sources.markdown: add extract_from_text helper

[seanbreckenridge]

related #233

even if the strategy of converting markdown into HTML on the backend
isn't merged, can see this being useful for other people
writing their own sources for promnesia (e.g. me, for my discord module)

[seanbreckenridge]

Have integrated this with the sources.github as well, but not sure if you wanted to merge that here, or if you wanted to do something to render it in JS instead

As an example:

This random comment from HPI renders like this in the sidebar when converted from markdown to HTML:

Also worth mentioning that the <pre> blocks break the horizontal scrolling

)

[seanbreckenridge]

Wasnt sure where to add tests for this, so just to show it works;

In [1]: from promnesia.sources.markdown import extract_from_text

In [2]: import itertools

In [4]: import my.github.all
In [5]: for x in itertools.islice(my.github.all.events(),3):
   ...:     if x.body is not None:
   ...:         for url in extract_from_text(x.body):
   ...:             print(url)
   ...:
Parsed(url='https://github.com/seanbreckenridge/dotfiles/commit/4c57fd97cbb2605e63d0cf5d2af37039fe6e6d35', context='!html <p>Note: This is used for <a href="https://github.com/seanbreckenridge/dotfiles/commit/4c57fd97cbb2605e63d0cf5d2af37039fe6e6d35">gitopen</a></p>\n')

[karlicoss]

Nice, thanks!

Also worth mentioning that the pre blocks break the horizontal scrolling

Ah yeah -- I guess what I meant 'something like pre' -- it should be tweakable with an ad-hoc style maybe like nowrap or something like that.