extract-python 0.4.0__py3-none-any.whl → 0.4.2__py3-none-any.whl

extract_python/docling_.py

@@ -7,7 +7,7 @@ from pathlib import Path
 from typing import Annotated, Any, ClassVar, Self, TypeVar, get_type_hints
 
 from docling.backend.abstract_backend import AbstractDocumentBackend
-from docling.datamodel.backend_options import BackendOptions
+from docling.datamodel.backend_options import BackendOptions, BaseBackendOptions
 
 # Data model import are quick it's ok to leave it there
 from docling.datamodel.base_models import FormatToExtensions, InputFormat
@@ -26,7 +26,15 @@ from docling_core.types.doc import ImageRefMode
 from docling_core.types.io import DocumentStream
 from icij_common.pydantic_utils import to_lower_snake_case
 from icij_common.registrable import FromConfig
-from pydantic import AfterValidator, BeforeValidator, Field, model_validator
+from pydantic import (
+    AfterValidator,
+    BeforeValidator,
+    Field,
+    PlainSerializer,
+    WrapSerializer,
+    model_validator,
+)
+from pydantic_core.core_schema import SerializerFunctionWrapHandler
 
 from .constants import ARTIFACTS, DEFAULT_MD_PAGE_SEP
 from .objects import (
@@ -73,7 +81,7 @@ def _find_subcls(cls: type[T], name: str) -> type[T]:
 
 def _find_init_arg_type(cls: type[Any], arg: str) -> type:
     hints = get_type_hints(cls.__init__)
-    return hints[arg].__class__
+    return hints[arg]
 
 
 def _resolve_pipeline_cls(v: Any) -> Any:
@@ -82,6 +90,23 @@ def _resolve_pipeline_cls(v: Any) -> Any:
     return v
 
 
+def _ser_class_as_str(v: Any) -> Any:
+    if isinstance(v, type):
+        return v.__name__
+    return v
+
+
+def _ser_with_backend_option_kind(
+    v: Any, handler: SerializerFunctionWrapHandler
+) -> Any:
+    serialized = handler(v)
+    if isinstance(v, BaseBackendOptions):
+        kind = getattr(v, "kind", None)
+        if kind is not None:
+            serialized["kind"] = kind
+    return serialized
+
+
 def _resolve_backend(v: Any) -> Any:
     if isinstance(v, str):
         return _find_subcls(AbstractDocumentBackend, v)
@@ -90,15 +115,21 @@ def _resolve_backend(v: Any) -> Any:
 
 class DoclingFormatOption(FormatOption):
     pipeline_cls: Annotated[
-        str | type[BasePipeline], BeforeValidator(_resolve_pipeline_cls)
+        str | type[BasePipeline],
+        BeforeValidator(_resolve_pipeline_cls),
+        PlainSerializer(_ser_class_as_str),
     ]
     pipeline_options: Annotated[
         dict | PipelineOptions | None, AfterValidator(_validate_pipeline_opts)
     ] = None
     backend: Annotated[
-        str | type[AbstractDocumentBackend], BeforeValidator(_resolve_backend)
+        str | type[AbstractDocumentBackend],
+        BeforeValidator(_resolve_backend),
+        PlainSerializer(_ser_class_as_str),
     ]
-    backend_options: BackendOptions | None = None
+    backend_options: Annotated[
+        BackendOptions | None, WrapSerializer(_ser_with_backend_option_kind)
+    ] = None
 
     @model_validator(mode="after")
     def _resolve_pipeline_options(self) -> Self:

extract_python/objects.py

@@ -71,6 +71,7 @@ class SupportedExt(StrEnum):
     DOTM = ".dotm"
     DOCM = ".docm"
     EPUB = ".epub"
+    EML = ".eml"
     GIF = ".gif"
     HTLM = ".html"
     HTM = ".htm"

{extract_python-0.4.0.dist-info → extract_python-0.4.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: extract-python
-Version: 0.4.0
+Version: 0.4.2
 Summary: Structured content extraction
 Project-URL: Homepage, https://github.com/ICIJ/extract-python
 Project-URL: Repository, https://github.com/ICIJ/extract-python

{extract_python-0.4.0.dist-info → extract_python-0.4.2.dist-info}/RECORD

@@ -1,11 +1,11 @@
 extract_python/__init__.py,sha256=Y-lcFbJd5sX3wt1UnyYAbm7H81Sgg7p2fG07CuuJUco,945
 extract_python/constants.py,sha256=JmAkjXyQMYwpTod9DCLc11zOc3caK2j_2ji_r3hGZws,236
-extract_python/docling_.py,sha256=ZGlOVrgQw50bDh4B4DiRiRQSv5rGX-EFi8Z51mnAHpY,8620
+extract_python/docling_.py,sha256=ys2vK4zgpWsPObIZWRFhHM4fNkojMYUa9QRevl8bd3c,9342
 extract_python/marker_.py,sha256=ACk9wa-wrEwYv4D7SKW4KjpZxrp2hBIt9_pheRhV0go,5014
 extract_python/miner_u.py,sha256=EcTXfdvArkoSw3bKkiWLerYAhXMU6ssJFn9kOsFVDPE,8007
-extract_python/objects.py,sha256=kCxg6m7j01Z1sNAGi8vniokMnw6Ry0gU2lXXH2uan8A,8744
+extract_python/objects.py,sha256=MHCUZ9L8LVXlSlHyDMnbuWV1KHWMhUEJQMEDTc9hYD0,8761
 extract_python/pipeline.py,sha256=ijQ8wI5x3kAzTfx3T-V52qSoAA_8IA_ihK1NPWVMwFM,1162
 extract_python/utils.py,sha256=kXe0CyT6zGYDTOvxs3BJYxq2-cgJZ0d_IT3hdMbuXa8,1943
-extract_python-0.4.0.dist-info/METADATA,sha256=_cFyQr6erjdP5CxXtFI9lbyMIDJ8fVuU2LM-h1oyv7k,1132
-extract_python-0.4.0.dist-info/WHEEL,sha256=mffPy8wBnZQn2VnJUU5jE99KsxaSfiyMHV9Yt0aLVxs,87
-extract_python-0.4.0.dist-info/RECORD,,
+extract_python-0.4.2.dist-info/METADATA,sha256=95THYq0jZgY2-1X2s8hDoFEo9_aNeukdHPxlcd8_rmI,1132
+extract_python-0.4.2.dist-info/WHEEL,sha256=mffPy8wBnZQn2VnJUU5jE99KsxaSfiyMHV9Yt0aLVxs,87
+extract_python-0.4.2.dist-info/RECORD,,