mirror of
https://github.com/yt-dlp/yt-dlp.git
synced 2024-11-10 09:07:58 +01:00
e3ea479087
They were being caught by YoutubeUserIE, now it only extracts a url if the rest of extractors aren't suitable. Now the url tests check that the urls can only be extracted with an specific extractor.
100 lines
5.1 KiB
Python
100 lines
5.1 KiB
Python
#!/usr/bin/env python
|
|
|
|
import sys
|
|
import unittest
|
|
|
|
# Allow direct execution
|
|
import os
|
|
sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
|
|
|
|
from youtube_dl.extractor import YoutubeIE, YoutubePlaylistIE, YoutubeChannelIE, JustinTVIE, gen_extractors
|
|
from helper import get_testcases
|
|
|
|
class TestAllURLsMatching(unittest.TestCase):
|
|
def setUp(self):
|
|
self.ies = gen_extractors()
|
|
|
|
def matching_ies(self, url):
|
|
return [ie.IE_NAME for ie in self.ies if ie.suitable(url) and ie.IE_NAME != 'generic']
|
|
|
|
def assertMatch(self, url, ie_list):
|
|
self.assertEqual(self.matching_ies(url), ie_list)
|
|
|
|
def test_youtube_playlist_matching(self):
|
|
assertPlaylist = lambda url: self.assertMatch(url, ['youtube:playlist'])
|
|
assertPlaylist(u'ECUl4u3cNGP61MdtwGTqZA0MreSaDybji8')
|
|
assertPlaylist(u'UUBABnxM4Ar9ten8Mdjj1j0Q') #585
|
|
assertPlaylist(u'PL63F0C78739B09958')
|
|
assertPlaylist(u'https://www.youtube.com/playlist?list=UUBABnxM4Ar9ten8Mdjj1j0Q')
|
|
assertPlaylist(u'https://www.youtube.com/course?list=ECUl4u3cNGP61MdtwGTqZA0MreSaDybji8')
|
|
assertPlaylist(u'https://www.youtube.com/playlist?list=PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC')
|
|
assertPlaylist(u'https://www.youtube.com/watch?v=AV6J6_AeFEQ&playnext=1&list=PL4023E734DA416012') #668
|
|
self.assertFalse('youtube:playlist' in self.matching_ies(u'PLtS2H6bU1M'))
|
|
|
|
def test_youtube_matching(self):
|
|
self.assertTrue(YoutubeIE.suitable(u'PLtS2H6bU1M'))
|
|
self.assertFalse(YoutubeIE.suitable(u'https://www.youtube.com/watch?v=AV6J6_AeFEQ&playnext=1&list=PL4023E734DA416012')) #668
|
|
self.assertMatch('http://youtu.be/BaW_jenozKc', ['youtube'])
|
|
self.assertMatch('http://www.youtube.com/v/BaW_jenozKc', ['youtube'])
|
|
|
|
def test_youtube_channel_matching(self):
|
|
assertChannel = lambda url: self.assertMatch(url, ['youtube:channel'])
|
|
assertChannel('https://www.youtube.com/channel/HCtnHdj3df7iM')
|
|
assertChannel('https://www.youtube.com/channel/HCtnHdj3df7iM?feature=gb_ch_rec')
|
|
assertChannel('https://www.youtube.com/channel/HCtnHdj3df7iM/videos')
|
|
|
|
def test_youtube_user_matching(self):
|
|
self.assertMatch('www.youtube.com/NASAgovVideo/videos', ['youtube:user'])
|
|
|
|
def test_youtube_feeds(self):
|
|
self.assertMatch('https://www.youtube.com/feed/watch_later', ['youtube:watch_later'])
|
|
self.assertMatch('https://www.youtube.com/feed/subscriptions', ['youtube:subscriptions'])
|
|
self.assertMatch('https://www.youtube.com/feed/recommended', ['youtube:recommended'])
|
|
self.assertMatch('https://www.youtube.com/my_favorites', ['youtube:favorites'])
|
|
|
|
def test_youtube_show_matching(self):
|
|
self.assertMatch('http://www.youtube.com/show/airdisasters', ['youtube:show'])
|
|
|
|
def test_justin_tv_channelid_matching(self):
|
|
self.assertTrue(JustinTVIE.suitable(u"justin.tv/vanillatv"))
|
|
self.assertTrue(JustinTVIE.suitable(u"twitch.tv/vanillatv"))
|
|
self.assertTrue(JustinTVIE.suitable(u"www.justin.tv/vanillatv"))
|
|
self.assertTrue(JustinTVIE.suitable(u"www.twitch.tv/vanillatv"))
|
|
self.assertTrue(JustinTVIE.suitable(u"http://www.justin.tv/vanillatv"))
|
|
self.assertTrue(JustinTVIE.suitable(u"http://www.twitch.tv/vanillatv"))
|
|
self.assertTrue(JustinTVIE.suitable(u"http://www.justin.tv/vanillatv/"))
|
|
self.assertTrue(JustinTVIE.suitable(u"http://www.twitch.tv/vanillatv/"))
|
|
|
|
def test_justintv_videoid_matching(self):
|
|
self.assertTrue(JustinTVIE.suitable(u"http://www.twitch.tv/vanillatv/b/328087483"))
|
|
|
|
def test_justin_tv_chapterid_matching(self):
|
|
self.assertTrue(JustinTVIE.suitable(u"http://www.twitch.tv/tsm_theoddone/c/2349361"))
|
|
|
|
def test_youtube_extract(self):
|
|
self.assertEqual(YoutubeIE()._extract_id('http://www.youtube.com/watch?&v=BaW_jenozKc'), 'BaW_jenozKc')
|
|
self.assertEqual(YoutubeIE()._extract_id('https://www.youtube.com/watch?&v=BaW_jenozKc'), 'BaW_jenozKc')
|
|
self.assertEqual(YoutubeIE()._extract_id('https://www.youtube.com/watch?feature=player_embedded&v=BaW_jenozKc'), 'BaW_jenozKc')
|
|
self.assertEqual(YoutubeIE()._extract_id('https://www.youtube.com/watch_popup?v=BaW_jenozKc'), 'BaW_jenozKc')
|
|
|
|
def test_no_duplicates(self):
|
|
ies = gen_extractors()
|
|
for tc in get_testcases():
|
|
url = tc['url']
|
|
for ie in ies:
|
|
if type(ie).__name__ in ['GenericIE', tc['name'] + 'IE']:
|
|
self.assertTrue(ie.suitable(url), '%s should match URL %r' % (type(ie).__name__, url))
|
|
else:
|
|
self.assertFalse(ie.suitable(url), '%s should not match URL %r' % (type(ie).__name__, url))
|
|
|
|
def test_keywords(self):
|
|
self.assertMatch(':ytsubs', ['youtube:subscriptions'])
|
|
self.assertMatch(':ytsubscriptions', ['youtube:subscriptions'])
|
|
self.assertMatch(':thedailyshow', ['ComedyCentral'])
|
|
self.assertMatch(':tds', ['ComedyCentral'])
|
|
self.assertMatch(':colbertreport', ['ComedyCentral'])
|
|
self.assertMatch(':cr', ['ComedyCentral'])
|
|
|
|
|
|
if __name__ == '__main__':
|
|
unittest.main()
|