forked from 13o-bbr-bbq/SAIVS
-
Notifications
You must be signed in to change notification settings - Fork 0
/
MyEnvironment.py
533 lines (470 loc) · 28 KB
/
MyEnvironment.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
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
#!/usr/bin/python
# coding:utf-8
import numpy as np
import random
import string
from requests import Request, Session
from MyDecision import Decision
from MyWord2Vec import Word2Vec
PROXY = {'http': '127.0.0.1:8083'}
# CredentialsTBLのカラム情報
str_col_credentialstbl = "site_id, " \
"type, " \
"credential_info1, " \
"credential_info2, " \
"credential_info3, " \
"credential_info4, " \
"credential_info5, " \
"secret_question, " \
"secret_answer, " \
"temp"
# WordSimilarityTBLのカラム情報
str_col_wordsimilaritytbl = "page_type, " \
"word, " \
"value"
# 環境の定義
class Environment(object):
int_max_learning_episode = 30 # 学習回数
#int_max_learning_episode = 100
int_which_episode = 0
int_learning_episode = 0
int_total_reward = 0
int_total_step = 0
# multipartリクエスト用のダミーファイル
str_dummy_file_path = ".\\dummy.png"
# パラメータ値の集合(16種類)
str_3num = str(random.randint(100, 999))
str_6num = str(random.randint(100000, 999999))
str_8num = str(random.randint(10000000, 99999999))
str_9num = str(random.randint(100000000, 999999999))
str_12num = str(random.randint(100000000000, 999999999999))
str_16num = str(random.randint(1000000000000000, 9999999999999999))
#lst_param_value_collections = [str_3num, str_6num, str_9num, str_12num, str_16num,
# "abc", "abcdef", "abcdefghi",
# str_3num + "abc", str_6num + "abcdef",
# str_3num + "@@@", str_6num + "@@@@@@",
# "abc@@@", str_16num + "@hoge.com",
# str_3num + "ab@@", str_6num + "abcd@@@@"]
# パラメータ値の集合(4種類)
lst_param_value_collections = [str_8num, str_16num, str_3num + "abc", str_16num + "@hoge.com"]
def __init__(self, str_train_action='TRAIN', obj_browser=None):
self.obj_browser = obj_browser
self.bol_use_dqn_flag = False
self.lst_create_param = []
self.lst_default_param = []
self.str_train = str_train_action
# 対象URLの設定
def set_url(self, str_url):
self.str_target_url = str_url
# パラメータ値の集合の更新
def update_random_params(self):
self.str_3num = str(random.randint(100, 999))
self.str_6num = str(random.randint(100000, 999999))
self.str_8num = str(random.randint(10000000, 99999999))
self.str_9num = str(random.randint(100000000, 999999999))
self.str_12num = str(random.randint(100000000000, 999999999999))
self.str_16num = str(random.randint(1000000000000000, 9999999999999999))
#self.lst_param_value_collections = [self.str_3num, self.str_6num, self.str_9num, self.str_12num, self.str_16num,
# "abc", "abcdef", "abcdefghi",
# self.str_3num + "abc", self.str_6num + "abcdef",
# self.str_3num + "@@@", self.str_6num + "@@@@@@",
# "abc@@@", self.str_16num + "@hoge.com",
# self.str_3num + "ab@@", self.str_6num + "abcd@@@@"]
# パラメータ値の集合(4種類)
self.lst_param_value_collections = [self.str_8num, self.str_16num,
self.str_3num + "abc", self.str_16num + "@hoge.com"]
# 最適パラメータの取得
def index_to_action(self, int_index_action, dic_post_param):
lst_param_name = dic_post_param.keys()
for str_param_name in lst_param_name:
# 値が未設定の場合
if dic_post_param[str_param_name] == '':
# ε-greedyで選択した値をパラメータ値に設定
dic_post_param[str_param_name] = self.lst_param_value_collections[int_index_action]
return dic_post_param
# 初回アクセス時のパラメータ構成を定義
def create_init_param(self, dic_post_params=''):
# 行動パターンの算出：パラメータと全候補パラメータ値の組み合わせ
# 但し、元値を持っているパラメータには候補パラメータ値を設定しない
lst_param_name = dic_post_params.keys()
lst_param_collections = []
