Check for page.find_tables returning None

PyMuPDF
1

Hi there,

I am running into an issue where we modify the pdf (deleting text) before running extraction, and this seems to corrupt the pdf in a way that crashes pymupdf4llm, because page.find_tables() returns None instead of raising, but then the None is not handled in the next line:

pymupdf_rag.py:1031-1032

tabs = page.find_tables(clip=parms.clip,
strategy=table_strategy)
for t in tabs.tables: # No None check!

The fix would be:
tabs = page.find_tables(clip=parms.clip,
strategy=table_strategy)
if tabs is not None:
for t in tabs.tables:

or maybe there should be a flag to either raise or ignore the tables?

2

Thanks @Oege_Dijk can you confirm which version of pymupdf_rag.py you are using? The latest one here: pymupdf4llm/pymupdf4llm/pymupdf4llm/helpers/pymupdf_rag.py at main · pymupdf/pymupdf4llm · GitHub seems to be different? Perhaps the latest version doesn’t exhibit this problem?