Files
web-scraper/test_helpers.py
2018-09-11 13:42:15 +01:00

37 lines
1.2 KiB
Python

#!/usr/bin/env python
import unittest
from utils.helpers import (sanitise_url)
class TestUrls(unittest.TestCase):
base_url_list = (('eu.httpbin.org', 'http://eu.httpbin.org'),
('www.simonweald.com', 'http://www.simonweald.com'),
('http://www.github.com/', 'http://www.github.com'),
('https://www.github.com', 'https://www.github.com'))
urls_to_clean = (('https://www.github.com/', 'https://www.github.com/'),
('https://github.com/?foo=bar', 'https://github.com/'),
('https://github.com/#anchor', 'https://github.com/'))
def test_sanitise_base_url(self):
'''
Tests whether a URL's protocol can be discovered if not provided.
'''
for url, target in self.base_url_list:
result = sanitise_url(url, base_url=True)
self.assertEqual(result, target)
def test_sanitise_url(self):
'''
Tests whether a URL's protocol can be discovered if not provided.
'''
for url, target in self.urls_to_clean:
result = sanitise_url(url)
self.assertEqual(result, target)
if __name__ == '__main__':
unittest.main()