Markdown conversion stops when using pymupdf4llm

[ozgurnsahin]

I am using pymudf4llm actively for my RAG based product so for it did not create any issues for the pdf but for a single pdf to_markdown stucks on a page of the pdf and process runs continuously I will share the pdf and my code below

Multiple-Input Variational Auto-Encoder.pdf

my code:
markdown_pages = pymupdf4llm.to_markdown("docs\domain1\Multiple-Input Variational Auto-Encoder.pdf",margins=0)

process stucks at the page 12 and my pymupdf version is 0.0.17

[robvandijk]

Note: this is the same issue as #215 and is solved by PR #216 . It is simply very slow - if you keep it running long enough it will finish. Once PR #216 has been merged, it will finish much sooner.

FYI this occurs when the array path_rects in pymupdf4llm/pymupdf4llm/helpers/multi_column.py becomes very large. For your PDF, the length of this array per page:

1. 651
2. 0
3. 37
4. 4
5. 1309
6. 17
7. 59
8. 154
9. 317
10. 387
11. 384
12. 64960

[ozgurnsahin]

Thank you for the information I will close the issue with this comment and wait for the patch.