Add a parser

[fake-name]

Fair bit of changes here.

First, there's the ability to optionally (default off) use a proper parser for formatting. This is MUCH more robust, and about 10+ times faster then the pure-regex solution.
It also should be much more correct, in that it much closer approaches how an actual browser understands HTML (it does full parsing).

I made some other minor changes that should probably improve the normal performance a little tiny bit. Mostly, the original codebase was using rawcode_flat_list = re.split('\n', rawcode_flat) in several places. I don't see any reason to use re for that, since your pattern is a fixed single char. I replaced all instances like that with the plain old rawcode_flat_list = rawcode_flat.split('\n'), which should push it down to the pure C layer.

Lastly, I added the ability to change the indentation character. It was previously hard-coded to a literal tab character, and while I personally agree with that, it's probably better to not start another holy war, at least unintentionally.

If the indent_with parameter is specified, each indentation level is created with one instance of the parameter string. Otherwise, it defaults to 4 spaces ( ).

Caveats

This change also requires introducing a dependency, namely on the bs4 package. I don't think this is a big deal, but that's just me.

---

Hopefully, this contains 100% less asshole bug reports.

[rareyman]

Awesome! I am looking forward to checking it out! I'm inundated with work project stuff, but I hope to check it out when I have some time to focus on all this. Thanks, so much!