]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - test/test_all_urls.py
debian/control: Update list of extractors/supported sites.
[youtubedl] / test / test_all_urls.py
index dd67286a79a6071317a4a6cc8bb6b4497a1c171a..465ce005097d246d6b77ae8c04b47fa94889dc41 100644 (file)
 #!/usr/bin/env python
 
-import sys
-import unittest
+from __future__ import unicode_literals
 
 # Allow direct execution
 import os
-sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+import sys
+import unittest
+import collections
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+
+from test.helper import gettestcases
+
+from youtube_dl.extractor import (
+    FacebookIE,
+    gen_extractors,
+    YoutubeIE,
+)
 
-from youtube_dl.InfoExtractors import YoutubeIE, YoutubePlaylistIE, YoutubeChannelIE, JustinTVIE
 
 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):
-        self.assertTrue(YoutubePlaylistIE.suitable(u'ECUl4u3cNGP61MdtwGTqZA0MreSaDybji8'))
-        self.assertTrue(YoutubePlaylistIE.suitable(u'UUBABnxM4Ar9ten8Mdjj1j0Q')) #585
-        self.assertTrue(YoutubePlaylistIE.suitable(u'PL63F0C78739B09958'))
-        self.assertTrue(YoutubePlaylistIE.suitable(u'https://www.youtube.com/playlist?list=UUBABnxM4Ar9ten8Mdjj1j0Q'))
-        self.assertTrue(YoutubePlaylistIE.suitable(u'https://www.youtube.com/course?list=ECUl4u3cNGP61MdtwGTqZA0MreSaDybji8'))
-        self.assertTrue(YoutubePlaylistIE.suitable(u'https://www.youtube.com/playlist?list=PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC'))
-        self.assertTrue(YoutubePlaylistIE.suitable(u'https://www.youtube.com/watch?v=AV6J6_AeFEQ&playnext=1&list=PL4023E734DA416012')) #668
-        self.assertFalse(YoutubePlaylistIE.suitable(u'PLtS2H6bU1M'))
+        assertPlaylist = lambda url: self.assertMatch(url, ['youtube:playlist'])
+        assertPlaylist('ECUl4u3cNGP61MdtwGTqZA0MreSaDybji8')
+        assertPlaylist('UUBABnxM4Ar9ten8Mdjj1j0Q')  # 585
+        assertPlaylist('PL63F0C78739B09958')
+        assertPlaylist('https://www.youtube.com/playlist?list=UUBABnxM4Ar9ten8Mdjj1j0Q')
+        assertPlaylist('https://www.youtube.com/course?list=ECUl4u3cNGP61MdtwGTqZA0MreSaDybji8')
+        assertPlaylist('https://www.youtube.com/playlist?list=PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC')
+        assertPlaylist('https://www.youtube.com/watch?v=AV6J6_AeFEQ&playnext=1&list=PL4023E734DA416012')  # 668
+        self.assertFalse('youtube:playlist' in self.matching_ies('PLtS2H6bU1M'))
+        # Top tracks
+        assertPlaylist('https://www.youtube.com/playlist?list=MCUS.20142101')
 
     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.assertTrue(YoutubeIE.suitable('PLtS2H6bU1M'))
+        self.assertFalse(YoutubeIE.suitable('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'])
+        self.assertMatch('https://youtube.googleapis.com/v/BaW_jenozKc', ['youtube'])
+        self.assertMatch('http://www.cleanvideosearch.com/media/action/yt/watch?videoId=8v_4O44sfjM', ['youtube'])
 
     def test_youtube_channel_matching(self):
-        self.assertTrue(YoutubeChannelIE.suitable('https://www.youtube.com/channel/HCtnHdj3df7iM'))
-        self.assertTrue(YoutubeChannelIE.suitable('https://www.youtube.com/channel/HCtnHdj3df7iM?feature=gb_ch_rec'))
-        self.assertTrue(YoutubeChannelIE.suitable('https://www.youtube.com/channel/HCtnHdj3df7iM/videos'))
-
-    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"))
+        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('http://www.youtube.com/NASAgovVideo/videos', ['youtube:user'])
+
+    def test_youtube_feeds(self):
+        self.assertMatch('https://www.youtube.com/feed/watch_later', ['youtube:watchlater'])
+        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_youtube_search_matching(self):
+        self.assertMatch('http://www.youtube.com/results?search_query=making+mustard', ['youtube:search_url'])
+        self.assertMatch('https://www.youtube.com/results?baz=bar&search_query=youtube-dl+test+video&filters=video&lclk=video', ['youtube:search_url'])
 
