polytext 0.2.1__tar.gz → 0.2.2__tar.gz

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 (87) hide show
  1. {polytext-0.2.1 → polytext-0.2.2}/PKG-INFO +4 -3
  2. {polytext-0.2.1 → polytext-0.2.2}/README.md +1 -1
  3. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/base.py +9 -0
  4. {polytext-0.2.1 → polytext-0.2.2}/polytext.egg-info/PKG-INFO +4 -3
  5. {polytext-0.2.1 → polytext-0.2.2}/pyproject.toml +2 -2
  6. {polytext-0.2.1 → polytext-0.2.2}/setup.py +3 -2
  7. {polytext-0.2.1 → polytext-0.2.2}/LICENSE +0 -0
  8. {polytext-0.2.1 → polytext-0.2.2}/polytext/__init__.py +0 -0
  9. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/__init__.py +0 -0
  10. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/audio_to_text.py +0 -0
  11. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/base.py +0 -0
  12. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/document_ocr_to_text.py +0 -0
  13. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/document_ocr_to_text_azure_oai.py +0 -0
  14. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/gemini_quality_guards.py +0 -0
  15. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/html_to_md.py +0 -0
  16. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/md_to_text.py +0 -0
  17. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/ocr_to_text.py +0 -0
  18. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/ocr_to_text_azure_oai.py +0 -0
  19. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/pdf.py +0 -0
  20. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/text_to_md.py +0 -0
  21. {polytext-0.2.1 → polytext-0.2.2}/polytext/converter/video_to_audio.py +0 -0
  22. {polytext-0.2.1 → polytext-0.2.2}/polytext/exceptions/__init__.py +0 -0
  23. {polytext-0.2.1 → polytext-0.2.2}/polytext/exceptions/base.py +0 -0
  24. {polytext-0.2.1 → polytext-0.2.2}/polytext/generator/__init__.py +0 -0
  25. {polytext-0.2.1 → polytext-0.2.2}/polytext/generator/pdf.py +0 -0
  26. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/__init__.py +0 -0
  27. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/audio.py +0 -0
  28. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/document.py +0 -0
  29. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/document_ocr.py +0 -0
  30. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/downloader/__init__.py +0 -0
  31. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/downloader/downloader.py +0 -0
  32. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/html.py +0 -0
  33. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/markdown.py +0 -0
  34. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/notebook.py +0 -0
  35. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/ocr.py +0 -0
  36. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/plain_text.py +0 -0
  37. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/video.py +0 -0
  38. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/xml_xbrl.py +0 -0
  39. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/youtube.py +0 -0
  40. {polytext-0.2.1 → polytext-0.2.2}/polytext/loader/youtube_llm.py +0 -0
  41. {polytext-0.2.1 → polytext-0.2.2}/polytext/processor/__init__.py +0 -0
  42. {polytext-0.2.1 → polytext-0.2.2}/polytext/processor/audio_chunker.py +0 -0
  43. {polytext-0.2.1 → polytext-0.2.2}/polytext/processor/text_merger.py +0 -0
  44. {polytext-0.2.1 → polytext-0.2.2}/polytext/processor/transcript_chunker.py +0 -0
  45. {polytext-0.2.1 → polytext-0.2.2}/polytext/prompts/__init__.py +0 -0
  46. {polytext-0.2.1 → polytext-0.2.2}/polytext/prompts/ocr.py +0 -0
  47. {polytext-0.2.1 → polytext-0.2.2}/polytext/prompts/text_merging.py +0 -0
  48. {polytext-0.2.1 → polytext-0.2.2}/polytext/prompts/text_to_md.py +0 -0
  49. {polytext-0.2.1 → polytext-0.2.2}/polytext/prompts/transcription.py +0 -0
  50. {polytext-0.2.1 → polytext-0.2.2}/polytext/utils/__init__.py +0 -0
  51. {polytext-0.2.1 → polytext-0.2.2}/polytext/utils/utils.py +0 -0
