Type Error during extracting pages in some pdfs

[psrubing]

Hello,

I've encountered bug during extrating pages using extract_pages() function from pdfminer.high_level module. This only happens to some pdf-s.
Image below provides this bug:

Below pdf implies this bug:
pdf_bug.pdf

Environment:
Python - 3.7.11
pdfminer.six - 20201018

[pietermarsman]

Can replicate:

$ PYTHONPATH=. python tools/pdf2txt.py ~/Downloads/pdf_bug.pdf 
Traceback (most recent call last):
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/pdffont.py", line 920, in char_width
    return cast(Dict[int, float], self.widths)[cid] * self.hscale
KeyError: 67

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/pdffont.py", line 924, in char_width
    return str_widths[self.to_unichr(cid)] * self.hscale
KeyError: 'a'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/pieter/projects/pdfminer-upstream/tools/pdf2txt.py", line 313, in <module>
    sys.exit(main())
  File "/home/pieter/projects/pdfminer-upstream/tools/pdf2txt.py", line 307, in main
    outfp = extract_text(**vars(parsed_args))
  File "/home/pieter/projects/pdfminer-upstream/tools/pdf2txt.py", line 62, in extract_text
    pdfminer.high_level.extract_text_to_fp(fp, **locals())
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/high_level.py", line 121, in extract_text_to_fp
    interpreter.process_page(page)
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/pdfinterp.py", line 991, in process_page
    self.render_contents(page.resources, page.contents, ctm=ctm)
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/pdfinterp.py", line 1010, in render_contents
    self.execute(list_value(streams))
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/pdfinterp.py", line 1036, in execute
    func(*args)
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/pdfinterp.py", line 896, in do_TJ
    self.device.render_string(
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/pdfdevice.py", line 133, in render_string
    textstate.linematrix = self.render_string_horizontal(
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/pdfdevice.py", line 173, in render_string_horizontal
    x += self.render_char(
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/converter.py", line 206, in render_char
    textwidth = font.char_width(cid)
  File "/home/pieter/projects/pdfminer-upstream/pdfminer/pdffont.py", line 926, in char_width
    return self.default_width * self.hscale
TypeError: unsupported operand type(s) for *: 'PDFObjRef' and 'float'

[pietermarsman]

Probably the solution is to call resolve1 when getting the default width.

[psrubing]

Hi, thanks for response, but I don't understand your comment. Function extract_pages() doesn't take any parameter related to resolve1 as from my knowledge based on documentation:

Or did I missed something?
Best regards

[pietermarsman]

I mean, to fix this issue we have to make a change to pdfminer.six, using resolve1(). This is a bug in the current code.

[psrubing]

Okey, I understand now :) Do you know approximate time of release with this fix?

[pietermarsman]

Nobody is working on it as far as I know.

Do you have time to work on this?

[psrubing]

Unfortunately I don't :/ Have to work on different projects, but if something change I will update and could look at this bug.

[gosiafilipek]

I found another file that generate similar error.
I already found solution so I will create pull request.

Below pdf implies this bug
pdf_bug2.pdf

[datatalking]

@pietermarsman @psrubing This was an issue for me in a past project and they ended up using an OCR solution. I was going to say I could take a look at this to debug it but noticed @gosiafilipek created a PR solution?

If those are done I have time in the next 2 months to contribute, but didn't see a 'good first issue' icon or whatever its called so I looked back and found these I could start with. Does anyone have requests or recommendations on where I should start?

#470
#154
#499
#497

[pietermarsman]

Hi @datatalking,

Thanks for reaching out! And for wanting to help! You can get in touch on gitter.im. In the private or group chat. We can have a sync about what to work on.

I'll try and see if I can create a good-first-issue label.

[pietermarsman]

Fixed by #772