docling 2.1.0__py3-none-any.whl → 2.4.1__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 (35) hide show
  1. docling/backend/abstract_backend.py +1 -0
  2. docling/backend/asciidoc_backend.py +431 -0
  3. docling/backend/docling_parse_backend.py +4 -4
  4. docling/backend/docling_parse_v2_backend.py +12 -4
  5. docling/backend/html_backend.py +61 -57
  6. docling/backend/md_backend.py +346 -0
  7. docling/backend/mspowerpoint_backend.py +62 -39
  8. docling/backend/msword_backend.py +12 -25
  9. docling/backend/pypdfium2_backend.py +1 -1
  10. docling/cli/main.py +38 -8
  11. docling/datamodel/base_models.py +16 -10
  12. docling/datamodel/document.py +36 -6
  13. docling/datamodel/pipeline_options.py +3 -3
  14. docling/datamodel/settings.py +15 -1
  15. docling/document_converter.py +38 -12
  16. docling/models/base_model.py +4 -1
  17. docling/models/base_ocr_model.py +21 -4
  18. docling/models/ds_glm_model.py +27 -11
  19. docling/models/easyocr_model.py +49 -39
  20. docling/models/layout_model.py +87 -61
  21. docling/models/page_assemble_model.py +102 -100
  22. docling/models/page_preprocessing_model.py +25 -7
  23. docling/models/table_structure_model.py +125 -90
  24. docling/models/tesseract_ocr_cli_model.py +62 -52
  25. docling/models/tesseract_ocr_model.py +76 -52
  26. docling/pipeline/base_pipeline.py +68 -69
  27. docling/pipeline/simple_pipeline.py +8 -11
  28. docling/pipeline/standard_pdf_pipeline.py +59 -56
  29. docling/utils/profiling.py +62 -0
  30. {docling-2.1.0.dist-info → docling-2.4.1.dist-info}/METADATA +27 -22
  31. docling-2.4.1.dist-info/RECORD +45 -0
  32. docling-2.1.0.dist-info/RECORD +0 -42
  33. {docling-2.1.0.dist-info → docling-2.4.1.dist-info}/LICENSE +0 -0
  34. {docling-2.1.0.dist-info → docling-2.4.1.dist-info}/WHEEL +0 -0
  35. {docling-2.1.0.dist-info → docling-2.4.1.dist-info}/entry_points.txt +0 -0
docling-2.1.0.dist-info/RECORD DELETED
@@ -1,42 +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=8Lh1gf1P9AnzlwB989OVBgLmokTpfI0LxYRfuvYTqoo,1646
4
- docling/backend/docling_parse_backend.py,sha256=UgBpopZIP5YkhwhybiqDnqVsSqv9DAAPFkafhfL0pPo,7623
5
- docling/backend/docling_parse_v2_backend.py,sha256=VY7MsiyqjN3Vl0UkyezriiVJMLbLRrQVuKjWaTgIUwY,8336
6
- docling/backend/html_backend.py,sha256=MlhEXaA0tgX_tLuQLnkex43gsKqpqHWnbkssxY4n_kc,14753
7
- docling/backend/mspowerpoint_backend.py,sha256=2UYfMMeWwgDtvIKQELCA-bYv5Z-rGvbMiBNcidNL_uE,14332
8
- docling/backend/msword_backend.py,sha256=4SDqZAZxLr6VV50OU3MRBAV8SwZMCyJCUbNVMVUpitc,17659
9
- docling/backend/pdf_backend.py,sha256=unnw7QiRE1VXg6Pj-eYrtnFGrp5SSYiI324OlFxyv6c,2050
10
- docling/backend/pypdfium2_backend.py,sha256=MJX6fQqwK3r967fyAAs-RA_YIkeQvhgsLkQAgaBTgaE,8995
11
- docling/cli/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
