Closed ZacharyDuBois closed 8 years ago
fwiw the consume command can be used to read in .json (with values for 'tweet'), .csv (with values for column named 'tweet'), and even lines from a text file (one line = one 'tweet'): https://github.com/mispy/twitter_ebooks/blob/master/lib/twitter_ebooks/model.rb#L142
(Oh and then use append to assemble them one at a time into the model https://github.com/mispy/twitter_ebooks/blob/master/bin/ebooks#L123 )
@negatendo That was extremely helpful :+1: Thanks :)
It would be awesome if there was support for multiple data sources to feed some text into the system to speed up the learning.
Example: You archive the user's tweets and grep and filter their IRC logs and then consume them both into one model for one bot.