diff --git a/CHANGELOG.md b/CHANGELOG.md index 43cae6c9..f881ec47 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,5 +1,6 @@ -## 0.0.61-dev0 +## 0.0.61 +* Bump unstructured to 0.11.6 * Handle invalid hi_res_model_name kwarg ## 0.0.60 diff --git a/requirements/base.txt b/requirements/base.txt index 2b67a6c9..4eef8870 100644 --- a/requirements/base.txt +++ b/requirements/base.txt @@ -17,7 +17,9 @@ backoff==2.2.1 beautifulsoup4==4.12.2 # via unstructured certifi==2023.11.17 - # via requests + # via + # requests + # unstructured-client cffi==1.16.0 # via cryptography chardet==5.2.0 @@ -26,6 +28,7 @@ charset-normalizer==3.3.2 # via # pdfminer-six # requests + # unstructured-client click==8.1.3 # via # -r requirements/base.in @@ -35,21 +38,25 @@ coloredlogs==15.0.1 # via onnxruntime contourpy==1.2.0 # via matplotlib -cryptography==41.0.5 +cryptography==41.0.7 # via pdfminer-six cycler==0.12.1 # via matplotlib -dataclasses-json==0.6.2 - # via unstructured +dataclasses-json==0.6.3 + # via + # unstructured + # unstructured-client +deprecated==1.2.14 + # via pikepdf effdet==0.4.1 # via layoutparser -emoji==2.8.0 +emoji==2.9.0 # via unstructured et-xmlfile==1.1.0 # via openpyxl -exceptiongroup==1.1.3 +exceptiongroup==1.2.0 # via anyio -fastapi==0.104.1 +fastapi==0.105.0 # via -r requirements/base.in filelock==3.13.1 # via @@ -60,15 +67,15 @@ filetype==1.2.0 # via unstructured flatbuffers==23.5.26 # via onnxruntime -fonttools==4.45.0 +fonttools==4.47.0 # via matplotlib -fsspec==2023.10.0 +fsspec==2023.12.2 # via # huggingface-hub # torch h11==0.14.0 # via uvicorn -huggingface-hub==0.19.4 +huggingface-hub==0.20.1 # via # timm # tokenizers @@ -76,24 +83,28 @@ huggingface-hub==0.19.4 # unstructured-inference humanfriendly==10.0 # via coloredlogs -idna==3.4 +idna==3.6 # via # anyio # requests + # unstructured-client iopath==0.1.10 # via layoutparser jinja2==3.1.2 # via torch joblib==1.3.2 # via nltk +jsonpath-python==1.0.6 + # via unstructured-client kiwisolver==1.4.5 # via matplotlib langdetect==1.0.9 # via unstructured layoutparser[layoutmodels,tesseract]==0.3.4 # via unstructured-inference -lxml==4.9.3 +lxml==4.9.4 # via + # pikepdf # python-docx # python-pptx # unstructured @@ -102,7 +113,9 @@ markdown==3.5.1 markupsafe==2.1.3 # via jinja2 marshmallow==3.20.1 - # via dataclasses-json + # via + # dataclasses-json + # unstructured-client matplotlib==3.8.2 # via pycocotools mpmath==1.3.0 @@ -110,7 +123,9 @@ mpmath==1.3.0 msg-parser==1.2.0 # via unstructured mypy-extensions==1.0.0 - # via typing-inspect + # via + # typing-inspect + # unstructured-client networkx==3.2.1 # via # torch @@ -131,7 +146,7 @@ numpy==1.26.2 # torchvision # transformers # unstructured -olefile==0.46 +olefile==0.47 # via msg-parser omegaconf==2.3.0 # via effdet @@ -153,10 +168,12 @@ packaging==23.2 # marshmallow # matplotlib # onnxruntime + # pikepdf # pytesseract # transformers + # unstructured-client # unstructured-pytesseract -pandas==2.1.3 +pandas==2.1.4 # via # layoutparser # unstructured @@ -170,12 +187,15 @@ pdfminer-six==20221105 # unstructured pdfplumber==0.10.3 # via layoutparser +pikepdf==8.10.1 + # via