def topicFilter():
dts.setSize(14000000)
dts.setFile("/home/server2103/dump/twitter.tweet.json",
"../entityOutput/topictwitter", "../log/matchtwitter")
dts.openFiles()
dts.loop(filterHashtags, 'filterHashtags')
dts.closeFiles()
def __testfind():
dts.setSize(3830000)
dts.setFile('../emojiOutput/afterPre.txt', '../emojiOutput/EmoAll.txt',
'../log/divideEmoticons')
dts.openFiles()
dts.loop(__testEmo, 'test emoji')
dts.closeFiles()
def featureVectorParse():
dts.setSize(10000)
dts.setFile('../data/featvect', '../emojiOutput/featureWang10000_no01',
'../log/featureWang')
dts.openFiles()
dts.loop(__lineParse, 'parse featvect')
dts.writeL(str(name_dict))
dts.closeFiles()
def divideHashtag():
dts.setSize(1000000)
dts.setFile('../hashOutput/afterPre.txt',
'../hashOutput/divideHashtag.txt', '../log/divideHashtag.log')
dts.openFiles()
dts.loop(__divide, 'divide by Hashtag')
for emo in EmoList:
print 'label %d \t: %d' % (emo['label'], emo['cnt'])
dts.writeL('label %d \t: %d\n' % (emo['label'], emo['cnt']))
dts.closeFiles()
def __cleanDup():
dts.openFiles()
tw = set()
def __push():
text = dts.readlineI()
tw.add(text)
dts.loop(__push, 'push into set')
print 'start write to file %s' % dts.ofileName
cnt = 0
for text in tw:
dts.writeO(text)
cnt += 1
print 'write finished, tot tweet left: %d' % cnt
dts.closeFiles()
def featureVectorParse():
topicList = [ur"#emabiggestfans1d", ur"#emabiggestfansjustinbieber", ur"#p**n", ur"#ipad", ur"#halloween", ur"#emabiggestfans5sos", ur"#stealmygirl", ur"#thewalkingdead", ur"#ebola", ur"#emabiggestfansarianagrande", ur"#lol"]
dfile = codecs.open( '../log/featureWang', 'r', 'utf-8' )
line = dfile.readline()
global name_dict
name_dict = eval( line )
dfile.close()
for topic in topicList:
ifilename = '../entityOutput/topicTwitter_' + topic[1:]
ofilename = '../entityOutput/topicFeat_' + topic[1:]
lfilename = '../log/featureVectorParse_entity'
dts.setSize( 50000 )
dts.setFile( ifilename, ofilename, lfilename )
dts.openFiles()
dts.loop( __lineParse, 'parse featvect:' + topic )
dts.closeFiles()
def labelCounter():
dts.setSize(100000)
dts.setFile('../data/featvect', '', '../log/featvectLabelCount')
dts.openFiles()
global counter
for x in range(9):
counter[x] = 0
dts.loop(__line, 'parse featvect')
sum = 0
for x in range(9):
sum += counter[x]
for x in range(9):
print 'Label\t%d\t:%d (%.2f%%)' % (
x, counter[x], float(counter[x] * 100.0) / float(sum))
dts.writeL('Label\t%d\t:%d (%.2f%%)\n' %
(x, counter[x], float(counter[x] * 100.0) / float(sum)))
print 'Sum\t\t:%d' % sum
dts.closeFiles()
loop_lfilename = '../Compare_Output/ans_unihash_'
all_ofilename = '../emojiOutput/feautre_unihash_all'
all_lfilename = '../Compare_Output/ans_unihash_all'
elif __type == 'UnigramEmoticon_run':
DictDir = '../emojiOutput/UnigramEmoticonDict'
loop_ofilename = '../emojiOutput/feautre_uniemo_'
loop_lfilename = '../Compare_Output/ans_uniemo_'
all_ofilename = '../emojiOutput/feautre_uniemo_all'
all_lfilename = '../Compare_Output/ans_uniemo_all'
load_Index()
for Emo in divideByEmoji.Emotions:
ifilename = divideByEmoji.OutputDir + Emo['name']
ofilename = loop_ofilename + Emo['name']
