+    # via -r ./test.in
 grpcio==1.73.0
     # via
-    #   -c requirements/./deps/constraints.txt
+    #   -c requirements/deps/constraints.txt
-    #   -r requirements/test.in
+    #   -r ./test.in
 iniconfig==2.1.0
     # via pytest
 liccheck==0.9.2
-    # via -r requirements/test.in
+    # via -r ./test.in
 mccabe==0.7.0
     # via flake8
-mypy==1.16.0
+mypy==1.16.1
-    # via -r requirements/test.in
+    # via -r ./test.in
 mypy-extensions==1.1.0
     # via
     #   -c requirements/base.txt
@@ -68,8 +68,8 @@ pycodestyle==2.13.0
     # via
     #   flake8
     #   flake8-print
-pydantic==2.11.5
+pydantic==2.11.7
-    # via -r requirements/test.in
+    # via -r ./test.in
 pydantic-core==2.33.2
     # via pydantic
 pyflakes==3.3.2
@@ -84,17 +84,17 @@ pytest==8.4.0
     #   pytest-mock
     #   pytest-xdist
 pytest-cov==6.2.1
-    # via -r requirements/test.in
+    # via -r ./test.in
 pytest-mock==3.14.1
-    # via -r requirements/test.in
+    # via -r ./test.in
 pytest-xdist==3.7.0
-    # via -r requirements/test.in
+    # via -r ./test.in
 python-dateutil==2.9.0.post0
     # via
     #   -c requirements/base.txt
     #   freezegun
 ruff==0.11.13
-    # via -r requirements/test.in
+    # via -r ./test.in
 semantic-version==2.10.0
     # via liccheck
 six==1.17.0
@@ -111,13 +111,13 @@ tomli==2.2.1
     #   mypy
     #   pytest
 types-click==7.1.8
-    # via -r requirements/test.in
+    # via -r ./test.in
 types-markdown==3.8.0.20250415
-    # via -r requirements/test.in
+    # via -r ./test.in
 types-requests==2.32.4.20250611
-    # via -r requirements/test.in
+    # via -r ./test.in
 types-tabulate==0.9.0.20241207
-    # via -r requirements/test.in
+    # via -r ./test.in
 typing-extensions==4.14.0
     # via
     #   -c requirements/base.txt
@@ -131,6 +131,6 @@ typing-inspection==0.4.1
     # via pydantic
 urllib3==2.4.0
     # via
-    #   -c requirements/./deps/constraints.txt
     #   -c requirements/base.txt
+    #   -c requirements/deps/constraints.txt
     #   types-requests

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.17.11-dev3"  # pragma: no cover
+__version__ = "0.18.0"  # pragma: no cover