chore: bump unstructured-inference pin (#2913)

**Summary**
Update dependencies to use the new version of `unstructured-inference`
released yesterday. Remedy a few small problems with `make pip-compile`
that stood in the way.

Author: scanny

Diff for: CHANGELOG.md

@@ -1,4 +1,4 @@
-## 0.13.3-dev9
+## 0.13.3
 
 ### Enhancements
 

Diff for: requirements/Makefile

@@ -24,11 +24,11 @@ compile-base:
 
 .PHONY: compile-all-base
 compile-all-base: compile-base compile-test compile-dev
-	@$(foreach file,$(BASE_REQUIREMENTS),echo "compiling: $(file)" && pip-compile --upgrade $(file);)
+	@$(foreach file,$(BASE_REQUIREMENTS),echo -e "\n\ncompiling: $(file)" && pip-compile --no-strip-extras --upgrade $(file) || exit;)
 
 .PHONY: compile-ingest
 compile-ingest:
-	@$(foreach file,$(INGEST_REQUIREMENTS),echo "compiling: $(file)" && pip-compile --upgrade $(file);)
+	@$(foreach file,$(INGEST_REQUIREMENTS),echo -e "\n\ncompiling: $(file)" && pip-compile --no-strip-extras --upgrade $(file) || exit;)
 
 .PHONY: clean
 clean: clean-base clean-ingest

Diff for: requirements/base.txt

@@ -69,7 +69,7 @@ python-magic==0.4.27
     # via -r ./base.in
 rapidfuzz==3.8.1
     # via -r ./base.in
-regex==2023.12.25
+regex==2024.4.16
     # via nltk
 requests==2.31.0
     # via

Diff for: requirements/deps/constraints.txt

@@ -22,8 +22,6 @@ Office365-REST-Python-Client<2.4.3
 # unstructured-inference to be upgraded when unstructured library is upgraded
 # https://github.com/Unstructured-IO/unstructured/issues/1458
 # unstructured-inference
-# unable to build wheel for arm on 0.3.3+
-safetensors<=0.3.2
 # use the known compatible version of weaviate and unstructured.pytesseract
 unstructured.pytesseract>=0.3.12
 weaviate-client>3.25.0
@@ -38,7 +36,6 @@ torch>2
 # pinned in unstructured paddleocr
 opencv-python==4.8.0.76
 opencv-contrib-python==4.8.0.76
-onnxruntime==1.15.1
 platformdirs==3.10.0
 
 # TODO: Constraint due to langchain, remove when that gets updated:

Diff for: requirements/dev.txt

@@ -73,7 +73,7 @@ defusedxml==0.7.1
     # via nbconvert
 distlib==0.3.8
     # via virtualenv
-exceptiongroup==1.2.0
+exceptiongroup==1.2.1
     # via
     #   -c ./test.txt
     #   anyio
@@ -91,7 +91,7 @@ httpcore==1.0.5
     # via httpx
 httpx==0.27.0
     # via jupyterlab
-identify==2.5.35
+identify==2.5.36
     # via pre-commit
 idna==3.7
     # via
@@ -195,7 +195,7 @@ markupsafe==2.1.5
     # via
     #   jinja2
     #   nbconvert
-matplotlib-inline==0.1.6
+matplotlib-inline==0.1.7
     # via
     #   ipykernel
     #   ipython
@@ -216,7 +216,7 @@ nest-asyncio==1.6.0
     # via ipykernel
 nodeenv==1.8.0
     # via pre-commit
-notebook==7.1.2
+notebook==7.1.3
     # via jupyter
 notebook-shim==0.2.4
     # via
@@ -294,7 +294,7 @@ pyyaml==6.0.1
     #   -c ./test.txt
     #   jupyter-events
     #   pre-commit
-pyzmq==25.1.2
+pyzmq==26.0.2
     # via
     #   ipykernel
     #   jupyter-client
@@ -368,7 +368,7 @@ tornado==6.4
     #   jupyterlab
     #   notebook
     #   terminado
-traitlets==5.14.2
+traitlets==5.14.3
     # via
     #   comm
     #   ipykernel
@@ -401,7 +401,7 @@ urllib3==1.26.18
     #   -c ./base.txt
     #   -c ./test.txt
     #   requests
-virtualenv==20.25.1
+virtualenv==20.25.3
     # via pre-commit
 wcwidth==0.2.13
     # via prompt-toolkit

