chore: bump unstructured-inference 0.7.33 (#3074)

Summary:
- bump unstructured-inference to `0.7.33`
- cut a release for `0.14.2`
- add some dependencies that previously came through from the
layoutparser extras.

Author: christinestraub

CHANGELOG.md

@@ -1,10 +1,12 @@
-## 0.14.2-dev1
+## 0.14.2
 
 ### Enhancements
 
+* **Bump unstructured-inference==0.7.33**.
+
 ### Features
 
-  * Add attribution to the `pinecone` connector
+* **Add attribution to the `pinecone` connector**.
 
 ### Fixes
 

docs/requirements.txt

@@ -66,7 +66,7 @@ pygments==2.18.0
     #   sphinx-tabs
 pyyaml==6.0.1
     # via myst-parser
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./base.txt
     #   sphinx

requirements/base.txt

@@ -71,7 +71,7 @@ rapidfuzz==3.9.1
     # via -r ./base.in
 regex==2024.5.15
     # via nltk
-requests==2.32.1
+requests==2.32.2
     # via
     #   -r ./base.in
     #   unstructured-client

requirements/build.txt

@@ -66,7 +66,7 @@ pygments==2.18.0
     #   sphinx-tabs
 pyyaml==6.0.1
     # via myst-parser
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./base.txt
     #   sphinx

requirements/dev.txt

@@ -189,7 +189,7 @@ jupyterlab==4.2.0
     # via notebook
 jupyterlab-pygments==0.3.0
     # via nbconvert
-jupyterlab-server==2.27.1
+jupyterlab-server==2.27.2
     # via
     #   jupyterlab
     #   notebook
@@ -315,7 +315,7 @@ referencing==0.35.1
     #   jsonschema
     #   jsonschema-specifications
     #   jupyter-events
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./base.txt
     #   -c ./test.txt

requirements/extra-paddleocr.txt

@@ -174,7 +174,7 @@ rapidfuzz==3.9.1
     #   unstructured-paddleocr
 rarfile==4.2
     # via visualdl
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./base.txt
     #   premailer

requirements/extra-pdf-image.in

@@ -7,10 +7,12 @@ pdfminer.six
 pikepdf
 pillow_heif
 pypdf
+pytesseract
+google-cloud-vision
+effdet
 # Do not move to constraints.in, otherwise unstructured-inference will not be upgraded
 # when unstructured library is.
-unstructured-inference==0.7.31
+unstructured-inference==0.7.33
 # unstructured fork of pytesseract that provides an interface to allow for multiple output formats
 # from one tesseract call
 unstructured.pytesseract>=0.3.12
-google-cloud-vision

requirements/extra-pdf-image.txt

@@ -31,7 +31,7 @@ cycler==0.12.1
 deprecated==1.2.14
     # via pikepdf
 effdet==0.4.1
-    # via layoutparser
+    # via -r ./extra-pdf-image.in
 filelock==3.14.0
     # via
     #   huggingface-hub
@@ -84,7 +84,7 @@ jinja2==3.1.4
     # via torch
 kiwisolver==1.4.5
     # via matplotlib
-layoutparser[layoutmodels,tesseract]==0.3.4
+layoutparser==0.3.4
     # via unstructured-inference
 lxml==5.2.2
     # via
@@ -96,6 +96,7 @@ matplotlib==3.7.2
     # via
     #   -c ././deps/constraints.txt
     #   pycocotools
+    #   unstructured-inference
 mpmath==1.3.0
     # via sympy
 networkx==3.2.1
@@ -203,7 +204,7 @@ pypdf==4.2.0
 pypdfium2==4.30.0
     # via pdfplumber
 pytesseract==0.3.10
-    # via layoutparser
+    # via -r ./extra-pdf-image.in
 python-dateutil==2.9.0.post0
     # via
     #   -c ./base.txt
@@ -228,7 +229,7 @@ regex==2024.5.15
     # via
     #   -c ./base.txt
     #   transformers
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./base.txt
     #   google-api-core
@@ -253,28 +254,29 @@ sympy==1.12
     #   onnxruntime
     #   torch
 timm==1.0.3
-    # via effdet
+    # via
+    #   effdet
+    #   unstructured-inference
 tokenizers==0.19.1
     # via transformers
 torch==2.3.0
     # via
     #   -c ././deps/constraints.txt
     #   effdet
-    #   layoutparser
     #   timm
     #   torchvision
+    #   unstructured-inference
 torchvision==0.18.0
     # via
     #   effdet
-    #   layoutparser
     #   timm
 tqdm==4.66.4
     # via
     #   -c ./base.txt
     #   huggingface-hub
     #   iopath
     #   transformers
-transformers==4.41.0
+transformers==4.41.1
     # via unstructured-inference
 typing-extensions==4.11.0
     # via
@@ -285,7 +287,7 @@ typing-extensions==4.11.0
     #   torch
 tzdata==2024.1
     # via pandas
-unstructured-inference==0.7.31
+unstructured-inference==0.7.33
     # via -r ./extra-pdf-image.in
 unstructured-pytesseract==0.3.12
     # via

requirements/huggingface.txt

@@ -71,7 +71,7 @@ regex==2024.5.15
     #   -c ./base.txt
     #   sacremoses
     #   transformers
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./base.txt
     #   huggingface-hub
@@ -100,7 +100,7 @@ tqdm==4.66.4
     #   huggingface-hub
     #   sacremoses
     #   transformers
-transformers==4.41.0
+transformers==4.41.1
     # via -r ./huggingface.in
 typing-extensions==4.11.0
     # via

