ignore blocks without index.html and save date info

[enjalot]

adding the created_at and updated_at fields to the blocks makes it possible to analyze the corpus by date. also removing blocks that don't have an index.html gets rid of cruft as well as tributaries (which shouldn't really show up imo)