[fix] html escape

This commit is contained in:
Adam Tauber 2014-08-06 14:43:44 +02:00
parent ea414c6574
commit 111a86d355

View File

@ -1,5 +1,6 @@
from urllib import urlencode from urllib import urlencode
from lxml import html from lxml import html
from cgi import escape
base_url = None base_url = None
search_url = None search_url = None
@ -35,11 +36,11 @@ def response(resp):
if url.startswith('http://www.google.')\ if url.startswith('http://www.google.')\
or url.startswith('https://www.google.'): or url.startswith('https://www.google.'):
continue continue
title = link.text_content() title = escape(link.text_content())
content = '' content = ''
if result.xpath('./p[@class="desc"]'): if result.xpath('./p[@class="desc"]'):
content = result.xpath('./p[@class="desc"]')[0].text_content() content = escape(result.xpath('./p[@class="desc"]')[0].text_content())
results.append({'url': url, 'title': title, 'content': content}) results.append({'url': url, 'title': title, 'content': content})