Stall when extracting using ocr on macos from pdf with embedded images

yobix-ai / extractous

Fast and efficient unstructured data extraction. Written in Rust with bindings for many languages.

Apache License 2.0

416 stars 17 forks source link

Open nmammeri opened 2 weeks ago

nmammeri commented 2 weeks ago

Process stalls until killed, when running on MacOS with OCR enabled with PDF documents that has embedded images in them.
OCR works fine with direct images but the bug is seen only on PDFs with embedded images
Tests failing (stalling forever) eng-ocr.pdf and deu-ocr.pdf

Initial investigation points to the bad handling of AWT within GraalVM . I think it's due to the AWT dispose thread deadlocking.

nmammeri commented 2 weeks ago

This bug is only experienced with extract to stream api on macos and not extract to string