extract-python 0.3.1__tar.gz → 0.3.2__tar.gz

{extract_python-0.3.1 → extract_python-0.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: extract-python
-Version: 0.3.1
+Version: 0.3.2
 Summary: Structured content extraction
 Project-URL: Homepage, https://github.com/ICIJ/extract-python
 Project-URL: Repository, https://github.com/ICIJ/extract-python

{extract_python-0.3.1 → extract_python-0.3.2}/extract_python/docling_.py

@@ -32,7 +32,7 @@ if TYPE_CHECKING:
 
 
 def _validate_pipeline_opts(opts: "PipelineOptions") -> None:
-    from docling.datamodel.pipeline_options import PdfPipelineOptions
+    from docling.datamodel.pipeline_options import PdfPipelineOptions  # noqa: PLC0415
 
     if isinstance(opts, PdfPipelineOptions) and not opts.generate_picture_images:
         msg = "generate_picture_images should be set to true"
@@ -49,11 +49,11 @@ def _validate_options(
 
 @cache
 def _default_format_opts() -> dict["InputFormat", "FormatOption"]:
-    from docling.datamodel.pipeline_options import (
+    from docling.datamodel.pipeline_options import (  # noqa: PLC0415
         EasyOcrOptions,
         PdfPipelineOptions,
     )
-    from docling.document_converter import PdfFormatOption
+    from docling.document_converter import PdfFormatOption  # noqa: PLC0415
 
     return {
         InputFormat.PDF: PdfFormatOption(
@@ -80,23 +80,21 @@ class DoclingPipelineConfig(PipelineConfig):
     task_group: ClassVar[str] = Field(frozen=True, default=CPU_GROUP)
 
     format_options: Annotated[
-        dict[InputFormat, FormatOption] | None, AfterValidator(_validate_options)
+        dict["InputFormat", "FormatOption"] | None, AfterValidator(_validate_options)
     ] = Field(default_factory=_default_format_opts)
 
-    _unsupported_input_formats: ClassVar[set[InputFormat]] = {
-        InputFormat.AUDIO,
-        InputFormat.METS_GBS,
-        InputFormat.VTT,
-    }
-
     @classmethod
     @cache
     def supported_exts(cls) -> set[SupportedExt]:
-        from docling.datamodel.base_models import FormatToExtensions, InputFormat
+        from docling.datamodel.base_models import (  # noqa: PLC0415
+            FormatToExtensions,
+            InputFormat,
+        )
 
+        unsupported = {InputFormat.AUDIO, InputFormat.METS_GBS, InputFormat.VTT}
         supported = set()
         for f in InputFormat:
-            if f in cls._unsupported_input_formats:
+            if f in unsupported:
                 continue
             for ext in FormatToExtensions[f]:
                 supported.add(SupportedExt(f".{ext.lower()}"))
@@ -105,8 +103,10 @@ class DoclingPipelineConfig(PipelineConfig):
 
 @Pipeline.register(PipelineType.DOCLING)
 class DoclingPipeline(Pipeline):
-    def __init__(self, format_options: dict[InputFormat, FormatOption] | None = None):
-        from docling.document_converter import DocumentConverter
+    def __init__(
+        self, format_options: dict["InputFormat", "FormatOption"] | None = None
+    ):
+        from docling.document_converter import DocumentConverter  # noqa: PLC0415
 
         allowed_format = [
             f.to_docling() for f in DoclingPipelineConfig.supported_exts()
@@ -128,7 +128,7 @@ class DoclingPipeline(Pipeline):
         return cls(config.format_options)
 
 
-def _to_docling(docs: Iterable[InputDoc]) -> Iterator[Path | "DocumentStream"]:
+def _to_docling(docs: Iterable[InputDoc]) -> Iterator["Path | DocumentStream"]:
     for d in docs:
         yield d.to_docling()
 
@@ -160,7 +160,7 @@ def _to_markdown_doc(
     page_sep: str = DEFAULT_MD_PAGE_SEP,
     **kwargs,
 ) -> MarkdownDoc:
-    from docling_core.types.doc import ImageRefMode
+    from docling_core.types.doc import ImageRefMode  # noqa: PLC0415
 