12
- docling/cli/main.py,sha256=NRVGz0z-3EBwYNMJGVnLtDBcfOeutaUyYdkM0ymRnGA,8008
13
- docling/datamodel/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
14
- docling/datamodel/base_models.py,sha256=Ha-DoRZoksjHSZHWqUSiQ79MTBEfY5ur8U_LVtyBRYU,5153
15
- docling/datamodel/document.py,sha256=FZQyJtHSeGBrZwFf-GGXDu-Dyp4iIl7VbVnTupmlUqk,19532
16
- docling/datamodel/pipeline_options.py,sha256=WNjluKC-Ww63ifkGMHwws8zIDHnOS1z5Hw7_j3S0qao,2446
17
- docling/datamodel/settings.py,sha256=KBFVeQviR1hoCFjA1ZwuLuQ6EAAYR7saIa6EUYiOkHI,767
18
- docling/document_converter.py,sha256=S_t9hs2uZfXC38LC0hTaAihrSJIrCvnTiuY5SvUccgk,9587
19
- docling/models/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
20
- docling/models/base_model.py,sha256=wSBGAIAbLqrqP_SMtkzXMuyFvvzjVU6iCqgSNnGIR4Y,603
21
- docling/models/base_ocr_model.py,sha256=SYelQRValiUo6M_p_9-J7CqNIOFO-EkK58j90SMsKQY,5028
22
- docling/models/ds_glm_model.py,sha256=vJLngchZonqFzGWbUr2izFSXk9DloPDhAfN2c3nkzNU,11254
23
- docling/models/easyocr_model.py,sha256=YfvdodjZ20WuOfouQXJmDyQL78QDOqWYsWSs2zSxWFc,3327
24
- docling/models/layout_model.py,sha256=zd2ULW3U6v9OJl4TnjWFEY6Q2O-lBfrIqtvrnDzF7HU,12596
25
- docling/models/page_assemble_model.py,sha256=LOKHho-r-RpeIVh8CpJ9tid_QIp5um3ukcrucZsyUlY,6645
26
- docling/models/page_preprocessing_model.py,sha256=cfhUIlGAGaX1RxILi69ZEV9Kmhhd3Y0XaSlQnGo18o4,1964
27
- docling/models/table_structure_model.py,sha256=YWSZKOz56gvicjTzVgSE-8Z_hI3NcRD5EN0yOUoM-_g,6979
28
- docling/models/tesseract_ocr_cli_model.py,sha256=fKc05V73ibMvAeuA4PForhYNtunpT5rR0k_xHZsew-E,5980
29
- docling/models/tesseract_ocr_model.py,sha256=v6td0vq8NogePuRTJRZhKF0DtZXITj70r9rKJKO5u9k,4984
30
- docling/pipeline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
31
- docling/pipeline/base_pipeline.py,sha256=7DTzVvM_jVHCxyY-BuuGRhmUsD_sgX4DD00oBFJWdB8,6723
32
- docling/pipeline/simple_pipeline.py,sha256=pxce0-3He5Lqa-xXT-7h173XVOSMZiMHl6HOfAJmQ7o,2162
33
- docling/pipeline/standard_pdf_pipeline.py,sha256=AVNSxGc6kPmBPDLWDc9eI8fryc25eOtiIVrOyVhZMZM,7527
34
- docling/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
35
- docling/utils/export.py,sha256=KyGF1BVDHPFfHVXZc8vegsWlFfOgGPP2YckWpTadyI8,4694
36
- docling/utils/layout_utils.py,sha256=vlN0rc8i0ayRGn3WnaG-pdmqEL00KKGl2zez3Gj-hrk,32074
37
- docling/utils/utils.py,sha256=llhXSbIDNZ1MHOwBEfLHBAoJIAYI7QlPIonlI1jLUJ0,1208
38
- docling-2.1.0.dist-info/LICENSE,sha256=mBb7ErEcM8VS9OhiGHnQ2kk75HwPhr54W1Oiz3965MY,1088
39
- docling-2.1.0.dist-info/METADATA,sha256=SorLD4OMK1dU3bX5eqnw5GHqPrPwdhQ7JfYvOyajE20,6109
40
- docling-2.1.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
41
- docling-2.1.0.dist-info/entry_points.txt,sha256=VOSzV77znM52dz5ysaDuJ0ijl1cnfrh1ZPg8od5OcTs,48
42
- docling-2.1.0.dist-info/RECORD,,