mirror of https://github.com/searx/searx
Stackoverflow's unit test
This commit is contained in:
Cqoicebordel
parent
787fee6a09
commit
d20ddf9da1
|
|
@ -12,6 +12,7 @@ from urlparse import urljoin
|
||||||
from cgi import escape
|
from cgi import escape
|
||||||
from urllib import urlencode
|
from urllib import urlencode
|
||||||
from lxml import html
|
from lxml import html
|
||||||
|
from searx.engines.xpath import extract_text
|
||||||
|
|
||||||
# engine dependent config
|
# engine dependent config
|
||||||
categories = ['it']
|
categories = ['it']
|
||||||
|
|
@ -24,8 +25,7 @@ search_url = url+'search?{query}&page={pageno}'
|
||||||
# specific xpath variables
|
# specific xpath variables
|
||||||
results_xpath = '//div[contains(@class,"question-summary")]'
|
results_xpath = '//div[contains(@class,"question-summary")]'
|
||||||
link_xpath = './/div[@class="result-link"]//a|.//div[@class="summary"]//h3//a'
|
link_xpath = './/div[@class="result-link"]//a|.//div[@class="summary"]//h3//a'
|
||||||
title_xpath = './/text()'
|
content_xpath = './/div[@class="excerpt"]'
|
||||||
content_xpath = './/div[@class="excerpt"]//text()'
|
|
||||||
|
|
||||||
|
|
||||||
# do search-request
|
# do search-request
|
||||||
|
|
@ -46,8 +46,8 @@ def response(resp):
|
||||||
for result in dom.xpath(results_xpath):
|
for result in dom.xpath(results_xpath):
|
||||||
link = result.xpath(link_xpath)[0]
|
link = result.xpath(link_xpath)[0]
|
||||||
href = urljoin(url, link.attrib.get('href'))
|
href = urljoin(url, link.attrib.get('href'))
|
||||||
title = escape(' '.join(link.xpath(title_xpath)))
|
title = escape(extract_text(link))
|
||||||
content = escape(' '.join(result.xpath(content_xpath)))
|
content = escape(extract_text(result.xpath(content_xpath)))
|
||||||
|
|
||||||
# append result
|
# append result
|
||||||
results.append({'url': href,
|
results.append({'url': href,
|
||||||
|
|
|
||||||
|
|
@ -0,0 +1,106 @@
|
||||||
|
from collections import defaultdict
|
||||||
|
import mock
|
||||||
|
from searx.engines import stackoverflow
|
||||||
|
from searx.testing import SearxTestCase
|
||||||
|
|
||||||
|
|
||||||
|
class TestStackoverflowEngine(SearxTestCase):
|
||||||
|
|
||||||
|
def test_request(self):
|
||||||
|
query = 'test_query'
|
||||||
|
dicto = defaultdict(dict)
|
||||||
|
dicto['pageno'] = 0
|
||||||
|
params = stackoverflow.request(query, dicto)
|
||||||
|
self.assertTrue('url' in params)
|
||||||
|
self.assertTrue(query in params['url'])
|
||||||
|
self.assertTrue('stackoverflow.com' in params['url'])
|
||||||
|
|
||||||
|
def test_response(self):
|
||||||
|
self.assertRaises(AttributeError, stackoverflow.response, None)
|
||||||
|
self.assertRaises(AttributeError, stackoverflow.response, [])
|
||||||
|
self.assertRaises(AttributeError, stackoverflow.response, '')
|
||||||
|
self.assertRaises(AttributeError, stackoverflow.response, '[]')
|
||||||
|
|
||||||
|
response = mock.Mock(text='<html></html>')
|
||||||
|
self.assertEqual(stackoverflow.response(response), [])
|
||||||
|
|
||||||
|
html = """
|
||||||
|
<div class="question-summary search-result" id="answer-id-1783426">
|
||||||
|
<div class="statscontainer">
|
||||||
|
<div class="statsarrow"></div>
|
||||||
|
<div class="stats">
|
||||||
|
<div class="vote">
|
