forked from Toperlock/sing-box-subscribe
-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.py
571 lines (517 loc) · 23.6 KB
/
main.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
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
import json, os, tool, time, requests, sys, urllib, importlib, argparse, yaml, ruamel.yaml
import re
from datetime import datetime
from urllib.parse import urlparse
from api.app import TEMP_DIR
from parsers.clash2base64 import clash2v2ray
parsers_mod = {}
providers = None
color_code = [31, 32, 33, 34, 35, 36, 91, 92, 93, 94, 95, 96]
def loop_color(text):
text = '\033[1;{color}m{text}\033[0m'.format(color=color_code[0], text=text)
color_code.append(color_code.pop(0))
return text
def init_parsers():
b = os.walk('parsers')
for path, dirs, files in b:
for file in files:
f = os.path.splitext(file)
if f[1] == '.py':
parsers_mod[f[0]] = importlib.import_module('parsers.' + f[0])
def get_template():
template_dir = 'config_template' # 配置模板文件夹路径
template_files = os.listdir(template_dir) # 获取文件夹中的所有文件
template_list = [os.path.splitext(file)[0] for file in template_files if
file.endswith('.json')] # 移除扩展名并过滤出以.json结尾的文件
template_list.sort() # 对文件名进行排序
return template_list
def load_json(path):
return json.loads(tool.readFile(path))
def process_subscribes(subscribes):
nodes = {}
for subscribe in subscribes:
if 'enabled' in subscribe and not subscribe['enabled']:
continue
_nodes = get_nodes(subscribe['url'])
if _nodes and len(_nodes) > 0:
add_prefix(_nodes, subscribe)
add_emoji(_nodes, subscribe)
if subscribe.get('subgroup'):
subscribe['tag'] = subscribe['tag'] + '-' + subscribe['subgroup'] + '-' + 'subgroup'
if not nodes.get(subscribe['tag']):
nodes[subscribe['tag']] = []
nodes[subscribe['tag']] += _nodes
else:
print('没有在此订阅下找到节点,跳过')
# print('Không tìm thấy proxy trong link thuê bao này, bỏ qua')
tool.proDuplicateNodeName(nodes)
return nodes
def nodes_filter(nodes, filter, group):
for a in filter:
if a.get('for') and group not in a['for']:
continue
nodes = action_keywords(nodes, a['action'], a['keywords'])
return nodes
def action_keywords(nodes, action, keywords):
# filter将按顺序依次执行
# "filter":[
# {"action":"include","keywords":[""]},
# {"action":"exclude","keywords":[""]}
# ]
temp_nodes = []
flag = False
if action == 'exclude':
flag = True
'''
# 空关键字过滤
'''
# Join the patterns list into a single pattern, separated by '|'
combined_pattern = '|'.join(keywords)
# If the combined pattern is empty or only contains whitespace, return the original nodes
if not combined_pattern or combined_pattern.isspace():
return nodes
# Compile the combined regex pattern
compiled_pattern = re.compile(combined_pattern)
for node in nodes:
name = node['tag']
# Use regex to check for a match
match_flag = bool(compiled_pattern.search(name))
# Use XOR to decide if the node should be included based on the action
if match_flag ^ flag:
temp_nodes.append(node)
return temp_nodes
def add_prefix(nodes, subscribe):
if subscribe.get('prefix'):
for node in nodes:
node['tag'] = subscribe['prefix'] + node['tag']
def add_emoji(nodes, subscribe):
if subscribe.get('emoji'):
for node in nodes:
node['tag'] = tool.rename(node['tag'])
def get_nodes(url):
if url.startswith('sub://'):
url = tool.b64Decode(url[6:]).decode('utf-8')
urlstr = urlparse(url)
if not urlstr.scheme:
try:
content = tool.b64Decode(url).decode('utf-8')
data = parse_content(content)
processed_list = []
for item in data:
if isinstance(item, tuple):
processed_list.extend([item[0], item[1]]) # 处理shadowtls
