AIL/ail-framework
7
0
Fork 0
mirror of https://github.com/ail-project/ail-framework.git synced 2024-11-27 00:07:16 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix: [Crawler] unpack_url

Browse source
This commit is contained in:
Terrtia 2019-04-25 13:54:06 +02:00
parent 39d0bfa65a
commit fc2c1422ff
No known key found for this signature in database
GPG key ID: 1E1B1F50D84613D0
1 changed files with 17 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

21
bin/Crawler.py
View file

@ -47,12 +47,16 @@ def unpack_url(url):
if url_unpack['scheme'] is None: if url_unpack['scheme'] is None:
to_crawl['scheme'] = 'http' to_crawl['scheme'] = 'http'
url= 'http://{}'.format(url_unpack['url'].decode())
else: else:
scheme = url_unpack['scheme'].decode() scheme = url_unpack['scheme'].decode()
if scheme in default_proto_map: if scheme in default_proto_map:
to_crawl['scheme'] = scheme to_crawl['scheme'] = scheme
url = url_unpack['url'].decode()
else: else:
redis_crawler.sadd('new_proto', '{} {}'.format(scheme, url_unpack['url'].decode()))
to_crawl['scheme'] = 'http' to_crawl['scheme'] = 'http'
url= 'http://{}'.format(url_unpack['url'].decode().replace(scheme, '', 1))
if url_unpack['port'] is None: if url_unpack['port'] is None:
to_crawl['port'] = default_proto_map[to_crawl['scheme']] to_crawl['port'] = default_proto_map[to_crawl['scheme']]
@ -66,11 +70,20 @@ def unpack_url(url):
port = default_proto_map[to_crawl['scheme']] port = default_proto_map[to_crawl['scheme']]
to_crawl['port'] = port to_crawl['port'] = port
if url_unpack['query_string'] is None: #if url_unpack['query_string'] is None:
to_crawl['url']= '{}://{}:{}'.format(to_crawl['scheme'], url_unpack['host'].decode(), to_crawl['port']) # if to_crawl['port'] == 80:
# to_crawl['url']= '{}://{}'.format(to_crawl['scheme'], url_unpack['host'].decode())
# else:
# to_crawl['url']= '{}://{}:{}'.format(to_crawl['scheme'], url_unpack['host'].decode(), to_crawl['port'])
#else:
# to_crawl['url']= '{}://{}:{}{}'.format(to_crawl['scheme'], url_unpack['host'].decode(), to_crawl['port'], url_unpack['query_string'].decode())
to_crawl['url'] = url
if to_crawl['port'] == 80:
to_crawl['domain_url'] = '{}://{}'.format(to_crawl['scheme'], url_unpack['host'].decode())
else: else:
to_crawl['url']= '{}://{}:{}{}'.format(to_crawl['scheme'], url_unpack['host'].decode(), to_crawl['port'], url_unpack['query_string'].decode()) to_crawl['domain_url'] = '{}://{}:{}'.format(to_crawl['scheme'], url_unpack['host'].decode(), to_crawl['port'])
to_crawl['domain_url'] = '{}://{}:{}'.format(to_crawl['scheme'], to_crawl['domain'], to_crawl['port'])
to_crawl['tld'] = url_unpack['tld'].decode() to_crawl['tld'] = url_unpack['tld'].decode()
return to_crawl return to_crawl