Diff for: requirements/extra-paddleocr.txt

@@ -63,7 +63,7 @@ importlib-metadata==7.1.0
     # via flask
 importlib-resources==6.4.0
     # via matplotlib
-itsdangerous==2.1.2
+itsdangerous==2.2.0
     # via flask
 jinja2==3.1.3
     # via
@@ -188,7 +188,7 @@ scipy==1.10.1
     #   -c ././deps/constraints.txt
     #   imgaug
     #   scikit-image
-shapely==2.0.3
+shapely==2.0.4
     # via
     #   imgaug
     #   unstructured-paddleocr
@@ -200,7 +200,7 @@ six==1.16.0
     #   imgaug
     #   python-dateutil
     #   visualdl
-tifffile==2024.2.12
+tifffile==2024.4.18
     # via scikit-image
 tqdm==4.66.2
     # via

Diff for: requirements/extra-pdf-image.in

@@ -9,7 +9,7 @@ pillow_heif
 pypdf
 # Do not move to constraints.in, otherwise unstructured-inference will not be upgraded
 # when unstructured library is.
-unstructured-inference==0.7.25
+unstructured-inference==0.7.27
 # unstructured fork of pytesseract that provides an interface to allow for multiple output formats
 # from one tesseract call
 unstructured.pytesseract>=0.3.12

Diff for: requirements/extra-pdf-image.txt

@@ -100,10 +100,8 @@ onnx==1.16.0
     # via
     #   -r ./extra-pdf-image.in
     #   unstructured-inference
-onnxruntime==1.15.1
-    # via
-    #   -c ././deps/constraints.txt
-    #   unstructured-inference
+onnxruntime==1.17.3
+    # via unstructured-inference
 opencv-python==4.8.0.76
     # via
     #   -c ././deps/constraints.txt
@@ -132,7 +130,7 @@ pdfminer-six==20231228
     #   pdfplumber
 pdfplumber==0.11.0
     # via layoutparser
-pikepdf==8.15.0
+pikepdf==8.15.1
     # via -r ./extra-pdf-image.in
 pillow==10.3.0
     # via
@@ -190,7 +188,7 @@ rapidfuzz==3.8.1
     # via
     #   -c ./base.txt
     #   unstructured-inference
-regex==2023.12.25
+regex==2024.4.16
     # via
     #   -c ./base.txt
     #   transformers
@@ -199,9 +197,8 @@ requests==2.31.0
     #   -c ./base.txt
     #   huggingface-hub
     #   transformers
-safetensors==0.3.2
+safetensors==0.4.3
     # via
-    #   -c ././deps/constraints.txt
     #   timm
     #   transformers
 scipy==1.10.1
@@ -218,7 +215,7 @@ sympy==1.12
     #   torch
 timm==0.9.16
     # via effdet
-tokenizers==0.15.2
+tokenizers==0.19.1
     # via transformers
 torch==2.2.2
     # via
@@ -238,7 +235,7 @@ tqdm==4.66.2
     #   huggingface-hub
     #   iopath
     #   transformers
-transformers==4.37.1
+transformers==4.40.0
     # via unstructured-inference
 typing-extensions==4.11.0
     # via
@@ -249,7 +246,7 @@ typing-extensions==4.11.0
     #   torch
 tzdata==2024.1
     # via pandas
-unstructured-inference==0.7.25
+unstructured-inference==0.7.27
     # via -r ./extra-pdf-image.in
 unstructured-pytesseract==0.3.12
     # via

Diff for: requirements/huggingface.txt

@@ -64,7 +64,7 @@ pyyaml==6.0.1
     # via
     #   huggingface-hub
     #   transformers
-regex==2023.12.25
+regex==2024.4.16
     # via
     #   -c ./base.txt
     #   sacremoses
@@ -76,10 +76,8 @@ requests==2.31.0
     #   transformers
 sacremoses==0.1.1
     # via -r ./huggingface.in
-safetensors==0.3.2
-    # via
-    #   -c ././deps/constraints.txt
-    #   transformers
+safetensors==0.4.3
+    # via transformers
 sentencepiece==0.2.0
     # via -r ./huggingface.in
 six==1.16.0
@@ -88,7 +86,7 @@ six==1.16.0
     #   langdetect
 sympy==1.12
     # via torch
-tokenizers==0.15.2
+tokenizers==0.19.1
     # via transformers
 torch==2.2.2
     # via
@@ -100,7 +98,7 @@ tqdm==4.66.2
     #   huggingface-hub
     #   sacremoses
     #   transformers
-transformers==4.37.1
+transformers==4.40.0
     # via -r ./huggingface.in
 typing-extensions==4.11.0
     # via