lst_value_collections2 = []
lst_param_collections_no_value = []
int_idx = 0
int_repeat = 0
# パラメータ毎の取り得る値を設定
for str_param_name in lst_param_name:
# 値が未設定のパラメータを抽出
if dic_post_params[str_param_name] == '':
lst_param_collections_no_value.append(str_param_name)
int_repeat += 1
# 値が設定されているパラメータを抽出
else:
lst_param_collections.append(tuple([str_param_name, dic_post_params[str_param_name]]))
int_idx += 1
# 値が未設定のパラメータが存在する場合
if int_repeat != 0:
lst_param_temp = []
for str_value_collection in self.lst_param_value_collections:
for int_idx in range(int_repeat):
lst_param_temp.append(tuple([lst_param_collections_no_value[int_idx], str_value_collection]))
lst_value_collections2.append(lst_param_temp)
lst_param_temp = []
# 全パラメータ組み合わせを行動パターンとする
for int_idx in range(len(lst_value_collections2)):
lst_value_collections2[int_idx] += lst_param_collections
self.bol_use_dqn_flag = True
return lst_value_collections2, self.bol_use_dqn_flag
# 既存のパラメータに全て値が入力されている、または、POSTパラメータが無い場合
else:
# print "Not Use DQN."
self.lst_default_param = lst_param_collections
self.bol_use_dqn_flag = False
return lst_param_collections, self.bol_use_dqn_flag
# HTTPレスポンス内容から状態を判定
def judge_state(self, obj_response, int_group_id, int_parent_seq):
obj_decision = Decision()
# 正常遷移の可否を判定
#int_result, int_score = obj_decision.decide_flow_okng(obj_response, int_group_id, int_parent_seq)
int_result = obj_decision.decide_flow_okng(obj_response, int_group_id, int_parent_seq)
int_next_state = 0
if int_result == 1:
int_reward = 1000 # 正常遷移した場合は報酬「100」を与える
int_next_state = 10 # 正常遷移した場合は状態を成功「1」にする
elif int_result == 0:
int_reward = -1000 # 正常遷移しない場合は報酬「-100」を与える
int_next_state = 1 # 正常遷移しない場合は状態を変化なし「0」にする
else:
int_reward = -1000 # エラーが発生した場合は報酬「-1000」を与える
int_next_state = 0 # エラーが発生した場合は状態をエラー「-1」にする
return int_reward, int_next_state
# 当該ページへの遷移方法を学習
def flow_learning(self,
obj_db_control,
obj_session,
obj_agent,
obj_web_inspect,
lst_row,
lst_flow,
dic_post_params):
obj_decision = Decision()
lst_temp_action = []
int_reward = 0
# 学習済みデータの有無を確認
# エージェントの学習
int_learning_episode = 0
int_total_cost = 0.0 # 10/1 追加
int_total_reward = 0.0 # 10/1 追加
int_frame = 0 # 10/1 追加
while int_learning_episode < self.int_max_learning_episode:
int_frame += 1
int_reward = 0
obj_request = None
# POSTリクエストの送信
if lst_row[19] == "multipart/form-data":
# マルチパートリクエスト
dic_post_files = {}
dic_post_data = {}
# ファイルを保持するパラメータと通常のパラメータに切り分ける
dic_post_files, dic_post_data = obj_decision.divide_params(lst_row[21],
dic_post_params,
self.str_dummy_file_path)
obj_request = Request("POST",
obj_decision.assemble_url(lst_row),
files=dic_post_files,
data=dic_post_data
)
else:
# 通常のリクエスト
if lst_row[8].upper() == 'POST':
obj_request = Request("POST",
obj_decision.assemble_url(lst_row),
data=dic_post_params
)
else:
obj_request = Request("GET",
obj_decision.assemble_url(lst_row),
params=dic_post_params
)
obj_prepped = obj_session.prepare_request(obj_request)
obj_response = obj_session.send(obj_prepped,
verify=True,
timeout=60,
proxies=PROXY,
allow_redirects=False
)
# レスポンスをブラウザに表示（デモ用）
if obj_response is not None:
self.obj_browser.write_response_to_html(obj_response.text, obj_response.encoding)
self.obj_browser.refresh_browser()
# レスポンスがリダイレクトの場合
# リダイレクトが終了するまで繰り返しリダイレクト
while obj_response.is_redirect is True:
# Locationヘッダから遷移先URLを取得
dic_res_headers = obj_response.headers._store
tpl_location_header = dic_res_headers["location"]
str_redirect_url = tpl_location_header[1]
if 'http://' not in str_redirect_url and 'https://' not in str_redirect_url:
