diff options
author | Adam Tauber <asciimoo@gmail.com> | 2014-02-27 10:59:39 +0100 |
---|---|---|
committer | Adam Tauber <asciimoo@gmail.com> | 2014-02-27 10:59:39 +0100 |
commit | 46277888340ceaa76233c57be23fb21591188e26 (patch) | |
tree | ab5485d258ebea7cf56d483330915923c42d5728 /searx/engines/yahoo.py | |
parent | 8c8d9d10cb69234f295fc9d0f1ba6c739c3e7147 (diff) | |
download | searxng-46277888340ceaa76233c57be23fb21591188e26.tar.gz searxng-46277888340ceaa76233c57be23fb21591188e26.zip |
[fix] yahoo url mods
Diffstat (limited to 'searx/engines/yahoo.py')
-rw-r--r-- | searx/engines/yahoo.py | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/searx/engines/yahoo.py b/searx/engines/yahoo.py index 3dba4c432..f83b4b967 100644 --- a/searx/engines/yahoo.py +++ b/searx/engines/yahoo.py @@ -1,6 +1,7 @@ #!/usr/bin/env python from urllib import urlencode +from urlparse import unquote from lxml import html from searx.engines.xpath import extract_text, extract_url @@ -33,7 +34,10 @@ def response(resp): dom = html.fromstring(resp.text) for result in dom.xpath(results_xpath): - url = extract_url(result.xpath(url_xpath), search_url) + url_string = extract_url(result.xpath(url_xpath), search_url) + start = url_string.find('/RU=')+4 + end = url_string.rfind('/RS') + url = unquote(url_string[start:end]) title = extract_text(result.xpath(title_xpath)[0]) content = extract_text(result.xpath(content_xpath)[0]) results.append({'url': url, 'title': title, 'content': content}) |