-
Notifications
You must be signed in to change notification settings - Fork 0
/
pos_tag.py
executable file
·172 lines (139 loc) · 5.52 KB
/
pos_tag.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
#!/usr/bin/env python
def prepareModel():
'''Reading data'''
data = open('brown_bigrams_tagged.txt', 'r')
inputLines = data.readlines()
# lines = [line.rstrip().split('\t') for line in inputLines]
lines = [[s.lower() for s in line.rstrip('\n').split('\t')] for line in inputLines]
'''Filling frequences info'''
tagPairFrequency = dict()
wordTagFrequency = dict()
wordFrequency = dict()
tagFrequency = dict()
tagPairProbability = dict()
wordTagProbability = dict()
tags = set()
for line in lines:
tag1 = line[1]
tag2 = line[3]
word1 = line[0]
word2 = line[2]
tags.add(tag1)
tags.add(tag2)
frequency = int(line[4])
tagBigram = tag1 + ' ' + tag2
wordTagBigram1 = word1 + ' ' + tag1
wordTagBigram2 = word2 + ' ' + tag2
tagFrequency[tag1] = tagFrequency.get(tag1, 0) + frequency
wordFrequency[word1] = wordFrequency.get(word1, 0) + frequency
tagPairFrequency[tagBigram] = tagPairFrequency.get(tagBigram, 0) + frequency
wordTagFrequency[wordTagBigram1] = wordTagFrequency.get(wordTagBigram1, 0) + frequency
V = len(tagFrequency)
tags = list(tags)
'''Filling probabilities info'''
for key in tagPairFrequency.keys():
tagPairProbability[key] = float(tagPairFrequency.get(key, 0)) / (tagFrequency.get(key.split()[0], 0))
for key in wordTagFrequency.keys():
wordTagProbability[key] = float(wordTagFrequency[key] + 1) / (tagFrequency[key.split()[1]] + V)
return {"tagPair": tagPairProbability, "wordTag": wordTagProbability, "tags": tags}
# log = Logger()
'''Viterbi algorithm'''
def viterbi(O, tagPairProbability, wordTagProbability, tags):
# log.debug('Tags: \'{0}\''.format(str(tags)))
statesNumber = len(tags)
stepsNumber = len(O) + 1
stateProbability = []
for i in range(0, statesNumber):
stateProbability.append([0.0] * stepsNumber)
# stateProbability = np.zeros((statesNumber, stepsNumber), dtype=np.float32)
backtrack = []
for i in range(0, statesNumber):
backtrack.append([0] * stepsNumber)
# backtrack = np.zeros((statesNumber, stepsNumber), dtype=np.int32)
for i in range(0, statesNumber):
stateProbability[i][0] = 1.0 / statesNumber
# stateProbability[:, 0] = 1.0 / statesNumber
for step in range(1, stepsNumber):
word = O[step - 1]
for currentState in range(0, statesNumber):
currentTag = tags[currentState]
wordTagPair = word + ' ' + currentTag
currentWordTagProbability = wordTagProbability.get(wordTagPair, 0)
for previousState in range(0, statesNumber):
if currentState == previousState:
continue
previousTag = tags[previousState]
tagPair = previousTag + ' ' + currentTag
currentTagPairProbability = tagPairProbability.get(tagPair, 0)
previousProbability = stateProbability[previousState][step - 1]
probability = previousProbability * currentTagPairProbability * currentWordTagProbability
if stateProbability[currentState][step] < probability:
stateProbability[currentState][step] = probability
backtrack[currentState][step] = previousState
mostProbableState = 0
mostProbableStateProbability = stateProbability[0][stepsNumber - 1]
for currentState in range(0, statesNumber):
if stateProbability[currentState][stepsNumber - 1] > mostProbableStateProbability:
mostProbableStateProbability = stateProbability[currentState][stepsNumber - 1]
mostProbableState = currentState
states = [0] * (stepsNumber - 1)
# states = np.zeros(stepsNumber - 1, dtype=np.int32)
currentState = mostProbableState
for step in reversed(range(1, stepsNumber)):
states[step - 1] = currentState
currentState = backtrack[currentState][step]
return [tags[s] for s in states]
# print(mostProbableStateProbability)
# print([tags[s] for s in states])
def posTagPhrase(phrase):
return viterbi(phrase.split(), tagPairProbability, wordTagProbability, tags)
def test():
model = prepareModel()
message = "i love birds"
tagging = viterbi(message.split(), model["tagPair"], model["wordTag"], model["tags"]);
print(tagging)
import msgpack
from cocaine.decorators import http
from cocaine.worker import Worker
from cocaine.logging import Logger
@http
def main(request, response):
# response.write_head(200, [("Content-Type", "plain/text")])
response.write_head(200, [("Content-Type", "text/html")])
request = yield request.read()
message = request.request.get('message', None)
tagging = []
if (message == None):
message = ""
else:
model = prepareModel()
tagging = viterbi(message.split(), model["tagPair"], model["wordTag"], model["tags"]);
html = """
<html>
<body>
<h1>Part of the speech tagging!</h1>
"""
if len(tagging) > 0:
html += """
<h2>%s ---> %s</h2>
""" % (str(message.split()), str(tagging))
html += """
<form name="submitform" method="get" onSubmit="window.location.replace('/?message=' + document.submitform['sentence'])">
Sentence: <input type="text" name="message"><br>
<input type="submit" value="Tag!">
</form>
"""
html += """
Please, send your suggestions to: kashin.andrej@gmail.com
</body>
</html>
"""
# message = "i love birds"
# response.write(str(message) + " " + str(tagging))
response.write(html)
response.close()
W = Worker()
W.run({
'posTag': main,
})
# test()