kreuzberg 3.6.2__py3-none-any.whl → 3.8.0__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 (36) hide show
  1. kreuzberg/_extractors/_base.py +40 -0
  2. kreuzberg/_extractors/_email.py +149 -0
  3. kreuzberg/_extractors/_html.py +15 -3
  4. kreuzberg/_extractors/_image.py +17 -18
  5. kreuzberg/_extractors/_pdf.py +68 -14
  6. kreuzberg/_extractors/_presentation.py +62 -10
  7. kreuzberg/_extractors/_spread_sheet.py +179 -4
  8. kreuzberg/_extractors/_structured.py +148 -0
  9. kreuzberg/_gmft.py +2 -2
  10. kreuzberg/_mcp/__init__.py +5 -0
  11. kreuzberg/_mcp/server.py +227 -0
  12. kreuzberg/_mime_types.py +27 -1
  13. kreuzberg/_multiprocessing/__init__.py +2 -3
  14. kreuzberg/_ocr/__init__.py +30 -0
  15. kreuzberg/{_multiprocessing/tesseract_pool.py → _ocr/_pool.py} +3 -5
  16. kreuzberg/_ocr/_sync.py +566 -0
  17. kreuzberg/_ocr/_tesseract.py +6 -2
  18. kreuzberg/_registry.py +4 -0
  19. kreuzberg/_types.py +131 -0
  20. kreuzberg/_utils/_cache.py +17 -2
  21. kreuzberg/_utils/_process_pool.py +178 -1
  22. kreuzberg/_utils/_quality.py +237 -0
  23. kreuzberg/_utils/_serialization.py +4 -2
  24. kreuzberg/_utils/_string.py +153 -10
  25. kreuzberg/_utils/_sync.py +5 -2
  26. kreuzberg/_utils/_table.py +261 -0
  27. {kreuzberg-3.6.2.dist-info → kreuzberg-3.8.0.dist-info}/METADATA +116 -48
  28. kreuzberg-3.8.0.dist-info/RECORD +57 -0
  29. {kreuzberg-3.6.2.dist-info → kreuzberg-3.8.0.dist-info}/entry_points.txt +1 -0
  30. kreuzberg/_multiprocessing/process_manager.py +0 -189
  31. kreuzberg/_multiprocessing/sync_easyocr.py +0 -235
  32. kreuzberg/_multiprocessing/sync_paddleocr.py +0 -199
  33. kreuzberg/_multiprocessing/sync_tesseract.py +0 -261
  34. kreuzberg-3.6.2.dist-info/RECORD +0 -54
  35. {kreuzberg-3.6.2.dist-info → kreuzberg-3.8.0.dist-info}/WHEEL +0 -0
  36. {kreuzberg-3.6.2.dist-info → kreuzberg-3.8.0.dist-info}/licenses/LICENSE +0 -0
@@ -1,54 +0,0 @@
1
- kreuzberg/__init__.py,sha256=wVxbug-w1cO2xHcP04Bf6QeIKmT2Ep6aeenb8EOYLA0,1534
2
- kreuzberg/__main__.py,sha256=s2qM1nPEkRHAQP-G3P7sf5l6qA_KJeIEHS5LpPz04lg,183
3
- kreuzberg/_chunker.py,sha256=2eHSRHcZdJ2ZjR3in49y3o9tPl5HMO3vkbnMqaVCbHI,1887
4
- kreuzberg/_cli_config.py,sha256=WD_seFjbuay_NJv77vGLBW6BVV9WZNujdzf3zQkhzPc,5691
5
- kreuzberg/_constants.py,sha256=Bxc8oiN-wHwnWXT9bEiJhTUcu1ygPpra5qHirAif3b4,191
6
- kreuzberg/_entity_extraction.py,sha256=EIasBGpkZ-3FwivjEpisz23LilTwx8os-IbfrDtzNl4,7815
7
- kreuzberg/_gmft.py,sha256=e-UpYwizRX_V-dn0a7ja0Z9nShAmDKA1Q7HThJy8cyA,14856
8
- kreuzberg/_language_detection.py,sha256=22-uXoOu_ws0K8Hz2M7U_SF9QX3npRYLhntAE1dNLFU,3283
9
- kreuzberg/_mime_types.py,sha256=QgX-k8aI4lTKArObDM0TFPt7DUjUVwWrdIaIZDh_XQY,7815
10
- kreuzberg/_playa.py,sha256=rU6ii2Qnrj8tkDYlSiab5h-BCYLJnUg4QwSLVDEXF5g,11883
11
- kreuzberg/_registry.py,sha256=c2B_PJbaL0q3ab2eNmj_0jldeyMaqgvRwkZqUU4MM5Q,3290
12
- kreuzberg/_types.py,sha256=U72a4SXS1e-zV8cXG0tiozMy9mX9wFM1ma6sVz7HpJo,9936
13
- kreuzberg/cli.py,sha256=S0w2nGXBWPFn1NhxppW7dpUwB9f_3ymFuWSAB2aRu9g,12465
14
- kreuzberg/exceptions.py,sha256=xRaiJh11i8E6Nc-gAQPgNW5xvhiiFBhRS-CBbCEbHQM,2881
15
- kreuzberg/extraction.py,sha256=mdH45bMAAUUNXYT7UrNyWJ2oD_gXuLUU-NyuYxQM884,17459
16
- kreuzberg/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
17
- kreuzberg/_api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
18
