mirrors/searxng
1
0
Fork 0
mirror of https://github.com/searxng/searxng.git synced 2024-11-20 09:51:00 +00:00
Code Issues Projects Releases Wiki Activity

Yahoo's unit test

Browse source
This commit is contained in:
Cqoicebordel 2015-02-04 20:41:40 +01:00
parent ff2ad57a87
commit 9f13af8d3c
3 changed files with 157 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
searx/engines/yahoo.py
View file

@ -35,7 +35,7 @@ suggestion_xpath = '//div[@id="satat"]//a'
def parse_url(url_string):
endings = ['/RS', '/RK']
endpositions = []
start = url_string.find('http', url_string.find('/RU=')+1)
start = url_string.find('http', url_string.find('/RU=') + 1)
for ending in endings:
endpos = url_string.rfind(ending)
@ -91,7 +91,7 @@ def response(resp):
'content': content})
# if no suggestion found, return results
if not suggestion_xpath:
if not dom.xpath(suggestion_xpath):
return results
# parse suggestion

154
searx/tests/engines/test_yahoo.py Normal file
View file

@ -0,0 +1,154 @@
# -*- coding: utf-8 -*-
from collections import defaultdict
import mock
from searx.engines import yahoo
from searx.testing import SearxTestCase
class TestYahooEngine(SearxTestCase):
def test_parse_url(self):
test_url = 'http://r.search.yahoo.com/_ylt=A0LEb9JUSKcAEGRXNyoA;_ylu=X3oDMTEzZm1qazYwBHNlYwNzcgRwb3MDMQRjb' +\
'2xvA2Jm2dGlkA1NNRTcwM18x/RV=2/RE=1423106085/RO=10/RU=https%3a%2f%2fthis.is.the.url%2f/RK=0/RS=' +\
'dtcJsfP4mEeBOjnVfUQ-'
url = yahoo.parse_url(test_url)
self.assertEqual('https://this.is.the.url/', url)
test_url = 'http://r.search.yahoo.com/_ylt=A0LElb9JUSKcAEGRXNyoA;_ylu=X3oDMTEzZm1qazYwBHNlYwNzcgRwb3MDMQRjb' +\
'2xvA2Jm2dGlkA1NNRTcwM18x/RV=2/RE=1423106085/RO=10/RU=https%3a%2f%2fthis.is.the.url%2f/RS=' +\
'dtcJsfP4mEeBOjnVfUQ-'
url = yahoo.parse_url(test_url)
self.assertEqual('https://this.is.the.url/', url)
test_url = 'https://this.is.the.url/'
url = yahoo.parse_url(test_url)
self.assertEqual('https://this.is.the.url/', url)
def test_request(self):
query = 'test_query'
dicto = defaultdict(dict)
dicto['pageno'] = 1
dicto['language'] = 'fr_FR'
params = yahoo.request(query, dicto)
self.assertIn('url', params)
self.assertIn(query, params['url'])
self.assertIn('search.yahoo.com', params['url'])
self.assertIn('fr', params['url'])
self.assertIn('cookies', params)
self.assertIn('sB', params['cookies'])
self.assertIn('fr', params['cookies']['sB'])
dicto['language'] = 'all'
params = yahoo.request(query, dicto)
self.assertIn('cookies', params)
self.assertIn('sB', params['cookies'])
self.assertIn('en', params['cookies']['sB'])
self.assertIn('en', params['url'])
def test_response(self):
self.assertRaises(AttributeError, yahoo.response, None)
self.assertRaises(AttributeError, yahoo.response, [])
self.assertRaises(AttributeError, yahoo.response, '')
self.assertRaises(AttributeError, yahoo.response, '[]')
response = mock.Mock(text='<html></html>')
self.assertEqual(yahoo.response(response), [])
html = """
<div class="res">
<div>