else:
processed_list.append(item)
return processed_list
except:
content = get_content_form_file(url)
else:
content = get_content_from_url(url)
# print (content)
if type(content) == dict:
if 'proxies' in content:
share_links = []
for proxy in content['proxies']:
share_links.append(clash2v2ray(proxy))
data = '\n'.join(share_links)
data = parse_content(data)
processed_list = []
for item in data:
if isinstance(item, tuple):
processed_list.extend([item[0], item[1]]) # 处理shadowtls
else:
processed_list.append(item)
return processed_list
elif 'outbounds' in content:
outbounds = []
excluded_types = {"selector", "urltest", "direct", "block", "dns"}
filtered_outbounds = [outbound for outbound in content['outbounds'] if outbound.get("type") not in excluded_types]
outbounds.extend(filtered_outbounds)
return outbounds
else:
data = parse_content(content)
processed_list = []
for item in data:
if isinstance(item, tuple):
processed_list.extend([item[0], item[1]]) # 处理shadowtls
else:
processed_list.append(item)
return processed_list
def parse_content(content):
# firstline = tool.firstLine(content)
# # print(firstline)
# if not get_parser(firstline):
# return None
nodelist = []
for t in content.splitlines():
t = t.strip()
if len(t) == 0:
continue
factory = get_parser(t)
if not factory:
continue
node = factory(t)
if node:
nodelist.append(node)
return nodelist
def get_parser(node):
proto = tool.get_protocol(node)
if providers.get('exclude_protocol'):
eps = providers['exclude_protocol'].split(',')
if len(eps) > 0:
eps = [protocol.strip() for protocol in eps]
if proto in eps:
return None
if not proto or proto not in parsers_mod.keys():
return None
return parsers_mod[proto].parse
def get_content_from_url(url, n=6):
UA = ''
print('处理: \033[31m' + url + '\033[0m')
# print('Đang tải link đăng ký: \033[31m' + url + '\033[0m')
prefixes = ["vmess://", "vless://", "ss://", "ssr://", "trojan://", "tuic://", "hysteria://", "hysteria2://",
"hy2://", "wg://", "http2://", "socks://", "socks5://"]
if any(url.startswith(prefix) for prefix in prefixes):
response_text = tool.noblankLine(url)
return response_text
for subscribe in providers["subscribes"]:
if 'enabled' in subscribe and not subscribe['enabled']:
continue
if subscribe['url'] == url:
UA = subscribe.get('User-Agent', '')
response = tool.getResponse(url, custom_user_agent=UA)
concount = 1
while concount <= n and not response:
print('连接出错,正在进行第 ' + str(concount) + ' 次重试,最多重试 ' + str(n) + ' 次...')
# print('Lỗi kết nối, đang thử lại '+str(concount)+'/'+str(n)+'...')
response = tool.getResponse(url)
concount = concount + 1
time.sleep(1)
if not response:
print('获取错误,跳过此订阅')
# print('Lỗi khi tải link đăng ký, bỏ qua link đăng ký này')
print('----------------------------')
pass
response_content = response.content
response_text = response_content.decode('utf-8-sig') # utf-8-sig 可以忽略 BOM
#response_encoding = response.encoding
if response_text.isspace():
print('没有从订阅链接获取到任何内容')
# print('Không nhận được proxy nào từ link đăng ký')
return None
if not response_text:
response = tool.getResponse(url, custom_user_agent='clashmeta')
response_text = response.text
if any(response_text.startswith(prefix) for prefix in prefixes):
response_text = tool.noblankLine(response_text)
return response_text
elif 'proxies' in response_text:
yaml_content = response.content.decode('utf-8')
yaml = ruamel.yaml.YAML()
try:
response_text = dict(yaml.load(yaml_content))
return response_text
except:
pass
elif 'outbounds' in response_text:
try:
