Browse Source

[fix] prevent google engine to redirect

nid/pref cookies are also removed
Adam Tauber 9 years ago
parent
commit
5cea4f9445
2 changed files with 2 additions and 34 deletions
  1. 1
    29
      searx/engines/google.py
  2. 1
    5
      searx/tests/engines/test_google.py

+ 1
- 29
searx/engines/google.py View File

@@ -13,7 +13,6 @@ from cgi import escape
13 13
 from urllib import urlencode
14 14
 from urlparse import urlparse, parse_qsl
15 15
 from lxml import html, etree
16
-from searx.poolrequests import get
17 16
 from searx.engines.xpath import extract_text, extract_url
18 17
 from searx.search import logger
19 18
 
@@ -91,7 +90,7 @@ url_map = 'https://www.openstreetmap.org/'\
91 90
 search_path = '/search'
92 91
 search_url = ('https://{hostname}' +
93 92
               search_path +
94
-              '?{query}&start={offset}&gbv=1')
93
+              '?{query}&start={offset}&gbv=1&gws_rd=cr')
95 94
 
96 95
 # other URLs
97 96
 map_hostname_start = 'maps.google.'
@@ -129,27 +128,6 @@ image_img_src_xpath = './img/@src'
129 128
 property_address = "Address"
130 129
 property_phone = "Phone number"
131 130
 
132
-# cookies
133
-pref_cookie = ''
134
-nid_cookie = {}
135
-
136
-
137
-# see https://support.google.com/websearch/answer/873?hl=en
138
-def get_google_pref_cookie():
139
-    global pref_cookie
140
-    if pref_cookie == '':
141
-        resp = get('https://www.google.com/ncr', allow_redirects=False)
142
-        pref_cookie = resp.cookies["PREF"]
143
-    return pref_cookie
144
-
145
-
146
-def get_google_nid_cookie(google_hostname):
147
-    global nid_cookie
148
-    if google_hostname not in nid_cookie:
149
-        resp = get('https://' + google_hostname)
150
-        nid_cookie[google_hostname] = resp.cookies.get("NID", None)
151
-    return nid_cookie[google_hostname]
152
-
153 131
 
154 132
 # remove google-specific tracking-url
155 133
 def parse_url(url_string, google_hostname):
@@ -201,12 +179,6 @@ def request(query, params):
201 179
 
202 180
     params['headers']['Accept-Language'] = language
203 181
     params['headers']['Accept'] = 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8'
204
-    if google_hostname == default_hostname:
205
-        try:
206
-            params['cookies']['PREF'] = get_google_pref_cookie()
207
-        except:
208
-            logger.warning('cannot fetch PREF cookie')
209
-    params['cookies']['NID'] = get_google_nid_cookie(google_hostname)
210 182
 
211 183
     params['google_hostname'] = google_hostname
212 184
 

+ 1
- 5
searx/tests/engines/test_google.py View File

@@ -9,7 +9,7 @@ from searx.testing import SearxTestCase
9 9
 class TestGoogleEngine(SearxTestCase):
10 10
 
11 11
     def mock_response(self, text):
12
-        response = mock.Mock(text=text, url='https://www.google.com/search?q=test&start=0&gbv=1')
12
+        response = mock.Mock(text=text, url='https://www.google.com/search?q=test&start=0&gbv=1&gws_rd=cr')
13 13
         response.search_params = mock.Mock()
14 14
         response.search_params.get = mock.Mock(return_value='www.google.com')
15 15
         return response
@@ -23,16 +23,12 @@ class TestGoogleEngine(SearxTestCase):
23 23
         self.assertIn('url', params)
24 24
         self.assertIn(query, params['url'])
25 25
         self.assertIn('google.fr', params['url'])
26
-        self.assertNotIn('PREF', params['cookies'])
27
-        self.assertIn('NID', params['cookies'])
28 26
         self.assertIn('fr', params['headers']['Accept-Language'])
29 27
 
30 28
         dicto['language'] = 'all'
31 29
         params = google.request(query, dicto)
32 30
         self.assertIn('google.com', params['url'])
33 31
         self.assertIn('en', params['headers']['Accept-Language'])
34
-        # self.assertIn('PREF', params['cookies'])
35
-        self.assertIn('NID', params['cookies'])
36 32
 
37 33
     def test_response(self):
38 34
         self.assertRaises(AttributeError, google.response, None)