add missing extra indicies for analogy

update cognates

update cognates

unify analogies and train data

add analogies as a checked dataset

Fix vocabulary disprepancies.

fix keys

change from csv to jsonl

add cognates

update analogies

add ARPabet transcription to all words

update dataset

add incomplete version of train