fix(extract-python): avoid blocking the io loop when processing document

Author: ClemDoum

extract-python/extract_python/docling_.py

@@ -1,3 +1,4 @@
+import asyncio
 import shutil
 import tempfile
 from collections.abc import AsyncGenerator, Iterable, Iterator
@@ -55,7 +56,13 @@ async def extract_content(
     ) -> AsyncGenerator[Result, None]:
         docs, path_or_streams = map_and_preserve(_to_docling, docs)
         outputs = self._converter.convert_all(path_or_streams, raises_on_error=False)
-        for doc, res in zip(docs, outputs, strict=True):
+
+        sentinel = object()
+        while True:
+            res = await asyncio.to_thread(next, outputs, sentinel)
+            if res is sentinel:
+                return
+            doc = next(docs)
             yield _to_result(res, doc, output_format, output_path=output_path)
 
     @classmethod

extract-python/extract_python/marker_.py

@@ -1,3 +1,4 @@
+import asyncio
 import gc
 from collections.abc import AsyncGenerator, Iterable
 from copy import deepcopy
@@ -86,23 +87,23 @@ async def extract_content(
             renderer=renderer,
         )
         for doc in docs:
-            yield _process_doc(doc, converter, output_format, output_path)
+            yield await _process_doc(doc, converter, output_format, output_path)
 
     @classmethod
     def _from_config(cls, config: MarkerPipelineConfig) -> Self:
         return cls(config.config)
 
 
 @report_recoverable_errors(_MARKER_CONVERSION_ERRORS)
-def _process_doc(
+async def _process_doc(
     doc: InputDoc,
     converter: "PdfConverter",
     output_format: OutputFormat,
     output_path: Path,
 ) -> Result:
     from marker.output import text_from_rendered  # noqa: PLC0415
 
-    rendered = converter(str(doc.path))
+    rendered = await asyncio.to_thread(converter, str(doc.path))
     content, _, images = text_from_rendered(rendered)
     match output_format:
         case OutputFormat.MARKDOWN: