kreuzberg 3.1.6__tar.gz → 3.1.7__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 (38) hide show
  1. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/PKG-INFO +11 -10
  2. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/README.md +1 -0
  3. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg.egg-info/PKG-INFO +11 -10
  4. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg.egg-info/requires.txt +9 -9
  5. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/pyproject.toml +13 -13
  6. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/LICENSE +0 -0
  7. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/__init__.py +0 -0
  8. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_chunker.py +0 -0
  9. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_constants.py +0 -0
  10. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_extractors/__init__.py +0 -0
  11. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_extractors/_base.py +0 -0
  12. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_extractors/_html.py +0 -0
  13. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_extractors/_image.py +0 -0
  14. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_extractors/_pandoc.py +0 -0
  15. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_extractors/_pdf.py +0 -0
  16. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_extractors/_presentation.py +0 -0
  17. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_extractors/_spread_sheet.py +0 -0
  18. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_gmft.py +0 -0
  19. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_mime_types.py +0 -0
  20. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_ocr/__init__.py +0 -0
  21. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_ocr/_base.py +0 -0
  22. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_ocr/_easyocr.py +0 -0
  23. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_ocr/_paddleocr.py +0 -0
  24. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_ocr/_tesseract.py +0 -0
  25. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_playa.py +0 -0
  26. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_registry.py +0 -0
  27. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_types.py +0 -0
  28. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_utils/__init__.py +0 -0
  29. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_utils/_string.py +0 -0
  30. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_utils/_sync.py +0 -0
  31. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_utils/_tmp.py +0 -0
  32. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/exceptions.py +0 -0
  33. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/extraction.py +0 -0
  34. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/py.typed +0 -0
  35. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg.egg-info/SOURCES.txt +0 -0
  36. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg.egg-info/dependency_links.txt +0 -0
  37. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg.egg-info/top_level.txt +0 -0
  38. {kreuzberg-3.1.6 → kreuzberg-3.1.7}/setup.cfg +0 -0
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: kreuzberg
3
- Version: 3.1.6
3
+ Version: 3.1.7
4
4
  Summary: A text extraction library supporting PDFs, images, office documents and more
5
5
  Author-email: Na'aman Hirschfeld <nhirschfed@gmail.com>
6
6
  License: MIT
@@ -27,33 +27,34 @@ License-File: LICENSE
27
27
  Requires-Dist: anyio>=4.9.0
28
28
  Requires-Dist: charset-normalizer>=3.4.2
29
29
  Requires-Dist: exceptiongroup>=1.2.2; python_version < "3.11"
30
- Requires-Dist: html-to-markdown>=1.3.2
31
- Requires-Dist: playa-pdf==0.4.3
30
+ Requires-Dist: html-to-markdown>=1.3.3
31
+ Requires-Dist: playa-pdf>=0.5.1
32
32
  Requires-Dist: pypdfium2==4.30.0
33
33
  Requires-Dist: python-calamine>=0.3.2
34
34
  Requires-Dist: python-pptx>=1.0.2
35
- Requires-Dist: typing-extensions>=4.13.2; python_version < "3.12"
35
+ Requires-Dist: typing-extensions>=4.14.0; python_version < "3.12"
36
36
  Provides-Extra: all
37
37
  Requires-Dist: easyocr>=1.7.2; extra == "all"
38
38
  Requires-Dist: gmft>=0.4.1; extra == "all"
39
- Requires-Dist: paddleocr>=2.10.0; extra == "all"
39
+ Requires-Dist: paddleocr>=3.0.1; extra == "all"
40
40
  Requires-Dist: paddlepaddle>=3.0.0; extra == "all"
41
- Requires-Dist: semantic-text-splitter>=0.26.0; extra == "all"
42
- Requires-Dist: setuptools>=80.4.0; extra == "all"
41
+ Requires-Dist: semantic-text-splitter>=0.27.0; extra == "all"
42
+ Requires-Dist: setuptools>=80.9.0; extra == "all"
43
43
  Provides-Extra: chunking
44
- Requires-Dist: semantic-text-splitter>=0.26.0; extra == "chunking"
44
+ Requires-Dist: semantic-text-splitter>=0.27.0; extra == "chunking"
45
45
  Provides-Extra: easyocr
46
46
  Requires-Dist: easyocr>=1.7.2; extra == "easyocr"
47
47
  Provides-Extra: gmft
48
48
  Requires-Dist: gmft>=0.4.1; extra == "gmft"
49
49
  Provides-Extra: paddleocr
50
- Requires-Dist: paddleocr>=2.10.0; extra == "paddleocr"
50
+ Requires-Dist: paddleocr>=3.0.1; extra == "paddleocr"
51
51
  Requires-Dist: paddlepaddle>=3.0.0; extra == "paddleocr"
52
- Requires-Dist: setuptools>=80.4.0; extra == "paddleocr"
52
+ Requires-Dist: setuptools>=80.9.0; extra == "paddleocr"
53
53
  Dynamic: license-file
54
54
 