response_text = json.loads(response.text)
return response_text
except:
pass
else:
try:
response_text = tool.b64Decode(response_text)
response_text = response_text.decode(encoding="utf-8")
# response_text = bytes.decode(response_text,encoding=response_encoding)
except:
pass
# traceback.print_exc()
return response_text
def get_content_form_file(url):
print('处理: \033[31m' + url + '\033[0m')
# print('Đang tải link đăng ký: \033[31m' + url + '\033[0m')
# encoding = tool.get_encoding(url)
file_extension = os.path.splitext(url)[1] # 获取文件的后缀名
if file_extension.lower() == '.yaml':
with open(url, 'rb') as file:
content = file.read()
yaml_data = dict(yaml.safe_load(content))
share_links = []
for proxy in yaml_data['proxies']:
share_links.append(clash2v2ray(proxy))
node = '\n'.join(share_links)
processed_list = tool.noblankLine(node)
return processed_list
else:
data = tool.readFile(url)
data = bytes.decode(data, encoding='utf-8')
data = tool.noblankLine(data)
return data
def save_config(path, nodes):
try:
if 'auto_backup' in providers and providers['auto_backup']:
now = datetime.now().strftime('%Y%m%d%H%M%S')
if os.path.exists(path):
os.rename(path, f'{path}.{now}.bak')
if os.path.exists(path):
os.remove(path)
print(f"已删除文件,并重新保存:\033[33m{path}\033[0m")
# print(f"File cấu hình đã được lưu vào: \033[33m{path}\033[0m")
else:
print(f"文件不存在,正在保存:\033[33m{path}\033[0m")
# print(f"File không tồn tại, đang lưu tại: \033[33m{path}\033[0m")
tool.saveFile(path, json.dumps(nodes, indent=2, ensure_ascii=False))
except Exception as e:
print(f"保存配置文件时出错:{str(e)}")
# print(f"Lỗi khi lưu file cấu hình: {str(e)}")
# 如果保存出错,尝试使用 config_file_path 再次保存
config_path = json.loads(temp_json_data).get("save_config_path", "config.json")
CONFIG_FILE_NAME = config_path
config_file_path = os.path.join('/tmp', CONFIG_FILE_NAME)
try:
if os.path.exists(config_file_path):
os.remove(config_file_path)
print(f"已删除文件,并重新保存:\033[33m{config_file_path}\033[0m")
# print(f"File cấu hình đã được lưu vào: \033[33m{config_file_path}\033[0m")
else:
print(f"文件不存在,正在保存:\033[33m{config_file_path}\033[0m")
# print(f"File không tồn tại, đang lưu tại: \033[33m{config_file_path}\033[0m")
tool.saveFile(config_file_path, json.dumps(nodes, indent=2, ensure_ascii=False))
# print(f"配置文件已保存到 {config_file_path}")
# print(f"Tập tin cấu hình đã được lưu vào {config_file_path}")
except Exception as e:
os.remove(config_file_path)
print(f"已删除文件:\033[33m{config_file_path}\033[0m")
# print(f"Các file đã bị xóa: \033[33m{config_file_path}\033[0m")
print(f"再次保存配置文件时出错:{str(e)}")
# print(f"Lỗi khi lưu lại file cấu hình: {str(e)}")
def set_proxy_rule_dns(config):
# dns_template = {
# "tag": "remote",
# "address": "tls://1.1.1.1",
# "detour": ""
# }
config_rules = config['route']['rules']
outbound_dns = []
dns_rules = config['dns']['rules']
asod = providers["auto_set_outbounds_dns"]
for rule in config_rules:
if rule['outbound'] not in ['block', 'dns-out']:
if rule['outbound'] != 'direct':
outbounds_dns_template = \
list(filter(lambda server: server['tag'] == asod["proxy"], config['dns']['servers']))[0]
dns_obj = outbounds_dns_template.copy()
dns_obj['tag'] = rule['outbound'] + '_dns'
dns_obj['detour'] = rule['outbound']
if dns_obj not in outbound_dns:
outbound_dns.append(dns_obj)
if rule.get('type') and rule['type'] == 'logical':
dns_rule_obj = {
'type': 'logical',
'mode': rule['mode'],
'rules': [],
'server': rule['outbound'] + '_dns' if rule['outbound'] != 'direct' else asod["direct"]
}
for _rule in rule['rules']:
