kreuzberg 3.20.1__py3-none-any.whl → 3.20.2__py3-none-any.whl

kreuzberg/_gmft.py

@@ -193,9 +193,14 @@ async def extract_tables(
                 await run_sync(doc.close)
 
         except ImportError as e:  # pragma: no cover
-            raise MissingDependencyError.create_for_package(
+            error = MissingDependencyError.create_for_package(
                 dependency_group="gmft", functionality="table extraction", package_name="gmft"
-            ) from e
+            )
+            error.context = {
+                "file_path": str(Path(file_path)),
+                "error_message": str(e),
+            }
+            raise error from e
     finally:
         table_cache.mark_complete(**cache_kwargs)
 
@@ -294,9 +299,14 @@ def extract_tables_sync(
             doc.close()  # type: ignore[no-untyped-call]
 
     except ImportError as e:  # pragma: no cover
-        raise MissingDependencyError.create_for_package(
+        error = MissingDependencyError.create_for_package(
             dependency_group="gmft", functionality="table extraction", package_name="gmft"
-        ) from e
+        )
+        error.context = {
+            "file_path": str(Path(file_path)),
+            "error_message": str(e),
+        }
+        raise error from e
 
 
 def _extract_tables_in_process(
@@ -449,6 +459,17 @@ def _extract_tables_isolated(
             return tables
 
         error_info = result
+        if error_info.get("type") == "ImportError":
+            error = MissingDependencyError.create_for_package(
+                dependency_group="gmft", functionality="table extraction", package_name="gmft"
+            )
+            error.context = {
+                "file_path": str(Path(file_path)),
+                "error_message": error_info["error"],
+                "traceback": error_info.get("traceback"),
+            }
+            raise error from ImportError(error_info["error"])
+
         raise ParsingError(
             f"GMFT table extraction failed: {error_info['error']}",
             context={
@@ -536,6 +557,17 @@ async def _extract_tables_isolated_async(
             return tables
 
         error_info = result
+        if error_info.get("type") == "ImportError":
+            error = MissingDependencyError.create_for_package(
+                dependency_group="gmft", functionality="table extraction", package_name="gmft"
+            )
+            error.context = {
+                "file_path": str(Path(file_path)),
+                "error_message": error_info["error"],
+                "traceback": error_info.get("traceback"),
+            }
+            raise error from ImportError(error_info["error"])
+
         raise ParsingError(
             f"GMFT table extraction failed: {error_info['error']}",
             context={

kreuzberg/_types.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 import sys
+import warnings
 from collections.abc import Awaitable, Callable, Mapping
 from dataclasses import asdict, dataclass, field
 from enum import Enum
@@ -262,6 +263,15 @@ class PaddleOCRConfig(ConfigDict):
 
 @dataclass(unsafe_hash=True, frozen=True, slots=True)
 class GMFTConfig(ConfigDict):
+    def __post_init__(self) -> None:
+        warnings.warn(
+            "GMFTConfig is deprecated and will be removed in Kreuzberg v4.0. "
+            "Install `kreuzberg[gmft]` only if you still rely on GMFT. "
+            "Future versions use native TATR-based table extraction via TableExtractionConfig.",
+            FutureWarning,
+            stacklevel=2,
+        )
+
     verbosity: int = 0
     """
     Verbosity level for logging.

{kreuzberg-3.20.1.dist-info → kreuzberg-3.20.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: kreuzberg
-Version: 3.20.1
+Version: 3.20.2
 Summary: Document intelligence framework for Python - Extract text, metadata, and structured data from diverse file formats
 Keywords: async,document-analysis,document-classification,document-intelligence,document-processing,extensible,information-extraction,mcp,metadata-extraction,model-context-protocol,ocr,pandoc,pdf-extraction,pdfium,plugin-architecture,rag,retrieval-augmented-generation,structured-data,table-extraction,tesseract,text-extraction
 Author: Na'aman Hirschfeld
@@ -28,7 +28,7 @@ Classifier: Typing :: Typed
 Requires-Dist: anyio>=4.11.0
 Requires-Dist: chardetng-py>=0.3.5
 Requires-Dist: exceptiongroup>=1.2.2 ; python_full_version < '3.11'
-Requires-Dist: html-to-markdown>=2.1.0
+Requires-Dist: html-to-markdown>=2.1.2
 Requires-Dist: langcodes>=3.5.0
 Requires-Dist: mcp>=1.17.0
 Requires-Dist: msgspec>=0.18.0

