TypeError while loading some documents, caused by _cmap.py, line 93

[elhele]

While using the library I'm getting the following error:
TypeError: unsupported operand type(s) for /: 'IndirectObject' and 'int'

Environment

Which environment were you using when you encountered the problem?

macOS-10.16-x86_64-i386-64bit
pypdf==3.17.0, crypt_provider=('cryptography', '37.0.4'), PIL=9.0.1

It happens locally as well as during Azure-deployment.

Code + PDF

This is a minimal, complete example that shows the issue:

reader = PdfReader(file)
pages = reader.pages

Unfortunately I cannot share the document that causes the problem as it contains sensitive information. I also couldn't reproduce it with other documents. This adjustment after line 89, however, solves the problem:

import pypdf
...
sp_width = compute_space_width(ft, sp, space_width)
sp_width = sp_width if type(sp_width) != pypdf.generic._base.IndirectObject else sp_width.get_object()

Traceback

This is the complete Traceback I see:

  File "..././scripts/prepdocs.py", line 262, in <module>
    loop.run_until_complete(main(file_strategy, azd_credential, args))
  File ".../opt/anaconda3/lib/python3.9/asyncio/base_events.py", line 642, in run_until_complete
    return future.result()
  File "..././scripts/prepdocs.py", line 137, in main
    await strategy.run(search_info)
  File ".../scripts/prepdocslib/filestrategy.py", line 58, in run
    pages = [page async for page in self.pdf_parser.parse(content=file.content)]
  File ".../scripts/prepdocslib/filestrategy.py", line 58, in <listcomp>
    pages = [page async for page in self.pdf_parser.parse(content=file.content)]
  File ".../scripts/prepdocslib/pdfparser.py", line 52, in parse
    page_text = p.extract_text()
  File ".../scripts/.venv/lib/python3.9/site-packages/pypdf/_page.py", line 2284, in extract_text
    return self._extract_text(
  File ".../scripts/.venv/lib/python3.9/site-packages/pypdf/_page.py", line 1903, in _extract_text
    cmaps[f] = build_char_map(f, space_width, obj)
  File ".../scripts/.venv/lib/python3.9/site-packages/pypdf/_cmap.py", line 29, in build_char_map
    font_subtype, font_halfspace, font_encoding, font_map = build_char_map_from_dict(
  File ".../scripts/.venv/lib/python3.9/site-packages/pypdf/_cmap.py", line 93, in build_char_map_from_dict
    float(sp_width / 2),
TypeError: unsupported operand type(s) for /: 'IndirectObject' and 'int'

[pubpub-zz]

@elhele
can you in _cmap.py file, at the end of function compute_space_width (line460 normally)

    if isinstance(sp_width,indirect_object):    ## to be added
        sp_width = sp_width.get_object()      ## to be added
    return sp_width

and check weither it fixes the error

[elhele]

Hello @pubpub-zz thank you so much for such a fast reply!

"indirect_object" in the code above doesn't seem to be defined. If I change it in such a way, it works perfectly and fixes the error:

    indirect_object = pypdf.generic._base.IndirectObject ## added

    if isinstance(sp_width, indirect_object):  ## added

        sp_width = sp_width.get_object()  ## added

    return sp_width

[MartinThoma]

@elhele A PR with the fix was created. I guess until end of the week we will have the fix on PyPI.

I'm curious which program created the bad PDF. Can you share that?

from pypdf import PdfReader
reader = PdfReader('example.pdf')
print(reader.metadata)

It should be something like:

{'/ModDate': "D:20220901234405+02'00'", 
'/Creator': 'pdftk-java 3.2.2',
 '/CreationDate': "D:20220901234405+02'00'",
 '/Producer': 'itext-paulo-155 (itextpdf.sf.net-lowagie.com)'}

I'm especially interested in the /Creator and /Producer.

[elhele]

@MartinThoma thank you very much!

That's what I'm getting:

{'/Creator': 'Sejda Console 3.2.75',
'/Producer': 'SAMBox 1.1.53 (www.sejda.org)',
'/ModDate': "D:20230531143456+02'00'"}

I'm also getting this error only with one page of one document in my database. This page looks kind of misplaced and I don't think that this problem occurs very often.