-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathContextEncoder.py
More file actions
303 lines (254 loc) · 11 KB
/
ContextEncoder.py
File metadata and controls
303 lines (254 loc) · 11 KB
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
301
302
303
'''
Author: your name
Date: 2020-12-02 11:10:04
LastEditTime: 2020-12-20 09:03:58
LastEditors: superlova
Description: In User Settings Edit
FilePath: \codeclf\preprocessing\ContextEncoder.py
'''
# -*- coding: utf-8 -*-
# @Author : superlova
# @Time : 2020/11/24 8:56
# @Function:
import pandas as pd
import logging
import sys, os
# project_dir = 'C:/Users/zyt/Documents/GitHub Repositories/codeclf_gui/codeclf'
# sys.path.append(project_dir)
from preprocessing.FSM import FSM
from utils.Utils import timethis
class ContextEncoder(object):
def __init__(self):
pass
def context_encode(self, text, before=1, after=1):
"""
输入（一个python文件内的所有）文本，输出每行的上下文、每行的label的字典构成的列表。其中上下文是列表形式。
保持了上下文的顺序
:param text:
:param before:
:param after:
:return:
"""
codes = []
docs = []
# 逐个扫描每个py文件，标记每行的类别
text = text.split('\n')
text = [line for line in text if len(line) > 0] # 只留下非空行
length = len(text)
fsm = FSM(text)
fsm.scan()
for index in fsm.codes:
# 每一行代码都查找它的上下文：
context_before = []
context_after = []
if index < before: # index不够大导致before不够
for i in range(index):
context_before.append(text[i])
else:
for i in range(index - before, index):
context_before.append(text[i])
if length - index - 1 < after: # index过大导致after不够
for i in range(index + 1, length):
context_after.append(text[i])
else:
for i in range(index + 1, index + after + 1):
context_after.append(text[i])
item = {'before': context_before, 'text': text[index], 'after': context_after, 'label': 0}
codes.append(item)
for index in fsm.docs:
# 每一行代码都查找它的上下文：
context_before = []
context_after = []
if index < before: # index不够大导致before不够
for i in range(index):
context_before.append(text[i])
else:
for i in range(index - before, index):
context_before.append(text[i])
if length - index - 1 < after: # index过大导致after不够
for i in range(index + 1, length):
context_after.append(text[i])
else:
for i in range(index + 1, index + after + 1):
context_after.append(text[i])
item = {'before': context_before, 'text': text[index], 'after': context_after, 'label': 1}
docs.append(item)
return codes, docs
def context_encode_merge(self, text, before=1, after=1):
"""
输入（一个python文件内的所有）文本，输出每行的上下文、每行的label的字典构成的列表。其中上下文是字符串形式。
保持了上下文的顺序
:param text:
:param before:
:param after:
:return:
"""
codes = []
docs = []
# 逐个扫描每个py文件，标记每行的类别
text = text.split('\n')
text = [line for line in text if len(line) > 0] # 只留下非空行
length = len(text)
fsm = FSM(text)
fsm.scan()
for index in fsm.codes:
# 每一行代码都查找它的上下文：
context_before = []
context_after = []
if index < before: # index不够大导致before不够
for i in range(index):
context_before.append(text[i])
else:
for i in range(index - before, index):
context_before.append(text[i])
if length - index - 1 < after: # index过大导致after不够
for i in range(index + 1, length):
context_after.append(text[i])
else:
for i in range(index + 1, index + after + 1):
context_after.append(text[i])
item = {'before': "\n".join(context_before), 'text': text[index], 'after': "\n".join(context_after), 'label': 0}
codes.append(item)
for index in fsm.docs:
# 每一行代码都查找它的上下文：
context_before = []
context_after = []
if index < before: # index不够大导致before不够
for i in range(index):
context_before.append(text[i])
else:
for i in range(index - before, index):
context_before.append(text[i])
if length - index - 1 < after: # index过大导致after不够
for i in range(index + 1, length):
context_after.append(text[i])
else:
for i in range(index + 1, index + after + 1):
context_after.append(text[i])
item = {'before': "\n".join(context_before), 'text': text[index], 'after': "\n".join(context_after), 'label': 1}
docs.append(item)
return codes, docs
def context_divide_all(self, file_texts, before=1, after=1):
codes = []
docs = []
for corpus in file_texts:
sub_codes, sub_docs = self.context_encode(corpus, before, after)
codes.extend(sub_codes)
docs.extend(sub_docs)
return codes, docs
def context_merge_all(self, file_texts, before=1, after=1):
codes = []
docs = []
for corpus in file_texts:
sub_codes, sub_docs = self.context_encode_merge(corpus, before, after)
codes.extend(sub_codes)
docs.extend(sub_docs)
return codes, docs
def context_encode_all(self, text):
