mirror of https://github.com/CIRCL/AIL-framework
35 lines
1.0 KiB
Python
Executable File
35 lines
1.0 KiB
Python
Executable File
#!/usr/bin/env python3
|
|
# -*-coding:UTF-8 -*
|
|
|
|
import os
|
|
import sys
|
|
import configparser
|
|
from TorSplashCrawler import TorSplashCrawler
|
|
|
|
if __name__ == '__main__':
|
|
|
|
if len(sys.argv) != 5:
|
|
print('usage:', 'tor_crawler.py', 'url', 'domain', 'paste', 'super_father')
|
|
exit(1)
|
|
|
|
configfile = os.path.join(os.environ['AIL_BIN'], 'packages/config.cfg')
|
|
if not os.path.exists(configfile):
|
|
raise Exception('Unable to find the configuration file. \
|
|
Did you set environment variables? \
|
|
Or activate the virtualenv.')
|
|
|
|
cfg = configparser.ConfigParser()
|
|
cfg.read(configfile)
|
|
|
|
splash_url = cfg.get("Crawler", "splash_url")
|
|
http_proxy = cfg.get("Crawler", "http_proxy")
|
|
crawler_depth_limit = cfg.getint("Crawler", "crawler_depth_limit")
|
|
|
|
url = sys.argv[1]
|
|
domain = sys.argv[2]
|
|
paste = sys.argv[3]
|
|
super_father = sys.argv[4]
|
|
|
|
crawler = TorSplashCrawler(splash_url, http_proxy, crawler_depth_limit)
|
|
crawler.crawl(url, domain, paste, super_father)
|