Error when unicode character with add_image

Hello, The first code below is working, but the second raises an error. Of course, both /home/mth/nounicode.jpg and /home/mth/unicodé.jpg exist (and both are the same duplicate picture).

#!/usr/bin/env python
# coding: utf-8

from pylatex import Document, Figure

doc = Document()

with doc.create(Figure(position='h!')) as kitten_pic:
    kitten_pic.add_image('/home/mth/nounicode.jpg')     # No problem with this name

doc.generate_pdf('toto', clean_tex=False)

#!/usr/bin/env python
# coding: utf-8

from pylatex import Document, Figure

doc = Document()

with doc.create(Figure(position='h!')) as kitten_pic:
    kitten_pic.add_image('/home/mth/unicodé.jpg')        # The name with é raises an error

doc.generate_pdf('toto', clean_tex=False)

>>> 
Traceback (most recent call last):
  File "/usr/local/lib/python3.4/dist-packages/pylatex/document.py", line 228, in generate_pdf
    stderr=subprocess.STDOUT)
  File "/usr/lib/python3.4/subprocess.py", line 620, in check_output
    raise CalledProcessError(retcode, process.args, output=output)
subprocess.CalledProcessError: Command '['latexmk', '--pdf', '--interaction=nonstopmode', 'toto.tex']' returned non-zero exit status 12

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/mth/Pb accents.py", line 11, in <module>
    doc.generate_pdf('toto', clean_tex=False)
  File "/usr/local/lib/python3.4/dist-packages/pylatex/document.py", line 239, in generate_pdf
    print(e.output.decode())
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xe9 in position 3116: invalid continuation byte

In the log file, I find this :

LaTeX Warning: File `/home/mth/unicod\IeC {\'e}.jpg' not found on input lin
e 17.


! Package pdftex.def Error: File `/home/mth/unicod\E9.jpg' not found.

See the pdftex.def package documentation for explanation.
Type  H <return>  for immediate help.
 ...                                              
                                                  
l.17 ...0.8\textwidth]{/home/mth/unicodé.jpg}
                                                  %
Using draft setting for this image.

Would you have a clue to how I could manage this kind of situation ? By the way, thank you for your work !

[JelteF]

No problem. Thanks for the info. I’ll leave the issue open for now and will maybe fix it myself using your snippets in the future when I have some time.

[mth31]

I’m afraid I can’t do that. I have no idea how to create a pull request, and I am an amateur in python. However I watched in the code, and I would suggest to add this lines in the class StandAloneGraphic from figure.py, first in the def __init__ function :

if not all(ord(char) < 128 for char in filename):
    packages.add(Package('grffile', options=['encoding', 'filenameencoding=utf8']))

But I’m not sure if packages.add will work (in my case, I was using doc.packages.add instead).

Also, what if someone doesn’t use utf8 as encoding but something else ? (might be rare though).

I’m afraid that’s all I can do myself. Thank you very much, sorry I can’t help more.