extract-python 0.11.1__tar.gz → 0.11.3__tar.gz

{extract_python-0.11.1 → extract_python-0.11.3}/PKG-INFO

@@ -1,13 +1,13 @@
 Metadata-Version: 2.5
 Name: extract-python
-Version: 0.11.1
+Version: 0.11.3
 Summary: Structured content extraction
 Project-URL: Homepage, https://github.com/ICIJ/extract-python
 Project-URL: Repository, https://github.com/ICIJ/extract-python
 Project-URL: Issues, https://github.com/ICIJ/extract-python/issues
 Author-email: Clément Doumouro <cdoumouro@icij.org>
 Requires-Python: <3.15,>=3.13
-Requires-Dist: extract-core~=0.11.0
+Requires-Dist: extract-core~=0.11.3
 Requires-Dist: icij-common~=0.8.2
 Requires-Dist: pympler~=1.1
 Provides-Extra: benches

{extract_python-0.11.1 → extract_python-0.11.3}/extract_python/miner_u.py

@@ -16,7 +16,8 @@ from extract_core import (
     Result,
     Status,
 )
-from mineru import MiddleJson, ParseResult
+from mineru import MiddleJson, MinerUParser, ParseResult
+from mineru.config import VlmConfig
 
 from .constants import ARTIFACTS, DEFAULT_MD_PAGE_SEP
 from .utils import path_to_artifacts_dirname, reset_env, write_pages
@@ -34,18 +35,20 @@ class MinerUPipeline(Pipeline):
     async def extract_content(
         self, docs: Iterable[InputDoc], output_format: OutputFormat, output_path: Path
     ) -> AsyncIterable[Result]:
-        from mineru import MinerUParser  # noqa: PLC0415
 
         with reset_env():
             os.environ["MINERU_DEVICE_MODE"] = self._device
             # TODO: exclude files which are not pdf and return an error
             # TODO: we should only process valid PDFs
             config = self._config.config
+            vlm_config = None
+            if config.vlm_config is not None:
+                vlm_config = VlmConfig.model_validate(config.vlm_config)
             parser = MinerUParser(
                 tier=config.tier,
                 parse_mode=config.parse_mode,
                 image_analysis=config.image_analysis,
-                vlm_config=config.vlm_config,
+                vlm_config=vlm_config,
             )
             for doc in docs:
                 res = await parser.parse_async(doc.path)

{extract_python-0.11.1 → extract_python-0.11.3}/pyproject.toml

@@ -8,7 +8,7 @@ authors = [
 readme = "README.md"
 requires-python = ">=3.13,<3.15"
 dependencies = [
-  "extract-core~=0.11.0",
+  "extract-core~=0.11.3",
   "icij-common~=0.8.2",
   "pympler~=1.1",
 ]