diff options
| author | asciimoo <asciimoo@gmail.com> | 2014-01-20 02:31:20 +0100 |
|---|---|---|
| committer | asciimoo <asciimoo@gmail.com> | 2014-01-20 02:31:20 +0100 |
| commit | b2492c94f422e18cb8954ec983134f4fa5c7cdc0 (patch) | |
| tree | 969ea30e5dc642d896fa7b744571110ebfe13e7a /searx/engines/startpage.py | |
| parent | 692c0bf5f0b353bfbb46aaee1af54afb164dedbc (diff) | |
[fix] pep/flake8 compatibility
Diffstat (limited to 'searx/engines/startpage.py')
| -rw-r--r-- | searx/engines/startpage.py | 6 |
1 files changed, 2 insertions, 4 deletions
diff --git a/searx/engines/startpage.py b/searx/engines/startpage.py index 87c091e2d..d6d7cf44d 100644 --- a/searx/engines/startpage.py +++ b/searx/engines/startpage.py @@ -1,11 +1,10 @@ from urllib import urlencode from lxml import html -from urlparse import urlparse -from cgi import escape base_url = 'https://startpage.com/' search_url = base_url+'do/search' + def request(query, params): global search_url query = urlencode({'q': query})[2:] @@ -20,11 +19,10 @@ def response(resp): results = [] 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"] + # not ads: div[@class="result"] are the direct childs of div[@id="results"] for result in dom.xpath('//div[@id="results"]/div[@class="result"]'): link = result.xpath('.//h3/a')[0] url = link.attrib.get('href') - parsed_url = urlparse(url) title = link.text_content() content = result.xpath('./p[@class="desc"]')[0].text_content() results.append({'url': url, 'title': title, 'content': content}) |