natural-pdf 0.1.38__py3-none-any.whl → 0.2.0__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 (55) hide show
  1. natural_pdf/__init__.py +11 -6
  2. natural_pdf/analyzers/__init__.py +6 -1
  3. natural_pdf/analyzers/guides.py +354 -258
  4. natural_pdf/analyzers/layout/layout_analyzer.py +2 -3
  5. natural_pdf/analyzers/layout/layout_manager.py +18 -4
  6. natural_pdf/analyzers/layout/paddle.py +11 -0
  7. natural_pdf/analyzers/layout/surya.py +2 -3
  8. natural_pdf/analyzers/shape_detection_mixin.py +25 -34
  9. natural_pdf/analyzers/text_structure.py +2 -2
  10. natural_pdf/classification/manager.py +1 -1
  11. natural_pdf/collections/mixins.py +3 -2
  12. natural_pdf/core/highlighting_service.py +743 -32
  13. natural_pdf/core/page.py +252 -399
  14. natural_pdf/core/page_collection.py +1249 -0
  15. natural_pdf/core/pdf.py +231 -89
  16. natural_pdf/{collections → core}/pdf_collection.py +18 -11
  17. natural_pdf/core/render_spec.py +335 -0
  18. natural_pdf/describe/base.py +1 -1
  19. natural_pdf/elements/__init__.py +1 -0
  20. natural_pdf/elements/base.py +108 -83
  21. natural_pdf/elements/{collections.py → element_collection.py} +575 -1372
  22. natural_pdf/elements/line.py +0 -1
  23. natural_pdf/elements/rect.py +0 -1
  24. natural_pdf/elements/region.py +405 -280
  25. natural_pdf/elements/text.py +9 -7
  26. natural_pdf/exporters/base.py +2 -2
  27. natural_pdf/exporters/original_pdf.py +1 -1
  28. natural_pdf/exporters/paddleocr.py +2 -4
  29. natural_pdf/exporters/searchable_pdf.py +3 -2
  30. natural_pdf/extraction/mixin.py +1 -3
  31. natural_pdf/flows/collections.py +1 -69
  32. natural_pdf/flows/element.py +25 -0
  33. natural_pdf/flows/flow.py +1658 -19
  34. natural_pdf/flows/region.py +757 -263
  35. natural_pdf/ocr/ocr_options.py +0 -2
  36. natural_pdf/ocr/utils.py +2 -1
  37. natural_pdf/qa/document_qa.py +21 -5
  38. natural_pdf/search/search_service_protocol.py +1 -1
  39. natural_pdf/selectors/parser.py +35 -2
  40. natural_pdf/tables/result.py +35 -1
  41. natural_pdf/text_mixin.py +101 -0
  42. natural_pdf/utils/debug.py +2 -1
  43. natural_pdf/utils/highlighting.py +1 -0
  44. natural_pdf/utils/layout.py +2 -2
  45. natural_pdf/utils/packaging.py +4 -3
  46. natural_pdf/utils/text_extraction.py +15 -12
  47. natural_pdf/utils/visualization.py +385 -0
  48. {natural_pdf-0.1.38.dist-info → natural_pdf-0.2.0.dist-info}/METADATA +7 -3
  49. {natural_pdf-0.1.38.dist-info → natural_pdf-0.2.0.dist-info}/RECORD +55 -52
  50. optimization/memory_comparison.py +1 -1
  51. optimization/pdf_analyzer.py +2 -2
  52. {natural_pdf-0.1.38.dist-info → natural_pdf-0.2.0.dist-info}/WHEEL +0 -0
  53. {natural_pdf-0.1.38.dist-info → natural_pdf-0.2.0.dist-info}/entry_points.txt +0 -0
  54. {natural_pdf-0.1.38.dist-info → natural_pdf-0.2.0.dist-info}/licenses/LICENSE +0 -0
  55. {natural_pdf-0.1.38.dist-info → natural_pdf-0.2.0.dist-info}/top_level.txt +0 -0
natural_pdf/elements/line.py CHANGED
@@ -102,7 +102,6 @@ class LineElement(Element):
102
102
  elif self.is_vertical:
103
103
  return "vertical"
104
104
 
105
-
106
105
  def extract_text(self, keep_blank_chars=True, apply_exclusions=True, **kwargs) -> str:
107
106
  """
108
107
  Lines don't have text, so this returns an empty string.
natural_pdf/elements/rect.py CHANGED
@@ -88,7 +88,6 @@ class RectangleElement(Element):
88
88
  """Get the stroke width of the rectangle."""
89
89
  return self._obj.get("linewidth", 0)
90
90
 
91
-
92
91
  def extract_text(self, **kwargs) -> str:
93
92
  """
94
93
  Extract text from inside this rectangle.