Add lemmas to token table

[JMankewitz]

Right now we're live stemming each word using the wordnet lemmatizer to match wordnet senses. We should consider storing the lemmas as a column in the tokens table instead.

WordSense/ws_web/public_views.py

Line 259 in 73430f0

word = lemmatizer.lemmatize(gloss, pos_map[pos])