Diff for: requirements/ingest/astra.txt

@@ -14,7 +14,7 @@ bson==0.5.10
     # via astrapy
 cassandra-driver==3.29.1
     # via cassio
-cassio==0.1.5
+cassio==0.1.6
     # via astrapy
 certifi==2024.2.2
     # via
@@ -33,7 +33,7 @@ click==8.1.7
     #   geomet
 deprecation==2.1.0
     # via astrapy
-exceptiongroup==1.2.0
+exceptiongroup==1.2.1
     # via anyio
 geomet==0.2.1.post1
     # via cassandra-driver
@@ -46,7 +46,9 @@ hpack==4.0.0
 httpcore==1.0.5
     # via httpx
 httpx[http2]==0.27.0
-    # via astrapy
+    # via
+    #   astrapy
+    #   httpx
 hyperframe==6.0.1
     # via h2
 idna==3.7

Diff for: requirements/ingest/azure.txt

@@ -4,9 +4,9 @@
 #
 #    pip-compile ./ingest/azure.in
 #
-adlfs==2024.2.0
+adlfs==2024.4.1
     # via -r ./ingest/azure.in
-aiohttp==3.9.4
+aiohttp==3.9.5
     # via adlfs
 aiosignal==1.3.1
     # via aiohttp
@@ -80,7 +80,9 @@ portalocker==2.8.2
 pycparser==2.22
     # via cffi
 pyjwt[crypto]==2.8.0
-    # via msal
+    # via
+    #   msal
+    #   pyjwt
 requests==2.31.0
     # via
     #   -c ./ingest/../base.txt

Diff for: requirements/ingest/box.txt

@@ -9,7 +9,9 @@ attrs==23.2.0
 boxfs==0.3.0
     # via -r ./ingest/box.in
 boxsdk[jwt]==3.9.2
-    # via boxfs
+    # via
+    #   boxfs
+    #   boxsdk
 certifi==2024.2.2
     # via
     #   -c ./ingest/../base.txt

Diff for: requirements/ingest/chroma.txt

@@ -44,9 +44,9 @@ coloredlogs==15.0.1
     # via onnxruntime
 deprecated==1.2.14
     # via opentelemetry-api
-exceptiongroup==1.2.0
+exceptiongroup==1.2.1
     # via anyio
-fastapi==0.110.1
+fastapi==0.110.2
     # via chromadb
 filelock==3.13.4
     # via huggingface-hub
@@ -58,7 +58,7 @@ google-auth==2.29.0
     # via kubernetes
 googleapis-common-protos==1.63.0
     # via opentelemetry-exporter-otlp-proto-grpc
-grpcio==1.62.1
+grpcio==1.62.2
     # via
     #   chromadb
     #   opentelemetry-exporter-otlp-proto-grpc
@@ -95,10 +95,8 @@ oauthlib==3.2.2
     # via
     #   kubernetes
     #   requests-oauthlib
-onnxruntime==1.15.1
-    # via
-    #   -c ./ingest/../deps/constraints.txt
-    #   chromadb
+onnxruntime==1.17.3
+    # via chromadb
 opentelemetry-api==1.16.0
     # via
     #   chromadb
@@ -131,7 +129,7 @@ protobuf==4.23.4
     #   googleapis-common-protos
     #   onnxruntime
     #   opentelemetry-proto
-pulsar-client==3.4.0
+pulsar-client==3.5.0
     # via chromadb
 pyasn1==0.6.0
     # via
@@ -186,7 +184,7 @@ sympy==1.12
     # via onnxruntime
 tenacity==8.2.3
     # via chromadb
-tokenizers==0.15.2
+tokenizers==0.19.1
     # via chromadb
 tqdm==4.66.2
     # via
@@ -216,7 +214,9 @@ urllib3==1.26.18
     #   kubernetes
     #   requests
 uvicorn[standard]==0.29.0
-    # via chromadb
+    # via
+    #   chromadb
+    #   uvicorn
 uvloop==0.19.0
     # via uvicorn
 watchfiles==0.21.0

Diff for: requirements/ingest/clarifai.txt

@@ -13,15 +13,15 @@ charset-normalizer==3.3.2
     # via
     #   -c ./ingest/../base.txt
     #   requests
-clarifai==10.3.0
+clarifai==10.3.1
     # via -r ./ingest/clarifai.in
 clarifai-grpc==10.2.3
     # via clarifai
 contextlib2==21.6.0
     # via schema
 googleapis-common-protos==1.63.0
     # via clarifai-grpc
-grpcio==1.62.1
+grpcio==1.62.2
     # via clarifai-grpc
 idna==3.7
     # via