-
Notifications
You must be signed in to change notification settings - Fork 17
/
extraction.py
157 lines (126 loc) · 5.27 KB
/
extraction.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
'''
Copyright (c) 2017 Hanze Dong. All Rights Reserved.
extraction.py: a backup of main.py
It's the main function to perform a HMM to extract events
We embeded the POS information to improve the model and provide some punish options
'''
from read_data import *
import collections as cc
from HMM import *
import numpy as np
import time
from add_pos import *
import matplotlib.pyplot as plt
train_arg_sent,train_tri_sent,test_arg_sent,\
test_tri_sent,train_trigger,train_argument,set_arg,set_tri, \
argument_tuple, trigger_tuple,vocab_arg,vocab_tri=read_data()
param_arg = {'type':'arg','conf_prob':cc.Counter(argument_tuple),'arg':np.array(list(set_arg)),
'vocab':np.array(list(vocab_arg)),'Naive_bayes':0,'k':10**5,'O_punish':0.5,'diag_punish':0.5,
'conf_mat':np.zeros([(len(vocab_arg)),(len(set_arg))]),'k_conf_mat':10**-5,
'lambda':5,'lambda2':2
}
param_tri = {'type':'tri','conf_prob':cc.Counter(trigger_tuple),'tri':np.array(list(set_tri)),
'vocab':np.array(list(vocab_tri)),'Naive_bayes':0,'k':1,'O_punish':0,
'conf_mat':np.zeros([(len(vocab_tri)),(len(set_tri))]),'k_conf_mat':10**-5,'diag_punish':0.5,
'lambda':1,'lambda2':2
}
argument_test_pos,set_arg_te_pos = readpos('argument_test')
argument_train_pos,set_arg_tr_pos = readpos('argument_train')
trigger_test_pos,set_tri_te_pos = readpos('trigger_test')
trigger_train_pos,set_tri_tr_pos = readpos('trigger_train')
arg_pos_tuple = replace_tuple(argument_tuple,argument_train_pos)
tri_pos_tuple = replace_tuple(trigger_tuple,trigger_train_pos)
param_arg_pos = {'type':'arg','conf_prob':cc.Counter(arg_pos_tuple),'arg':np.array(list(set_arg)),
'vocab':np.array(list(set_arg_tr_pos.union(set_arg_te_pos))),'Naive_bayes':0,'k':1,'O_punish':0,
'conf_mat':np.zeros([len(set_arg_tr_pos.union(set_arg_te_pos)),(len(set_arg))]),'k_conf_mat':10**-5,
'lambda':0,'diag_punish':0
}
param_tri_pos = {'type':'tri','conf_prob':cc.Counter(tri_pos_tuple),'tri':np.array(list(set_tri)),
'vocab':np.array(list(set_tri_tr_pos.union(set_tri_te_pos))),'Naive_bayes':0,'k':1,'O_punish':0,
'conf_mat':np.zeros([len(set_tri_tr_pos.union(set_tri_te_pos)),(len(set_tri))]),'k_conf_mat':10**-5,
'lambda':0,'diag_punish':0.5
}
def build_conf_mat(conf_mat,conf_prob,vocab,att,add_param):
for i in range(len(vocab)):
for j in range(len(att)):
conf_mat[i, j] = conf_prob[(vocab[i], att[j])]
conf_mat += add_param
conf_mat = conf_mat.T / np.sum(conf_mat, axis=1)
conf_mat = conf_mat.T
return conf_mat
t1 = time.clock()
param_arg['conf_mat']=build_conf_mat(param_arg['conf_mat'],
param_arg['conf_prob'],param_arg['vocab'],param_arg['arg'],param_arg['k_conf_mat'])
param_arg_pos['conf_mat']=build_conf_mat(param_arg_pos['conf_mat'],
param_arg_pos['conf_prob'],param_arg_pos['vocab'],param_arg_pos['arg'],param_arg_pos['k_conf_mat'])
param_tri['conf_mat']=build_conf_mat(param_tri['conf_mat'],
param_tri['conf_prob'],param_tri['vocab'],param_tri['tri'],param_tri['k_conf_mat'])
param_tri_pos['conf_mat']=build_conf_mat(param_tri_pos['conf_mat'],
param_tri_pos['conf_prob'],param_tri_pos['vocab'],param_tri_pos['tri'],param_tri_pos['k_conf_mat'])
hmm_arg = HMM(param_arg)
hmm_arg_pos = HMM(param_arg_pos)
hmm_tri = HMM(param_tri)
hmm_tri_pos = HMM(param_tri_pos)
hmm_arg.word_count(train_arg_sent)
hmm_arg_pos.word_count(argument_train_pos)
hmm_tri.word_count(train_tri_sent)
hmm_tri_pos.word_count(trigger_train_pos)
hmm_arg.trans_prob(train_argument)
hmm_arg_pos.trans_prob(train_argument)
hmm_tri.trans_prob(train_trigger)
hmm_tri_pos.trans_prob(train_trigger)
f1 = codecs.open('trigger_test.txt', 'r', 'utf8')
trigger_test = f1.readlines()
f1.close()
f2 = codecs.open('argument_test.txt', 'r', 'utf8')
argument_test = f2.readlines()
f2.close()
for i in range(len(trigger_test)):trigger_test[i]=trigger_test[i].strip()
for i in range(len(argument_test)):argument_test[i]=argument_test[i].strip()
f = codecs.open('trigger_result.txt', 'w', 'utf8')
res = []
for i in range(len(test_tri_sent)):
x1 = hmm_tri_pos.predict(trigger_test_pos[i],'trigger')
x = hmm_tri.predict(test_tri_sent[i], 'trigger',hmm_tri_pos.inter_mat)
if len(set(x))==1:x = x1
res.append(param_tri['tri'][x.astype('int')])
res2 =[]
for i in range(len(test_arg_sent)):
x2 = hmm_arg_pos.predict(argument_test_pos[i])
x1 = hmm_arg.predict(test_arg_sent[i],'viterbi',hmm_arg_pos.inter_mat)
#for i in range(len(x1)):
# if param_arg['arg'][int(x1[i])]=='O':x1[i]=x2[i]
res2.append(param_arg['arg'][x1.astype('int')])
t2 = time.clock()
k=0
j=0
for i in range(len(trigger_test)):
if trigger_test[i]=='':
f.write('\n')
k=k+1
j=0
else:
f.write(trigger_test[i])
f.write('\t')
f.write(res[k][j])
f.write('\n')
j+=1
f.close()
k=0
j=0
f = codecs.open('argument_result.txt', 'w', 'utf8')
for i in range(len(argument_test)):
if argument_test[i]=='':
f.write('\n')
k=k+1
j=0
else:
f.write(argument_test[i])
f.write('\t')
f.write(res2[k][j])
f.write('\n')
j+=1
#plt.imshow(hmm_arg.trans_mat)
#plt.colorbar()
#plt.show()
print('Elapsed time',t2-t1)