mirror of
https://github.com/netzbegruenung/green-spider.git
synced 2024-05-06 02:43:42 +02:00
Add basic test for check_content
This commit is contained in:
parent
9bea186008
commit
e09bf7a4d4
35
test.py
35
test.py
|
@ -1,4 +1,6 @@
|
||||||
import unittest
|
import unittest
|
||||||
|
import requests
|
||||||
|
import responses
|
||||||
import spider
|
import spider
|
||||||
|
|
||||||
class TestSpider(unittest.TestCase):
|
class TestSpider(unittest.TestCase):
|
||||||
|
@ -24,6 +26,39 @@ class TestSpider(unittest.TestCase):
|
||||||
result = spider.reduce_urls(testdata)
|
result = spider.reduce_urls(testdata)
|
||||||
self.assertEqual(result, expected_result)
|
self.assertEqual(result, expected_result)
|
||||||
|
|
||||||
|
@responses.activate
|
||||||
|
def test_check_content1(self):
|
||||||
|
"""
|
||||||
|
Very basic test of our content analysis function
|
||||||
|
"""
|
||||||
|
url = 'http://my.url'
|
||||||
|
responses.add(responses.GET, url, status=200,
|
||||||
|
content_type='text/html',
|
||||||
|
body='''
|
||||||
|
<!DOCTYPE html>
|
||||||
|
<html>
|
||||||
|
<head>
|
||||||
|
<meta charset="UTF-8">
|
||||||
|
<title>The title</title>
|
||||||
|
</head>
|
||||||
|
</html>
|
||||||
|
''')
|
||||||
|
r = requests.get(url)
|
||||||
|
result = spider.check_content(r)
|
||||||
|
|
||||||
|
del result['html'] # don't want to have the messy HTML part in comparison
|
||||||
|
|
||||||
|
expected_result = {
|
||||||
|
'icon': None,
|
||||||
|
'title': 'The title',
|
||||||
|
'generator': None,
|
||||||
|
'feeds': [],
|
||||||
|
'encoding': 'ISO-8859-1',
|
||||||
|
'canonical_link': None,
|
||||||
|
'opengraph': None
|
||||||
|
}
|
||||||
|
self.assertDictEqual(result, expected_result)
|
||||||
|
|
||||||
|
|
||||||
if __name__ == '__main__':
|
if __name__ == '__main__':
|
||||||
unittest.main()
|
unittest.main()
|
||||||
|
|
Loading…
Reference in a new issue