kreuzberg 3.1.0__py3-none-any.whl → 3.1.2__py3-none-any.whl

kreuzberg/_extractors/_pandoc.py

@@ -228,7 +228,7 @@ class PandocExtractor(Extractor):
             command = ["pandoc", "--version"]
             result = await run_process(command)
 
-            version_match = re.search(r"pandoc\s+v?(\d+)\.\d+\.\d+", result.stdout.decode())
+            version_match = re.search(r"pandoc\s+v?(\d+)\.\d+(\.\d+)?", result.stdout.decode())
             if not version_match or int(version_match.group(1)) < MINIMAL_SUPPORTED_PANDOC_VERSION:
                 raise MissingDependencyError(
                     "Pandoc version 2 or above is a required system dependency. Please install it on your system and make sure its available in $PATH."

kreuzberg/_ocr/_easyocr.py

@@ -173,6 +173,10 @@ class EasyOCRBackend(OCRBackend[EasyOCRConfig]):
         await self._init_easyocr(**kwargs)
 
         beam_width = kwargs.pop("beam_width")
+
+        kwargs.pop("language", None)
+        kwargs.pop("use_gpu", None)
+
         try:
             result = await run_sync(
                 self._reader.readtext,

{kreuzberg-3.1.0.dist-info → kreuzberg-3.1.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: kreuzberg
-Version: 3.1.0
+Version: 3.1.2
 Summary: A text extraction library supporting PDFs, images, office documents and more
 Author-email: Na'aman Hirschfeld <nhirschfed@gmail.com>
 License: MIT
@@ -27,12 +27,12 @@ License-File: LICENSE
 Requires-Dist: anyio>=4.9.0
 Requires-Dist: charset-normalizer>=3.4.1
 Requires-Dist: exceptiongroup>=1.2.2; python_version < "3.11"
-Requires-Dist: html-to-markdown>=1.2.1
+Requires-Dist: html-to-markdown>=1.3.0
 Requires-Dist: playa-pdf>=0.4.1
 Requires-Dist: pypdfium2==4.30.0
-Requires-Dist: python-calamine>=0.3.1
+Requires-Dist: python-calamine>=0.3.2
 Requires-Dist: python-pptx>=1.0.2
-Requires-Dist: typing-extensions>=4.12.2; python_version < "3.12"
+Requires-Dist: typing-extensions>=4.13.1; python_version < "3.12"
 Provides-Extra: all
 Requires-Dist: easyocr>=1.7.2; extra == "all"
 Requires-Dist: gmft>=0.4.1; extra == "all"
@@ -140,7 +140,7 @@ For comprehensive documentation, visit our [GitHub Pages](https://goldziher.gith
 
 Kreuzberg supports a wide range of document formats:
 
-- **Documents**: PDF, DOCX, DOC, RTF, TXT, EPUB, etc.
+- **Documents**: PDF, DOCX, RTF, TXT, EPUB, etc.
 - **Images**: JPG, PNG, TIFF, BMP, GIF, etc.
 - **Spreadsheets**: XLSX, XLS, CSV, etc.
 - **Presentations**: PPTX, PPT, etc.
@@ -162,11 +162,11 @@ This library is open to contribution. Feel free to open issues or submit PRs. It
 
 ### Local Development
 
-1. Clone the repo
-1. Install the system dependencies
-1. Install the full dependencies with `uv sync`
-1. Install the pre-commit hooks with: `pre-commit install && pre-commit install --hook-type commit-msg`
-1. Make your changes and submit a PR
+- Clone the repo
+- Install the system dependencies
+- Install the full dependencies with `uv sync`
+- Install the pre-commit hooks with: `pre-commit install && pre-commit install --hook-type commit-msg`
+- Make your changes and submit a PR
 