{kreuzberg-3.20.1.dist-info → kreuzberg-3.20.2.dist-info}/RECORD

@@ -19,7 +19,7 @@ kreuzberg/_extractors/_pdf.py,sha256=_MPtO_8BCpyAXyIWusmfqOaEsPMDxucjTQKz3cTaj8o
 kreuzberg/_extractors/_presentation.py,sha256=2g6PJnpgUpUfMjQJh-7_gHywDulE8QE8ypH__BrEUTQ,10692
 kreuzberg/_extractors/_spread_sheet.py,sha256=TJOM70DLN0HzcOkAowZJogAx7QFrouohvU5V0OIliag,12738
 kreuzberg/_extractors/_structured.py,sha256=thpXhsBnvaHzGQX4sy6eVHowFv0yaYxLGHwxx4DouCI,8947
-kreuzberg/_gmft.py,sha256=gfRXOsv-K9R7Y0zZ2SUa5wid3FpP2eFIlg5nepWcz1Q,20827
+kreuzberg/_gmft.py,sha256=0jmCTEs9EV78awlpCvmfLWAEiiq9n9Qx3n3FnCboTgg,22119
 kreuzberg/_language_detection.py,sha256=4JzQldcDIVZRWUzRFc9AOFiq6Wfl9858mip1ZnrD2Ks,1143
 kreuzberg/_mcp/__init__.py,sha256=h6DgLFO4TMUk7_wCJ2jn2Y6IkFmfzb-Z7jX-G5UCYVc,43
 kreuzberg/_mcp/server.py,sha256=71MhjiFDwgFROdGejf0djgO1eG370qudWmZsN59CUeA,16743
@@ -99,7 +99,7 @@ kreuzberg/_token_reduction/stopwords/vi_stopwords.json,sha256=UOyAEKBwMcQV65QGpQ
 kreuzberg/_token_reduction/stopwords/yo_stopwords.json,sha256=60liY89h7KReEvHEPxe-hCWLPuqr4U89aQDCi7iRCfo,651
 kreuzberg/_token_reduction/stopwords/zh_stopwords.json,sha256=rouSTCkXun90Q1aCvLjHyt4I7pGrtlcruDpNVybpAMI,8934
 kreuzberg/_token_reduction/stopwords/zu_stopwords.json,sha256=hfm4E2EDI_VWyR0GUOVjcMQA7ZDH7FsV4FUMcns1H28,324
-kreuzberg/_types.py,sha256=eh4bZFG3jIw5GhfC3u4R0aa_y9niKZDI4O93j0MCZGw,53672
+kreuzberg/_types.py,sha256=qh4S9PhIZkvueTCgvO1d3lKmy5pWYfuhHpoGcoIEyYY,54061
 kreuzberg/_utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 kreuzberg/_utils/_cache.py,sha256=AtANbs1MWR4WLB2MhatVGhlh7kM-yjSfFuDnSVSNp50,14110
 kreuzberg/_utils/_device.py,sha256=o03rLiHiRX6TKhJ55LO1Vj2Map1Po5YdjuMdA63tGOE,8249
@@ -121,7 +121,7 @@ kreuzberg/cli.py,sha256=P_dqOHbGh-fFYZ4WErjngTKq7wbqaUmTD1Gjw2lIsDI,15242
 kreuzberg/exceptions.py,sha256=KiGAfIX3_TkGYG1h9eTZ_E_pALsAqhZ_A3XfhwxwaS0,2909
 kreuzberg/extraction.py,sha256=jMsomvg7SPnuXLGZKQl0YH64D0AhczSNDM4CKORd9d0,24185
 kreuzberg/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-kreuzberg-3.20.1.dist-info/WHEEL,sha256=X16MKk8bp2DRsAuyteHJ-9qOjzmnY0x1aj0P1ftqqWA,78
-kreuzberg-3.20.1.dist-info/entry_points.txt,sha256=nFqXF3_dNlhHuvtNX26GKKutjOjH2a7TUiaYSu2eOAk,92
-kreuzberg-3.20.1.dist-info/METADATA,sha256=FIG_Res9xL_wupiTlRzQ6sdofF-UeJ_z9_8uArvjCO0,11923
-kreuzberg-3.20.1.dist-info/RECORD,,
+kreuzberg-3.20.2.dist-info/WHEEL,sha256=X16MKk8bp2DRsAuyteHJ-9qOjzmnY0x1aj0P1ftqqWA,78
+kreuzberg-3.20.2.dist-info/entry_points.txt,sha256=nFqXF3_dNlhHuvtNX26GKKutjOjH2a7TUiaYSu2eOAk,92
+kreuzberg-3.20.2.dist-info/METADATA,sha256=vR4TfFvm5tx66RV7dS0CXRCTmCFksHb8Jyhy-3FHNUQ,11923
+kreuzberg-3.20.2.dist-info/RECORD,,