mirror of
https://github.com/Johnshall/Shadowrocket-ADBlock-Rules-Forever.git
synced 2025-12-17 23:34:48 +08:00
96 lines
2.2 KiB
Python
96 lines
2.2 KiB
Python
# -*- coding: utf-8 -*-
|
|
|
|
#
|
|
# 提取广告规则，并且只提取对全域禁止的那种规则
|
|
#
|
|
|
|
import time
|
|
import sys
|
|
import requests
|
|
import re
|
|
|
|
|
|
rules_url = [
|
|
# EasyList China
|
|
#'https://easylist-downloads.adblockplus.org/easylistchina.txt',
|
|
# EasyList + China
|
|
'https://easylist-downloads.adblockplus.org/easylistchina+easylist.txt',
|
|
# 乘风 广告过滤规则
|
|
'https://raw.githubusercontent.com/xinggsf/Adblock-Plus-Rule/master/ABP-FX.txt'
|
|
]
|
|
|
|
# contain both domains and ips
|
|
domains = []
|
|
|
|
|
|
for rule_url in rules_url:
|
|
print('loading... ' + rule_url)
|
|
|
|
# get rule text
|
|
success = False
|
|
try_times = 0
|
|
r = None
|
|
while try_times < 5 and not success:
|
|
r = requests.get(rule_url)
|
|
if r.status_code != 200:
|
|
time.sleep(1)
|
|
try_times = try_times + 1
|
|
else:
|
|
success = True
|
|
break
|
|
|
|
if not success:
|
|
sys.exit('error in request %s\n\treturn code: %d' % (rule_url, r.status_code) )
|
|
|
|
rule = r.text
|
|
|
|
# parse rule
|
|
rule = rule.split('\n')
|
|
for row in rule:
|
|
row = row.strip()
|
|
|
|
# 直接跳过
|
|
if row.startswith('!') or row.startswith('@@') or "$" in row:
|
|
continue
|
|
|
|
# 清除前缀
|
|
row = re.sub(r'^\|?https?:\/\/', '', row)
|
|
row = re.sub(r'^\|\|', '', row)
|
|
row = row.lstrip('.*')
|
|
|
|
# 清除后缀
|
|
row = row.rstrip('/^*')
|
|
|
|
# 不能含有的字符
|
|
if re.search(r'[\/\^:\*]', row):
|
|
continue
|
|
|
|
# 只匹配域名或 IP
|
|
if re.match(r'^([a-zA-Z0-9]([a-zA-Z0-9\-]{0,61}[a-zA-Z0-9])?\.)+[a-zA-Z]{2,9}$', row) or re.match(r'^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}$', row):
|
|
domains.append(row)
|
|
|
|
print('done.')
|
|
|
|
|
|
# write into files
|
|
|
|
domains.sort()
|
|
|
|
file_ad = sys.stdout
|
|
try:
|
|
if sys.version_info.major == 3:
|
|
file_ad = open('resultant/ad.list', 'w', encoding='utf-8')
|
|
else:
|
|
file_ad = open('resultant/ad.list', 'w')
|
|
except:
|
|
pass
|
|
|
|
file_ad.write('# adblock rules refresh time: ' + time.strftime("%Y-%m-%d %H:%M:%S") + '\n')
|
|
|
|
last = ''
|
|
for item in domains:
|
|
if last == item:
|
|
continue
|
|
file_ad.write(item + '\n')
|
|
last = item
|