devol
/
searx
mirror of https://github.com/searx/searx
1
0
Fork 0
Code Issues Releases Wiki Activity

[fix] google engine - ignore new useless result type

This commit is contained in:
Adam Tauber 2015-10-29 12:47:12 +01:00
parent a59d171385
commit 5d49c15f79
1 changed files with 6 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
searx/engines/google.py
View File

@ -12,9 +12,12 @@ import re
from cgi import escape from cgi import escape
from urllib import urlencode from urllib import urlencode
from urlparse import urlparse, parse_qsl from urlparse import urlparse, parse_qsl
from lxml import html from lxml import html, etree
from searx.poolrequests import get from searx.poolrequests import get
from searx.engines.xpath import extract_text, extract_url from searx.engines.xpath import extract_text, extract_url
from searx.search import logger
logger = logger.getChild('google engine')
# engine dependent config # engine dependent config
@ -225,8 +228,8 @@ def response(resp):
# parse results # parse results
for result in dom.xpath(results_xpath): for result in dom.xpath(results_xpath):
title = extract_text(result.xpath(title_xpath)[0])
try: try:
title = extract_text(result.xpath(title_xpath)[0])
url = parse_url(extract_url(result.xpath(url_xpath), google_url), google_hostname) url = parse_url(extract_url(result.xpath(url_xpath), google_url), google_hostname)
parsed_url = urlparse(url, google_hostname) parsed_url = urlparse(url, google_hostname)
@ -269,6 +272,7 @@ def response(resp):
'content': content 'content': content
}) })
except: except:
logger.debug('result parse error in:\n%s', etree.tostring(result, pretty_print=True))
continue continue
# parse suggestion # parse suggestion