refactor: Get the taxonomy support back to a functional state. There's now a taxonomy content source that wraps another normal content source like a blog posts' source. It works in tandem with a taxonomy content pipeline that will do the heavy lifting of figuring out what kind of terms exist and need to be baked.

[pytest]
norecursedirs = .* _darcs CVS .svn .git .hg *.egg-info dist build bower_components node_modules venv tmp* _cache {args}