docling-ocr-onnxtr 0.1.3__tar.gz → 0.2.1__tar.gz

{docling_ocr_onnxtr-0.1.3 → docling_ocr_onnxtr-0.2.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-ocr-onnxtr
-Version: 0.1.3
+Version: 0.2.1
 Summary: Onnx Text Recognition (OnnxTR) OCR plugin for docling
 Author-email: Felix Dittrich <felixdittrich92@gmail.com>
 Maintainer: Felix Dittrich
@@ -266,7 +266,7 @@ Dynamic: license-file
 [![codecov](https://codecov.io/gh/felixdittrich92/docling-OCR-OnnxTR/graph/badge.svg?token=L3AHXKV86A)](https://codecov.io/gh/felixdittrich92/docling-OCR-OnnxTR)
 [![Codacy Badge](https://app.codacy.com/project/badge/Grade/0d250447650240ee9ca573950fea8b99)](https://app.codacy.com/gh/felixdittrich92/docling-OCR-OnnxTR/dashboard?utm_source=gh&utm_medium=referral&utm_content=&utm_campaign=Badge_grade)
 [![CodeFactor](https://www.codefactor.io/repository/github/felixdittrich92/docling-ocr-onnxtr/badge)](https://www.codefactor.io/repository/github/felixdittrich92/docling-ocr-onnxtr)
-[![Pypi](https://img.shields.io/badge/pypi-v0.1.3-blue.svg)](https://pypi.org/project/docling-ocr-onnxtr/)
+[![Pypi](https://img.shields.io/badge/pypi-v0.2.0-blue.svg)](https://pypi.org/project/docling-ocr-onnxtr/)
 ![PyPI - Downloads](https://img.shields.io/pypi/dm/docling-ocr-onnxtr)
 
 The `docling-OCR-OnnxTR` repository provides a plugin that integrates the [OnnxTR OCR engine](https://github.com/felixdittrich92/OnnxTR) into the [Docling framework](https://github.com/docling-project/docling), enhancing document processing capabilities with efficient and accurate text recognition.
@@ -358,6 +358,62 @@ def main():
     print(md)
 
