feat(ocr): Add OnnxTR as possible OCR engine

Signed-off-by: felix <felixdittrich92@gmail.com>

.github/actions/setup-poetry/action.yml

@@ -14,6 +14,9 @@ runs:
       with:
         python-version: ${{ inputs.python-version }}
         cache: 'poetry'
+    - name: Update lockfile
+      run: poetry update
+      shell: bash
     - name: Install dependencies
       run: poetry install --all-extras
       shell: bash

pyproject.toml

@@ -44,7 +44,7 @@ packages = [{ include = "docling" }]
 ######################
 # actual dependencies:
 ######################
-python = "^3.10"
+python = "^3.9"
 pydantic = "^2.0.0"
 docling-core = {extras = ["chunking"], version = "^2.24.1"}
 docling-ibm-models = "^3.4.0"

tests/test_e2e_ocr_conversion.py

@@ -63,7 +63,6 @@ def test_e2e_conversions():
         TesseractOcrOptions(),
         TesseractCliOcrOptions(),
         EasyOcrOptions(force_full_page_ocr=True),
-        OnnxtrOcrOptions(force_full_page_ocr=True),
         TesseractOcrOptions(force_full_page_ocr=True),
         TesseractOcrOptions(force_full_page_ocr=True, lang=["auto"]),
         TesseractCliOcrOptions(force_full_page_ocr=True),
@@ -75,6 +74,11 @@ def test_e2e_conversions():
         engines.append(RapidOcrOptions())
         engines.append(RapidOcrOptions(force_full_page_ocr=True))
 
+    # onnxtr is only available for Python >=3.10
+    if sys.version_info >= (3, 10):
+        engines.append(OnnxtrOcrOptions())
+        engines.append(OnnxtrOcrOptions(force_full_page_ocr=True))
+
     # only works on mac
     if "darwin" == sys.platform:
         engines.append(OcrMacOptions())