OCR requests Tesseract 4 when I have 5

[xantof]

New install on Ubuntu 24.04.
I get
(1, 'Error opening data file /usr/share/tesseract-ocr/4.00/tessdata/eng.traineddata Please make sure the TESSDATA_PREFIX environment variable is set to your "tessdata" directory. Failed loading language 'eng' Tesseract couldn't load any languages! Could not initialize tesseract.')

My system Path is
/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/usr/share/tesseract-ocr/5/

[nelsonduarte]

Thanks @xantof — your PATH detail (/usr/share/tesseract-ocr/5/) made the root cause obvious.

The bug was that the OCR tool only looked for tessdata adjacent to the tesseract binary (the Windows / Homebrew layout). On Debian/Ubuntu the binary lives at /usr/bin/tesseract while the data is under /usr/share/tesseract-ocr/<version>/tessdata, so TESSDATA_PREFIX never got set and tesseract fell back to its compiled-in default — which on Ubuntu 24.04 still points at .../4.00/tessdata even though the package ships v5.

Fixed on main in 55ee16a: a new _find_tessdata() in app/tools/ocr.py globs /usr/share/tesseract-ocr/*/tessdata sorted descending (so v5 wins over v4.00), with /usr/share/tessdata, /usr/local/share/tessdata and the snap path as fallbacks.

Validated against the exact filesystem layout you reported (single /usr/share/tesseract-ocr/5/tessdata, no /4.00): the fix picks /usr/share/tesseract-ocr/5/tessdata and sets TESSDATA_PREFIX explicitly.

Ship target: v1.13.4, currently gated on a macOS CI fix that just went green (PR #30). The issue will auto-close when the release publishes — likely within a day or two depending on review pace. If you'd rather not wait, you can run from source today:

git clone https://github.com/nelsonduarte/PDFApps.git
cd PDFApps
pip install -r requirements.txt
python pdfapps.py

[nelsonduarte]

Released as v1.13.5 (bumped from 1.13.4 because the previous tag's macOS DMG build failed before publishing): https://github.com/nelsonduarte/PDFApps/releases/tag/v1.13.5

If you're on the AppImage or .deb you can grab it directly from the release page; the Snap, AUR and Copr packages will refresh automatically over the next few minutes via the publish workflow. Let me know if it works on your setup!