#!/bin/bash
# makeData: create data files for sentiment analysis
# usage: bin/makeData
# 20120312 erikt(at)xs4all.nl

# extract tweets with agreed annotation
bin/select < logfile > logfile.agreed

# select first tweet with one party from each user
../bin/onePartyPerTweet < logfile.agreed |\
   ../bin/oneTweetPerUser > logfile.selected

# extract negative and nonnegative tweet
grep '\-$' logfile.selected > logfile.selected.negative
grep '\.$' logfile.selected > logfile.selected.nonnegative

# sanity checks
wc logfile.selected.negative logfile.selected.nonnegative
wc logfile.selected

exit 0
