Tables recognition on PDF

[bjmvercelli]

Hello folks.

I've been trying to extract tables from a PDF file but i'm having this issue when there are multiple tables in the same page. In the following image you can see that it is grouping all tables as a single one...

You can replicate it in my Collab

And you can find the (public) PDF here: edital_convocacao_matricula_res__medica_2024.pdf

[JorjMcKie]

This is an error in the logic that recognizes clusters of vector graphics. I have developed a fix.
[Ping me if you urgently need it].

I am going to open an issue here.

[JorjMcKie]

Thank you for the feedback!
Again the question: If you have access to your PyMuPDF installation (e.g. it is on your computer), simply replace its table.py ... if urgent.

[bjmvercelli]

If it's not a problem for you, i would really appreciate it :)

[JorjMcKie]

no problem at all, here is the file
table.zip

Just unpack and copy it over to site-packages/fitz/, replacing the file present there. Note that there also currently still is /fitz_old/, the "classic" implememtnation of PyMuPDF. We are maintaining this for a while more as a fallback. It contains the same file, so you might want to do the same replacement there.

[bjmvercelli]

Looks perfect! Thank you so much.

I'll be keeping an eye on PR's, looking foward to check on this one 🏆