summaryrefslogtreecommitdiff
path: root/searx/engines/dailymotion.py
diff options
context:
space:
mode:
authorDalf <alex@al-f.net>2014-01-05 13:55:17 +0100
committerDalf <alex@al-f.net>2014-01-05 13:55:17 +0100
commit49c85fce51c2b8431013d9269999eac5a6a9c038 (patch)
tree5ca4a97f0d5b9fe640b5c1e8e8140feecb98a546 /searx/engines/dailymotion.py
parente88cf0a0a877458e2babc8af43c5a77654036af6 (diff)
downloadsearxng-49c85fce51c2b8431013d9269999eac5a6a9c038.tar.gz
searxng-49c85fce51c2b8431013d9269999eac5a6a9c038.zip
[fix] dailymotion engine : no more html tag in the description
Diffstat (limited to 'searx/engines/dailymotion.py')
-rw-r--r--searx/engines/dailymotion.py12
1 files changed, 9 insertions, 3 deletions
diff --git a/searx/engines/dailymotion.py b/searx/engines/dailymotion.py
index 7046132f3..655485957 100644
--- a/searx/engines/dailymotion.py
+++ b/searx/engines/dailymotion.py
@@ -1,16 +1,17 @@
from urllib import urlencode
+from lxml import html
from json import loads
from cgi import escape
categories = ['videos']
-localization = 'en'
+locale = 'en_US'
# see http://www.dailymotion.com/doc/api/obj-video.html
search_url = 'https://api.dailymotion.com/videos?fields=title,description,duration,url,thumbnail_360_url&sort=relevance&limit=25&page=1&{query}'
def request(query, params):
global search_url
- params['url'] = search_url.format(query=urlencode({'search': query, 'localization': localization }))
+ params['url'] = search_url.format(query=urlencode({'search': query, 'localization': locale }))
return params
@@ -27,6 +28,11 @@ def response(resp):
else:
content = ''
if res['description']:
- content += escape(res['description'][:500])
+ description = text_content_from_html(res['description'])
+ content += description[:500]
results.append({'url': url, 'title': title, 'content': content})
return results
+
+def text_content_from_html(html_string):
+ desc_html = html.fragment_fromstring(html_string, create_parent=True)
+ return desc_html.text_content()