Unhandled Exception in the library crashes code binascii.Error: Odd-length string

[change-is-constant]

What happened? What were you trying to achieve?
I am trying to read text from a pdf file. But I am getting an error which I don't think is because of my code. The code crashes for the particular pdf mentioned below.

Environment

Which environment were you using when you encountered the problem?

$ python -m platform
Windows-10-10.0.22000-SP0

$ python -c "import PyPDF2;print(PyPDF2.__version__)"
2.10.9

Code + PDF

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

    if afile.endswith(".pdf"):
        pdfReader = PyPDF2.PdfFileReader(open(afile, "rb"), strict=False)
        texts = pdfReader.getPage(0).extract_text() # the ERROR happens on this line !!!!!
        for kw in keywords:
            if kw not in texts: return False
            # pbar.write(f'{kw} found')
        return True

PDF file: https://d4duas3s44z1s.cloudfront.net/Dat2020Live/document/test/question/Question_Report_299.pdf
don't add them to your tests!

Traceback

This is the complete Traceback I see:

File "path\to\there\Python\Python310\lib\site-packages\PyPDF2\_cmap.py", line 291, in parse_bfrange
    unhexlify(fmt % a).decode(
binascii.Error: Odd-length string

[change-is-constant]

Also, I am getting these unwanted console.logs for some files even after mentioning strict=False:

Multiple definitions in dictionary at byte 0x83b for key /Filter
Multiple definitions in dictionary at byte 0x358a for key /Filter
Multiple definitions in dictionary at byte 0x3b57 for key /Filter
Multiple definitions in dictionary at byte 0x42fe for key /Filter
Multiple definitions in dictionary at byte 0x4709 for key /Filter
Multiple definitions in dictionary at byte 0x4b08 for key /Filter

[pubpub-zz]

Thanks for this example.
Your file show some very odd but acceptable cmap definitions. I've produced a fix that improve the robustness.
I've noted also that the non roman characters are not properly read, but the characters can not been copied neither using acrobat reader : the 'Translation table' is not correct within the file.

About the logs, I've not been able to reproduce them after the fix, not tried before : can you confirm ?

[pubpub-zz]

adding test sample with similar results:
cmap1370.pdf

[MatteoRiva95]

Hello everyone, I know this is an old topic, but I am facing the same issue (binascii.Error: Odd-length string + Multiple definitions in dictionary at byte 0x3f42b for key /MediaBox) and I do not understand how it was fixed. I am using PyPDFDirectoryLoader from LangChain and the exact error is:

File "/usr/local/lib/python3.10/dist-packages/langchain_community/document_loaders/pdf.py", line 247, in load
raise e
File "/usr/local/lib/python3.10/dist-packages/langchain_community/document_loaders/pdf.py", line 239, in load
sub_docs = loader.load()
File "/usr/local/lib/python3.10/dist-packages/langchain_community/document_loaders/pdf.py", line 162, in load
return list(self.lazy_load())
File "/usr/local/lib/python3.10/dist-packages/langchain_community/document_loaders/pdf.py", line 172, in lazy_load
yield from self.parser.parse(blob)
File "/usr/local/lib/python3.10/dist-packages/langchain_community/document_loaders/base.py", line 102, in parse
return list(self.lazy_parse(blob))
File "/usr/local/lib/python3.10/dist-packages/langchain_community/document_loaders/parsers/pdf.py", line 95, in lazy_parse
yield from [
File "/usr/local/lib/python3.10/dist-packages/langchain_community/document_loaders/parsers/pdf.py", line 97, in
page_content=page.extract_text()
File "/usr/local/lib/python3.10/dist-packages/pypdf/_page.py", line 2076, in extract_text
return self._extract_text(
File "/usr/local/lib/python3.10/dist-packages/pypdf/_page.py", line 1588, in _extract_text
cmaps[f] = build_char_map(f, space_width, obj)
File "/usr/local/lib/python3.10/dist-packages/pypdf/_cmap.py", line 33, in build_char_map
font_subtype, font_halfspace, font_encoding, font_map = build_char_map_from_dict(
File "/usr/local/lib/python3.10/dist-packages/pypdf/_cmap.py", line 58, in build_char_map_from_dict
map_dict, space_code, int_entry = parse_to_unicode(ft, space_code)
File "/usr/local/lib/python3.10/dist-packages/pypdf/_cmap.py", line 237, in parse_to_unicode
process_rg, process_char, multiline_rg = process_cm_line(
File "/usr/local/lib/python3.10/dist-packages/pypdf/_cmap.py", line 313, in process_cm_line
multiline_rg = parse_bfrange(line, map_dict, int_entry, multiline_rg)
File "/usr/local/lib/python3.10/dist-packages/pypdf/_cmap.py", line 372, in parse_bfrange
] = unhexlify(fmt2 % c).decode("utf-16-be", "surrogatepass")
binascii.Error: Odd-length string

Can you explain, please? Thank you so much in advance!

[stefan6419846]

See #2216 for the current state.

[MatteoRiva95]

@stefan6419846 thank you for your reply. I have already commented the blog you posted, but the author could not give me a real solution to the problem. I wondered if someone here could give me a detailed one.

[stefan6419846]

Just commenting multiple threads of the same issue will most likely not resolve your issue. In the case of #1370, this already could be solved by some further change, see details in #1372. You are always invited to further debug/analyze #2216 and propose a corresponding PR to improve pypdf.