searx

commit: 7db4558de707d62e85ab8b68fa4198d92cf537d7
parent: 6f86b96a3536d9c99654e5be261437a4ed519ee8
Author: asciimoo <asciimoo@gmail.com>
Date:   Tue, 18 Feb 2014 16:14:31 +0100

[mod][fix] startpage engine updates

Diffstat:
M
searx/engines/startpage.py
5
++++-

1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/searx/engines/startpage.py b/searx/engines/startpage.py
@@ -29,9 +29,12 @@ def response(resp):
     dom = html.fromstring(resp.content)
     # ads xpath //div[@id="results"]/div[@id="sponsored"]//div[@class="result"]
     # not ads: div[@class="result"] are the direct childs of div[@id="results"]
-    for result in dom.xpath('//div[@id="results"]/div[@class="result"]'):
+    for result in dom.xpath('//div[@class="result"]'):
         link = result.xpath('.//h3/a')[0]
         url = link.attrib.get('href')
+        if url.startswith('http://www.google.')\
+           or url.startswith('https://www.google.'):
+            continue
         title = link.text_content()
 
         content = ''