<h3>
<a id="link-1" class="yschttl spt" href="http://r.search.yahoo.com/_ylt=A0LEVzClb9JUSKcAEGRXNyoA;
_ylu=X3oDMTEzZm1qazYwBHNlYwNzcgRwb3MDMQRjb2xvA2JmMQR2dGlkA1NNRTcwM18x/RV=2/RE=1423106085/RO=10
/RU=https%3a%2f%2fthis.is.the.url%2f/RK=0/RS=dtcJsfP4mEeBOjnVfUQ-"target="_blank" data-bk="5063.1">
<b>This</b> is the title
</a>
</h3>
</div>
<span class="url" dir="ltr">www.<b>test</b>.com</span>
<div class="abstr">
<b>This</b> is the content
</div>
</div>
<div id="satat" data-bns="Yahoo" data-bk="124.1">
<h2>Also Try</h2>
<table>
<tbody>
<tr>
<td>
<a id="srpnat0" class="" href="https://search.yahoo.com/search=rs-bottom" >
<span>
<b></b>This is <b>the suggestion</b>
</span>
</a>
</td>
</tr>
</tbody>
</table>
</div>
"""
response = mock.Mock(text=html)
results = yahoo.response(response)
self.assertEqual(type(results), list)
self.assertEqual(len(results), 2)
self.assertEqual(results[0]['title'], 'This is the title')
self.assertEqual(results[0]['url'], 'https://this.is.the.url/')
self.assertEqual(results[0]['content'], 'This is the content')
self.assertEqual(results[1]['suggestion'], 'This is the suggestion')
html = """
<div class="res">
<div>
<h3>
<a id="link-1" class="yschttl spt" href="http://r.search.yahoo.com/_ylt=A0LEVzClb9JUSKcAEGRXNyoA;
_ylu=X3oDMTEzZm1qazYwBHNlYwNzcgRwb3MDMQRjb2xvA2JmMQR2dGlkA1NNRTcwM18x/RV=2/RE=1423106085/RO=10
/RU=https%3a%2f%2fthis.is.the.url%2f/RK=0/RS=dtcJsfP4mEeBOjnVfUQ-"target="_blank" data-bk="5063.1">
<b>This</b> is the title
</a>
</h3>
</div>
<span class="url" dir="ltr">www.<b>test</b>.com</span>
<div class="abstr">
<b>This</b> is the content
</div>
</div>
<div class="res">
<div>
<h3>
<a id="link-1" class="yschttl spt">
<b>This</b> is the title
</a>
</h3>
</div>
<span class="url" dir="ltr">www.<b>test</b>.com</span>
<div class="abstr">
<b>This</b> is the content
</div>
</div>
<div class="res">
<div>
<h3>
</h3>
</div>
<span class="url" dir="ltr">www.<b>test</b>.com</span>
<div class="abstr">
<b>This</b> is the content
</div>
</div>
"""
response = mock.Mock(text=html)
results = yahoo.response(response)
self.assertEqual(type(results), list)
self.assertEqual(len(results), 1)
self.assertEqual(results[0]['title'], 'This is the title')
self.assertEqual(results[0]['url'], 'https://this.is.the.url/')
self.assertEqual(results[0]['content'], 'This is the content')
html = """
<li class="b_algo" u="0|5109|4755453613245655|UAGjXgIrPH5yh-o5oNHRx_3Zta87f_QO">
</li>
"""
response = mock.Mock(text=html)
results = yahoo.response(response)
self.assertEqual(type(results), list)
self.assertEqual(len(results), 0)

1
searx/tests/test_engines.py
View file

@ -27,4 +27,5 @@ from searx.tests.engines.test_twitter import * # noqa
from searx.tests.engines.test_vimeo import * # noqa
from searx.tests.engines.test_www500px import * # noqa
from searx.tests.engines.test_youtube import * # noqa
from searx.tests.engines.test_yahoo import * # noqa
from searx.tests.engines.test_yahoo_news import * # noqa