markitdown-pro 2.1.1__tar.gz → 2.1.2__tar.gz

{markitdown_pro-2.1.1 → markitdown_pro-2.1.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: markitdown-pro
-Version: 2.1.1
+Version: 2.1.2
 Summary: A package that converts almost any file format to Markdown.
 Author: Developer
 License-Expression: MIT

{markitdown_pro-2.1.1 → markitdown_pro-2.1.2}/markitdown_pro/converters/gotenberg_converter.py

@@ -54,8 +54,11 @@ def _retry_transient(
     def decorator(
         fn: Callable[..., Awaitable[T | None]],
     ) -> Callable[..., Awaitable[T | None]]:
+        """Wrap ``fn`` with the retry-on-transient-httpx-errors policy above."""
+
         @functools.wraps(fn)
         async def wrapper(self, *args, **kwargs) -> T | None:
+            """Call ``fn`` up to ``attempts`` times, sleeping per ``delays``."""
             tag = args[1] if len(args) >= 2 else kwargs.get("tag", _COMPONENT)
             for i in range(attempts):
                 try:
@@ -141,7 +144,7 @@ class GotenbergConverter(Converter):
         try:
             fd, tmp_pdf = tempfile.mkstemp(suffix=".pdf", prefix="mkdpro_")
             os.close(fd)
-            Path(tmp_pdf).write_bytes(pdf_bytes)
+            await asyncio.to_thread(Path(tmp_pdf).write_bytes, pdf_bytes)
 
             logger.info(f"{tag} | PDF received ({len(pdf_bytes)} bytes), delegating to PDFHandler")
 
@@ -167,27 +170,15 @@ class GotenbergConverter(Converter):
 
     @_retry_transient(attempts=3, delays=(1.0, 3.0))
     async def _convert_to_pdf(self, file_path: str, tag: str) -> bytes | None:
-        """Send file to Gotenberg and return PDF bytes."""
-        try:
-            async with httpx.AsyncClient(timeout=120) as client:
-                with open(file_path, "rb") as f:
-                    response = await client.post(
-                        self._convert_url,
-                        files={"files": (Path(file_path).name, f)},
-                        data={
-                            "losslessImageCompression": "true",
-                            "quality": "100",
-                        },
-                    )
-
-                if response.status_code == 200:
-                    return response.content
-
-                logger.error(
-                    f"{tag} | Gotenberg returned {response.status_code}: {response.text[:200]}"
-                )
-                return None
+        """Send file to Gotenberg and return PDF bytes.
 
+        The HTTP call runs inside ``asyncio.to_thread`` with a sync
+        ``httpx.Client`` + open file handle so httpx streams the multipart
+        body from disk instead of holding the entire upload in memory.
+        Keeps peak memory bounded under concurrent large-file conversions.
+        """
+        try:
+            return await asyncio.to_thread(self._post_sync, file_path, tag)
         except httpx.ConnectError:
             logger.info(f"{tag} | Gotenberg not reachable at {self.gotenberg_url}, skipping")
             return None
@@ -198,3 +189,21 @@ class GotenbergConverter(Converter):
         except Exception as e:
             logger.error(f"{tag} | Gotenberg error: {e}")
             return None
+
+    def _post_sync(self, file_path: str, tag: str) -> bytes | None:
+        """Blocking multipart upload to Gotenberg. Called via ``to_thread``."""
+        with httpx.Client(timeout=120) as client, open(file_path, "rb") as f:
+            response = client.post(
+                self._convert_url,
+                files={"files": (Path(file_path).name, f)},
+                data={
+                    "losslessImageCompression": "true",
+                    "quality": "100",
+                },
+            )
+
+        if response.status_code == 200:
+            return response.content
+
+        logger.error(f"{tag} | Gotenberg returned {response.status_code}: {response.text[:200]}")
+        return None

{markitdown_pro-2.1.1 → markitdown_pro-2.1.2}/markitdown_pro/services/azure_doc_intelligence.py

@@ -114,7 +114,7 @@ class DocumentIntelligenceService:
             return None
 
         try:
-            data = file_path.read_bytes()
+            data = await asyncio.to_thread(file_path.read_bytes)
 
             poller = await self._client.begin_analyze_document(
                 model_id="prebuilt-layout",