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
|
#! /usr/bin/env python3
#
# SPDX-FileCopyrightText: (c) 2020-2021 CokeMine & Its repository contributors
# SPDX-FileCopyrightText: (c) 2021 A beam of light
#
# SPDX-License-Identifier: GPL-3.0-or-later
#
"""
euserv auto-renew script
v2021.09.30
* Captcha automatic recognition using TrueCaptcha API
* Email notification
* Add login failure retry mechanism
* reformat log info
v2021.11.06
* Receive renew PIN(6-digits) using mailparser parsed data download url
workflow: auto-forward your EUserv PIN email to your mailparser inbox
-> parsing PIN via mailparser
-> get PIN from mailparser
* Update kc2_security_password_get_token request
"""
import os
import re
import json
import time
import base64
from email.mime.application import MIMEApplication
from email.mime.multipart import MIMEMultipart
from email.mime.text import MIMEText
from smtplib import SMTP_SSL, SMTPDataError
import requests
from bs4 import BeautifulSoup
# 多个账户请使用空格隔开
USERNAME = "" # 用户名或邮箱
PASSWORD = "" # 密码
# default value is TrueCaptcha demo credential,
# you can use your own credential via set environment variables:
# TRUECAPTCHA_USERID and TRUECAPTCHA_APIKEY
# demo: https://apitruecaptcha.org/demo
# demo2: https://apitruecaptcha.org/demo2
# demo apikey also has a limit of 100 times per day
# {
# 'error': '101.0 above free usage limit 100 per day and no balance',
# 'requestId': '7690c065-70e0-4757-839b-5fd8381e65c7'
# }
TRUECAPTCHA_USERID = os.environ.get("TRUECAPTCHA_USERID", "arun56")
TRUECAPTCHA_APIKEY = os.environ.get("TRUECAPTCHA_APIKEY", "wMjXmBIcHcdYqO2RrsVN")
# Extract key data from your emails, automatically. https://mailparser.io
# 30 Emails/Month, 10 inboxes and unlimited downloads for free.
# 多个mailparser下载链接id请使用空格隔开, 顺序与 EUserv 账号/邮箱一一对应
MAILPARSER_DOWNLOAD_URL_ID = "这里填写我下方加粗链接中的仓库里获取到的mailparser的json文件链接ID"
# mailparser.io parsed data download base url
MAILPARSER_DOWNLOAD_BASE_URL = "https://files.mailparser.io/d/"
# Telegram Bot Push https://core.telegram.org/bots/api#authorizing-your-bot
TG_BOT_TOKEN = "TG的机器人TOKEN" # 通过 @BotFather 申请获得,示例:1077xxx4424:AAFjv0FcqxxxxxxgEMGfi22B4yh15R5uw
TG_USER_ID = "你的TGID" # 用户、群组或频道 ID,示例:129xxx206
TG_API_HOST = "https://api.telegram.org" # 自建 API 反代地址,供网络环境无法访问时使用,网络正常则保持默认
# Email notification
RECEIVER_EMAIL = os.environ.get("RECEIVER_EMAIL", "")
YD_EMAIL = os.environ.get("YD_EMAIL", "")
YD_APP_PWD = os.environ.get("YD_APP_PWD", "") # yandex mail 使用第三方 APP 授权码
# Magic internet access
PROXIES = {"http": "http://127.0.0.1:10808", "https": "http://127.0.0.1:10808"}
# Maximum number of login retry
LOGIN_MAX_RETRY_COUNT = 5
# Waiting time of receiving PIN, units are seconds.
WAITING_TIME_OF_PIN = 15
# options: True or False
CHECK_CAPTCHA_SOLVER_USAGE = True
user_agent = (
"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) "
"Chrome/97.0.4692.71 Safari/537.36"
)
desp = "" # 空值
def log(info: str):
print(info)
global desp
desp = desp + info + "\n\n"
def login_retry(*args, **kwargs):
def wrapper(func):
def inner(username, password):
ret, ret_session = func(username, password)
max_retry = kwargs.get("max_retry")
# default retry 3 times
if not max_retry:
max_retry = 3
number = 0
if ret == "-1":
while number < max_retry:
number += 1
if number > 1:
log("[EUserv] Login tried the {}th time".format(number))
sess_id, session = func(username, password)
if sess_id != "-1":
return sess_id, session
else:
if number == max_retry:
return sess_id, session
else:
return ret, ret_session
return inner
return wrapper
def captcha_solver(captcha_image_url: str, session: requests.session) -> dict:
"""
TrueCaptcha API doc: https://apitruecaptcha.org/api
Free to use 100 requests per day.
