Add support for r.search.yahoo.com

boneyao · Dec 16, 2014 · 9f8310d · 9f8310d
1 parent 72985c5
commit 9f8310d
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 2 deletions.
diff --git a/serpextract/serpextract.py b/serpextract/serpextract.py
@@ -414,15 +414,18 @@ def parse(self, url_parts):
                     keyword = query[extractor][-1]
 
                 # Now we have to check for a tricky case where it is a SERP
-                # but just with no keyword as can be the case with Google
-                # Images or DuckDuckGo
+                # but just with no keyword as can be the case with Google,
+                # DuckDuckGo or Yahoo!
                 if keyword is None and extractor == 'q' and \
                    engine_name in ('Google Images', 'DuckDuckGo'):
                     keyword = ''
                 elif keyword is None and extractor == 'q' and \
                      engine_name == 'Google' and \
                      _is_url_without_path_query_or_fragment(url_parts):
                     keyword = ''
+                elif keyword is None and engine_name == 'Yahoo!' and \
+                     url_parts.netloc.lower() == 'r.search.yahoo.com':
+                    keyword = ''
 
         if keyword is not None:
             return ExtractResult(engine_name, keyword, self)

diff --git a/tests/test_serps.py b/tests/test_serps.py
@@ -72,6 +72,7 @@ def test_yahoo(self):
         serps = (
             ('http://ca.search.yahoo.com/search;_ylt=At9vKXZDJTDsQ6o7bDQPLBUt17V_;_ylc=X1MDMjE0MjYyMzUzMwRfcgMyBGZyA3lmcC10LTcxNQRuX2dwcwMxMARvcmlnaW4DY2EueWFob28uY29tBHF1ZXJ5A2hlbGxvBHNhbwMx?p=hello&toggle=1&cop=mss&ei=UTF-8&fr=yfp-t-715', 'Yahoo!', u'hello'),
             ('http://search.yahoo.com/search;_ylt=AnQcoCW29caK.8RLkGgSiqGbvZx4?p=united+states&toggle=1&cop=mss&ei=UTF-8&fr=yfp-t-900', 'Yahoo!', u'united states'),
+            ('http://r.search.yahoo.com/_ylt=A0LEVy5UeJBUUzgAR2FXNyoA;_ylu=X3oDMTEzaGpsYTZuBHNlYwNzcgRwb3MDMgRjb2xvA2JmMQR2dGlkA1ZJUDU1OF8x/RV=2/RE=1418782933/RO=10/RU=http%3a%2f%2fen.wikipedia.org%2fwiki%2fToronto/RK=0/RS=cUOWJ12k59iqbScMA1r6sQedikc-', 'Yahoo!', u''),
         )
         self.assertValidSERPs(serps)