PGVector gotcha #70

New Issue

2026-02-15T17:05:51-05:00

yindo commented

2026-02-15 17:05:51 -05:00

Originally created by @humbertovnavarro on GitHub (Mar 28, 2025).

By default postgres uses ASCII encoding which causes document uploads to fail if they contain unicode characters.

Originally created by @humbertovnavarro on GitHub (Mar 28, 2025). By default postgres uses ASCII encoding which causes document uploads to fail if they contain unicode characters.

yindo added the documentation help wanted labels 2026-02-15 17:05:51 -05:00

yindo closed this issue

2026-02-15 17:05:51 -05:00

yindo commented

2026-02-15 17:05:52 -05:00

@mihirahlawat commented on GitHub (Apr 17, 2025):

I spent a lot of time and will try to save you some.

The issue: If PDFs have ligatures in them, PyPDF does not recognize them and gives you an empty character (in unicode) which throws error in a PGVector Database.

The solution: I used PyMuPDF for loading pdf documents. It is able to recognize ligatures seamlessly.

@mihirahlawat commented on GitHub (Apr 17, 2025): I spent a lot of time and will try to save you some. The issue: If PDFs have ligatures in them, PyPDF does not recognize them and gives you an empty character (in unicode) which throws error in a PGVector Database. The solution: I used PyMuPDF for loading pdf documents. It is able to recognize ligatures seamlessly.

yindo commented

2026-02-15 17:05:52 -05:00

@Classic298 commented on GitHub (Sep 14, 2025):

Hi, what improvement to the docs do you propose based on these observations? Not sure where to add it?

@Classic298 commented on GitHub (Sep 14, 2025): Hi, what improvement to the docs do you propose based on these observations? Not sure where to add it?

yindo referenced this issue

2026-02-15 17:15:51 -05:00

[PR #70] [CLOSED] added Dalle3 image generation tutorial #197

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: open-webui/docs#70