"""
输入（一个python文件内的所有）文本，输出每行的上下文、每行的label的字典构成的列表。其中上下文是字符串形式。
:param text:
:param before:
:param after:
:return:
"""
codes = []
docs = []
# 逐个扫描每个py文件，标记每行的类别
text = text.split('\n')
text = [line for line in text if len(line) > 0] # 只留下非空行
fsm = FSM(text)
fsm.scan()
for index in fsm.codes:
item = {'context': "\n".join(text), 'text': text[index], 'label': 0}
codes.append(item)
for index in fsm.docs:
item = {'context': "\n".join(text), 'text': text[index], 'label': 1}
docs.append(item)
return codes, docs
def context_encode_allfile(self, file_texts):
codes = []
docs = []
for corpus in file_texts:
sub_codes, sub_docs = self.context_encode_all(corpus)
codes.extend(sub_codes)
docs.extend(sub_docs)
return codes, docs
@timethis
def process(corpus_path, output_path, before=1, after=1):
df_data = pd.read_pickle(corpus_path)
data = df_data['code']
processor = ContextEncoder()
codes, docs = processor.context_divide_all(data, before=before, after=after)
df_codes = pd.DataFrame(data=codes, columns=['before', 'text', 'after', 'label'])
df_docs = pd.DataFrame(data=docs, columns=['before', 'text', 'after', 'label'])
print(f"df_codes:{len(df_codes)}, df_docs:{len(df_docs)}")
df = pd.concat([df_codes, df_docs], ignore_index=True)
df.to_pickle(output_path, protocol=4)
###########################################################
def test_process():
process('../datasets/df_train_corpus.tar.bz2', '../datasets/df_train_context.tar.bz2')
process('../datasets/df_test_corpus.tar.bz2', '../datasets/df_test_context.tar.bz2')
process('../datasets/df_valid_corpus.tar.bz2', '../datasets/df_valid_context.tar.bz2')
process('../datasets/df_train_corpus.tar.bz2', '../datasets/df_train_context_2.tar.bz2', before=2, after=2)
process('../datasets/df_test_corpus.tar.bz2', '../datasets/df_test_context_2.tar.bz2', before=2, after=2)
process('../datasets/df_valid_corpus.tar.bz2', '../datasets/df_valid_context_2.tar.bz2', before=2, after=2)
def test_context_merge():
corpus_path = '../datasets/df_test_corpus.tar.bz2'
df_data = pd.read_pickle(corpus_path)
data = df_data['code'][90]
processor = ContextEncoder()
context_data = processor.context_encode(data, before=1, after=1)
df = pd.DataFrame(data=context_data, columns=['before', 'text', 'after', 'label'])
print(df)
def test_context_divide_all():
corpus_path = '../datasets/df_test_corpus.tar.bz2'
df_data = pd.read_pickle(corpus_path)
data = df_data['code'][:90]
processor = ContextEncoder()
# context_data = processor.context_divide_all(data, before=1, after=1)
codes, docs = processor.context_divide_all(data, before=1, after=1)
df_codes = pd.DataFrame(data=codes, columns=['before', 'text', 'after', 'label'])
df_docs = pd.DataFrame(data=docs, columns=['before', 'text', 'after', 'label'])
df = pd.concat([df_codes, df_docs], ignore_index=True)
# df = pd.DataFrame(data=context_data, columns=['before', 'text', 'after', 'label'])
print(df.head())
def test_context_merge():
corpus_path = '../datasets/df_test_corpus.tar.bz2'
df_data = pd.read_pickle(corpus_path)
data = df_data['code'][:90]
processor = ContextEncoder()
# context_data = processor.context_divide_all(data, before=1, after=1)
codes, docs = processor.context_merge_all(data, before=1, after=1)
df_codes = pd.DataFrame(data=codes, columns=['before', 'text', 'after', 'label'])
df_docs = pd.DataFrame(data=docs, columns=['before', 'text', 'after', 'label'])
df = pd.concat([df_codes, df_docs], ignore_index=True)
# df = pd.DataFrame(data=context_data, columns=['before', 'text', 'after', 'label'])
print(df_data['code'][0])
for index, row in df.iterrows():
print("before", row[0])
print("text", row[1])
print("after", row[2])
def test_context_encode_allfile():
corpus_path = '../datasets/df_test_corpus.tar.bz2'
df_data = pd.read_pickle(corpus_path)
data = df_data['code'][:90]
processor = ContextEncoder()
codes, docs = processor.context_encode_allfile(data)
print(codes[0])
print(docs[0])
def test_context_encode_allfile_filelen():
df_train = pd.read_pickle('../datasets/df_train_corpus.tar.bz2')
df_test = pd.read_pickle('../datasets/df_test_corpus.tar.bz2')
df_valid = pd.read_pickle('../datasets/df_valid_corpus.tar.bz2')
def main():
logging.basicConfig(
level=logging.INFO
)
# test_process()
# test_context_merge()
# test_context_divide_all()
# test_context_merge()
test_context_encode_allfile()
if __name__ == '__main__':
main()