 ## License
 

{kreuzberg-3.1.0.dist-info → kreuzberg-3.1.2.dist-info}/RECORD

@@ -13,21 +13,21 @@ kreuzberg/_extractors/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3h
 kreuzberg/_extractors/_base.py,sha256=YUr6A2n34LlFzbYQkiKqhXAphL9RYrvAls5SlkoQqNg,3028
 kreuzberg/_extractors/_html.py,sha256=L_vcKyv1ObI6irPaD3-WTKqxeRfZA4Rhsl3zUiAe_ws,1312
 kreuzberg/_extractors/_image.py,sha256=VQgSFSzXIMX3A52-DyvuKgfTRXUJIjYn6IX4-sQWWdg,2626
-kreuzberg/_extractors/_pandoc.py,sha256=a6cYQxoh5G9EMrDWVcQhrTkE4Mar24sNiGCY0zOOzw4,20121
+kreuzberg/_extractors/_pandoc.py,sha256=U5CUrVilKdJqXJFUUT5xzcpy2jfJ26h7kde3p1N_N4w,20124
 kreuzberg/_extractors/_pdf.py,sha256=eNFws_UxLgWSTC_VC_zJmVojpyQvioOXgNjSHQzBq5c,6607
 kreuzberg/_extractors/_presentation.py,sha256=K4ALrpmZ0EWyp2O-3oEmTRCS7yAET9xjinrzo13rpWo,8764
 kreuzberg/_extractors/_spread_sheet.py,sha256=1ejRZk8AE1dXS1tRIdg2S0J9Vo0wG81iKkW2IF6PjlE,4445
 kreuzberg/_ocr/__init__.py,sha256=VTqwKDlIRbjve71Y11Ztygyhv5aWG9LWTj8iX66ANxE,533
 kreuzberg/_ocr/_base.py,sha256=lNT0Tin4hzbmaamqqySxvYEwNtrJB5gGlStrANQQcyc,1637
-kreuzberg/_ocr/_easyocr.py,sha256=VfYW66SkB2Bigbrtd7WEeJ6QZ_1Y5d8Z_rZYBPMsuk0,11037
+kreuzberg/_ocr/_easyocr.py,sha256=IUX5AGMp3C2u3Byiz8BADLMlgoNEpFhwswmdeifMcIo,11112
 kreuzberg/_ocr/_paddleocr.py,sha256=NDKXiMtHjIy-Uq4hXe4qm5oUWwOrhjJaibyC708Cw5E,10422
 kreuzberg/_ocr/_tesseract.py,sha256=cdnVxNpaKjxtBN4xy0Timz-uYtPA9wq9kc6kyYVeDug,9779
 kreuzberg/_utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 kreuzberg/_utils/_string.py,sha256=oNO0cmwjVNG0jAzaqNCjYtzvM_nxH5TW2KV-Uh3oEUU,978
 kreuzberg/_utils/_sync.py,sha256=lycobEMXk0tBMWLwkuMdOuNMStDwPKMC0V1Qgp_oi6k,4071
 kreuzberg/_utils/_tmp.py,sha256=5rqG_Nlb9xweaLqJA8Kc5csHDase9_eY_Fq93rNQGWc,1044
-kreuzberg-3.1.0.dist-info/licenses/LICENSE,sha256=-8caMvpCK8SgZ5LlRKhGCMtYDEXqTKH9X8pFEhl91_4,1066
-kreuzberg-3.1.0.dist-info/METADATA,sha256=YemIJR6aygDxNgz9aoeg2oIGRHJjm897jD8sHuJYdMY,6651
-kreuzberg-3.1.0.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
-kreuzberg-3.1.0.dist-info/top_level.txt,sha256=rbGkygffkZiyKhL8UN41ZOjLfem0jJPA1Whtndne0rE,10
-kreuzberg-3.1.0.dist-info/RECORD,,
+kreuzberg-3.1.2.dist-info/licenses/LICENSE,sha256=-8caMvpCK8SgZ5LlRKhGCMtYDEXqTKH9X8pFEhl91_4,1066
+kreuzberg-3.1.2.dist-info/METADATA,sha256=7cx9eSl0NAfeu18rvYT4BtwVdVOA1ZgInDx8KcpXlw8,6641
+kreuzberg-3.1.2.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
+kreuzberg-3.1.2.dist-info/top_level.txt,sha256=rbGkygffkZiyKhL8UN41ZOjLfem0jJPA1Whtndne0rE,10
+kreuzberg-3.1.2.dist-info/RECORD,,