55
55
  # Kreuzberg
56
56
 
57
+ [![Discord](https://img.shields.io/badge/Discord-Join%20our%20community-7289da)](https://discord.gg/pXxagNK2zN)
57
58
  [![PyPI version](https://badge.fury.io/py/kreuzberg.svg)](https://badge.fury.io/py/kreuzberg)
58
59
  [![Documentation](https://img.shields.io/badge/docs-GitHub_Pages-blue)](https://goldziher.github.io/kreuzberg/)
59
60
  [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/README.md RENAMED
@@ -1,5 +1,6 @@
1
1
  # Kreuzberg
2
2
 
3
+ [![Discord](https://img.shields.io/badge/Discord-Join%20our%20community-7289da)](https://discord.gg/pXxagNK2zN)
3
4
  [![PyPI version](https://badge.fury.io/py/kreuzberg.svg)](https://badge.fury.io/py/kreuzberg)
4
5
  [![Documentation](https://img.shields.io/badge/docs-GitHub_Pages-blue)](https://goldziher.github.io/kreuzberg/)
5
6
  [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg.egg-info/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: kreuzberg
3
- Version: 3.1.6
3
+ Version: 3.1.7
4
4
  Summary: A text extraction library supporting PDFs, images, office documents and more
5
5
  Author-email: Na'aman Hirschfeld <nhirschfed@gmail.com>
6
6
  License: MIT
@@ -27,33 +27,34 @@ License-File: LICENSE
27
27
  Requires-Dist: anyio>=4.9.0
28
28
  Requires-Dist: charset-normalizer>=3.4.2
29
29
  Requires-Dist: exceptiongroup>=1.2.2; python_version < "3.11"
30
- Requires-Dist: html-to-markdown>=1.3.2
31
- Requires-Dist: playa-pdf==0.4.3
30
+ Requires-Dist: html-to-markdown>=1.3.3
31
+ Requires-Dist: playa-pdf>=0.5.1
32
32
  Requires-Dist: pypdfium2==4.30.0
33
33
  Requires-Dist: python-calamine>=0.3.2
34
34
  Requires-Dist: python-pptx>=1.0.2
35
- Requires-Dist: typing-extensions>=4.13.2; python_version < "3.12"
35
+ Requires-Dist: typing-extensions>=4.14.0; python_version < "3.12"
36
36
  Provides-Extra: all
37
37
  Requires-Dist: easyocr>=1.7.2; extra == "all"
38
38
  Requires-Dist: gmft>=0.4.1; extra == "all"
39
- Requires-Dist: paddleocr>=2.10.0; extra == "all"
39
+ Requires-Dist: paddleocr>=3.0.1; extra == "all"
40
40
  Requires-Dist: paddlepaddle>=3.0.0; extra == "all"
41
- Requires-Dist: semantic-text-splitter>=0.26.0; extra == "all"
42
- Requires-Dist: setuptools>=80.4.0; extra == "all"
41
+ Requires-Dist: semantic-text-splitter>=0.27.0; extra == "all"
42
+ Requires-Dist: setuptools>=80.9.0; extra == "all"
43
43
  Provides-Extra: chunking
44
- Requires-Dist: semantic-text-splitter>=0.26.0; extra == "chunking"
44
+ Requires-Dist: semantic-text-splitter>=0.27.0; extra == "chunking"
45
45
  Provides-Extra: easyocr
46
46
  Requires-Dist: easyocr>=1.7.2; extra == "easyocr"
47
47
  Provides-Extra: gmft
48
48
  Requires-Dist: gmft>=0.4.1; extra == "gmft"
49
49
  Provides-Extra: paddleocr
50
- Requires-Dist: paddleocr>=2.10.0; extra == "paddleocr"
50
+ Requires-Dist: paddleocr>=3.0.1; extra == "paddleocr"
51
51
  Requires-Dist: paddlepaddle>=3.0.0; extra == "paddleocr"
52
- Requires-Dist: setuptools>=80.4.0; extra == "paddleocr"
52
+ Requires-Dist: setuptools>=80.9.0; extra == "paddleocr"
53
53
  Dynamic: license-file
54
54
 
55
55
  # Kreuzberg
56
56
 
57
+ [![Discord](https://img.shields.io/badge/Discord-Join%20our%20community-7289da)](https://discord.gg/pXxagNK2zN)
57
58
  [![PyPI version](https://badge.fury.io/py/kreuzberg.svg)](https://badge.fury.io/py/kreuzberg)
58
59
  [![Documentation](https://img.shields.io/badge/docs-GitHub_Pages-blue)](https://goldziher.github.io/kreuzberg/)
59
60
  [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg.egg-info/requires.txt RENAMED
@@ -1,7 +1,7 @@
1
1
  anyio>=4.9.0
2
2
  charset-normalizer>=3.4.2
3
- html-to-markdown>=1.3.2
4
- playa-pdf==0.4.3
3
+ html-to-markdown>=1.3.3
4
+ playa-pdf>=0.5.1
5
5
  pypdfium2==4.30.0
6
6
  python-calamine>=0.3.2
7
7
  python-pptx>=1.0.2
@@ -10,18 +10,18 @@ python-pptx>=1.0.2
10
10
  exceptiongroup>=1.2.2
11
11
 
