Structure for test documents

Initial commit

Parses an EPUB archive into HTML. Currently dumps everything to STDOUT,
but can push into array and then sort according to table of contents.

Blockquotes are not handled correctly. Need to hack on (un)marshaler for
that to work correctly.

Titles are not always handled correctly. Need to look for more than just
h3 tags (i.e. h1, h2) and need to let these bubble up from div tags.