||||||
|
<div class="votes answered">
|
||||||
|
<span class="vote-count-post "><strong>2583</strong></span>
|
||||||
|
<div class="viewcount">votes</div>
|
||||||
|
</div>
|
||||||
|
</div>
|
||||||
|
</div>
|
||||||
|
</div>
|
||||||
|
<div class="summary">
|
||||||
|
<div class="result-link">
|
||||||
|
<span>
|
||||||
|
<a href="/questions/this.is.the.url"
|
||||||
|
data-searchsession="/questions"
|
||||||
|
title="Checkout remote Git branch">
|
||||||
|
This is the title
|
||||||
|
</a>
|
||||||
|
</span>
|
||||||
|
</div>
|
||||||
|
<div class="excerpt">
|
||||||
|
This is the content
|
||||||
|
</div>
|
||||||
|
<div class="tags user-tags t-git t-git-checkout t-remote-branch">
|
||||||
|
</div>
|
||||||
|
<div class="started fr">
|
||||||
|
answered <span title="2009-11-23 14:26:08Z" class="relativetime">nov 23 '09</span> by
|
||||||
|
<a href="/users/214090/hallski">hallski</a>
|
||||||
|
</div>
|
||||||
|
</div>
|
||||||
|
</div>
|
||||||
|
"""
|
||||||
|
response = mock.Mock(text=html)
|
||||||
|
results = stackoverflow.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'], 'http://stackoverflow.com/questions/this.is.the.url')
|
||||||
|
self.assertEqual(results[0]['content'], 'This is the content')
|
||||||
|
|
||||||
|
html = """
|
||||||
|
<div class="statscontainer">
|
||||||
|
<div class="statsarrow"></div>
|
||||||
|
<div class="stats">
|
||||||
|
<div class="vote">
|
||||||
|
<div class="votes answered">
|
||||||
|
<span class="vote-count-post "><strong>2583</strong></span>
|
||||||
|
<div class="viewcount">votes</div>
|
||||||
|
</div>
|
||||||
|
</div>
|
||||||
|
</div>
|
||||||
|
</div>
|
||||||
|
<div class="summary">
|
||||||
|
<div class="result-link">
|
||||||
|
<span>
|
||||||
|
<a href="/questions/this.is.the.url"
|
||||||
|
data-searchsession="/questions"
|
||||||
|
title="Checkout remote Git branch">
|
||||||
|
This is the title
|
||||||
|
</a>
|
||||||
|
</span>
|
||||||
|
</div>
|
||||||
|
<div class="excerpt">
|
||||||
|
This is the content
|
||||||
|
</div>
|
||||||
|
<div class="tags user-tags t-git t-git-checkout t-remote-branch">
|
||||||
|
</div>
|
||||||
|
<div class="started fr">
|
||||||
|
answered <span title="2009-11-23 14:26:08Z" class="relativetime">nov 23 '09</span> by
|
||||||
|
<a href="/users/214090/hallski">hallski</a>
|
||||||
|
</div>
|
||||||
|
</div>
|
||||||
|
"""
|
||||||
|
response = mock.Mock(text=html)
|
||||||
|
results = stackoverflow.response(response)
|
||||||
|
self.assertEqual(type(results), list)
|
||||||
|
self.assertEqual(len(results), 0)
|
||||||
|
|
@ -16,4 +16,5 @@ from searx.tests.engines.test_mixcloud import * # noqa
|
||||||
from searx.tests.engines.test_searchcode_code import * # noqa
|
from searx.tests.engines.test_searchcode_code import * # noqa
|
||||||
from searx.tests.engines.test_searchcode_doc import * # noqa
|
from searx.tests.engines.test_searchcode_doc import * # noqa
|
||||||
from searx.tests.engines.test_soundcloud import * # noqa
|
from searx.tests.engines.test_soundcloud import * # noqa
|
||||||
|
from searx.tests.engines.test_stackoverflow import * # noqa
|
||||||
from searx.tests.engines.test_youtube import * # noqa
|
from searx.tests.engines.test_youtube import * # noqa
|
||||||
|
|
|
||||||
Loading…
Reference in New Issue