- kreuzberg/_api/main.py,sha256=kZCMPPzP4BGzEege9pdhQTJPKKVjCaC6kZdMMeaqP2M,2599
19
- kreuzberg/_extractors/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
20
- kreuzberg/_extractors/_base.py,sha256=YUr6A2n34LlFzbYQkiKqhXAphL9RYrvAls5SlkoQqNg,3028
21
- kreuzberg/_extractors/_html.py,sha256=L_vcKyv1ObI6irPaD3-WTKqxeRfZA4Rhsl3zUiAe_ws,1312
22
- kreuzberg/_extractors/_image.py,sha256=pYfh3x9CkiIxOLvp0jkkZcmLbB_FpdfDo01klSc6OzQ,4819
23
- kreuzberg/_extractors/_pandoc.py,sha256=oQ4DgQSPoX1LXjGAKh_A40JHqiKWb91LeRBYSS_6EUA,26750
24
- kreuzberg/_extractors/_pdf.py,sha256=R33ggTd0IU6NsEnzgHFTr9ScgcnM8nIIstDq7XMVcvg,14792
25
- kreuzberg/_extractors/_presentation.py,sha256=ZX-EKQppHwvKtyKk0-IQVF6QAqJi0SfGgCiiyqMQh0w,8701
26
- kreuzberg/_extractors/_spread_sheet.py,sha256=HOzCeYQc6kaMveAHfi80LrsF0yU7Kn74aKQ7lrMAlo8,6480
27
- kreuzberg/_multiprocessing/__init__.py,sha256=nwYQpKH7ixHwzkQbTMFCstOCBKktmbNq5dTrwI2Mn94,203
28
- kreuzberg/_multiprocessing/gmft_isolated.py,sha256=ZfbhiL5bhBEJnibUSls3WV-FECrnU9VvKfq5O2foHcc,11191
29
- kreuzberg/_multiprocessing/process_manager.py,sha256=_qtB8y9td2coJevlIl4z6F__jau320RdI1lqdyuaeD4,6061
30
- kreuzberg/_multiprocessing/sync_easyocr.py,sha256=-3_Ol0H8G6RhPxTbTPvoe8fTsTz3e-dg2QbHHnoJL48,7693
31
- kreuzberg/_multiprocessing/sync_paddleocr.py,sha256=5558iTjPXCyJWuyhZckmuJLadUwJDb5YVC8Cv-FOaWg,6090
32
- kreuzberg/_multiprocessing/sync_tesseract.py,sha256=Ck1PvHGWOMQWUcC7RyVrBt8K9VDFQ0lQcwFkwYzl3rE,8240
33
- kreuzberg/_multiprocessing/tesseract_pool.py,sha256=UN7BtS_ib1ux9xuR6d6AB3PY7UEUhd-5Ti1n1H0UnYw,10945
34
- kreuzberg/_ocr/__init__.py,sha256=VTqwKDlIRbjve71Y11Ztygyhv5aWG9LWTj8iX66ANxE,533
35
- kreuzberg/_ocr/_base.py,sha256=lNT0Tin4hzbmaamqqySxvYEwNtrJB5gGlStrANQQcyc,1637
36
- kreuzberg/_ocr/_easyocr.py,sha256=90Dv1xaLXbpG7EtmRQE5ykvnhqZJR3xSFXlxFMCSVSI,13740
37
- kreuzberg/_ocr/_paddleocr.py,sha256=UvugDdZd7RojHUiFeBaI8aqz36ecegPLj2v6oT6c42g,13776
38
- kreuzberg/_ocr/_tesseract.py,sha256=3s3MkZN9xA_Uedx4s2p5m4IEIMhGjs9gYHxan9Iz-2g,13044
39
- kreuzberg/_utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
40
- kreuzberg/_utils/_cache.py,sha256=JGiwwcNBoD950IbsPUUAD5gAGS7byUuz0BqYSneVakc,13088
41
- kreuzberg/_utils/_device.py,sha256=rnaSSB5ibf2wr7EDxrcmOUZ4Ocor0pHkwb3N1pC46EY,10276
42
- kreuzberg/_utils/_document_cache.py,sha256=z8irioKsOu8xve1YgHatm__wIFvs9I1gDK3tLNsNyqM,6926
43
- kreuzberg/_utils/_errors.py,sha256=AV3oaRQDgJxe1YUZd9pCQUysUv9KW8Ib37MvnyFOZ4o,6386
44
- kreuzberg/_utils/_pdf_lock.py,sha256=nqxAYCNlfWDrJtP4ZNu57st1YnkDl-gYXdr0q8nv0kA,1961
45
- kreuzberg/_utils/_process_pool.py,sha256=-0SNP01Qz21D7hgJmN0eHoqKusSygwPbi1U7IzJlPio,2895
46
- kreuzberg/_utils/_serialization.py,sha256=AhZvyAu4KsjAqyZDh--Kn2kSWGgCuH7udio8lTklO0g,2132
47
- kreuzberg/_utils/_string.py,sha256=owIVkUtP0__GiJD9RIJzPdvyIigT5sQho3mOXPbsnW0,958
48
- kreuzberg/_utils/_sync.py,sha256=oT4Y_cDBKtE_BFEoLTae3rSisqlYXzW-jlUG_x-dmLM,4725
49
- kreuzberg/_utils/_tmp.py,sha256=hVn-VVijIg2FM7EZJ899gc7wZg-TGoJZoeAcxMX-Cxg,1044
50
- kreuzberg-3.6.2.dist-info/METADATA,sha256=shguv5yge8FkD9aT0x02dRdLpuLi1PW4SmczFYiILmU,9910
51
- kreuzberg-3.6.2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
52
- kreuzberg-3.6.2.dist-info/entry_points.txt,sha256=VdoFaTl3QSvVWOZcIlPpDd47o6kn7EvmXSs8FI0ExLc,48
53
- kreuzberg-3.6.2.dist-info/licenses/LICENSE,sha256=-8caMvpCK8SgZ5LlRKhGCMtYDEXqTKH9X8pFEhl91_4,1066
54
- kreuzberg-3.6.2.dist-info/RECORD,,