"""
response = session.get(captcha_image_url)
encoded_string = base64.b64encode(response.content)
url = "https://api.apitruecaptcha.org/one/gettext"
data = {
"userid": TRUECAPTCHA_USERID,
"apikey": TRUECAPTCHA_APIKEY,
# case sensitivity of text (upper | lower| mixed)
"case": "lower",
# use human or AI (human | default)
"mode": "default",
"data": str(encoded_string)[2:-1],
}
r = requests.post(url=url, json=data)
j = json.loads(r.text)
return j
def handle_captcha_solved_result(solved: dict) -> str:
"""Since CAPTCHA sometimes appears as a very simple binary arithmetic expression.
But since recognition sometimes doesn't show the result of the calculation directly,
that's what this function is for.
"""
if "result" in solved:
solved_text = solved["result"]
if "RESULT IS" in solved_text:
log("[Captcha Solver] You are using the demo apikey.")
print("There is no guarantee that demo apikey will work in the future!")
# because using demo apikey
text = re.findall(r"RESULT IS . (.*) .", solved_text)[0]
else:
# using your own apikey
log("[Captcha Solver] You are using your own apikey.")
text = solved_text
operators = ["X", "x", "+", "-"]
if any(x in text for x in operators):
for operator in operators:
operator_pos = text.find(operator)
if operator == "x" or operator == "X":
operator = "*"
if operator_pos != -1:
left_part = text[:operator_pos]
right_part = text[operator_pos + 1 :]
if left_part.isdigit() and right_part.isdigit():
return eval(
"{left} {operator} {right}".format(
left=left_part, operator=operator, right=right_part
)
)
else:
# Because these symbols("X", "x", "+", "-") do not appear at the same time,
# it just contains an arithmetic symbol.
return text
else:
return text
else:
print(solved)
raise KeyError("Failed to find parsed results.")
def get_captcha_solver_usage() -> dict:
url = "https://api.apitruecaptcha.org/one/getusage"
params = {
"username": TRUECAPTCHA_USERID,
"apikey": TRUECAPTCHA_APIKEY,
}
r = requests.get(url=url, params=params)
j = json.loads(r.text)
return j
def get_pin_from_mailparser(url_id: str) -> str:
"""
response format:
[
{
"id": "83b95f50f6202fb03950afbe00975eab",
"received_at": "2021-11-06 02:30:07", ==> up to mailparser account timezone setting, here is UTC 0000.
"processed_at": "2021-11-06 02:30:07",
"pin": "123456"
}
]
"""
response = requests.get(
f"{MAILPARSER_DOWNLOAD_BASE_URL}{url_id}",
# Mailparser parsed data download using Basic Authentication.
# auth=("<your mailparser username>", "<your mailparser password>")
)
pin = response.json()[0]["pin"]
return pin
@login_retry(max_retry=LOGIN_MAX_RETRY_COUNT)
def login(username: str, password: str) -> (str, requests.session):
headers = {"user-agent": user_agent, "origin": "https://www.euserv.com"}
url = "https://support.euserv.com/index.iphp"
captcha_image_url = "https://support.euserv.com/securimage_show.php"
session = requests.Session()
sess = session.get(url, headers=headers)
sess_id = re.findall("PHPSESSID=(\\w{10,100});", str(sess.headers))[0]
# visit png
logo_png_url = "https://support.euserv.com/pic/logo_small.png"
session.get(logo_png_url, headers=headers)
login_data = {
"email": username,
"password": password,
"form_selected_language": "en",
"Submit": "Login",
"subaction": "login",
"sess_id": sess_id,
}
f = session.post(url, headers=headers, data=login_data)
f.raise_for_status()
if (
f.text.find("Hello") == -1
and f.text.find("Confirm or change your customer data here") == -1
):
if (
f.text.find(
"To finish the login process please solve the following captcha."
