docling-project / docling-project/docling-parse
Wrong text cell positions in some PDF
Open
bug
- Dominant language
- C++
- Stars
- 333
- Forks
- 80
- Avg merge
- 1d 14h
- Merged PRs (30d)
- 10
Description
There is apparently a scaling issue with the page cells for particular PDFs, one such sample is attached. It works correctly with pypdfium backend, but breaks on docling-parse-v2.
Source file:
[S111072430100016X.pdf](https://github.com/user-attachments/files/18348077/S111072430100016X.pdf)
Contributor guide
Assessment
This issue has not been assessed yet.