 
+if __name__ == "__main__":
+    main()
+```
+
+It is also possible to load the models from local files instead of using the Hugging Face Hub or downloading them from the repo:
+
+```python
+from docling.datamodel.pipeline_options import PdfPipelineOptions
+from docling.document_converter import (
+    ConversionResult,
+    DocumentConverter,
+    InputFormat,
+    PdfFormatOption,
+)
+from docling_ocr_onnxtr import OnnxtrOcrOptions
+from onnxtr.models import db_mobilenet_v3_large, parseq
+
+
+def main():
+    # Source document to convert
+    source = "https://arxiv.org/pdf/2408.09869v4"
+
+    # Load models from local files
+    # NOTE: You need to download the models first and then adjust the paths accordingly.
+    det_model = db_mobilenet_v3_large("/home/felix/.cache/onnxtr/models/db_mobilenet_v3_large-1866973f.onnx")
+    reco_model = parseq("/home/felix/.cache/onnxtr/models/parseq-00b40714.onnx")
+
+    ocr_options = OnnxtrOcrOptions(
+        # Text detection model
+        det_arch=det_model,
+        # Text recognition model
+        reco_arch=reco_model,
+        # This can be set to `True` to auto-correct the orientation of the pages
+        auto_correct_orientation=False,
+    )
+
+    pipeline_options = PdfPipelineOptions(
+        ocr_options=ocr_options,
+    )
+    pipeline_options.allow_external_plugins = True  # <-- enabled the external plugins
+
+    # Convert the document
+    converter = DocumentConverter(
+        format_options={
+            InputFormat.PDF: PdfFormatOption(
+                pipeline_options=pipeline_options,
+            ),
+        },
+    )
+
+    conversion_result: ConversionResult = converter.convert(source=source)
+    doc = conversion_result.document
+    md = doc.export_to_markdown()
+    print(md)
+
+
 if __name__ == "__main__":
     main()
 ```

{docling_ocr_onnxtr-0.1.3 → docling_ocr_onnxtr-0.2.1}/README.md

@@ -7,7 +7,7 @@
 [![codecov](https://codecov.io/gh/felixdittrich92/docling-OCR-OnnxTR/graph/badge.svg?token=L3AHXKV86A)](https://codecov.io/gh/felixdittrich92/docling-OCR-OnnxTR)
 [![Codacy Badge](https://app.codacy.com/project/badge/Grade/0d250447650240ee9ca573950fea8b99)](https://app.codacy.com/gh/felixdittrich92/docling-OCR-OnnxTR/dashboard?utm_source=gh&utm_medium=referral&utm_content=&utm_campaign=Badge_grade)
 [![CodeFactor](https://www.codefactor.io/repository/github/felixdittrich92/docling-ocr-onnxtr/badge)](https://www.codefactor.io/repository/github/felixdittrich92/docling-ocr-onnxtr)
-[![Pypi](https://img.shields.io/badge/pypi-v0.1.3-blue.svg)](https://pypi.org/project/docling-ocr-onnxtr/)
+[![Pypi](https://img.shields.io/badge/pypi-v0.2.0-blue.svg)](https://pypi.org/project/docling-ocr-onnxtr/)
 ![PyPI - Downloads](https://img.shields.io/pypi/dm/docling-ocr-onnxtr)
 
 The `docling-OCR-OnnxTR` repository provides a plugin that integrates the [OnnxTR OCR engine](https://github.com/felixdittrich92/OnnxTR) into the [Docling framework](https://github.com/docling-project/docling), enhancing document processing capabilities with efficient and accurate text recognition.
@@ -99,6 +99,62 @@ def main():
     print(md)
 
 
+if __name__ == "__main__":
+    main()
+```
+
+It is also possible to load the models from local files instead of using the Hugging Face Hub or downloading them from the repo:
+
+```python
+from docling.datamodel.pipeline_options import PdfPipelineOptions
+from docling.document_converter import (
+    ConversionResult,
+    DocumentConverter,
+    InputFormat,
+    PdfFormatOption,
+)
+from docling_ocr_onnxtr import OnnxtrOcrOptions
+from onnxtr.models import db_mobilenet_v3_large, parseq
+
+
+def main():
+    # Source document to convert
+    source = "https://arxiv.org/pdf/2408.09869v4"
+
+    # Load models from local files
+    # NOTE: You need to download the models first and then adjust the paths accordingly.
+    det_model = db_mobilenet_v3_large("/home/felix/.cache/onnxtr/models/db_mobilenet_v3_large-1866973f.onnx")
+    reco_model = parseq("/home/felix/.cache/onnxtr/models/parseq-00b40714.onnx")
+
+    ocr_options = OnnxtrOcrOptions(
+        # Text detection model
+        det_arch=det_model,
+        # Text recognition model
+        reco_arch=reco_model,
+        # This can be set to `True` to auto-correct the orientation of the pages
+        auto_correct_orientation=False,
+    )
+
+    pipeline_options = PdfPipelineOptions(
+        ocr_options=ocr_options,
+    )
+    pipeline_options.allow_external_plugins = True  # <-- enabled the external plugins
+
+    # Convert the document
+    converter = DocumentConverter(
+        format_options={
+            InputFormat.PDF: PdfFormatOption(
+                pipeline_options=pipeline_options,
+            ),
+        },
+    )
+
+    conversion_result: ConversionResult = converter.convert(source=source)
+    doc = conversion_result.document
+    md = doc.export_to_markdown()
+    print(md)
+
+
 if __name__ == "__main__":
     main()
 ```

{docling_ocr_onnxtr-0.1.3 → docling_ocr_onnxtr-0.2.1}/docling_ocr_onnxtr/onnxtr_model.py

@@ -1,4 +1,4 @@
-# Copyright (C) 2021-2025, Felix Dittrich.
+# Copyright (C) 2021-2026, Felix Dittrich.
 
 # This program is licensed under the Apache License 2.0.
 # See LICENSE or go to <https://opensource.org/licenses/Apache-2.0> for full license details.
@@ -12,7 +12,7 @@ import numpy
 import numpy as np
 from docling.datamodel.base_models import Page
 from docling.datamodel.document import ConversionResult
-from docling.datamodel.pipeline_options import (
+from docling.datamodel.pipeline_options import (  # type: ignore[attr-defined]
     AcceleratorOptions,
     OcrOptions,
 )
@@ -91,11 +91,13 @@ class OnnxtrOcrModel(BaseOcrModel):
 
             self.reader = ocr_predictor(
                 det_arch=(
-                    from_hub(self.options.det_arch) if self.options.det_arch.count("/") == 1 else self.options.det_arch
+                    from_hub(self.options.det_arch)
+                    if isinstance(self.options.det_arch, str) and self.options.det_arch.count("/") == 1
+                    else self.options.det_arch
                 ),
                 reco_arch=(
                     from_hub(self.options.reco_arch)
-                    if self.options.reco_arch.count("/") == 1
+                    if isinstance(self.options.reco_arch, str) and self.options.reco_arch.count("/") == 1
                     else self.options.reco_arch
                 ),
                 det_bs=1,  # NOTE: Should be always 1, because docling handles batching

{docling_ocr_onnxtr-0.1.3 → docling_ocr_onnxtr-0.2.1}/docling_ocr_onnxtr/options.py

@@ -1,4 +1,4 @@
-# Copyright (C) 2021-2025, Felix Dittrich.
+# Copyright (C) 2021-2026, Felix Dittrich.
 