child_rule = pro_dns_from_route_rules(_rule)
if child_rule:
dns_rule_obj['rules'].append(child_rule)
if len(dns_rule_obj['rules']) == 0:
dns_rule_obj = None
else:
dns_rule_obj = pro_dns_from_route_rules(rule)
if dns_rule_obj:
dns_rules.append(dns_rule_obj)
# 清除重复规则
_dns_rules = []
for dr in dns_rules:
if dr not in _dns_rules:
_dns_rules.append(dr)
config['dns']['rules'] = _dns_rules
config['dns']['servers'].extend(outbound_dns)
def pro_dns_from_route_rules(route_rule):
dns_route_same_list = ["inbound", "ip_version", "network", "protocol", 'domain', 'domain_suffix', 'domain_keyword',
'domain_regex', 'geosite', "source_geoip", "source_ip_cidr", "source_port",
"source_port_range", "port", "port_range", "process_name", "process_path", "package_name",
"user", "user_id", "clash_mode", "invert"]
dns_rule_obj = {}
for key in route_rule:
if key in dns_route_same_list:
dns_rule_obj[key] = route_rule[key]
if len(dns_rule_obj) == 0:
return None
if route_rule.get('outbound'):
dns_rule_obj['server'] = route_rule['outbound'] + '_dns' if route_rule['outbound'] != 'direct' else \
providers["auto_set_outbounds_dns"]['direct']
return dns_rule_obj
def pro_node_template(data_nodes, config_outbound, group):
if config_outbound.get('filter'):
data_nodes = nodes_filter(data_nodes, config_outbound['filter'], group)
return [node.get('tag') for node in data_nodes]
def combin_to_config(config, data):
config_outbounds = config["outbounds"] if config.get("outbounds") else None
i = 0
for group in data:
if 'subgroup' in group:
i += 1
for out in config_outbounds:
if out.get("outbounds"):
if out['tag'] == 'proxy':
out["outbounds"] = [out["outbounds"]] if isinstance(out["outbounds"], str) else out["outbounds"]
if '{all}' in out["outbounds"]:
index_of_all = out["outbounds"].index('{all}')
out["outbounds"][index_of_all] = (group.rsplit("-", 1)[0]).rsplit("-", 1)[-1]
i += 1
else:
out["outbounds"].insert(i - 1, (group.rsplit("-", 1)[0]).rsplit("-", 1)[-1])
new_outbound = {'tag': (group.rsplit("-", 1)[0]).rsplit("-", 1)[-1], 'type': 'selector', 'outbounds': ['{' + group + '}']}
config_outbounds.insert(-4, new_outbound)
else:
for out in config_outbounds:
if out.get("outbounds"):
if out['tag'] == 'proxy':
out["outbounds"] = [out["outbounds"]] if isinstance(out["outbounds"], str) else out["outbounds"]
out["outbounds"].append('{' + group + '}')
temp_outbounds = []
if config_outbounds:
# 提前处理all模板
for po in config_outbounds:
# 处理出站
if po.get("outbounds"):
if '{all}' in po["outbounds"]:
o1 = []
for item in po["outbounds"]:
if item.startswith('{') and item.endswith('}'):
_item = item[1:-1]
if _item == 'all':
o1.append(item)
else:
o1.append(item)
po['outbounds'] = o1
t_o = []
check_dup = []
for oo in po["outbounds"]:
# 避免添加重复节点
if oo in check_dup:
continue
else:
check_dup.append(oo)
# 处理模板
if oo.startswith('{') and oo.endswith('}'):
oo = oo[1:-1]
if data.get(oo):
nodes = data[oo]
t_o.extend(pro_node_template(nodes, po, oo))
else:
if oo == 'all':
for group in data:
nodes = data[group]
t_o.extend(pro_node_template(nodes, po, group))
else:
t_o.append(oo)
if len(t_o) == 0:
print('发现 {} 出站下的节点数量为 0 ,会导致sing-box无法运行,请检查config模板是否正确。'.format(
po['tag']))
# print('Sing-Box không chạy được vì không tìm thấy bất kỳ proxy nào trong outbound của {}. Vui lòng kiểm tra xem mẫu cấu hình có đúng không!!'.format(po['tag']))
config_path = json.loads(temp_json_data).get("save_config_path", "config.json")
CONFIG_FILE_NAME = config_path
config_file_path = os.path.join('/tmp', CONFIG_FILE_NAME)