str_fqdn = lst_row[9].encode() + "://" + lst_row[10].encode() + ":" + str(lst_row[11])
if str_redirect_url.startswith('/') is False:
str_redirect_url = '/' + str_redirect_url
str_redirect_url = str_fqdn + str_redirect_url
# リダイレクト
obj_request = Request("GET", str_redirect_url)
obj_prepped = obj_session.prepare_request(obj_request)
obj_response = obj_session.send(obj_prepped,
verify=True,
timeout=60,
proxies=PROXY,
allow_redirects=False
)
# レスポンスをブラウザに表示（デモ用）
if obj_response is not None:
self.obj_browser.write_response_to_html(obj_response.text, obj_response.encoding)
self.obj_browser.refresh_browser()
# レスポンス内容から報酬を決定
int_reward, int_next_state = self.judge_state(obj_response, lst_row[2], lst_row[6])
# 対象が会員情報変更の場合、かつ、遷移に成功した場合、認証情報を入れ替える
if lst_row[3] == 5 and int_reward > 0:
# 認証情報の取得
lst_credentials, str_temp_label_value = obj_decision.decide_get_credentials(
obj_decision.change_dictionary_to_list(dic_post_params),
lst_row[22])
# 認証情報が取得できた場合、認証情報を入れ替える
if lst_credentials[0] is not None and lst_credentials[1] is not None:
str_sql = "DELETE FROM CredentialsTBL WHERE site_id = 1;"
obj_db_control.delete(str_sql)
str_sql = "INSERT INTO CredentialsTBL(" + str_col_credentialstbl + ") " \
"VALUES (1, 1, ?, ?, ?, ?, ?, '', '', '')"
lst_value = [lst_credentials[0],
lst_credentials[1],
lst_credentials[2],
lst_credentials[3],
lst_credentials[4]]
obj_db_control.insert(str_sql, lst_value)
# 認証情報が取得できない場合、一旦temp領域にパラメータ情報を入れておく
else:
str_sql = "DELETE FROM CredentialsTBL WHERE site_id = 1;"
obj_db_control.delete(str_sql)
str_sql = "INSERT INTO CredentialsTBL(" + str_col_credentialstbl + ") " \
"VALUES (1, 1, '', '', '', '', '', '', '', ?)"
lst_value = [str_temp_label_value]
obj_db_control.insert(str_sql, lst_value)
# 次の状態を設定
lst_state_dash = [lst_row[6], lst_row[2], int_next_state]
int_total_cost += obj_agent.observe(int_reward)
int_total_reward += int_reward
obj_agent.new_episode()
nd_state = np.array(lst_state_dash, dtype=np.uint8)
int_index_action, nd_values = obj_agent.act(nd_state)
print('frame:%d / total reward:%d / total cost:%f / action:%d / reward:%d' %
(int_frame, int_total_reward, int_total_cost, int_index_action, int_reward))
# 次の学習の準備：ルートから当該ページの一つ手前まで遷移
obj_session = None
obj_session = Session() # 新しいセッションで遷移を再生
if obj_web_inspect.flow_replay(obj_session, obj_db_control, obj_agent, lst_row, lst_flow) is False:
continue
# 次の学習の準備：パラメータ値の更新
# POSTパラメータが存在する場合
if lst_row[13] != '':
# 最新のパラメータ構成を取得
self.update_random_params()
dic_post_params, bol_skip_flag = obj_decision.update_post_parameter(obj_db_control, lst_row)
dic_post_params = self.index_to_action(int_index_action, dic_post_params)
int_learning_episode += 1
# 学習済みの単語・値のセットを単語類似度管理テーブルに格納
dic_post_params, bol_skip_flag = obj_decision.update_post_parameter(obj_db_control, lst_row)
lst_state = [lst_row[6], lst_row[2], 0]
int_action_index, nd_values = obj_agent.act(np.array(lst_state, dtype=np.uint8))
lst_params = np.argsort(nd_values)[0].tolist()
int_action_index = len(lst_params) - 1
self.update_random_params()
dic_post_params = self.index_to_action(int_action_index, dic_post_params)
int_idx = 0
lst_param_names = lst_row[13].split('&')
lst_param_types = lst_row[21].split(',')
lst_label_names = lst_row[22].split(',')
while int_idx < int(lst_row[14]):
if lst_label_names[int_idx] != '@':
lst_param_names_tmp = lst_param_names[int_idx].split('=')
str_param_value = dic_post_params[lst_param_names_tmp[0]]
str_sql = "INSERT INTO WordSimilarityTBL(" + str_col_wordsimilaritytbl + ") " \
"VALUES (?, ?, ?)"
