chakert

[Harut]

Hi!

Please take a look at my typography library https://github.com/Harut/chakert . It uses lxml to parse HTML and ignore tags during replacement. And it allows to write replacement rules based on token objects, not on raw regexps.

If you want you can join by implementing rules for French. Or just if you allow, I can take your test cases and implement them in future (onfortunately, you didn't provide a license, so I can't just take them).

Thanks for your reply!