requirements/ingest/airtable.txt

@@ -27,7 +27,7 @@ pydantic==2.7.1
     # via pyairtable
 pydantic-core==2.18.2
     # via pydantic
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   pyairtable

requirements/ingest/astra.txt

@@ -8,7 +8,7 @@ anyio==3.7.1
     # via
     #   -c ./ingest/../deps/constraints.txt
     #   httpx
-astrapy==1.1.0
+astrapy==1.2.0
     # via -r ./ingest/astra.in
 bson==0.5.10
     # via astrapy
@@ -69,7 +69,7 @@ python-dateutil==2.9.0.post0
     # via
     #   -c ./ingest/../base.txt
     #   bson
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   cassio

requirements/ingest/azure-cognitive-search.txt

@@ -25,7 +25,7 @@ idna==3.7
     #   requests
 isodate==0.6.1
     # via azure-search-documents
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   azure-core

requirements/ingest/azure.txt

@@ -82,7 +82,7 @@ pycparser==2.22
     # via cffi
 pyjwt[crypto]==2.8.0
     # via msal
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   azure-core

requirements/ingest/box.txt

@@ -8,7 +8,7 @@ attrs==23.2.0
     # via boxsdk
 boxfs==0.3.0
     # via -r ./ingest/box.in
-boxsdk[jwt]==3.9.2
+boxsdk[jwt]==3.10.0
     # via boxfs
 certifi==2024.2.2
     # via
@@ -40,7 +40,7 @@ python-dateutil==2.9.0.post0
     # via
     #   -c ./ingest/../base.txt
     #   boxsdk
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   boxsdk

requirements/ingest/chroma.txt

@@ -161,7 +161,7 @@ pyyaml==6.0.1
     #   huggingface-hub
     #   kubernetes
     #   uvicorn
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   chromadb

requirements/ingest/clarifai.txt

@@ -56,7 +56,7 @@ python-rapidjson==1.17
     # via tritonclient
 pyyaml==6.0.1
     # via clarifai
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   clarifai-grpc

requirements/ingest/confluence.txt

@@ -31,7 +31,7 @@ oauthlib==3.2.2
     # via
     #   atlassian-python-api
     #   requests-oauthlib
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   atlassian-python-api

requirements/ingest/databricks-volumes.txt

@@ -29,7 +29,7 @@ pyasn1==0.6.0
     #   rsa
 pyasn1-modules==0.4.0
     # via google-auth
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   databricks-sdk

requirements/ingest/dropbox.txt

@@ -28,7 +28,7 @@ idna==3.7
     #   requests
 ply==3.11
     # via stone
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   dropbox

requirements/ingest/embed-aws-bedrock.txt

@@ -60,14 +60,14 @@ langchain==0.2.0
     # via langchain-community
 langchain-community==0.2.0
     # via -r ./ingest/embed-aws-bedrock.in
-langchain-core==0.2.0
+langchain-core==0.2.1
     # via
     #   langchain
     #   langchain-community
     #   langchain-text-splitters
 langchain-text-splitters==0.2.0
     # via langchain
-langsmith==0.1.60
+langsmith==0.1.61
     # via
     #   langchain
     #   langchain-community
@@ -114,7 +114,7 @@ pyyaml==6.0.1
     #   langchain
     #   langchain-community
     #   langchain-core
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   langchain

requirements/ingest/embed-huggingface.txt

@@ -72,14 +72,14 @@ langchain==0.2.0
     # via langchain-community
 langchain-community==0.2.0
     # via -r ./ingest/embed-huggingface.in
-langchain-core==0.2.0
+langchain-core==0.2.1
     # via
     #   langchain
     #   langchain-community
     #   langchain-text-splitters
 langchain-text-splitters==0.2.0
     # via langchain
-langsmith==0.1.60
+langsmith==0.1.61
     # via
     #   langchain
     #   langchain-community
@@ -142,7 +142,7 @@ regex==2024.5.15
     # via
     #   -c ./ingest/../base.txt
     #   transformers
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   huggingface-hub
@@ -186,7 +186,7 @@ tqdm==4.66.4
     #   huggingface-hub
     #   sentence-transformers
     #   transformers
-transformers==4.41.0
+transformers==4.41.1
     # via sentence-transformers
 typing-extensions==4.11.0
     # via

requirements/ingest/embed-octoai.txt

@@ -48,7 +48,7 @@ regex==2024.5.15
     # via
     #   -c ./ingest/../base.txt
     #   tiktoken
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   tiktoken

requirements/ingest/embed-openai.txt

@@ -68,14 +68,14 @@ langchain==0.2.0
     # via langchain-community
 langchain-community==0.2.0
     # via -r ./ingest/embed-openai.in
-langchain-core==0.2.0
+langchain-core==0.2.1
     # via
     #   langchain
     #   langchain-community
     #   langchain-text-splitters
 langchain-text-splitters==0.2.0
     # via langchain
-langsmith==0.1.60
+langsmith==0.1.61
     # via
     #   langchain
     #   langchain-community
@@ -125,7 +125,7 @@ regex==2024.5.15
     # via
     #   -c ./ingest/../base.txt
     #   tiktoken
-requests==2.32.1
+requests==2.32.2
     # via
     #   -c ./ingest/../base.txt
     #   langchain