build(deps): Remove unstructured.paddlepaddle fork (#3506)

This PR aims to remove "unstructured.paddlepaddle" fork. Previously, we
used `unstructured.paddlepaddle` fork to support
`unstructured.paddleocr` on arm64 architecture. But currently,
`unstructured.paddleocr` with `unstructured.paddlepaddle` fails to work
on `arm64` architecture. Also, `unstructured.paddleocr` with the latest
version of the original `paddlepaddle` works on both `amd64` and `arm64`
architectures.

### Testing
```
os.environ["OCR_AGENT"] = "unstructured.partition.utils.ocr_models.paddle_ocr.OCRAgentPaddle"

elements = partition_pdf(
    filename=<file_path>,
    strategy="hi_res",
    infer_table_structure=True,
)
```

.github/workflows/ci.yml

@@ -72,7 +72,6 @@ jobs:
       - name: Install all doc and test dependencies
         run: |
           make install-ci
-          make install-paddleocr
           make install-all-ingest
           make check-licenses
 

Dockerfile

@@ -17,7 +17,6 @@ RUN chown -R notebook-user:notebook-user /app && \
 USER notebook-user
 
 RUN find requirements/ -type f -name "*.txt" -exec pip3.11 install --no-cache-dir --user -r '{}' ';' && \
-  pip3.11 install unstructured.paddlepaddle && \
   python3.11 -c "from unstructured.nlp.tokenize import download_nltk_packages; download_nltk_packages()" && \
   python3.11 -c "from unstructured.partition.model_init import initialize; initialize()" && \
   python3.11 -c "from unstructured_inference.models.tables import UnstructuredTableTransformerModel; model = UnstructuredTableTransformerModel(); model.initialize('microsoft/table-transformer-structure-recognition')"

Makefile

@@ -277,10 +277,6 @@ install-local-inference: install install-all-docs
 install-pandoc:
 	ARCH=${ARCH} ./scripts/install-pandoc.sh
 
-.PHONY: install-paddleocr
-install-paddleocr:
-	ARCH=${ARCH} ./scripts/install-paddleocr.sh
-
 ## pip-compile:             compiles all base/dev/test requirements
 .PHONY: pip-compile
 pip-compile:

requirements/extra-paddleocr.in

@@ -1,4 +1,5 @@
 -c ./deps/constraints.txt
 -c base.txt
 
+paddlepaddle==3.0.0b1 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/
 unstructured.paddleocr==2.8.0.1

requirements/extra-paddleocr.txt

@@ -4,6 +4,13 @@
 #
 #    pip-compile ./extra-paddleocr.in
 #
+anyio==3.7.1
+    # via
+    #   -c ././deps/constraints.txt
+    #   -c ./base.txt
+    #   httpx
+astor==0.8.1
+    # via paddlepaddle
 attrdict==2.0.1
     # via unstructured-paddleocr
 cachetools==5.4.0
@@ -12,6 +19,8 @@ certifi==2024.7.4
     # via
     #   -c ././deps/constraints.txt
     #   -c ./base.txt
+    #   httpcore
+    #   httpx
     #   requests
 charset-normalizer==3.3.2
     # via
@@ -27,13 +36,33 @@ cycler==0.12.1
     # via matplotlib
 cython==3.0.11
     # via unstructured-paddleocr
+decorator==5.1.1
+    # via paddlepaddle
 et-xmlfile==1.1.0
     # via openpyxl
+exceptiongroup==1.2.2
+    # via
+    #   -c ./base.txt
+    #   anyio
 fonttools==4.53.1
     # via matplotlib
+h11==0.14.0
+    # via
+    #   -c ./base.txt
+    #   httpcore
+httpcore==1.0.5
+    # via
+    #   -c ./base.txt
+    #   httpx
+httpx==0.27.0
+    # via
+    #   -c ./base.txt
+    #   paddlepaddle
 idna==3.7
     # via
     #   -c ./base.txt
+    #   anyio
+    #   httpx
     #   requests
 imageio==2.34.2
     # via
@@ -59,7 +88,9 @@ matplotlib==3.9.1.post1
 more-itertools==10.4.0
     # via cssutils
 networkx==3.2.1
-    # via scikit-image
+    # via
+    #   paddlepaddle
+    #   scikit-image
 numpy==1.26.4
     # via
     #   -c ./base.txt
@@ -69,6 +100,8 @@ numpy==1.26.4
     #   matplotlib
     #   opencv-contrib-python
     #   opencv-python
+    #   opt-einsum
+    #   paddlepaddle
     #   scikit-image
     #   scipy
     #   shapely
@@ -85,6 +118,8 @@ opencv-python==4.8.0.76
     #   unstructured-paddleocr
 openpyxl==3.1.5
     # via unstructured-paddleocr
+opt-einsum==3.3.0
+    # via paddlepaddle
 packaging==23.2
     # via
     #   -c ././deps/constraints.txt
@@ -92,6 +127,8 @@ packaging==23.2
     #   lazy-loader
     #   matplotlib
     #   scikit-image
+paddlepaddle==3.0.0b1
+    # via -r ./extra-paddleocr.in
 pdf2image==1.17.0
     # via unstructured-paddleocr
 pillow==10.4.0
@@ -99,11 +136,16 @@ pillow==10.4.0
     #   imageio
     #   imgaug
     #   matplotlib
+    #   paddlepaddle
     #   pdf2image
     #   scikit-image
     #   unstructured-paddleocr
 premailer==3.10.0
     # via unstructured-paddleocr
+protobuf==4.23.4
+    # via
+    #   -c ././deps/constraints.txt
+    #   paddlepaddle
 pyclipper==1.3.0.post5
     # via unstructured-paddleocr
 pyparsing==3.0.9
@@ -144,12 +186,21 @@ six==1.16.0
     #   attrdict
     #   imgaug
     #   python-dateutil
+sniffio==1.3.1
+    # via
+    #   -c ./base.txt
+    #   anyio
+    #   httpx
 tifffile==2024.7.24
     # via scikit-image
 tqdm==4.66.5
     # via
     #   -c ./base.txt
     #   unstructured-paddleocr
+typing-extensions==4.12.2
+    # via
+    #   -c ./base.txt
+    #   paddlepaddle
 unstructured-paddleocr==2.8.0.1
     # via -r ./extra-paddleocr.in
 urllib3==1.26.19

scripts/install-paddleocr.sh

@@ -1,9 +0,0 @@
-#!/usr/bin/env bash
-
-# aarch64 requires a custom build of paddlepaddle
-if [ "${ARCH}" = "aarch64" ]; then
-  python3 -m pip install unstructured.paddlepaddle
-else
-  python3 -m pip install paddlepaddle
-fi
-python3 -m pip install unstructured.paddleocr