support lang attribute for language

[snarfed]

https://github.com/microformats/microformats2-parsing/issues/3 is the current proposal for extracting and surfacing the lang attribute in parsed mf2 for a page's or item's language. that proposal isn't official yet, but it's been stable for over 2y, and it's implemented in PHP and JS, so it'd be nice for m2py to support it too!

example use case: apple/itunes require the <language> element in podcast RSS feeds. i'd like to convert HTML lang to RSS <language> in granary (snarfed/granary#124).