-
Notifications
You must be signed in to change notification settings - Fork 4
/
seva_dataset_utils.py
300 lines (272 loc) · 9.16 KB
/
seva_dataset_utils.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
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
import nltk, re
from nltk.corpus import stopwords
import enchant
from nltk import ngrams
d = enchant.Dict("en_US")
# create a dictionary
# input: accronym file
# output: dictionary
# key: acronym
# value: full form
def create_acronym_dict(fname):
d = {}
f = open(fname)
lines = f.readlines()
f.close()
for line in lines:
splits = line.strip().split()
if len(splits) < 2:
continue
d[splits[0].strip()] = " ".join(splits[1:]).strip()
return d
# create a dictionary
# input: accronym file
# output: dictionary
# key: full form
# value: accronym
def create_acronym_dict_inverse(fname):
d = {}
f = open(fname)
lines = f.readlines()
f.close()
for line in lines:
splits = line.strip().split()
if len(splits) < 2:
continue
d[" ".join(splits[1:]).strip().lower()] = splits[0].strip()
return d
# create a dictionary
# input: definition file
# output: dictionary
# key: definition
# value: description
def create_definition_dict(fname):
d = {}
f = open(fname)
lines = f.readlines()
f.close()
for line in lines:
splits = line.strip().split(":")
d[splits[0]] = " ".join(splits[1:])
return d
# dealing with definitions with '/' and accronyms
# Automated/Automation
# Configuration Items (CI)
# input: phrase/definition
# output: list of definitions extracted from the input
def get_def_combos(term, accro_dict):
in_brackets = re.findall(r"\([^\(\)]*\)",term)
term = re.sub(r"\([^\(\)]*\)", " ", term)
term = re.sub('\s+', ' ', term)
in_brackets = [b[1:len(b)-1] for b in in_brackets]
in_brackets = [b.strip() for b in in_brackets if accro_dict.get(b, None) is not None]
terms_plus = term.split("/")
terms_plus = [t.lower().strip() for t in terms_plus]
#in_brackets = [b.lower().strip() for b in in_brackets if accro_dict.get(b).lower() in terms_plus]
combo = terms_plus + in_brackets
return combo
# Fixing the first sentence of the explanation to create a coherent sentence
# input: accronym file, definition file
# output: dictionary
# key: definition
# value: full form
def fix_def(accr_fname, fname):
accro_dict = create_acronym_dict(accr_fname)
f = open(fname)
lines = f.readlines()
f.close()
definition_dict = {}
for line in lines:
sent = nltk.sent_tokenize(line.strip())
if len(sent) < 1:
continue
splits = sent[0].strip().split(":")
definition = splits[0]
explanation = splits[1:]
explanation = explanation[0]
combos = get_def_combos(definition, accro_dict)
new_def = " | ".join(combos)
Flag = False
for c in combos:
if c.lower() in explanation.lower():
Flag = True
if Flag:
if len(sent) > 1:
explanation += " ".join(sent[1:])
definition_dict[new_def.strip()] = explanation.strip()
else:
if explanation.split()[0].lower() in stopwords.words('english'):
explanation = explanation.split()
explanation[0] = explanation[0].lower()
else:
explanation = explanation.split()
explanation[0] = explanation[0].lower()
exp = definition + " is " + " ".join(explanation).strip()
if len(sent) > 1:
exp += " ".join(sent[1:])
definition_dict[new_def.strip()] = exp
return definition_dict
# put definitions in to buckets corresponding to number of words in them
# input: accronym filename, definition filename
# output: dictionary
# key: integer 'n' - number of words
# value: definitions that has 'n' number of words
def categorize_def_into_numwords(accr_location, definition_location):
accro_dict = create_acronym_dict(accr_location)
d = fix_def(accr_location, definition_location)
res = {}
for key in d.keys():
splits = key.split("|")
for s in splits:
s = s.strip()
if s == "":
continue
if not s.isupper() and s not in accro_dict.keys():
if res.get(len(s.split()), None) is None:
res[len(s.split())] = set()
res[len(s.split())].add(s)
return res
# write all keywords from buckets to a file
# input: accronym filename, definition filename, output filename
# output: a new file
def create_file_from_buckets_4_human_labelling(accr_location, definition_location, fname):
