-
Notifications
You must be signed in to change notification settings - Fork 8
/
createFeatures.py
201 lines (162 loc) · 7.2 KB
/
createFeatures.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
import numpy as np
from sklearn import linear_model
import pandas as pd
from collections import namedtuple
import os.path
import features
import utils
class ModelNewText(object):
def __init__(self, brnspace=None, brnclst=None, embeddings=None):
self.featurestest = {} ## <name, flist>
self.test = []
self.brnclst = brnclst
self.brnspace = brnspace
self.embeddings = embeddings
self.fileid = None
self.testLabels = []
self.userid = []
self.time = []
self.tweetid = []
def loadFromCSV(self, filename):
original_df = pd.read_csv(filename, encoding='utf-8')
### given userID. Will be deleted in the final version.
self.userid = list(original_df.loc[:, 'USERID'])
try:
self.time = list(original_df.loc[:, 'TweetTime'])
self.tweetid = list(original_df.loc[:, 'TweetID'])
self.testLabels = list(original_df.loc[:, 'Score'])
except KeyError:
pass
self.test = [features.ParseText(y) for y in
list(original_df.loc[:, 'Tweet'])]
print(len(self.test))
def loadFromTSV(self, filename):
original_df = pd.read_csv(filename, encoding='utf-8', sep="\t")
print(original_df.shape)
### given userID. Will be deleted in the final version.
self.userid = list(original_df.loc[:, 'user_id'])
try:
self.time = [int(y.split()[-1]) for y in
list(original_df.loc[:, 'time'])]
self.tweetid = list(original_df.loc[:, 'tweet_id'])
self.testLabels = list(original_df.loc[:, 'Score'])
except KeyError:
pass
self.test = [features.ParseText(y) for y in
list(original_df.loc[:, 'text'])]
def loadFromFile(self, filename):
self.test = []
self.fileid = os.path.basename(filename)
i = 0
with open(filename) as f:
for line in f:
if len(line.strip()) == 0: continue
self.test.append(features.ParseText(line))
i += 1
print(len(self.test))
f.close()
def loadSentences(self, identifier, sentlist):
## sentlist should be a list of sentence strings, tokenized;
## identifier is a string serving as the header of this sentlst
self.test = []
self.fileid = identifier
for i, sent in enumerate(sentlist):
self.test.append(
Instance(identifier + "." + str(i), 0, features.RawSent(sent)))
def _add_feature(self, key, values):
if key in self.featurestest: return
self.featurestest[key] = values
def numEmoji(self):
df = pd.DataFrame()
recs = [features.RawSent(r) for r in self.test]
df["numsymbols"] = features.numSymbols(recs, normalize=True)
df["numemoji"] = features.countEmoji(recs, normalize=True)
df.to_csv("symbol&emoji.csv", encoding="utf-8")
# Genereate word embeddings.
def fNeuralVec(self):
sentlst = [features.RawSent(r) for r in self.test]
keys = ["word_embed-" + str(i) for i in range(100)]
if keys[0] not in self.featurestest:
embeddingList = features.word_2_weights(sentlst, self.embeddings)
for fid, fname in enumerate(keys):
self.featurestest[fname] = [embeddingList[j][fid] for j in
range(len(embeddingList))]
print("Successfully generate word_embdding features")
# POS_Tags.
def fPostag(self):
sentlst = [features.RawSent(r) for r in self.test]
pos_tag = features.extractPOS(sentlst)
Useful_Tag = ['DT', 'NN', "VB", 'JJ', 'IN', '.', 'PRP', 'NNP', 'WP']
for i in Useful_Tag:
self._add_feature(i, pos_tag.loc[:, i])
# Brown CLuster, short 100 vectors.
def fBrownCluster_100(self):
sentlst = [features.RawSent(r) for r in self.test]
keys = ["brnclst_100-" + str(i) for i in range(100)]
if keys[0] not in self.featurestest:
print("Start initialize Browncluster ....")
brownClus, cluster_2_index = utils.readMetaOptimizeBrownCluster_100()
print("finished generating brownClusterlist !")
self.brnclst = brownClus
brownClusterList = features.brownCluster(sentlst, brownClus,
cluster_2_index, 100)
for fid, fname in enumerate(keys):
self.featurestest[fname] = [brownClusterList[j][fid] for j in
range(len(brownClusterList))]
# NE and Concrete words.
def NE_Concrete(self):
sentlst = [features.RawSent(r) for r in self.test]
pos_tag = features.NE_Concrete_Emo(sentlst)
Useful_Tag = ['ORGANIZATION', "PERCENT", 'PERSON', 'DATE', 'MONEY',
'TIME', 'LOCATION', 'Concrete']
for i in Useful_Tag:
self._add_feature(i, pos_tag.loc[:, i])
def transformEmoji(self):
recs = [features.RawSent(r) for r in self.test]
self._add_feature("numemoji", features.countEmoji(recs, normalize=True))
def fShallow(self):
normalize = True
recs = [features.RawSent(r) for r in self.test]
self._add_feature("avgwordlen", features.avgWordLen(recs))
self._add_feature("sentlen", features.sentLen(recs))
self._add_feature("numsymbols", features.numSymbols(recs, normalize))
self._add_feature("numcapltrs", features.numCapLetters(recs, normalize))
self._add_feature("numnumbers", features.numNumbers(recs, normalize))
################## 4 main feature groups. #########################
### Notice, in our best model, we did not use the transormTweet feature.
# Surface and Lexical features,
def transLexical(self):
self.fShallow()
self.fPostag()
self.NE_Concrete()
# Distributional word representations
def transEmbedding(self):
self.fNeuralVec()
self.fBrownCluster_100()
# Emotion features
def transEmotionFeature(self):
self.transformEmoji()
try:
f = pd.read_csv("NE_Concrete_Emo.csv")
self._add_feature("Negative", f.loc[:, 'Negative'])
self._add_feature("Positive", f.loc[:, 'Positive'])
except IOError:
sentlst = [features.RawSent(r) for r in self.test]
file = features.NE_Concrete_Emo(sentlst)
self._add_feature("Negative", file.loc[:, 'Negative'])
self._add_feature("Positive", file.loc[:, 'Positive'])
# def transformTweet(self):
# recs = [features.RawSent(r) for r in self.test]
# self._add_feature("numurl",features.numUrl(recs))
# tweet_begin, tweet_else = features.user_begin_or_else(recs)
# self._add_feature("user_beginning",tweet_begin)
# self._add_feature("user_else",tweet_else)
def transform_features(self):
df = pd.DataFrame()
df["userID"] = self.userid
df["Tweet"] = self.test
for feature in self.featurestest.keys():
df[feature] = self.featurestest[feature]
# print(len(self.feature))
df.to_csv("./output/test.csv", sep='\t')
print("DOne")