.ipynb to .py fails with encoding error

[lebigot]

A Jupyter notebook that contains Unicode characters seems to be failing when converted to Python code:

  …
  File "/…/anaconda/envs/py2/lib/python2.7/site-packages/ipymd/utils/utils.py", line 133, in _write_text
    f.write(contents)
UnicodeEncodeError: 'ascii' codec can't encode character u'\xe9' in position 15310: ordinal not in range(128)

Encoding the Python file in UTF-8 instead works (f.write(contents.encode("utf-8"))), but then probably this should also add a # coding=UTF-8 at the beginning of the generated Python file.

Also, the conversion back to a Jupyter notebook might have to be adapted so as to be compatible with this change.

[rossant]

does the bug also exists with Python 3?

[lebigot]

It's a bit difficult to test: pip install installs under Python 2. Running the ipymd script with Python 3 doesn't find the ipymd module…