  52. {polytext-0.2.1 → polytext-0.2.2}/polytext.egg-info/SOURCES.txt +0 -0
  53. {polytext-0.2.1 → polytext-0.2.2}/polytext.egg-info/dependency_links.txt +0 -0
  54. {polytext-0.2.1 → polytext-0.2.2}/polytext.egg-info/not-zip-safe +0 -0
  55. {polytext-0.2.1 → polytext-0.2.2}/polytext.egg-info/requires.txt +0 -0
  56. {polytext-0.2.1 → polytext-0.2.2}/polytext.egg-info/top_level.txt +0 -0
  57. {polytext-0.2.1 → polytext-0.2.2}/setup.cfg +0 -0
  58. {polytext-0.2.1 → polytext-0.2.2}/tests/test_audio_chunker.py +0 -0
  59. {polytext-0.2.1 → polytext-0.2.2}/tests/test_audio_comparison_helpers.py +0 -0
  60. {polytext-0.2.1 → polytext-0.2.2}/tests/test_audio_transcription_model_migration.py +0 -0
  61. {polytext-0.2.1 → polytext-0.2.2}/tests/test_compare_audio_models.py +0 -0
  62. {polytext-0.2.1 → polytext-0.2.2}/tests/test_compare_document_ocr_to_text_models.py +0 -0
  63. {polytext-0.2.1 → polytext-0.2.2}/tests/test_compare_ocr_to_text_models.py +0 -0
  64. {polytext-0.2.1 → polytext-0.2.2}/tests/test_compare_youtube_models.py +0 -0
  65. {polytext-0.2.1 → polytext-0.2.2}/tests/test_dowload_audio_from_youtube.py +0 -0
  66. {polytext-0.2.1 → polytext-0.2.2}/tests/test_dowload_audio_from_youtube_helpers.py +0 -0
  67. {polytext-0.2.1 → polytext-0.2.2}/tests/test_extracted_text_whitespace.py +0 -0
  68. {polytext-0.2.1 → polytext-0.2.2}/tests/test_get_audio_transcript_from_gcs.py +0 -0
  69. {polytext-0.2.1 → polytext-0.2.2}/tests/test_get_customized_pdf_from_markdown.py +0 -0
  70. {polytext-0.2.1 → polytext-0.2.2}/tests/test_get_document_ocr.py +0 -0
  71. {polytext-0.2.1 → polytext-0.2.2}/tests/test_get_document_ocr_azure_oai.py +0 -0
  72. {polytext-0.2.1 → polytext-0.2.2}/tests/test_get_document_text.py +0 -0
  73. {polytext-0.2.1 → polytext-0.2.2}/tests/test_get_document_text_from_gcs.py +0 -0
  74. {polytext-0.2.1 → polytext-0.2.2}/tests/test_get_ocr_from_image.py +0 -0
  75. {polytext-0.2.1 → polytext-0.2.2}/tests/test_get_text_from_markdown.py +0 -0
  76. {polytext-0.2.1 → polytext-0.2.2}/tests/test_get_video_transcript_from_gcs.py +0 -0
  77. {polytext-0.2.1 → polytext-0.2.2}/tests/test_library.py +0 -0
  78. {polytext-0.2.1 → polytext-0.2.2}/tests/test_markdown_loader_gzip.py +0 -0
  79. {polytext-0.2.1 → polytext-0.2.2}/tests/test_markitdown_html.py +0 -0
  80. {polytext-0.2.1 → polytext-0.2.2}/tests/test_notebook_loader.py +0 -0
  81. {polytext-0.2.1 → polytext-0.2.2}/tests/test_ocr_fallbacks.py +0 -0
  82. {polytext-0.2.1 → polytext-0.2.2}/tests/test_ocr_image_descriptions.py +0 -0
  83. {polytext-0.2.1 → polytext-0.2.2}/tests/test_pain_text.py +0 -0
  84. {polytext-0.2.1 → polytext-0.2.2}/tests/test_split_audio_with_llm.py +0 -0
  85. {polytext-0.2.1 → polytext-0.2.2}/tests/test_xml_xbrl_loader.py +0 -0
  86. {polytext-0.2.1 → polytext-0.2.2}/tests/test_youtube_gemini_minimal_check.py +0 -0
  87. {polytext-0.2.1 → polytext-0.2.2}/tests/test_youtube_transcript.py +0 -0
{polytext-0.2.1 → polytext-0.2.2}/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: polytext
3
- Version: 0.2.1
3
+ Version: 0.2.2
4
4
  Summary: Python utilities to simplify document files management
5
5
  Home-page: https://github.com/docsity/polytext
6
6
  Author: Matteo Senardi
@@ -8,11 +8,12 @@ Author-email: matteo.s@docsity.com
8
8
  License: MIT
9
9
  Classifier: License :: OSI Approved :: MIT License
10
10
  Classifier: Operating System :: OS Independent
11
+ Classifier: Programming Language :: Python :: 3.11
11
12
  Classifier: Programming Language :: Python :: 3.12
12
13
  Classifier: Programming Language :: Python :: 3.13
13
14
  Classifier: Topic :: Software Development :: Libraries :: Application Frameworks
14
15
  Classifier: Topic :: Software Development :: Libraries :: Python Modules
15
- Requires-Python: >=3.12
16
+ Requires-Python: >=3.11
16
17
  Description-Content-Type: text/markdown
17
18
  License-File: LICENSE
18
19
  Requires-Dist: pypdf==5.5.0
@@ -90,7 +91,7 @@ pip install polytext
90
91
 