unstructured pillow==10.1.0 # via # layoutparser # matplotlib # pdf2image # pdfplumber + # pikepdf # pytesseract # python-pptx # torchvision @@ -186,7 +206,7 @@ protobuf==4.25.1 # via # onnx # onnxruntime -psutil==5.9.6 +psutil==5.9.7 # via -r requirements/base.in pycocotools==2.0.7 # via effdet @@ -202,9 +222,11 @@ pypandoc==1.12 # via unstructured pyparsing==3.1.1 # via matplotlib -pypdf==3.17.1 - # via -r requirements/base.in -pypdfium2==4.24.0 +pypdf==3.17.3 + # via + # -r requirements/base.in + # unstructured +pypdfium2==4.25.0 # via pdfplumber pytesseract==0.3.10 # via layoutparser @@ -212,9 +234,10 @@ python-dateutil==2.8.2 # via # matplotlib # pandas + # unstructured-client python-docx==1.1.0 # via unstructured -python-iso639==2023.6.15 +python-iso639==2023.12.11 # via unstructured python-magic==0.4.27 # via unstructured @@ -248,6 +271,7 @@ requests==2.31.0 # torchvision # transformers # unstructured + # unstructured-client safetensors==0.3.2 # via # -c requirements/constraints.in @@ -259,6 +283,7 @@ six==1.16.0 # via # langdetect # python-dateutil + # unstructured-client sniffio==1.3.0 # via anyio soupsieve==2.5 @@ -271,17 +296,17 @@ sympy==1.12 # torch tabulate==0.9.0 # via unstructured -timm==0.9.10 +timm==0.9.12 # via effdet tokenizers==0.15.0 # via transformers -torch==2.1.1 +torch==2.1.2 # via # effdet # layoutparser # timm # torchvision -torchvision==0.16.1 +torchvision==0.16.2 # via # effdet # layoutparser @@ -292,9 +317,9 @@ tqdm==4.66.1 # iopath # nltk # transformers -transformers==4.35.2 +transformers==4.36.2 # via unstructured-inference -typing-extensions==4.8.0 +typing-extensions==4.9.0 # via # fastapi # huggingface-hub @@ -305,23 +330,32 @@ typing-extensions==4.8.0 # torch # typing-inspect # unstructured + # unstructured-client # uvicorn typing-inspect==0.9.0 - # via dataclasses-json + # via + # dataclasses-json + # unstructured-client tzdata==2023.3 # via pandas -unstructured[local-inference]==0.11.0 +unstructured[local-inference]==0.11.6 # via -r requirements/base.in -unstructured-inference==0.7.15 +unstructured-client==0.15.1 + # via unstructured +unstructured-inference==0.7.18 # via unstructured unstructured-pytesseract==0.3.12 # via unstructured urllib3==2.1.0 - # via requests -uvicorn==0.24.0.post1 + # via + # requests + # unstructured-client +uvicorn==0.25.0 # via -r requirements/base.in wrapt==1.16.0 - # via unstructured + # via + # deprecated + # unstructured xlrd==2.0.1 # via unstructured xlsxwriter==3.1.9 diff --git a/requirements/test.txt b/requirements/test.txt index f94b89ff..8677b6c7 100644 --- a/requirements/test.txt +++ b/requirements/test.txt @@ -16,9 +16,7 @@ anyio==3.7.1 # jupyter-server # starlette appnope==0.1.3 - # via - # ipykernel - # ipython + # via ipykernel argon2-cffi==23.1.0 # via jupyter-server argon2-cffi-bindings==21.2.0 @@ -37,7 +35,7 @@ attrs==23.1.0 # via # jsonschema # referencing -babel==2.13.1 +babel==2.14.0 # via jupyterlab-server backoff==2.2.1 # via @@ -48,7 +46,7 @@ beautifulsoup4==4.12.2 # -r requirements/base.txt # nbconvert # unstructured -black==23.11.0 +black==23.12.1 # via -r requirements/test.in bleach==6.1.0 # via nbconvert @@ -58,6 +56,7 @@ certifi==2023.11.17 # httpcore # httpx # requests + # unstructured-client cffi==1.16.0 # via # -r requirements/base.txt @@ -72,6 +71,7 @@ charset-normalizer==3.3.2 # -r requirements/base.txt # pdfminer-six # requests + # unstructured-client click==8.1.3 # via # -r requirements/base.txt @@ -89,9 +89,9 @@ contourpy==1.2.0 # via # -r requirements/base.txt # matplotlib -coverage[toml]==7.3.2 +coverage[toml]==7.3.4 # via pytest-cov -cryptography==41.0.5 +cryptography==41.0.7 # via # -r requirements/base.txt # pdfminer-six @@ -99,21 +99,26 @@ cycler==0.12.1 # via # -r requirements/base.txt # matplotlib -dataclasses-json==0.6.2 +dataclasses-json==0.6.3 # via # -r requirements/base.txt # unstructured + # unstructured-client debugpy==1.8.0 # via ipykernel decorator==5.1.1 # via ipython defusedxml==0.7.1 # via nbconvert +deprecated==1.2.14 + # via + # -r requirements/base.txt + # pikepdf effdet==0.4.1 # via # -r requirements/base.txt # layoutparser -emoji==2.8.0 +emoji==2.9.0 # via # -r requirements/base.txt # unstructured @@ -121,7 +126,7 @@ et-xmlfile==1.1.0 # via # -r requirements/base.txt # openpyxl -exceptiongroup==1.1.3 +exceptiongroup==1.2.0 # via # -r requirements/base.txt # anyio @@ -131,7 +136,7 @@ execnb==0.1.5 # via nbdev executing==2.0.1 # via stack-data -fastapi==0.104.1 +fastapi==0.105.0 # via -r requirements/base.txt fastcore==1.5.29 # via @@ -156,13 +161,13 @@ flatbuffers==23.5.26 # via # -r requirements/base.txt # onnxruntime -fonttools==4.45.0 +fonttools==4.47.0 # via # -r requirements/base.txt # matplotlib fqdn==1.5.1 # via jsonschema -fsspec==2023.10.0 +fsspec==2023.12.2 # via # -r requirements/base.txt # huggingface-hub @@ -176,9 +181,9 @@ h11==0.14.0 # uvicorn httpcore==1.0.2 # via httpx -httpx==0.25.1 +httpx==0.26.0 # via -r requirements/test.in -huggingface-hub==0.19.4 +huggingface-hub==0.20.1 # via # -r requirements/base.txt # timm @@ -189,27 +194,28 @@ humanfriendly==10.0 # via # -r requirements/base.txt # coloredlogs -idna==3.4 +idna==3.6 # via # -r requirements/base.txt # anyio # httpx # jsonschema # requests + # unstructured-client iniconfig==2.0.0 # via pytest iopath==0.1.10 # via # -r requirements/base.txt # layoutparser -ipykernel==6.26.0 +ipykernel==6.27.1 # via # ipywidgets # jupyter # jupyter-console # jupyterlab # qtconsole -ipython==8.17.2 +ipython==8.19.0 # via # execnb # ipykernel @@ -237,6 +243,10 @@ joblib==1.3.2 # nltk json5==0.9.14 # via jupyterlab-server +jsonpath-python==1.0.6 + # via + # -r requirements/base.txt + # unstructured-client jsonpointer==2.4 # via jsonschema jsonschema[format-nongpl]==4.20.0 @@ -244,7 +254,7 @@ jsonschema[format-nongpl]==4.20.0 # jupyter-events # jupyterlab-server # nbformat -jsonschema-specifications==2023.11.1 +jsonschema-specifications==2023.11.2 # via jsonschema jupyter==1.0.0 # via -r requirements/test.in @@ -257,7 +267,7 @@ jupyter-client==8.6.0 # qtconsole jupyter-console==6.6.3 # via jupyter -jupyter-core==5.5.0 +jupyter-core==5.5.1 # via # ipykernel # jupyter-client @@ -270,20 +280,20 @@ jupyter-core==5.5.0 # qtconsole jupyter-events==0.9.0 # via jupyter-server -jupyter-lsp==2.2.0 +jupyter-lsp==2.2.1 # via jupyterlab -jupyter-server==2.10.1 +jupyter-server==2.12.1 # via # jupyter-lsp # jupyterlab # jupyterlab-server # notebook # notebook-shim -jupyter-server-terminals==0.4.4 +jupyter-server-terminals==0.5.0 # via jupyter-server jupyterlab==4.0.9 # via notebook -jupyterlab-pygments==0.2.2 +jupyterlab-pygments==0.3.0 # via nbconvert jupyterlab-server==2.25.2 # via @@ -303,9 +313,10 @@ layoutparser[layoutmodels,tesseract]==0.3.4 # via # -r requirements/base.txt # unstructured-inference -lxml==4.9.3 +lxml==4.9.4 # via # -r requirements/base.txt + # pikepdf # python-docx # python-pptx # unstructured @@ -322,6 +333,7 @@ marshmallow==3.20.1 # via # -r requirements/base.txt # dataclasses-json + # unstructured-client matplotlib==3.8.2 # via # -r requirements/base.txt @@ -342,7 +354,7 @@ msg-parser==1.2.0 # via # -r requirements/base.txt # unstructured -mypy==1.7.0 +mypy==1.8.0 # via -r requirements/test.in mypy-extensions==1.0.0 # via @@ -350,9 +362,10 @@ mypy-extensions==1.0.0 # black # mypy # typing-inspect + # unstructured-client nbclient==0.9.0 # via nbconvert -nbconvert==7.11.0 +nbconvert==7.13.1 # via # jupyter # jupyter-server @@ -395,7 +408,7 @@ numpy==1.26.2 # torchvision # transformers # unstructured -olefile==0.46 +olefile==0.47 # via # -r requirements/base.txt # msg-parser @@ -438,13 +451,15 @@ packaging==23.2 # matplotlib # nbconvert # onnxruntime + # pikepdf # pytesseract # pytest # qtconsole # qtpy # transformers + # unstructured-client # unstructured-pytesseract -pandas==2.1.3 +pandas==2.1.4 # via # -r requirements/base.txt # layoutparser @@ -453,7 +468,7 @@ pandocfilters==1.5.0 # via nbconvert parso==0.8.3 # via jedi -pathspec==0.11.2 +pathspec==0.12.1 # via black pdf2image==1.16.3 # via @@ -469,8 +484,12 @@ pdfplumber==0.10.3 # via # -r requirements/base.txt # layoutparser -pexpect==4.8.0 +pexpect==4.9.0 # via ipython +pikepdf==8.10.1 + # via + # -r requirements/base.txt + # unstructured pillow==10.1.0 # via # -r requirements/base.txt @@ -478,11 +497,12 @@ pillow==10.1.0 # matplotlib # pdf2image # pdfplumber + # pikepdf # pytesseract # python-pptx # torchvision # unstructured-pytesseract -platformdirs==4.0.0 +platformdirs==4.1.0 # via # black # jupyter-core @@ -492,9 +512,9 @@ portalocker==2.8.2 # via # -r requirements/base.txt # iopath -prometheus-client==0.18.0 +prometheus-client==0.19.0 # via jupyter-server -prompt-toolkit==3.0.41 +prompt-toolkit==3.0.43 # via # ipython # jupyter-console @@ -503,7 +523,7 @@ protobuf==4.25.1 # -r requirements/base.txt # onnx # onnxruntime -psutil==5.9.6 +psutil==5.9.7 # via # -r requirements/base.txt # ipykernel @@ -531,7 +551,7 @@ pydantic==1.10.13 # fastapi pyflakes==3.1.0 # via flake8 -pygments==2.17.1 +pygments==2.17.2 # via # ipython # jupyter-console @@ -545,9 +565,11 @@ pyparsing==3.1.1 # via # -r requirements/base.txt # matplotlib -pypdf==3.17.1 - # via -r requirements/base.txt -pypdfium2==4.24.0 +pypdf==3.17.3 + # via + # -r requirements/base.txt + # unstructured +pypdfium2==4.25.0 # via # -r requirements/base.txt # pdfplumber @@ -570,11 +592,12 @@ python-dateutil==2.8.2 # jupyter-client # matplotlib # pandas + # unstructured-client python-docx==1.1.0 # via # -r requirements/base.txt # unstructured -python-iso639==2023.6.15 +python-iso639==2023.12.11 # via # -r requirements/base.txt # unstructured @@ -606,7 +629,7 @@ pyyaml==6.0.1 # omegaconf # timm # transformers -pyzmq==25.1.1 +pyzmq==25.1.2 # via # ipykernel # jupyter-client @@ -624,7 +647,7 @@ rapidfuzz==3.5.2 # unstructured-inference ratelimit==2.2.1 # via -r requirements/base.txt -referencing==0.31.0 +referencing==0.32.0 # via # jsonschema # jsonschema-specifications @@ -642,6 +665,7 @@ requests==2.31.0 # torchvision # transformers # unstructured + # unstructured-client rfc3339-validator==0.1.4 # via # jsonschema @@ -650,7 +674,7 @@ rfc3986-validator==0.1.1 # via # jsonschema # jupyter-events -rpds-py==0.13.1 +rpds-py==0.15.2 # via # jsonschema # referencing @@ -675,6 +699,7 @@ six==1.16.0 # langdetect # python-dateutil # rfc3339-validator + # unstructured-client sniffio==1.3.0 # via # -r requirements/base.txt @@ -703,7 +728,7 @@ terminado==0.18.0 # via # jupyter-server # jupyter-server-terminals -timm==0.9.10 +timm==0.9.12 # via # -r requirements/base.txt # effdet @@ -720,20 +745,20 @@ tomli==2.0.1 # jupyterlab # mypy # pytest -torch==2.1.1 +torch==2.1.2 # via # -r requirements/base.txt # effdet # layoutparser # timm # torchvision -torchvision==0.16.1 +torchvision==0.16.2 # via # -r requirements/base.txt # effdet # layoutparser # timm -tornado==6.3.3 +tornado==6.4 # via # ipykernel # jupyter-client @@ -748,7 +773,7 @@ tqdm==4.66.1 # iopath # nltk # transformers -traitlets==5.13.0 +traitlets==5.14.0 # via # comm # ipykernel @@ -765,13 +790,13 @@ traitlets==5.13.0 # nbconvert # nbformat # qtconsole -transformers==4.35.2 +transformers==4.36.2 # via # -r requirements/base.txt # unstructured-inference types-python-dateutil==2.8.19.14 # via arrow -typing-extensions==4.8.0 +typing-extensions==4.9.0 # via # -r requirements/base.txt # async-lru @@ -785,18 +810,24 @@ typing-extensions==4.8.0 # torch # typing-inspect # unstructured + # unstructured-client # uvicorn typing-inspect==0.9.0 # via # -r requirements/base.txt # dataclasses-json + # unstructured-client tzdata==2023.3 # via # -r requirements/base.txt # pandas -unstructured[local-inference]==0.11.0 +unstructured[local-inference]==0.11.6 # via -r requirements/base.txt -unstructured-inference==0.7.15 +unstructured-client==0.15.1 + # via + # -r requirements/base.txt + # unstructured +unstructured-inference==0.7.18 # via # -r requirements/base.txt # unstructured @@ -810,11 +841,12 @@ urllib3==2.1.0 # via # -r requirements/base.txt # requests -uvicorn==0.24.0.post1 + # unstructured-client +uvicorn==0.25.0 # via -r requirements/base.txt watchdog==3.0.0 # via nbdev -wcwidth==0.2.10 +wcwidth==0.2.12 # via prompt-toolkit webcolors==1.13 # via jsonschema @@ -822,15 +854,16 @@ webencodings==0.5.1 # via # bleach # tinycss2 -websocket-client==1.6.4 +websocket-client==1.7.0 # via jupyter-server -wheel==0.41.3 +wheel==0.42.0 # via astunparse widgetsnbextension==4.0.9 # via ipywidgets wrapt==1.16.0 # via # -r requirements/base.txt + # deprecated # unstructured xlrd==2.0.1 # via