res = categorize_def_into_numwords(accr_location, definition_location)
f = open(fname, 'w+')
keys = sorted(res.keys())
for key in keys:
f.write(str(key)+'\n')
for item in res[key]:
f.write(item+", \n")
f.close()
# read annotated keywords as dictionary
# input: filename
# ouput: dictionary
# key: tag
# value: terms
def read_annotated_keywords(fname):
tag_dict = {}
f = open(fname)
lines = f.readlines()
f.close()
for line in lines:
if len(line.split()) > 1:
tag = line.split(",")[-1].strip()
term = " ".join(line.split(",")[:-1]).strip()
if tag_dict.get(tag, None) is None:
tag_dict[tag] = set()
tag_dict[tag].add(term.strip())
'''
accro_dict = create_acronym_dict_inverse(accr_location)
for key,val in tag_dict.items():
new_item_set = set()
for item in val:
splits = item.split()
for index in range(len(splits)):
if accro_dict.get(splits[index].strip().lower(), None) is not None:
splits[index] = accro_dict.get(splits[index].strip().lower())
new_item = " ".join(splits)
new_item_set.add(new_item)
tag_dict[key] = tag_dict[key].union(new_item_set)
'''
return tag_dict
# read annotated keywords as dictionary
# input: filename
# ouput: dictionary
# key: term
# value: tag
def read_annotated_keywords_inverse(fname):
tag_dict = {}
f = open(fname)
lines = f.readlines()
f.close()
for line in lines:
if len(line.split()) > 1:
tag = line.split(",")[-1].strip()
term = " ".join(line.split(",")[:-1]).strip()
tag_dict[term.strip().lower()] = tag
return tag_dict
def is_ascii(s):
return all(ord(c) < 128 for c in s)
def update_vocab(location_of_vocab, *input_fname):
text = ""
for fname in input_fname:
f_read = open(fname)
text += " " + f_read.read()
f_read.close()
new_words = nltk.word_tokenize(text)
new_words = [w.strip().lower() for w in new_words]
new_words = set(new_words)
n = len(new_words)
f = open(location_of_vocab)
text = f.readlines()
f.close()
old_words = [w.strip() for w in text]
o_len = len(old_words)
old_words = set(old_words)
words = old_words.union(new_words)
num_words2remove = len(words) - o_len
while num_words2remove > 0 :
num_words = len(words)
count = 1
for item in words:
count += 1
if not is_ascii(item):
words.remove(item)
num_words2remove -= 1
break
if count >= num_words:
break
unused = 993
while num_words2remove > 0 and unused >=0 :
for item in words:
if 'unused' in item:
words.remove(item)
num_words2remove -= 1
unused -= 1
break
assert(o_len == len(words))
f = open(location_of_vocab, 'w+')
for word in words:
f.write(word)
f.write('\n')
f.close()
def sentence2tag_keyword(sentence, keyword_fname):
phrase_labels = []
d = read_annotated_keywords_inverse(keyword_fname)
for i in reversed(range(1, 6)):
words = nltk.word_tokenize(sentence)
igrams = ngrams(words, i)
for g in igrams:
phrase = " ".join(g).strip()
if d.get(phrase.lower(), None) != None:
phrase_labels.append((phrase,d[phrase.lower()]))
sentence = sentence.replace(phrase, "").strip()
return phrase_labels
def keywords2sentposlabel(keyword_fname):
d = read_annotated_keywords_inverse(keyword_fname)
sentences = []
poses = []
labels = []
for key,val in d.items():
words = nltk.word_tokenize(key)
#words.append(".")
pos = nltk.pos_tag(words)
pos = [p[1] for p in pos]
lab = []
for index in range(len(words)):
if index == 0:
lab.append('B-'+val)
else:
lab.append('I-'+val)
sentences.append(words)
poses.append(pos)
labels.append(lab)
return sentences, poses, labels
######## USAGE EXAMPLE ########
# accr_location = "se_data/acronyms.txt"
# definition_location = "se_data/definitions.txt"
# d = fix_def(accr_location, definition_location)
# res = categorize_def_into_numwords(accr_location, definition_location)
# create_file_from_buckets_4_human_labelling(accr_location, definition_location, "se_data/keywords2annotate.txt")
# d = read_annotated_keywords( accr_location, "se_data/keywords2annotate.txt") #text file should be annotated
# d = read_annotated_keywords_inverse("se_data/keywords2annotate.txt")