 # This program is licensed under the Apache License 2.0.
 # See LICENSE or go to <https://opensource.org/licenses/Apache-2.0> for full license details.
@@ -22,9 +22,9 @@ class OnnxtrOcrOptions(OcrOptions):
     # detection model objectness score threshold 'fast algorithm'
     objectness_score: float = 0.3
 
-    # NOTE: This can be also a hf hub model
-    det_arch: str = "fast_base"
-    reco_arch: str = "crnn_vgg16_bn"
+    # NOTE: This can be also a hf hub model or an instance of a model class.
+    det_arch: Any = "fast_base"
+    reco_arch: Any = "crnn_vgg16_bn"
     reco_bs: int = 512
     auto_correct_orientation: bool = False
     preserve_aspect_ratio: bool = True

{docling_ocr_onnxtr-0.1.3 → docling_ocr_onnxtr-0.2.1}/docling_ocr_onnxtr/plugin.py

@@ -1,4 +1,4 @@
-# Copyright (C) 2021-2025, Felix Dittrich.
+# Copyright (C) 2021-2026, Felix Dittrich.
 
 # This program is licensed under the Apache License 2.0.
 # See LICENSE or go to <https://opensource.org/licenses/Apache-2.0> for full license details.

docling_ocr_onnxtr-0.2.1/docling_ocr_onnxtr/version.py

@@ -0,0 +1 @@
+__version__ = 'v0.2.1'

{docling_ocr_onnxtr-0.1.3 → docling_ocr_onnxtr-0.2.1}/docling_ocr_onnxtr.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-ocr-onnxtr
-Version: 0.1.3
+Version: 0.2.1
 Summary: Onnx Text Recognition (OnnxTR) OCR plugin for docling
 Author-email: Felix Dittrich <felixdittrich92@gmail.com>
 Maintainer: Felix Dittrich
@@ -266,7 +266,7 @@ Dynamic: license-file
 [![codecov](https://codecov.io/gh/felixdittrich92/docling-OCR-OnnxTR/graph/badge.svg?token=L3AHXKV86A)](https://codecov.io/gh/felixdittrich92/docling-OCR-OnnxTR)
 [![Codacy Badge](https://app.codacy.com/project/badge/Grade/0d250447650240ee9ca573950fea8b99)](https://app.codacy.com/gh/felixdittrich92/docling-OCR-OnnxTR/dashboard?utm_source=gh&utm_medium=referral&utm_content=&utm_campaign=Badge_grade)
 [![CodeFactor](https://www.codefactor.io/repository/github/felixdittrich92/docling-ocr-onnxtr/badge)](https://www.codefactor.io/repository/github/felixdittrich92/docling-ocr-onnxtr)
-[![Pypi](https://img.shields.io/badge/pypi-v0.1.3-blue.svg)](https://pypi.org/project/docling-ocr-onnxtr/)
+[![Pypi](https://img.shields.io/badge/pypi-v0.2.0-blue.svg)](https://pypi.org/project/docling-ocr-onnxtr/)
 ![PyPI - Downloads](https://img.shields.io/pypi/dm/docling-ocr-onnxtr)
 
 The `docling-OCR-OnnxTR` repository provides a plugin that integrates the [OnnxTR OCR engine](https://github.com/felixdittrich92/OnnxTR) into the [Docling framework](https://github.com/docling-project/docling), enhancing document processing capabilities with efficient and accurate text recognition.
@@ -358,6 +358,62 @@ def main():
     print(md)
 
