diff options
| -rw-r--r-- | searx/engines/www1x.py | 3 | ||||
| -rw-r--r-- | searx/tests/engines/test_www1x.py | 57 | ||||
| -rw-r--r-- | searx/tests/test_engines.py | 1 |
3 files changed, 60 insertions, 1 deletions
diff --git a/searx/engines/www1x.py b/searx/engines/www1x.py index d10c4ca37..a68c105ce 100644 --- a/searx/engines/www1x.py +++ b/searx/engines/www1x.py @@ -36,7 +36,8 @@ def response(resp): results = [] # get links from result-text - results_parts = re.split(r'(</a>|<a)', resp.text) + regex = re.compile('(</a>|<a)') + results_parts = re.split(regex, resp.text) cur_element = '' diff --git a/searx/tests/engines/test_www1x.py b/searx/tests/engines/test_www1x.py new file mode 100644 index 000000000..ab4f282c1 --- /dev/null +++ b/searx/tests/engines/test_www1x.py @@ -0,0 +1,57 @@ +from collections import defaultdict +import mock +from searx.engines import www1x +from searx.testing import SearxTestCase + + +class TestWww1xEngine(SearxTestCase): + + def test_request(self): + query = 'test_query' + params = www1x.request(query, defaultdict(dict)) + self.assertTrue('url' in params) + self.assertTrue(query in params['url']) + self.assertTrue('1x.com' in params['url']) + + def test_response(self): + self.assertRaises(AttributeError, www1x.response, None) + self.assertRaises(AttributeError, www1x.response, []) + self.assertRaises(AttributeError, www1x.response, '') + self.assertRaises(AttributeError, www1x.response, '[]') + + response = mock.Mock(text='<html></html>') + self.assertEqual(www1x.response(response), []) + html = """ + <?xml version="1.0" encoding="UTF-8"?><!DOCTYPE characters + [ + <!ELEMENT characters (character*) > + <!ELEMENT character (#PCDATA ) > + + <!ENTITY iexcl "¡" > + <!ENTITY cent "¢" > + <!ENTITY pound "£" > + ] + ><root><searchresult><![CDATA[<table border="0" cellpadding="0" cellspacing="0" width="100%"> + <tr> + <td style="min-width: 220px;" valign="top"> + <div style="font-size: 30px; margin: 0px 0px 20px 0px;">Photos</div> + <div> + <a href="/photo/123456" class="dynamiclink"> +<img border="0" class="searchresult" src="/images/user/testimage-123456.jpg" style="width: 125px; height: 120px;"> + </a> + <a title="sjoerd lammers street photography" href="/member/sjoerdlammers" class="dynamiclink"> +<img border="0" class="searchresult" src="/images/profile/60c48b394c677d2fa4d9e7d263aabf44-square.jpg"> + </a> + </div> + </td> + </table> + ]]></searchresult></root> + """ + response = mock.Mock(text=html) + results = www1x.response(response) + self.assertEqual(type(results), list) + self.assertEqual(len(results), 1) + self.assertEqual(results[0]['url'], 'http://1x.com/photo/123456') + self.assertEqual(results[0]['thumbnail_src'], 'http://1x.com/images/user/testimage-123456.jpg') + self.assertEqual(results[0]['content'], '') + self.assertEqual(results[0]['template'], 'images.html') diff --git a/searx/tests/test_engines.py b/searx/tests/test_engines.py index 1ffdbe529..cfd7fa26a 100644 --- a/searx/tests/test_engines.py +++ b/searx/tests/test_engines.py @@ -1,2 +1,3 @@ from searx.tests.engines.test_dummy import * # noqa from searx.tests.engines.test_github import * # noqa +from searx.tests.engines.test_www1x import * # noqa |