-
Notifications
You must be signed in to change notification settings - Fork 0
/
get_language_model_for_positive_entities.py
240 lines (205 loc) · 8.73 KB
/
get_language_model_for_positive_entities.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
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
"""
get the language model surrounding some entities
"""
import os
import json
import sys
import re
import argparse
import codecs
from myUtility.corpus import Document, Sentence
def get_sentence_window(entity_map,sentence,windows):
"""
Use the whole sentence as the context
"""
#print sentence
for w in entity_map:
if sentence.find(w) != -1:
temp_sentence = sentence
#print "found sentence %s" %temp_sentence
for t in entity_map:
if entity_map[t]:
temp_sentence = temp_sentence.replace(entity_map[t],"")
elif temp_sentence.find(t) != -1:
temp_sentence = temp_sentence.replace(t,"")
#print "after process %s" %temp_sentence
if entity_map[w]:
w = entity_map[w]
if w not in windows:
windows[w] = Sentence(temp_sentence,remove_stopwords=True).stemmed_model
else:
windows[w] += Sentence(temp_sentence,remove_stopwords=True).stemmed_model
def get_entity_map(words):
entity_map = {}
multiple = []
for w in words:
entity_map[w] = None
for e in words:
if w != e:
if e.find(w)!=-1:
if entity_map[w] != None:
if entity_map[w].find(e) != -1:
pass
elif e.find(entity_map[w]) != -1:
entity_map[w] = e
else:
multiple.append(w)
#print "For %s, there are two none substring candidates: %s %s" %(w, e,entity_map[w] )
#sys.exit(-1)
else:
entity_map[w] = e
# delete the ones that have multiple possibilities
for w in multiple:
entity_map.pop(w, None)
#print json.dumps(entity_map)
#sys.exit(-1)
return entity_map
def get_nochange_map(words):
entity_map = {}
for w in words:
entity_map[w] = w
return entity_map
def get_text_window(entity_map,sentence,windows,window_size):
"""
Use a sized text window as the context
"""
for w in entity_map:
if sentence.find(w) != -1:
temp_sentence = sentence
#print "found sentence %s" %temp_sentence
for t in entity_map:
if t.find(w) != -1:
continue
if entity_map[t]:
temp_sentence = temp_sentence.replace(entity_map[t],"")
elif temp_sentence.find(t) != -1:
temp_sentence = temp_sentence.replace(t,"")
if entity_map[w]:
w = entity_map[w]
w_size = w.count(" ")+1
temp_sentence = re.sub(" +"," ",temp_sentence)
temp_sentence += ' ' #little trick to ensure that the last token of sentence is a space
spaces = [m.start() for m in re.finditer(' ', temp_sentence)]
for m in re.finditer(w,temp_sentence):
start = m.start()-1
if start in spaces:
w_start = max(0,spaces.index(start)-window_size)
w_end = min(len(spaces)-1,spaces.index(start)+window_size+w_size)
#window_string = document[spaces[w_start]:spaces[w_end]]
window_string = temp_sentence[spaces[w_start]:m.start()-1] +" "+ temp_sentence[m.end()+1:spaces[w_end]]
else:
w_end = min(len(spaces)-1,window_size+w_size-1)
#window_string = document[0:spaces[w_end]]
window_string = temp_sentence[m.end()+1:spaces[w_end]]
if w not in windows:
windows[w] = Sentence(window_string,remove_stopwords=True).stemmed_model
else:
windows[w] += Sentence(window_string,remove_stopwords=True).stemmed_model
def get_files(a_dir):
all_files = os.walk(a_dir).next()[2]
files = []
for f in all_files:
files.append( f )
return files
def show_documents(documents):
for instance in documents:
print "%s:" %instance
for single_file in documents[instance]:
for sentence in documents[instance][single_file].sentences:
print "%s:%s" %(single_file,sentence.text)
print "-"*20
def get_all_sentence_windows(documents,entities_judgement):
windows = {}
for instance in documents:
print "%s:" %instance
words = []
for entity_type in entities_judgement[instance]:
words += entities_judgement[instance][entity_type]
if entity_type not in windows:
windows[entity_type] = {}
entity_map = get_nochange_map(words)
temp_windows = {}
for single_file in documents[instance]:
print "process file %s" %single_file
for sentence in documents[instance][single_file].sentences:
get_sentence_window(entity_map,sentence.text,temp_windows)
for w in temp_windows:
for entity_type in entities_judgement[instance]:
if w in entities_judgement[instance][entity_type]:
windows[entity_type][w] = temp_windows[w]
break
return windows
def get_all_text_windows(documents,entities_judgement,window_size):
windows = {}
for instance in documents:
print "%s:" %instance
words = []
for entity_type in entities_judgement[instance]:
words += entities_judgement[instance][entity_type]
if entity_type not in windows:
windows[entity_type] = {}
entity_map = get_nochange_map(words)
temp_windows = {}
for single_file in documents[instance]:
print "process file %s" %single_file
for sentence in documents[instance][single_file].sentences:
get_text_window(entity_map,sentence.text,temp_windows,window_size)
for w in temp_windows:
for entity_type in entities_judgement[instance]:
if w in entities_judgement[instance][entity_type]:
windows[entity_type][w] = temp_windows[w]
break
return windows
def main():
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("disaster_name")
parser.add_argument("--top_dir",'-tp',default='/lustre/scratch/lukuang/Temporal_Summerization/TS-2013/data/disaster_profile/data')
parser.add_argument("dest_dir")
parser.add_argument("run_id",type=int)
parser.add_argument("--using_text_window","-u",action='store_true')
parser.add_argument("--window_size",'-wz',type=int,default=3)
parser.add_argument("--entity_judgement_file","-e",default="/lustre/scratch/lukuang/Temporal_Summerization/TS-2013/data/disaster_profile/data/src/new_judgement.json")
args=parser.parse_args()
data = ""
with open(args.entity_judgement_file) as f:
data = f.read()
entities_judgement_data = json.loads(data)
entities_judgement = {}
single = entities_judgement_data[args.run_id-1]
q = single["query_string"]
single.pop("query_string",None)
entities_judgement[q] = single
#for single in entities_judgement_data:
# q = single["query_string"]
# single.pop("query_string",None)
# entities_judgement[q] = single
args.top_dir = os.path.abspath(args.top_dir)
instance_names = entities_judgement.keys()
documents = {}
for instance in instance_names:
print "for %s" %instance
source_dir = os.path.join(args.top_dir,"clean_text",args.disaster_name,instance)
sub_dirs = os.walk(source_dir).next()[1]
documents[instance] = {}
for a_dir in sub_dirs:
date_dir = os.path.join(source_dir,a_dir)
print date_dir
for single_file in get_files(date_dir):
#print "open file %s" %os.path.join(date_dir,single_file)
single_file = os.path.join(date_dir,single_file)
documents[instance][single_file] = Document(single_file,file_path = single_file)
#show_documents(documents)#debug purpose
#print json.documents(files,indent=4)
if args.using_text_window:
windows = get_all_text_windows(documents,entities_judgement,args.window_size)
else:
windows = get_all_sentence_windows(documents,entities_judgement)
for entity_type in windows:
for w in windows[entity_type]:
windows[entity_type][w] = windows[entity_type][w].model
with codecs.open(os.path.join(args.dest_dir,q),"w","utf-8") as f:
f.write(json.dumps(windows))
#print json.dumps(windows,indent=4)
print "finished"
if __name__=="__main__":
main()