)
== -1
):
return "-1", session
else:
log("[Captcha Solver] 进行验证码识别...")
solved_result = captcha_solver(captcha_image_url, session)
captcha_code = handle_captcha_solved_result(solved_result)
log("[Captcha Solver] 识别的验证码是: {}".format(captcha_code))
if CHECK_CAPTCHA_SOLVER_USAGE:
usage = get_captcha_solver_usage()
log(
"[Captcha Solver] current date {0} api usage count: {1}".format(
usage[0]["date"], usage[0]["count"]
)
)
f2 = session.post(
url,
headers=headers,
data={
"subaction": "login",
"sess_id": sess_id,
"captcha_code": captcha_code,
},
)
if (
f2.text.find(
"To finish the login process please solve the following captcha."
)
== -1
):
log("[Captcha Solver] 验证通过")
return sess_id, session
else:
log("[Captcha Solver] 验证失败")
return "-1", session
else:
return sess_id, session
def get_servers(sess_id: str, session: requests.session) -> {}:
d = {}
url = "https://support.euserv.com/index.iphp?sess_id=" + sess_id
headers = {"user-agent": user_agent, "origin": "https://www.euserv.com"}
f = session.get(url=url, headers=headers)
f.raise_for_status()
soup = BeautifulSoup(f.text, "html.parser")
for tr in soup.select(
"#kc2_order_customer_orders_tab_content_1 .kc2_order_table.kc2_content_table tr"
):
server_id = tr.select(".td-z1-sp1-kc")
if not len(server_id) == 1:
continue
flag = (
True
if tr.select(".td-z1-sp2-kc .kc2_order_action_container")[0]
.get_text()
.find("Contract extension possible from")
== -1
else False
)
d[server_id[0].get_text()] = flag
return d
def renew(
sess_id: str, session: requests.session, password: str, order_id: str, mailparser_dl_url_id: str
) -> bool:
url = "https://support.euserv.com/index.iphp"
headers = {
"user-agent": user_agent,
"Host": "support.euserv.com",
"origin": "https://support.euserv.com",
"Referer": "https://support.euserv.com/index.iphp",
}
data = {
"Submit": "Extend contract",
"sess_id": sess_id,
"ord_no": order_id,
"subaction": "choose_order",
"choose_order_subaction": "show_contract_details",
}
session.post(url, headers=headers, data=data)
# pop up 'Security Check' window, it will trigger 'send PIN' automatically.
session.post(
url,
headers=headers,
data={
"sess_id": sess_id,
"subaction": "show_kc2_security_password_dialog",
"prefix": "kc2_customer_contract_details_extend_contract_",
"type": "1",
},
)
# # trigger 'Send new PIN to your Email-Address' (optional),
# new_pin = session.post(url, headers=headers, data={
# "sess_id": sess_id,
# "subaction": "kc2_security_password_send_pin",
# "ident": f"kc2_customer_contract_details_extend_contract_{order_id}"
# })
# if not json.loads(new_pin.text)["rc"] == "100":
# print("new PIN Not Sended")
# return False
# sleep WAITING_TIME_OF_PIN seconds waiting for mailparser email parsed PIN
time.sleep(WAITING_TIME_OF_PIN)
pin = get_pin_from_mailparser(mailparser_dl_url_id)
log(f"[MailParser] PIN: {pin}")
# using PIN instead of password to get token
data = {
"auth": pin,
"sess_id": sess_id,
"subaction": "kc2_security_password_get_token",
"prefix": "kc2_customer_contract_details_extend_contract_",
"type": 1,
"ident": f"kc2_customer_contract_details_extend_contract_{order_id}",
}
f = session.post(url, headers=headers, data=data)
f.raise_for_status()
if not json.loads(f.text)["rs"] == "success":
return False
token = json.loads(f.text)["token"]["value"]
data = {
"sess_id": sess_id,
"ord_id": order_id,
"subaction": "kc2_customer_contract_details_extend_contract_term",
"token": token,
}
session.post(url, headers=headers, data=data)
time.sleep(5)
return True
def check(sess_id: str, session: requests.session):
print("Checking.......")
