kreuzberg 3.11.2__py3-none-any.whl → 3.11.4__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.
kreuzberg/cli.py CHANGED
@@ -19,11 +19,9 @@ except ImportError as e: # pragma: no cover
19
19
 
20
20
  from kreuzberg import __version__, extract_bytes_sync, extract_file_sync
21
21
  from kreuzberg._config import build_extraction_config, find_config_file, load_config_from_file
22
+ from kreuzberg._constants import DEFAULT_MAX_CHARACTERS, DEFAULT_MAX_OVERLAP
22
23
  from kreuzberg.exceptions import KreuzbergError, MissingDependencyError
23
24
 
24
- DEFAULT_MAX_CHARACTERS = 4000
25
- DEFAULT_MAX_OVERLAP = 200
26
-
27
25
  if TYPE_CHECKING:
28
26
  from kreuzberg._types import ExtractionConfig, ExtractionResult
29
27
 
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: kreuzberg
3
- Version: 3.11.2
3
+ Version: 3.11.4
4
4
  Summary: Document intelligence framework for Python - Extract text, metadata, and structured data from diverse file formats
5
5
  Project-URL: documentation, https://kreuzberg.dev
6
6
  Project-URL: homepage, https://github.com/Goldziher/kreuzberg
@@ -52,9 +52,9 @@ Requires-Dist: gmft>=0.4.2; extra == 'all'
52
52
  Requires-Dist: keybert>=0.9.0; extra == 'all'
53
53
  Requires-Dist: litestar[opentelemetry,standard,structlog]>=2.17.0; extra == 'all'
54
54
  Requires-Dist: mailparse>=1.0.15; extra == 'all'
55
- Requires-Dist: paddleocr>=3.1.1; extra == 'all'
56
- Requires-Dist: paddlepaddle>=3.1.0; extra == 'all'
57
- Requires-Dist: pandas>=2.3.1; extra == 'all'
55
+ Requires-Dist: paddleocr>=3.2.0; extra == 'all'
56
+ Requires-Dist: paddlepaddle>=3.1.1; extra == 'all'
57
+ Requires-Dist: pandas>=2.3.2; extra == 'all'
58
58
  Requires-Dist: playa-pdf[crypto]>=0.7.0; extra == 'all'
59
59
  Requires-Dist: rich>=14.1.0; extra == 'all'
60
60
  Requires-Dist: semantic-text-splitter>=0.27.0; extra == 'all'
@@ -73,7 +73,7 @@ Provides-Extra: crypto
73
73
  Requires-Dist: playa-pdf[crypto]>=0.7.0; extra == 'crypto'
74
74
  Provides-Extra: document-classification
75
75
  Requires-Dist: deep-translator>=1.11.4; extra == 'document-classification'
76
- Requires-Dist: pandas>=2.3.1; extra == 'document-classification'
76
+ Requires-Dist: pandas>=2.3.2; extra == 'document-classification'
77
77
  Provides-Extra: easyocr
78
78
  Requires-Dist: easyocr>=1.7.2; extra == 'easyocr'
79
79
  Provides-Extra: entity-extraction
@@ -84,8 +84,8 @@ Requires-Dist: gmft>=0.4.2; extra == 'gmft'
84
84
  Provides-Extra: langdetect
85
85
  Requires-Dist: fast-langdetect>=0.3.2; extra == 'langdetect'
86
86
  Provides-Extra: paddleocr
87
- Requires-Dist: paddleocr>=3.1.1; extra == 'paddleocr'
88
- Requires-Dist: paddlepaddle>=3.1.0; extra == 'paddleocr'
87
+ Requires-Dist: paddleocr>=3.2.0; extra == 'paddleocr'
88
+ Requires-Dist: paddlepaddle>=3.1.1; extra == 'paddleocr'
89
89
  Requires-Dist: setuptools>=80.9.0; extra == 'paddleocr'
90
90
  Description-Content-Type: text/markdown
91
91
 
@@ -11,7 +11,7 @@ kreuzberg/_mime_types.py,sha256=2warRVqfBUNIg8JBg8yP4pRqaMPvwINosHMkJwtH_Fc,8488
11
11
  kreuzberg/_playa.py,sha256=_IPrUSWwSfDQlWXOpKlauV0D9MhGrujGP5kmQ0U3L0g,12188
12
12
  kreuzberg/_registry.py,sha256=wGSlkS0U1zqruWQCLE95vj4a2mw1yyvf0j6rgz80sJg,3473
13
13
  kreuzberg/_types.py,sha256=bMaU6VuoqwOpW6ufshA-DWpNw6t9EokjEDEfFsznvdo,15389
14
- kreuzberg/cli.py,sha256=rJMdHg7FhUxefCrx-sf4c2qVGRXr8Xrpjgfx_DQSKMg,12558
14
+ kreuzberg/cli.py,sha256=nG1CD_h50EWLmDbrb0_DffRl25uTCKeCS6_gRVpjEdU,12578
15
15
  kreuzberg/exceptions.py,sha256=PTiAZgQwcG9hXbgYg2W7sfxksFhq5_wzOFgZGnTJAoc,2991
16
16
  kreuzberg/extraction.py,sha256=Kt1mOxdlOb35yVOdpdhiRPuTgA9BW_TTG9qwCkSxSkc,17332
17
17
  kreuzberg/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -47,8 +47,8 @@ kreuzberg/_utils/_string.py,sha256=bCzO3UO6nXupxvtMWvHqfp1Vd9CTzEH9jmpJXQ7upAU,6
47
47
  kreuzberg/_utils/_sync.py,sha256=7LSavBmxVKQUzdjfx9fYRAI9IbJtRw8iGf_Q8B7RX9g,4923
48
48
  kreuzberg/_utils/_table.py,sha256=IomrfQBP85DZI8RmQjOVs2Siq7VP9FUTYPaZR4t3yRw,8199
49
49
  kreuzberg/_utils/_tmp.py,sha256=hVn-VVijIg2FM7EZJ899gc7wZg-TGoJZoeAcxMX-Cxg,1044
50
- kreuzberg-3.11.2.dist-info/METADATA,sha256=J-UWkai5WTH0ECDX_kcp-1H45Qxa1rFgYlgcRbdE4zE,12136
51
- kreuzberg-3.11.2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
52
- kreuzberg-3.11.2.dist-info/entry_points.txt,sha256=GplGhFryCP7kyAG_k-Mdahznvo2fwi73qLFg5yQfH_A,91
53
- kreuzberg-3.11.2.dist-info/licenses/LICENSE,sha256=-8caMvpCK8SgZ5LlRKhGCMtYDEXqTKH9X8pFEhl91_4,1066
54
- kreuzberg-3.11.2.dist-info/RECORD,,
50
+ kreuzberg-3.11.4.dist-info/METADATA,sha256=l3d8PyVfX_aEgXl5ykkuRHJi-8Qzhu4_KcHDYOK2RYg,12136
51
+ kreuzberg-3.11.4.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
52
+ kreuzberg-3.11.4.dist-info/entry_points.txt,sha256=GplGhFryCP7kyAG_k-Mdahznvo2fwi73qLFg5yQfH_A,91
53
+ kreuzberg-3.11.4.dist-info/licenses/LICENSE,sha256=-8caMvpCK8SgZ5LlRKhGCMtYDEXqTKH9X8pFEhl91_4,1066
54
+ kreuzberg-3.11.4.dist-info/RECORD,,