 
+if __name__ == "__main__":
+    main()
+```
+
+It is also possible to load the models from local files instead of using the Hugging Face Hub or downloading them from the repo:
+
+```python
+from docling.datamodel.pipeline_options import PdfPipelineOptions
+from docling.document_converter import (
+    ConversionResult,
+    DocumentConverter,
+    InputFormat,
+    PdfFormatOption,
+)
+from docling_ocr_onnxtr import OnnxtrOcrOptions
+from onnxtr.models import db_mobilenet_v3_large, parseq
+
+
+def main():
+    # Source document to convert
+    source = "https://arxiv.org/pdf/2408.09869v4"
+
+    # Load models from local files
+    # NOTE: You need to download the models first and then adjust the paths accordingly.
+    det_model = db_mobilenet_v3_large("/home/felix/.cache/onnxtr/models/db_mobilenet_v3_large-1866973f.onnx")
+    reco_model = parseq("/home/felix/.cache/onnxtr/models/parseq-00b40714.onnx")
+
+    ocr_options = OnnxtrOcrOptions(
+        # Text detection model
+        det_arch=det_model,
+        # Text recognition model
+        reco_arch=reco_model,
+        # This can be set to `True` to auto-correct the orientation of the pages
+        auto_correct_orientation=False,
+    )
+
+    pipeline_options = PdfPipelineOptions(
+        ocr_options=ocr_options,
+    )
+    pipeline_options.allow_external_plugins = True  # <-- enabled the external plugins
+
+    # Convert the document
+    converter = DocumentConverter(
+        format_options={
+            InputFormat.PDF: PdfFormatOption(
+                pipeline_options=pipeline_options,
+            ),
+        },
+    )
+
+    conversion_result: ConversionResult = converter.convert(source=source)
+    doc = conversion_result.document
+    md = doc.export_to_markdown()
+    print(md)
+
+
 if __name__ == "__main__":
     main()
 ```

{docling_ocr_onnxtr-0.1.3 → docling_ocr_onnxtr-0.2.1}/setup.py

@@ -1,4 +1,4 @@
-# Copyright (C) 2021-2025, Felix Dittrich.
+# Copyright (C) 2021-2026, Felix Dittrich.
 
 # This program is licensed under the Apache License 2.0.
 # See LICENSE or go to <https://opensource.org/licenses/Apache-2.0> for full license details.
@@ -9,7 +9,7 @@ from pathlib import Path
 from setuptools import setup
 
 PKG_NAME = "docling_ocr_onnxtr"
-VERSION = os.getenv("BUILD_VERSION", "0.1.3a0")
+VERSION = os.getenv("BUILD_VERSION", "0.2.1a0")
 
 
 if __name__ == "__main__":

{docling_ocr_onnxtr-0.1.3 → docling_ocr_onnxtr-0.2.1}/tests/test_plugin.py

@@ -15,9 +15,8 @@ from docling.document_converter import DocumentConverter, PdfFormatOption
 from docling_ocr_onnxtr import OnnxtrOcrOptions
 
 from .test_data_gen_flag import GEN_TEST_DATA
-from .verify_utils import verify_conversion_result_v1, verify_conversion_result_v2
+from .verify_utils import verify_conversion_result_v2
 
-GENERATE_V1 = GEN_TEST_DATA
 GENERATE_V2 = GEN_TEST_DATA
 
 
@@ -73,7 +72,7 @@ def test_e2e_conversions(ocr_options: OcrOptions):
     print(f"Converting with ocr_engine: {ocr_options.kind}, language: {ocr_options.lang}")
     converter = get_converter(ocr_options=ocr_options)
     for pdf_path in pdf_paths:
-        if not ocr_options.auto_correct_orientation and "rotated" in pdf_path.name:
+        if not ocr_options.auto_correct_orientation and ("rotated" in pdf_path.name or "rotation" in pdf_path.name):
             # Skip rotated PDFs if orientation correction is disabled
             print(f"Skipping {pdf_path} due to orientation correction settings.")
             continue
@@ -82,12 +81,6 @@ def test_e2e_conversions(ocr_options: OcrOptions):
         doc_result: ConversionResult = converter.convert(pdf_path)
 
         try:
-            verify_conversion_result_v1(
-                input_path=pdf_path,
-                doc_result=doc_result,
-                generate=GENERATE_V1,
-                fuzzy=True,
-            )
             verify_conversion_result_v2(
                 input_path=pdf_path,
                 doc_result=doc_result,
@@ -95,7 +88,7 @@ def test_e2e_conversions(ocr_options: OcrOptions):
                 fuzzy=True,
             )
         except AssertionError as e:
-            if "rotated" in pdf_path.name:
+            if "rotated" in pdf_path.name or "rotation" in pdf_path.name or ocr_options.auto_correct_orientation:
                 pytest.xfail(f"Skipping {pdf_path} due to orientation correction settings: {e}")
             else:
                 raise  # Unexpected failure — re-raise the error

docling_ocr_onnxtr-0.1.3/docling_ocr_onnxtr/version.py

@@ -1 +0,0 @@
-__version__ = 'v0.1.3'