searx

commit: 0549f8c7c8d25d9764c83bc006fa32cdddc13c46
parent 1d5741954eb6786bbbc31e9bc4d540f73bccdb46
Author: Thomas Pointhuber <thomas.pointhuber@gmx.at>
Date:   Tue,  4 Mar 2014 13:11:04 +0100

Create google_news.py
Diffstat:
A
searx/engines/google_news.py
37
+++++++++++++++++++++++++++++++++++++

1 file changed, 37 insertions(+), 0 deletions(-)
diff --git a/searx/engines/google_news.py b/searx/engines/google_news.py
@@ -0,0 +1,37 @@
+#!/usr/bin/env python
+
+from urllib import urlencode
+from json import loads
+
+categories = ['news']
+
+url = 'https://ajax.googleapis.com/'
+search_url = url + 'ajax/services/search/news?v=2.0&start={offset}&rsz=large&safe=off&filter=off&{query}&hl={language}' # noqa
+
+paging = True
+language_support = True
+
+
+def request(query, params):
+    offset = (params['pageno'] - 1) * 8
+    language = 'en-US'
+    if params['language'] != 'all':
+        language = params['language'].replace('_', '-')
+    params['url'] = search_url.format(offset=offset,
+                                      query=urlencode({'q': query}),
+                                      language=language)
+    return params
+
+
+def response(resp):
+    results = []
+    search_res = loads(resp.text)
+
+    if not search_res.get('responseData', {}).get('results'):
+        return []
+
+    for result in search_res['responseData']['results']:
+        results.append({'url': result['unescapedUrl'],
+                        'title': result['titleNoFormatting'],
+                        'content': result['content']})
+    return results