d = get_servers(sess_id, session)
flag = True
for key, val in d.items():
if val:
flag = False
log("[EUserv] ServerID: %s Renew Failed!" % key)
if flag:
log("[EUserv] ALL Work Done! Enjoy~")
# Telegram Bot Push https://core.telegram.org/bots/api#authorizing-your-bot
def telegram():
data = (("chat_id", TG_USER_ID), ("text", "EUserv续费日志\n\n" + desp))
response = requests.post(
TG_API_HOST + "/bot" + TG_BOT_TOKEN + "/sendMessage", data=data
)
if response.status_code != 200:
print("Telegram Bot 推送失败")
else:
print("Telegram Bot 推送成功")
def send_mail_by_yandex(
to_email, from_email, subject, text, files, sender_email, sender_password
):
msg = MIMEMultipart()
msg["Subject"] = subject
msg["From"] = from_email
msg["To"] = to_email
msg.attach(MIMEText(text, _charset="utf-8"))
if files is not None:
for file in files:
file_name, file_content = file
# print(file_name)
part = MIMEApplication(file_content)
part.add_header(
"Content-Disposition", "attachment", filename=("gb18030", "", file_name)
)
msg.attach(part)
s = SMTP_SSL("smtp.yandex.ru", 465)
s.login(sender_email, sender_password)
try:
s.sendmail(msg["From"], msg["To"], msg.as_string())
except SMTPDataError as e:
raise e
finally:
s.close()
def email():
msg = "EUserv 续费日志\n\n" + desp
try:
send_mail_by_yandex(
RECEIVER_EMAIL, YD_EMAIL, "EUserv 续费日志", msg, None, YD_EMAIL, YD_APP_PWD
)
print("eMail 推送成功")
except requests.exceptions.RequestException as e:
print(str(e))
print("eMail 推送失败")
except SMTPDataError as e1:
print(str(e1))
print("eMail 推送失败")
def main_handler(event, context):
if not USERNAME or not PASSWORD:
log("[EUserv] 你没有添加任何账户")
exit(1)
user_list = [USERNAME]
passwd_list = [PASSWORD]
mailparser_dl_url_id_list = [MAILPARSER_DOWNLOAD_URL_ID]
if len(user_list) != len(passwd_list):
log("[EUserv] The number of usernames and passwords do not match!")
exit(1)
if len(mailparser_dl_url_id_list) != len(user_list):
log("[Mailparser] The number of mailparser_dl_url_ids and usernames do not match!")
exit(1)
for i in range(len(user_list)):
print("*" * 30)
log("[EUserv] 正在续费第 %d 个账号" % (i + 1))
sessid, s = login(user_list[i], passwd_list[i])
if sessid == "-1":
log("[EUserv] 第 %d 个账号登陆失败,请检查登录信息" % (i + 1))
continue
SERVERS = get_servers(sessid, s)
log("[EUserv] 检测到第 {} 个账号有 {} 台 VPS,正在尝试续期".format(i + 1, len(SERVERS)))
for k, v in SERVERS.items():
if v:
if not renew(sessid, s, passwd_list[i], k, mailparser_dl_url_id_list[i]):
log("[EUserv] ServerID: %s Renew Error!" % k)
else:
log("[EUserv] ServerID: %s has been successfully renewed!" % k)
else:
log("[EUserv] ServerID: %s does not need to be renewed" % k)
time.sleep(15)
check(sessid, s)
time.sleep(5)
TG_BOT_TOKEN and TG_USER_ID and TG_API_HOST and telegram()
RECEIVER_EMAIL and YD_EMAIL and YD_APP_PWD and email()
print("*" * 30)
if __name__ == "__main__":
main_handler(None, None)
|