12
12
  [:python_version < "3.12"]
13
- typing-extensions>=4.13.2
13
+ typing-extensions>=4.14.0
14
14
 
15
15
  [all]
16
16
  easyocr>=1.7.2
17
17
  gmft>=0.4.1
18
- paddleocr>=2.10.0
18
+ paddleocr>=3.0.1
19
19
  paddlepaddle>=3.0.0
20
- semantic-text-splitter>=0.26.0
21
- setuptools>=80.4.0
20
+ semantic-text-splitter>=0.27.0
21
+ setuptools>=80.9.0
22
22
 
23
23
  [chunking]
24
- semantic-text-splitter>=0.26.0
24
+ semantic-text-splitter>=0.27.0
25
25
 
26
26
  [easyocr]
27
27
  easyocr>=1.7.2
@@ -30,6 +30,6 @@ easyocr>=1.7.2
30
30
  gmft>=0.4.1
31
31
 
32
32
  [paddleocr]
33
- paddleocr>=2.10.0
33
+ paddleocr>=3.0.1
34
34
  paddlepaddle>=3.0.0
35
- setuptools>=80.4.0
35
+ setuptools>=80.9.0
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/pyproject.toml RENAMED
@@ -1,6 +1,6 @@
1
1
  [project]
2
2
  name = "kreuzberg"
3
- version = "3.1.6"
3
+ version = "3.1.7"
4
4
  description = "A text extraction library supporting PDFs, images, office documents and more"
5
5
  readme = "README.md"
6
6
  keywords = [
@@ -40,12 +40,12 @@ dependencies = [
40
40
  "anyio>=4.9.0",
41
41
  "charset-normalizer>=3.4.2",
42
42
  "exceptiongroup>=1.2.2; python_version<'3.11'",
43
- "html-to-markdown>=1.3.2",
44
- "playa-pdf==0.4.3", # pinned due to breaking changes in 0.5.0
43
+ "html-to-markdown>=1.3.3",
44
+ "playa-pdf>=0.5.1", # pinned due to breaking changes in 0.5.0
45
45
  "pypdfium2==4.30.0", # pinned due to bug in 4.30.1, until v5 is stable
46
46
  "python-calamine>=0.3.2",
47
47
  "python-pptx>=1.0.2",
48
- "typing-extensions>=4.13.2; python_version<'3.12'",
48
+ "typing-extensions>=4.14.0; python_version<'3.12'",
49
49
  ]
50
50
 
51
51
  optional-dependencies.all = [
@@ -54,14 +54,14 @@ optional-dependencies.all = [
54
54
  # gmft
55
55
  "gmft>=0.4.1",
56
56
  # paddle
57
- "paddleocr>=2.10.0",
57
+ "paddleocr>=3.0.1",
58
58
  "paddlepaddle>=3.0.0",
59
59
  # chunking
60
- "semantic-text-splitter>=0.26.0",
61
- "setuptools>=80.4.0",
60
+ "semantic-text-splitter>=0.27.0",
61
+ "setuptools>=80.9.0",
62
62
  ]
63
63
  optional-dependencies.chunking = [
64
- "semantic-text-splitter>=0.26.0",
64
+ "semantic-text-splitter>=0.27.0",
65
65
  ]
66
66
  optional-dependencies.easyocr = [
67
67
  "easyocr>=1.7.2",
@@ -70,22 +70,22 @@ optional-dependencies.gmft = [
70
70
  "gmft>=0.4.1",
71
71
  ]
72
72
  optional-dependencies.paddleocr = [
73
- "paddleocr>=2.10.0",
73
+ "paddleocr>=3.0.1",
74
74
  "paddlepaddle>=3.0.0",
75
- "setuptools>=80.4.0",
75
+ "setuptools>=80.9.0",
76
76
  ]
77
77
  urls.homepage = "https://github.com/Goldziher/kreuzberg"
78
78
 
79
79
  [dependency-groups]
80
80
  dev = [
81
81
  "covdefaults>=2.3.0",
82
- "mypy>=1.15.0",
82
+ "mypy>=1.16.0",
83
83
  "pre-commit>=4.2.0",
84
- "pytest>=8.3.5",
84
+ "pytest>=8.4.0",
85
85
  "pytest-cov>=6.1.1",
86
86
  "pytest-mock>=3.14.0",
87
87
  "pytest-timeout>=2.4.0",
88
- "ruff>=0.11.9",
88
+ "ruff>=0.11.13",
89
89
  "trio>=0.30.0",
90
90
  "uv-bump",
91
91
  ]
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/LICENSE RENAMED
File without changes
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_gmft.py RENAMED
File without changes
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_playa.py RENAMED
File without changes
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/_types.py RENAMED
File without changes
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/kreuzberg/py.typed RENAMED
File without changes
{kreuzberg-3.1.6 → kreuzberg-3.1.7}/setup.cfg RENAMED
File without changes