     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')
+        assertExtractId = lambda url, id: self.assertEqual(YoutubeIE.extract_id(url), id)
+        assertExtractId('http://www.youtube.com/watch?&v=BaW_jenozKc', 'BaW_jenozKc')
+        assertExtractId('https://www.youtube.com/watch?&v=BaW_jenozKc', 'BaW_jenozKc')
+        assertExtractId('https://www.youtube.com/watch?feature=player_embedded&v=BaW_jenozKc', 'BaW_jenozKc')
+        assertExtractId('https://www.youtube.com/watch_popup?v=BaW_jenozKc', 'BaW_jenozKc')
+        assertExtractId('http://www.youtube.com/watch?v=BaW_jenozKcsharePLED17F32AD9753930', 'BaW_jenozKc')
+        assertExtractId('BaW_jenozKc', 'BaW_jenozKc')
+
+    def test_facebook_matching(self):
+        self.assertTrue(FacebookIE.suitable('https://www.facebook.com/Shiniknoh#!/photo.php?v=10153317450565268'))
+        self.assertTrue(FacebookIE.suitable('https://www.facebook.com/cindyweather?fref=ts#!/photo.php?v=10152183998945793'))
+
+    def test_no_duplicates(self):
+        ies = gen_extractors()
+        for tc in gettestcases(include_onlymatching=True):
+            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 . That URL belongs to %s.' % (type(ie).__name__, url, tc['name']))
+
+    def test_keywords(self):
+        self.assertMatch(':ytsubs', ['youtube:subscriptions'])
+        self.assertMatch(':ytsubscriptions', ['youtube:subscriptions'])
+        self.assertMatch(':ythistory', ['youtube:history'])
+
+    def test_vimeo_matching(self):
+        self.assertMatch('https://vimeo.com/channels/tributes', ['vimeo:channel'])
+        self.assertMatch('https://vimeo.com/channels/31259', ['vimeo:channel'])
+        self.assertMatch('https://vimeo.com/channels/31259/53576664', ['vimeo'])
+        self.assertMatch('https://vimeo.com/user7108434', ['vimeo:user'])
+        self.assertMatch('https://vimeo.com/user7108434/videos', ['vimeo:user'])
+        self.assertMatch('https://vimeo.com/user21297594/review/75524534/3c257a1b5d', ['vimeo:review'])
+
+    # https://github.com/ytdl-org/youtube-dl/issues/1930
+    def test_soundcloud_not_matching_sets(self):
+        self.assertMatch('http://soundcloud.com/floex/sets/gone-ep', ['soundcloud:set'])
+
+    def test_tumblr(self):
+        self.assertMatch('http://tatianamaslanydaily.tumblr.com/post/54196191430/orphan-black-dvd-extra-behind-the-scenes', ['Tumblr'])
+        self.assertMatch('http://tatianamaslanydaily.tumblr.com/post/54196191430', ['Tumblr'])
+
+    def test_pbs(self):
+        # https://github.com/ytdl-org/youtube-dl/issues/2350
+        self.assertMatch('http://video.pbs.org/viralplayer/2365173446/', ['pbs'])
+        self.assertMatch('http://video.pbs.org/widget/partnerplayer/980042464/', ['pbs'])
+
+    def test_yahoo_https(self):
+        # https://github.com/ytdl-org/youtube-dl/issues/2701
+        self.assertMatch(
+            'https://screen.yahoo.com/smartwatches-latest-wearable-gadgets-163745379-cbs.html',
+            ['Yahoo'])
+
+    def test_no_duplicated_ie_names(self):
+        name_accu = collections.defaultdict(list)
+        for ie in self.ies:
+            name_accu[ie.IE_NAME.lower()].append(type(ie).__name__)
+        for (ie_name, ie_list) in name_accu.items():
+            self.assertEqual(
+                len(ie_list), 1,
+                'Multiple extractors with the same IE_NAME "%s" (%s)' % (ie_name, ', '.join(ie_list)))
+
 
 if __name__ == '__main__':
     unittest.main()