91
92
  | Requirement | Notes | macOS (Homebrew) | Ubuntu / Debian |
92
93
  |-------------|---------------------------------------------------------------------------------|------------------|-----------------|
93
- | **Python** | ✔️ Tested on **3.12**<br> Older versions may fail to locate WeasyPrint’s dylibs | `brew install python@3.12` | `sudo apt install python3.12` |
94
+ | **Python** | Supported on **3.11 – 3.13**<br> WeasyPrint still requires its native libraries | `brew install python@3.11` | `sudo apt install python3.11` |
94
95
  | **WeasyPrint – native stack** | installs Pango, Cairo, etc. | `brew install weasyprint` | `sudo apt install weasyprint` |
95
96
  | **LibreOffice** | used for Office → PDF conversion | `brew install --cask libreoffice` | `sudo apt install libreoffice` |
96
97
 
{polytext-0.2.1 → polytext-0.2.2}/README.md RENAMED
@@ -35,7 +35,7 @@ pip install polytext
35
35
 
36
36
  | Requirement | Notes | macOS (Homebrew) | Ubuntu / Debian |
37
37
  |-------------|---------------------------------------------------------------------------------|------------------|-----------------|
38
- | **Python** | ✔️ Tested on **3.12**<br> Older versions may fail to locate WeasyPrint’s dylibs | `brew install python@3.12` | `sudo apt install python3.12` |
38
+ | **Python** | Supported on **3.11 – 3.13**<br> WeasyPrint still requires its native libraries | `brew install python@3.11` | `sudo apt install python3.11` |
39
39
  | **WeasyPrint – native stack** | installs Pango, Cairo, etc. | `brew install weasyprint` | `sudo apt install weasyprint` |
40
40
  | **LibreOffice** | used for Office → PDF conversion | `brew install --cask libreoffice` | `sudo apt install libreoffice` |
41
41
 
{polytext-0.2.1 → polytext-0.2.2}/polytext/loader/base.py RENAMED
@@ -317,6 +317,15 @@ class BaseLoader:
317
317
  return YoutubeTranscriptLoaderWithLlm(llm_api_key=llm_api_key, markdown_output=self.markdown_output, temp_dir=self.temp_dir, timeout_minutes=self.timeout_minutes, **kwargs)
318
318
  else:
319
319
  return HtmlLoader(markdown_output=self.markdown_output)
320
+ # Handle markdown files based on extension or MIME type
321
+ if file_extension in [".md", ".markdown"] or (
322
+ mime_type and mime_type.startswith("text/markdown")
323
+ ):
324
+ return MarkdownLoader(
325
+ markdown_output=self.markdown_output,
326
+ temp_dir=self.temp_dir,
327
+ **kwargs,
328
+ )
320
329
  elif mime_type:
321
330
  if file_extension in [".pdf", ".xlsx", ".docx", ".txt", ".csv", ".odt", ".pptx", ".xls", ".doc", ".ppt", ".rtf"]:
322
331
  return DocumentLoader(markdown_output=self.markdown_output, temp_dir=self.temp_dir, timeout_minutes=self.timeout_minutes, **kwargs)
{polytext-0.2.1 → polytext-0.2.2}/polytext.egg-info/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: polytext
3
- Version: 0.2.1
3
+ Version: 0.2.2
4
4
  Summary: Python utilities to simplify document files management
5
5
  Home-page: https://github.com/docsity/polytext
6
6
  Author: Matteo Senardi
@@ -8,11 +8,12 @@ Author-email: matteo.s@docsity.com
8
8
  License: MIT
9
9
  Classifier: License :: OSI Approved :: MIT License
10
10
  Classifier: Operating System :: OS Independent
11
+ Classifier: Programming Language :: Python :: 3.11
11
12
  Classifier: Programming Language :: Python :: 3.12
12
13
  Classifier: Programming Language :: Python :: 3.13
13
14
  Classifier: Topic :: Software Development :: Libraries :: Application Frameworks
14
15
  Classifier: Topic :: Software Development :: Libraries :: Python Modules
15
- Requires-Python: >=3.12
16
+ Requires-Python: >=3.11
16
17
  Description-Content-Type: text/markdown
17
18
  License-File: LICENSE
18
19
  Requires-Dist: pypdf==5.5.0
@@ -90,7 +91,7 @@ pip install polytext
90
91
 
