[fix] improve result handling of startpage engine

This commit is contained in:
Thomas Pointhuber 2015-08-24 11:28:55 +02:00
parent 5bffa9ca33
commit 23b9095cbf
1 changed files with 5 additions and 1 deletions

View File

@ -66,7 +66,11 @@ def response(resp):
url = link.attrib.get('href')
# block google-ad url's
if re.match("^http(s|)://www.google.[a-z]+/aclk.*$", url):
if re.match("^http(s|)://(www\.)?google\.[a-z]+/aclk.*$", url):
continue
# block startpage search url's
if re.match("^http(s|)://(www\.)?startpage\.com/do/search\?.*$", url):
continue
title = escape(extract_text(link))