chore: <breaking> changed 'force_ocr' to kwarg

Goldziher · Feb 11, 2025 · 29731fc · 29731fc
1 parent 5cc770a
commit 29731fc
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 4 deletions.
diff --git a/kreuzberg/extraction.py b/kreuzberg/extraction.py
@@ -55,8 +55,8 @@ class ExtractionResult(NamedTuple):
 async def extract_bytes(
     content: bytes,
     mime_type: str,
-    force_ocr: bool = False,
     *,
+    force_ocr: bool = False,
     config: Config | None = None,
 ) -> ExtractionResult:
     """Extract the textual content from a given byte string representing a file's contents.
@@ -159,8 +159,8 @@ def extract_file_sync(
 async def extract_file(
     file_path: Path | str,
     mime_type: str | None = None,
-    force_ocr: bool = False,
     *,
+    force_ocr: bool = False,
     config: Config | None = None,
 ) -> ExtractionResult:
     """Extract the textual content from a given file.

diff --git a/tests/extraction_test.py b/tests/extraction_test.py
@@ -27,7 +27,7 @@ async def test_extract_bytes_pdf(pdf_document: Path) -> None:
 
 async def test_extract_bytes_force_ocr_pdf(non_ascii_pdf: Path) -> None:
     content = non_ascii_pdf.read_bytes()
-    result = await extract_bytes(content, PDF_MIME_TYPE, True)
+    result = await extract_bytes(content, PDF_MIME_TYPE, force_ocr=True)
     assert result.mime_type == PLAIN_TEXT_MIME_TYPE
     assert result.content.startswith("AMTSBLATT")
     assert isinstance(result.content, str)
@@ -104,7 +104,7 @@ async def test_extract_file_pdf(pdf_document: Path) -> None:
 
 
 async def test_extract_file_force_ocr_pdf(non_ascii_pdf: Path) -> None:
-    result = await extract_file(non_ascii_pdf, PDF_MIME_TYPE, True)
+    result = await extract_file(non_ascii_pdf, PDF_MIME_TYPE, force_ocr=True)
     assert result.mime_type == PLAIN_TEXT_MIME_TYPE
     assert result.content.startswith("AMTSBLATT")
     assert isinstance(result.content, str)