html5lib (HTML parser based on the HTML5 specification)

HTML parser designed to follow the HTML5 specification. The parser is
designed to handle all flavours of HTML and parses invalid documents
using well-defined error handling rules compatible with the behaviour of
major desktop web browsers.

Output is to a tree structure; the current release supports output
to DOM, ElementTree and lxml tree formats as well as a simple
custom format.

Optional: datrie, lxml, and genshi