lst_value = [int(lst_row[3]), lst_label_names[int_idx], str_param_value]
obj_db_control.insert(str_sql, lst_value)
int_idx += 1
# 対象URLにPOSTリクエストを送信し、正常遷移の可否を返却。
def send_message(self, obj_db_control, obj_session, obj_agent, obj_web_inspect, lst_state, lst_row, lst_flow):
obj_decision = Decision()
# DQNを使用する場合、かつ、ログイン処理ではない場合
if self.bol_use_dqn_flag is True:
# 最新のパラメータ構成を取得
dic_post_params, bol_skip_flag = obj_decision.update_post_parameter(obj_db_control, lst_row)
int_idx = 0
bol_relearn_flag = False
lst_param_names = lst_row[13].split('&')
lst_param_types = lst_row[21].split(',')
lst_label_names = lst_row[22].split(',')
while int_idx < int(lst_row[14]):
if lst_label_names[int_idx] != '@':
str_sql = "SELECT value from WordSimilarityTBL where word like '%"\
+ lst_label_names[int_idx] + "%';"
obj_cursor = obj_db_control.select(str_sql)
lst_all_row = obj_cursor.fetchall()
lst_param_names_tmp = lst_param_names[int_idx].split('=')
# 類似単語が類似単語管理テーブルに存在する
if len(lst_all_row) != 0:
lst_candidate = list(lst_all_row[0])
str_candidate_value = str(lst_candidate[0])
int_find_idx = str_candidate_value.find('@')
if int_find_idx > 0:
str_random_value = ''.join([random.choice(string.digits) for i in range(int_find_idx)])
str_candidate_value = str_random_value + str_candidate_value[int_find_idx:]
dic_post_params[lst_param_names_tmp[0]] = str_candidate_value
# 類似単語が類似単語管理テーブルに存在しない
else:
obj_word2vec = Word2Vec()
obj_result = obj_word2vec.get_candidate_word(lst_label_names[int_idx])
if obj_result is not False:
for r in obj_result:
str_sql = "SELECT value from WordSimilarityTBL where word like '%" \
+ r[0] + "%';"
obj_cursor = obj_db_control.select(str_sql)
lst_all_row = obj_cursor.fetchall()
if len(lst_all_row) != 0:
lst_candidate = list(lst_all_row[0])
str_candidate_value = str(lst_candidate[0])
int_find_idx = str_candidate_value.find('@')
if int_find_idx > 0:
str_random_value = \
''.join([random.choice(string.digits) for i in range(int_find_idx)])
str_candidate_value = str_random_value + str_candidate_value[int_find_idx:]
dic_post_params[lst_param_names_tmp[0]] = str_candidate_value
break
if dic_post_params[lst_param_names_tmp[0]] == '':
bol_relearn_flag = True
else:
bol_relearn_flag = True
int_idx += 1
if bol_relearn_flag is True and self.str_train == 'TRAIN':
# エージェントの学習
obj_response = self.flow_learning(obj_db_control,
obj_session,
obj_agent,
obj_web_inspect,
lst_row,
lst_flow,
dic_post_params
)
# 学習結果に基づき遷移に最適なパラメータ値を設定
int_action_index, nd_values = obj_agent.act(np.array(lst_state, dtype=np.uint8))
lst_params = np.argsort(nd_values)[0].tolist()
int_action_index = len(lst_params) - 1
self.update_random_params()
dic_post_params = self.index_to_action(int_action_index, dic_post_params)
# POSTリクエストの送信
if lst_row[19] == "multipart/form-data":
# マルチパートリクエスト
dic_post_files = {}
dic_post_data = {}
# ファイルを保持するパラメータと通常のパラメータに切り分ける
dic_post_files, dic_post_data = obj_decision.divide_params(lst_row[21],
dic_post_params,
self.str_dummy_file_path)
obj_request = Request("POST",
obj_decision.assemble_url(lst_row),