if os.path.exists(config_file_path):
os.remove(config_file_path)
print(f"已删除文件:{config_file_path}")
# print(f"Các tập tin đã bị xóa: {config_file_path}")
sys.exit()
po['outbounds'] = t_o
if po.get('filter'):
del po['filter']
for group in data:
temp_outbounds.extend(data[group])
config['outbounds'] = config_outbounds + temp_outbounds
# 自动配置路由规则到dns规则,避免dns泄露
dns_tags = [server.get('tag') for server in config['dns']['servers']]
asod = providers.get("auto_set_outbounds_dns")
if asod and asod.get('proxy') and asod.get('direct') and asod['proxy'] in dns_tags and asod['direct'] in dns_tags:
set_proxy_rule_dns(config)
return config
def updateLocalConfig(local_host, path):
header = {
'Content-Type': 'application/json'
}
r = requests.put(local_host + '/configs?force=false', json={"path": path}, headers=header)
print(r.text)
def display_template(tl):
print_str = ''
for i in range(len(tl)):
print_str += loop_color('{index}、{name} '.format(index=i + 1, name=tl[i]))
print(print_str)
def select_config_template(tl, selected_template_index=None):
if args.template_index is not None:
uip = args.template_index
else:
# print ('Nhập số để chọn mẫu cấu hình tương ứng (nhấn Enter để chọn mẫu cấu hình đầu tiên theo mặc định): ')
uip = input('输入序号,载入对应config模板(直接回车默认选第一个配置模板):')
try:
if uip == '':
return 0
uip = int(uip)
if uip < 1 or uip > len(tl):
print('输入了错误信息!重新输入')
# print('Nhập thông tin không chính xác! Vui lòng nhập lại')
return select_config_template(tl)
else:
uip -= 1
except:
print('输入了错误信息!重新输入')
# print('Nhập thông tin không chính xác! Vui lòng nhập lại')
return select_config_template(tl)
return uip
# 自定义函数,用于解析参数为 JSON 格式
def parse_json(value):
try:
return json.loads(value)
except json.JSONDecodeError:
raise argparse.ArgumentTypeError(f"Invalid JSON: {value}")
if __name__ == '__main__':
init_parsers()
parser = argparse.ArgumentParser()
parser.add_argument('--temp_json_data', type=parse_json, help='临时内容')
parser.add_argument('--template_index', type=int, help='模板序号')
args = parser.parse_args()
temp_json_data = args.temp_json_data
if temp_json_data and temp_json_data != '{}':
providers = json.loads(temp_json_data)
else:
providers = load_json('providers.json') # 加载本地 providers.json
if providers.get('config_template'):
config_template_path = providers['config_template']
print('选择: \033[33m' + config_template_path + '\033[0m')
# print ('Mẫu cấu hình sử dụng: \033[33m' + template_list[uip] + '.json\033[0m')
response = requests.get(providers['config_template'])
response.raise_for_status()
config = response.json()
else:
template_list = get_template()
if len(template_list) < 1:
print('没有找到模板文件')
# print('Không tìm thấy file mẫu')
sys.exit()
display_template(template_list)
uip = select_config_template(template_list, selected_template_index=args.template_index)
config_template_path = 'config_template/' + template_list[uip] + '.json'
print('选择: \033[33m' + template_list[uip] + '.json\033[0m')
# print ('Mẫu cấu hình sử dụng: \033[33m' + template_list[uip] + '.json\033[0m')
config = load_json(config_template_path)
nodes = process_subscribes(providers["subscribes"])
if providers.get('Only-nodes'):
combined_contents = []
for sub_tag, contents in nodes.items():
# 遍历每个机场的内容
for content in contents:
# 将内容添加到新列表中
combined_contents.append(content)
final_config = combined_contents # 只返回节点信息
else:
final_config = combin_to_config(config, nodes) # 节点信息添加到模板
save_config(providers["save_config_path"], final_config)
# updateLocalConfig('http://127.0.0.1:9090',providers['save_config_path'])