91
92
  | Requirement | Notes | macOS (Homebrew) | Ubuntu / Debian |
92
93
  |-------------|---------------------------------------------------------------------------------|------------------|-----------------|
93
- | **Python** | ✔️ Tested on **3.12**<br> Older versions may fail to locate WeasyPrint’s dylibs | `brew install python@3.12` | `sudo apt install python3.12` |
94
+ | **Python** | Supported on **3.11 – 3.13**<br> WeasyPrint still requires its native libraries | `brew install python@3.11` | `sudo apt install python3.11` |
94
95
  | **WeasyPrint – native stack** | installs Pango, Cairo, etc. | `brew install weasyprint` | `sudo apt install weasyprint` |
95
96
  | **LibreOffice** | used for Office → PDF conversion | `brew install --cask libreoffice` | `sudo apt install libreoffice` |
96
97
 
{polytext-0.2.1 → polytext-0.2.2}/pyproject.toml RENAMED
@@ -7,9 +7,9 @@ readme = "README.md"
7
7
  build = "setup.py"
8
8
 
9
9
  [tool.poetry.dependencies]
10
- python = ">=3.12,<3.14"
10
+ python = ">=3.11,<3.14"
11
11
  pybind11 = "^2.12"
12
12
 
13
13
  [build-system]
14
14
  requires = ["setuptools>=68", "wheel"]
15
- build-backend = "setuptools.build_meta:__legacy__"
15
+ build-backend = "setuptools.build_meta:__legacy__"
{polytext-0.2.1 → polytext-0.2.2}/setup.py RENAMED
@@ -51,7 +51,7 @@ def get_requirements(*requirements_file):
51
51
 
52
52
  setup(
53
53
  name='polytext',
54
- version='0.2.1',
54
+ version='0.2.2',
55
55
  url='https://github.com/docsity/polytext',
56
56
  # download_url='https://github.com/pualien/py-polytext/archive/0.1.23.tar.gz',
57
57
  license='MIT',
@@ -65,10 +65,11 @@ setup(
65
65
  },
66
66
  long_description=open('README.md').read(),
67
67
  long_description_content_type='text/markdown',
68
- python_requires='>=3.12',
68
+ python_requires='>=3.11',
69
69
  classifiers=[
70
70
  "License :: OSI Approved :: MIT License",
71
71
  "Operating System :: OS Independent",
72
+ "Programming Language :: Python :: 3.11",
72
73
  "Programming Language :: Python :: 3.12",
73
74
  "Programming Language :: Python :: 3.13",
74
75
  "Topic :: Software Development :: Libraries :: Application Frameworks",
{polytext-0.2.1 → polytext-0.2.2}/LICENSE RENAMED
File without changes
{polytext-0.2.1 → polytext-0.2.2}/polytext/__init__.py RENAMED
File without changes
{polytext-0.2.1 → polytext-0.2.2}/setup.cfg RENAMED
File without changes
{polytext-0.2.1 → polytext-0.2.2}/tests/test_library.py RENAMED
File without changes