files=dic_post_files,
data=dic_post_data
)
else:
# 通常のリクエスト
if (lst_row[8].upper() == 'POST'):
obj_request = Request(lst_row[8],
obj_decision.assemble_url(lst_row),
data=dic_post_params
)
else:
obj_request = Request(lst_row[8],
obj_decision.assemble_url(lst_row),
params=dic_post_params
)
obj_prepped = obj_session.prepare_request(obj_request)
obj_response = obj_session.send(obj_prepped,
verify=True,
timeout=60,
proxies=PROXY,
allow_redirects=False
)
# レスポンスをブラウザに表示（デモ用）
if obj_response is not None:
self.obj_browser.write_response_to_html(obj_response.text, obj_response.encoding)
self.obj_browser.refresh_browser()
# レスポンスがリダイレクトの場合
# リダイレクトが終了するまで繰り返しリダイレクト
while obj_response.is_redirect is True:
# Locationヘッダから遷移先URLを取得
dic_res_headers = obj_response.headers._store
tpl_location_header = dic_res_headers["location"]
str_redirect_url = tpl_location_header[1]
if 'http://' not in str_redirect_url and 'https://' not in str_redirect_url:
str_fqdn = lst_row[9].encode() + "://" + lst_row[10].encode() + ":" + str(lst_row[11])
if str_redirect_url.startswith('/') is False:
str_redirect_url = '/' + str_redirect_url
str_redirect_url = str_fqdn + str_redirect_url
# リダイレクト
obj_request = Request("GET", str_redirect_url)
obj_prepped = obj_session.prepare_request(obj_request)
obj_response = obj_session.send(obj_prepped,
verify=True,
timeout=60,
proxies=PROXY,
allow_redirects=False
)
# レスポンスをブラウザに表示（デモ用）
if obj_response is not None:
self.obj_browser.write_response_to_html(obj_response.text, obj_response.encoding)
self.obj_browser.refresh_browser()
return obj_response, dic_post_params
else:
# GET/POSTリクエストの送信
obj_request = Request(lst_row[8],
obj_decision.assemble_url(lst_row),
data=dict(self.lst_default_param)
)
obj_prepped = obj_session.prepare_request(obj_request)
obj_response = obj_session.send(obj_prepped,
verify=True,
timeout=60,
proxies=PROXY,
allow_redirects=False
)
# レスポンスをブラウザに表示（デモ用）
if obj_response is not None:
self.obj_browser.write_response_to_html(obj_response.text, obj_response.encoding)
self.obj_browser.refresh_browser()
# レスポンスがリダイレクトの場合
# リダイレクトが終了するまで繰り返しリダイレクト
while obj_response.is_redirect is True:
# Locationヘッダから遷移先URLを取得
dic_res_headers = obj_response.headers._store
tpl_location_header = dic_res_headers["location"]
str_redirect_url = tpl_location_header[1]
if 'http://' not in str_redirect_url and 'https://' not in str_redirect_url:
str_fqdn = lst_row[9].encode() + "://" + lst_row[10].encode() + ":" + str(lst_row[11])
if str_redirect_url.startswith('/') is False:
str_redirect_url = '/' + str_redirect_url
str_redirect_url = str_fqdn + str_redirect_url
# リダイレクト
obj_request = Request("GET", str_redirect_url)
obj_prepped = obj_session.prepare_request(obj_request)
obj_response = obj_session.send(obj_prepped,
verify=True,
timeout=60,
proxies=PROXY,
allow_redirects=False
)
# レスポンスをブラウザに表示（デモ用）
if obj_response is not None:
self.obj_browser.write_response_to_html(obj_response.text, obj_response.encoding)
self.obj_browser.refresh_browser()
return obj_response, dict(self.lst_default_param)