docling 1.19.1__py3-none-any.whl → 2.1.0__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (38) hide show
  1. docling/backend/abstract_backend.py +32 -37
  2. docling/backend/docling_parse_backend.py +16 -12
  3. docling/backend/docling_parse_v2_backend.py +240 -0
  4. docling/backend/html_backend.py +425 -0
  5. docling/backend/mspowerpoint_backend.py +375 -0
  6. docling/backend/msword_backend.py +509 -0
  7. docling/backend/pdf_backend.py +78 -0
  8. docling/backend/pypdfium2_backend.py +15 -10
  9. docling/cli/main.py +61 -60
  10. docling/datamodel/base_models.py +73 -193
  11. docling/datamodel/document.py +379 -324
  12. docling/datamodel/pipeline_options.py +16 -0
  13. docling/datamodel/settings.py +1 -0
  14. docling/document_converter.py +215 -252
  15. docling/models/base_model.py +25 -0
  16. docling/models/base_ocr_model.py +19 -6
  17. docling/models/ds_glm_model.py +220 -22
  18. docling/models/easyocr_model.py +45 -40
  19. docling/models/layout_model.py +130 -114
  20. docling/models/page_assemble_model.py +119 -95
  21. docling/models/page_preprocessing_model.py +61 -0
  22. docling/models/table_structure_model.py +122 -111
  23. docling/models/tesseract_ocr_cli_model.py +63 -56
  24. docling/models/tesseract_ocr_model.py +58 -50
  25. docling/pipeline/base_pipeline.py +190 -0
  26. docling/pipeline/simple_pipeline.py +59 -0
  27. docling/pipeline/standard_pdf_pipeline.py +198 -0
  28. docling/utils/export.py +4 -3
  29. docling/utils/layout_utils.py +17 -11
  30. docling-2.1.0.dist-info/METADATA +149 -0
  31. docling-2.1.0.dist-info/RECORD +42 -0
  32. docling/pipeline/base_model_pipeline.py +0 -18
  33. docling/pipeline/standard_model_pipeline.py +0 -66
  34. docling-1.19.1.dist-info/METADATA +0 -380
  35. docling-1.19.1.dist-info/RECORD +0 -34
  36. {docling-1.19.1.dist-info → docling-2.1.0.dist-info}/LICENSE +0 -0
  37. {docling-1.19.1.dist-info → docling-2.1.0.dist-info}/WHEEL +0 -0
  38. {docling-1.19.1.dist-info → docling-2.1.0.dist-info}/entry_points.txt +0 -0
docling-1.19.1.dist-info/RECORD DELETED
@@ -1,34 +0,0 @@
1
- docling/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
2
- docling/backend/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
3
- docling/backend/abstract_backend.py,sha256=clJtGxLedpLriEhpx7oyxjmlwMLPorkv-1tdfZm9GdA,1546
4
- docling/backend/docling_parse_backend.py,sha256=RUWWZbx2cUotZeeTkc-Lbg2k8MVFXFxaDjM4sPfaFZE,7475
5
- docling/backend/pypdfium2_backend.py,sha256=bIIImVM73wmcVcKMqjl4JF8CD-Qj2W5rZbI4G7clU4s,8877
6
- docling/cli/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
7
- docling/cli/main.py,sha256=Q_5HTL2O20bMlF-U8Ac8ev4iqgLPyrwlHEyLAq6rezg,7913
8
- docling/datamodel/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
9
- docling/datamodel/base_models.py,sha256=ihBfPvVhxnEGmrzBMma_2RlOyObpaBAb7O5oBFBlIks,8833
10
- docling/datamodel/document.py,sha256=8iHylLaty-ZVYGzYmP61dZPtduzEyiRM79Tx3gS9UEU,16195
11
- docling/datamodel/pipeline_options.py,sha256=KTjos_AheIUDlMm4xsV9V1cvM5cI3ZG43z-WXAP_FXU,1931
12
- docling/datamodel/settings.py,sha256=t5g6wrEJnPa9gBzMMl8ppgBRUYz-8xgopEtfMS0ZH28,733
13
- docling/document_converter.py,sha256=V0Q5hCRyTbmavhQBN48biBttxpvPyoAPmFRhSePv2vE,11149
14
- docling/models/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
15
- docling/models/base_ocr_model.py,sha256=a89dLaT2vV5lTSNeetlCY9OnUd4LP4My_ZRky_fpVgo,4530
16
- docling/models/ds_glm_model.py,sha256=VXGmj8cW0WKMz1He4tp1lZhXHkS8Z39U1G-ujkc7deU,3368
17
- docling/models/easyocr_model.py,sha256=fmfTvOfqo97n_xYQDPRMKlvMOs9QqgLgSTqwBDrjK_8,3010
18
- docling/models/layout_model.py,sha256=5wCohyzGK1p5F5cTRY5QWbW2AI4eevXobJDJdSLhX7k,11272
19
- docling/models/page_assemble_model.py,sha256=8eoG2WiFxPxq9TPvM-wkngb2gkr0tdtCRVXg1JcTETo,5550
20
- docling/models/table_structure_model.py,sha256=iHJjWdKCpTcH3l_ElMWnC5pt6tkUpIuByed304Fdq9w,6009
21
- docling/models/tesseract_ocr_cli_model.py,sha256=B7zPkpKgpfTTUWXm-_zvw7x0yvzVd85WguawFVDwdqI,5529
22
- docling/models/tesseract_ocr_model.py,sha256=PqQv1Hv3GC1FByjegWvjNFw15Jcw-mT25_MvFr3hAHQ,4575
23
- docling/pipeline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
24
- docling/pipeline/base_model_pipeline.py,sha256=rrMog3EuiR5Gx9OWtfMj24rQvHCrWkxZ3g9OIr7LPSQ,607
25
- docling/pipeline/standard_model_pipeline.py,sha256=SVyMYmP1YLzYqDqagjn-pWugnjS7bp9YbmlV-VcenpI,2589
26
- docling/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
27
- docling/utils/export.py,sha256=bKLdbeUcR-rQsGPV1IqJkCHKMCv7X2QOHyxmjNuH3HE,4655
28
- docling/utils/layout_utils.py,sha256=FOFbL0hKzUoWXdZaeUvEtFqKv0IkPifIr4sdGW4suKs,31804
29
- docling/utils/utils.py,sha256=llhXSbIDNZ1MHOwBEfLHBAoJIAYI7QlPIonlI1jLUJ0,1208
30
- docling-1.19.1.dist-info/LICENSE,sha256=mBb7ErEcM8VS9OhiGHnQ2kk75HwPhr54W1Oiz3965MY,1088
31
- docling-1.19.1.dist-info/METADATA,sha256=hCQeq3JVB16CfTwtjjwnX5u9bWYjD0CsSbn9h1tZZTM,16800
32
- docling-1.19.1.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
33
- docling-1.19.1.dist-info/entry_points.txt,sha256=VOSzV77znM52dz5ysaDuJ0ijl1cnfrh1ZPg8od5OcTs,48
34
- docling-1.19.1.dist-info/RECORD,,