[discord] add support

[zWolfrost]

Closes #454.
I haven't tested it completely so there could still be edge cases that could cause problems.
The extractor requires the user to retrieve their Discord Token following the provided instructions.

[zWolfrost]

I had the intention of making this extractor useful especially when writing messages metadata to files using postprocessors, to possibly rebuild whole channels starting from those files, but I recognize that many format keywords (like avatars, banners, messages content etc.) are kind of unnecessary when someone's intention is to just download the images. Should I just hide them by default by prefixing them with an underscore? Other than that, the extractor is now feature-wise completed. Sorry for the mess in commit history.

[mikf]

You should extract as much metadata as possible (or reasonable) as long as it does not require additional HTTP requests or a lot of computation that goes to waste most of the time (e.g. twitter.text-tweets), in which case they should be hidden behind a metadata option or similar.