Temporary disable multi threading for PDFs (#3196)

AndreaFrancis · web-flow · commit 4e78113aa873 · 2025-05-29T14:08:15.000-04:00
Temporary disable multithreading for PDFs
diff --git a/libs/libapi/src/libapi/rows_utils.py b/libs/libapi/src/libapi/rows_utils.py
@@ -66,7 +66,7 @@ async def transform_rows(
         offset=offset,
         row_idx_column=row_idx_column,
     )
-    if "Audio(" in str(features) or "Image(" in str(features) or "Pdf(" in str(features):
+    if "Audio(" in str(features) or "Image(" in str(features):
         # Use multithreading to parallelize image/audio files uploads.
         # Also multithreading is ok to convert audio data
         # (we use pydub which might spawn one ffmpeg process per conversion, which releases the GIL)

Original file line number	Diff line number	Diff line change
`@@ -66,7 +66,7 @@ async def transform_rows(`
`66`	`66`	`offset=offset,`
`67`	`67`	`row_idx_column=row_idx_column,`
`68`	`68`	`)`
`69`		`- if "Audio(" in str(features) or "Image(" in str(features) or "Pdf(" in str(features):`
	`69`	`+ if "Audio(" in str(features) or "Image(" in str(features):`
`70`	`70`	`# Use multithreading to parallelize image/audio files uploads.`
`71`	`71`	`# Also multithreading is ok to convert audio data`
`72`	`72`	`# (we use pydub which might spawn one ffmpeg process per conversion, which releases the GIL)`