lfilename = loop_lfilename + Emo['name']
dts.setSize(100000)
dts.setFile(ifilename, ofilename, lfilename)
dts.openFiles()
PC = 0
dts.loop(parse_line, 'generating ' + Emo['name'])
dts.closeFiles()
ifilename = '../emojiOutput/featre_all'
dts.setSize(100000)
dts.setFile(ifilename, all_ofilename, all_lfilename)
dts.openFiles()
dts.loop(parse_line, 'generating all')
dts.closeFiles()
pass
"""
import re
import sys
import codecs
import HTMLParser
import preprocess_func
import dealTweets as dts
import utilities
#from nltk.stem.wordnet import WordNetLemmatizer
dts.setSize(25770000)
dts.setFile('../data/tweet_noRT_noDup.txt', '../output/afterPre.txt',
'../log/pre.log')
dts.openFiles()
tokenizer = utilities.Tokenizer()
def __preprocess():
line = preprocess_func.preprocess(dts.readlineI())
dts.writeO(line)
# terms = [term for term in tokenizer.tokenize(line)]
# print terms
dts.loop(__preprocess, 'preprocess symbols')
dts.closeFiles()
#! /usr/bin/env python
# -*- coding: utf-8 -*-
# vim:fenc=utf-8
#
# Copyright © 2014 Carwest Sung <*****@*****.**>
#
# Distributed under terms of the MIT license.
"""
"""
import dealTweets as dts
import json
def __io():
line = dts.readlineI()
if not line:
return
obj = json.loads(line)
text = obj['text']
dts.writeO(text + '\n')
dts.setSize(300)
dts.setFile('../emojiOutput/afterPre.txt', '../emojiOutput/checkAfterPre.txt')
dts.openFiles()
dts.loop(__io, 'io')
dts.closeFiles()
def dealLine():
line = dts.readlineI()
for emo in Emotions:
if emo['cnt'] > MaxEmotionSize:
continue
flag = -2
for eicon in emo['Icons']:
flag = line.find( eicon )
if flag != -1 :
emo['fileptr'].write( line )
break
if flag >= 0:
emo['cnt'] = emo['cnt'] + 1
dts.loop( dealLine, 'check Emoticons' )
for emo in Emotions:
emo['fileptr'].close()
print '============='
print 'processed Tweets:' + str( dts.processSize )
for emo in Emotions:
print emo['filename'] + ':' + str( emo['cnt'] )
dts.writeL( emo['filename'] + ':' + str( emo['cnt'] ) + '\n' )
dts.closeFiles()
#tfile = open( '../data/tweets_small.txt', 'r' )
#
#for x in range( processTweetSize + 1 ):
if ans == int(emo['label']):
label = 1
Emotions[ans - 1]['cnt'] += 1
tmp = {u'text': text, u'label': label}
emo['fileptr'].write(json.dumps(tmp) + '\n')
else:
if Emotions[ans - 1]['ncnt'] < Emotions[ans - 1]['cnt']:
label = -1
Emotions[ans - 1]['ncnt'] += 1
tmp = {u'text': text, u'label': label}
emo['fileptr'].write(json.dumps(tmp) + '\n')
pass
if __name__ == "__main__":
dts.setSize(2000000)
dts.setFile('../emojiOutput/afterPre.txt',
'../emojiOutput/test_featre_all',
'../log/test_labeled_by_emoji_log')
dts.openFiles()
for emo in Emotions:
emo['fileptr'] = codecs.open(OutputDir + emo['name'], 'w', 'utf-8')
dts.loop(__divide, 'divide and label twiiters')
for emo in Emotions:
print '%s\t:\t%d' % (emo['name'], emo['cnt'])
dts.writeL('%s\t:\t%d\n' % (emo['name'], emo['cnt']))
emo['fileptr'].close()
dts.closeFiles()
pass
def filterEmoticons():
dts.setSize( 310000 )
dts.setFile( '../data/tweet_noRT_noDup.txt', '../tmp/filter.out', '../log/filterEmoticons.log' )
dts.openFiles()
dts.loop( __cleanTweet, 'clean Tweets' )
dts.closeFiles()