Better hash source

[OllyButters]

Was using a hash of the title to generate a unique ID, but some titles are the same! Switched to zotero ID, but this probably needs thinking through some more.

puma/source/clean/clean.py

Line 34 in d2c2450

this_paper['IDs']['hash'] = hashlib.md5(this_paper['IDs']['zotero'].encode('ascii', 'ignore')).hexdigest()