fix stage pipeline step to maintain extension

Unstructured-IO · Dec 16, 2024 · bc53e7e · bc53e7e
1 parent 438f9b9
commit bc53e7e
Showing 1 changed file with 3 additions and 1 deletion.
diff --git a/unstructured_ingest/v2/pipeline/steps/stage.py b/unstructured_ingest/v2/pipeline/steps/stage.py
@@ -39,11 +39,13 @@ async def _run_async(
         self, fn: Callable, path: str, file_data_path: str
     ) -> UploadStageStepResponse:
         path = Path(path)
+        # Maintain extension
+        output_filename = f"{self.get_hash(extras=[path.name])}{path.suffix}"
         fn_kwargs = {
             "elements_filepath": path,
             "file_data": FileData.from_file(path=file_data_path),
             "output_dir": self.cache_dir,
-            "output_filename": self.get_hash(extras=[path.name]),
+            "output_filename": output_filename,
         }
         if not asyncio.iscoroutinefunction(fn):
             staged_output_path = fn(**fn_kwargs)