fix: default font writer for character classification dataset

[charlesmindee]

This PR replaces the default PIL font to FreeMono, to avoid the font size issue and UnicodeEncode issue.

Closes both #417 and #416

Any feedback is welcome!

[fg-mindee]

Thanks for the PR! Considering the user is supposed to be the only one to know which fonts are installed on her/his computer, I don't think it's a good idea to move forward with this PR (cf. my comment)

[fg-mindee]

I added a suggestion to fix the CI and get a better user error/warning message

[codecov]

Codecov Report

Merging #418 (a9c32d6) into main (94e984c) will increase coverage by 0.03%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #418      +/-   ##
==========================================
+ Coverage   95.80%   95.84%   +0.03%     
==========================================
  Files          96       96              
  Lines        3937     3945       +8     
==========================================
+ Hits         3772     3781       +9     
+ Misses        165      164       -1     

Flag	Coverage Δ
unittests	95.84% <100.00%> (+0.03%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
doctr/datasets/classification/base.py	92.10% <100.00%> (+1.48%)	⬆️
doctr/models/backbones/mobilenet/tensorflow.py	97.50% <0.00%> (+0.06%)	⬆️
...dels/detection/differentiable_binarization/base.py	91.97% <0.00%> (+0.61%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 94e984c...a9c32d6. Read the comment docs.

[fg-mindee]

Thanks! I added a last improvement suggestion, but it's optional

[fg-mindee]

Thanks!