     # TODO: Should we add a hash to avoid collision between files with same names
     #  nested in the tree structured

{extract_python-0.3.1 → extract_python-0.3.2}/extract_python/marker_.py

@@ -73,9 +73,9 @@ class MarkerPipeline(Pipeline):
     async def extract_content(
         self, docs: Iterable[InputDoc], output_format: OutputFormat, output_path: Path
     ) -> AsyncGenerator[Result, None]:
-        from marker.config.parser import ConfigParser
-        from marker.converters.pdf import PdfConverter
-        from marker.models import create_model_dict
+        from marker.config.parser import ConfigParser  # noqa: PLC0415
+        from marker.converters.pdf import PdfConverter  # noqa: PLC0415
+        from marker.models import create_model_dict  # noqa: PLC0415
 
         config = deepcopy(self._marker_config)
         config["output_format"] = output_format.to_marker()
@@ -102,7 +102,7 @@ def _process_doc(
     output_format: OutputFormat,
     output_path: Path,
 ) -> Result:
-    from marker.output import text_from_rendered
+    from marker.output import text_from_rendered  # noqa: PLC0415
 
     rendered = converter(str(doc.path))
     content, _, images = text_from_rendered(rendered)
@@ -118,7 +118,7 @@ def _process_doc(
 def _to_markdown_doc(
     input_doc: InputDoc, content: str, images: dict[str, "Image"], output_path: Path
 ) -> MarkdownDoc:
-    from marker.renderers.markdown import MarkdownRenderer
+    from marker.renderers.markdown import MarkdownRenderer  # noqa: PLC0415
 
     # TODO: Should we add a hash to avoid collision between files with same names
     #  nested in the tree structured

{extract_python-0.3.1 → extract_python-0.3.2}/extract_python/miner_u.py

@@ -52,13 +52,12 @@ class MinerUConfig(BaseModel):
     @classmethod
     @cache
     def _get_default_kwargs(cls) -> dict[str, Any]:
-
-        from mineru.utils.enum_class import MakeMode
+        from mineru.utils.enum_class import MakeMode  # noqa: PLC0415
 
         return {
             "server_url": None,
-            # We don't dump md directly we process, we dump the middle json in order to be
-            # able to get page indexes
+            # We don't dump md directly we process, we dump the middle json in order
+            # to be able to get page indexes
             "parse_method": "auto",
             "dump_md": False,
             "dump_middle_json": True,
@@ -104,7 +103,7 @@ class MinerUPipeline(Pipeline):
     async def extract_content(
         self, docs: Iterable[InputDoc], output_format: OutputFormat, output_path: Path
     ) -> AsyncGenerator[Result, None]:
-        from mineru.cli.common import aio_do_parse
+        from mineru.cli.common import aio_do_parse  # noqa: PLC0415
 
         docs = list(docs)
         # TODO: exclude files which are not pdf and return an error
@@ -154,13 +153,15 @@ def _parse_md_make_fn(backend: MinerUBackend) -> MDMakeFunction:
 
     match backend:
         case MinerUBackend.PIPELINE:
-            from mineru.backend.pipeline.pipeline_middle_json_mkcontent import (
+            from mineru.backend.pipeline.pipeline_middle_json_mkcontent import (  # noqa: PLC0415
                 union_make,
             )
 
             return union_make
         case MinerUBackend.VLM:
-            from mineru.backend.vlm.vlm_middle_json_mkcontent import union_make
+            from mineru.backend.vlm.vlm_middle_json_mkcontent import (  # noqa: PLC0415
+                union_make,
+            )
 
             return union_make
         case _:
@@ -212,7 +213,7 @@ def _dump_md_content(
     im_dir: Path,
     md_make_mode: str | None = None,
 ) -> ConversionOutput:
-    from mineru.utils.enum_class import MakeMode
+    from mineru.utils.enum_class import MakeMode  # noqa: PLC0415
 
     if md_make_mode is None:
         md_make_mode = MakeMode.MM_MD