some bug found with version 3.3

[htc1977]

Hello @ciur,

I installed the 3.3 via docker with OCR worker, redis and PostgreSQL.
I found the following bugs so far:

• Popups are too far to the right
• There is no way to setup if uploaded / imported documents should be OCR'd or not or I haven't found it.
• documents are not indexed after upload and OCR leads to way bigger files due to missing JBIG2 compression and the fact, that text documents are converted to pictures and OCR'd afterwards. (Mentioned with earlier versions already)
• search doesn't find anything and the waiting bar is cycling forever

I haven't tested the API so far due to the fact, that there are too many bugs and no clear option to index the files without reOCRing them.

[ciur]

Are you testing on Windows ? If yes, then what version of windows ?

[htc1977]

I installed the dockers via Portainer on my Synology NAS and my OS with my browser is Linux / Ubuntu 24.10 and Brave [Version 1.73.104 Chromium: 131.0.6778.204 (Offizieller Build) (64-Bit)]

[ciur]

Thank you providing details. I was able to reproduce the issue with Brave browser version 1.73.
Next week I will provide a fix for the problem and release it as part of version 3.3.1

[ciur]

@htc1977 can you please try:

docker run -p 2025:80 \
    -e PAPERMERGE__SECURITY__SECRET_KEY=abc \
    -e PAPERMERGE__AUTH__PASSWORD=pass123 \
    papermerge/papermerge:3.3.1a2

Locally to double check if it fixes the problem ?
This is not yet official release. Just a quick check if solution I've came with - works.

Note the 3.3.1a2 tag.

[htc1977]

This fix works for me as well ...
Now let's move on to the other issues 😄

[ciur]

Ragarding:

There is no way to setup if uploaded / imported documents should be OCR'd or not or I haven't found it.

There is an option which may be relevant for you: https://docs.papermerge.io/3.3/settings/ocr/#ocr__automatic
OCR configurations via UI is part of roadmap for this year (2025): https://docs.papermerge.io/3.3/roadmap/#2-ui-settings
Once UI settings is in place, I will add configurations of the OCR via UI.

Regarding indexing/searching: did you setup SOLR + worker here: https://docs.papermerge.io/3.3/setup/docker-compose/#i3-ocr-workers ?
Could you please provide more context ? Or maybe open a separate ticket for search issue ?

[kashyn89]

This fix works for me as well ... Now let's move on to the other issues 😄

Fixed it for me as well, thanks, sorry for the late reply

[htc1977]

I changed my dockerfile and added i3worker and solr (both were missing) and now search seems to work 👍 , but only if imported documents were OCRd, so it seems that indexing of documents isn't done without OCR, is it?

I have around 5000 pdfs I'd like to import and they are text documents or OCRd scans and I don't want to OCR them again nor to I want to increase files by factors due to the missing jbig2 compression in the ocr-worker.