[youtube:playlist] Don't use the gdata api (closes #1508)
authorJaime Marquínez Ferrándiz <jaime.marquinez.ferrandiz@gmail.com>
Wed, 13 Nov 2013 15:21:24 +0000 (16:21 +0100)
committerJaime Marquínez Ferrándiz <jaime.marquinez.ferrandiz@gmail.com>
Wed, 13 Nov 2013 15:26:50 +0000 (16:26 +0100)
Parse the playlist pages instead

test/test_youtube_lists.py
youtube_dl/extractor/youtube.py

index 4b7a7847bd3a33a9a2bff3e99f9f4cff0de7eebf..50ad52695e37bd7cb4a950fc1da5d2e65790dc6e 100644 (file)
@@ -27,7 +27,7 @@ class TestYoutubeLists(unittest.TestCase):
     def test_youtube_playlist(self):
         dl = FakeYDL()
         ie = YoutubePlaylistIE(dl)
-        result = ie.extract('https://www.youtube.com/playlist?list=PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re')[0]
+        result = ie.extract('https://www.youtube.com/playlist?list=PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re')
         self.assertIsPlaylist(result)
         self.assertEqual(result['title'], 'ytdl test PL')
         ytie_results = [YoutubeIE()._extract_id(url['url']) for url in result['entries']]
@@ -44,13 +44,13 @@ class TestYoutubeLists(unittest.TestCase):
     def test_issue_673(self):
         dl = FakeYDL()
         ie = YoutubePlaylistIE(dl)
-        result = ie.extract('PLBB231211A4F62143')[0]
+        result = ie.extract('PLBB231211A4F62143')
         self.assertTrue(len(result['entries']) > 25)
 
     def test_youtube_playlist_long(self):
         dl = FakeYDL()
         ie = YoutubePlaylistIE(dl)
-        result = ie.extract('https://www.youtube.com/playlist?list=UUBABnxM4Ar9ten8Mdjj1j0Q')[0]
+        result = ie.extract('https://www.youtube.com/playlist?list=UUBABnxM4Ar9ten8Mdjj1j0Q')
         self.assertIsPlaylist(result)
         self.assertTrue(len(result['entries']) >= 799)
 
@@ -58,7 +58,7 @@ class TestYoutubeLists(unittest.TestCase):
         #651
         dl = FakeYDL()
         ie = YoutubePlaylistIE(dl)
-        result = ie.extract('https://www.youtube.com/playlist?list=PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC')[0]
+        result = ie.extract('https://www.youtube.com/playlist?list=PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC')
         ytie_results = [YoutubeIE()._extract_id(url['url']) for url in result['entries']]
         self.assertFalse('pElCt5oNDuI' in ytie_results)
         self.assertFalse('KdPEApIVdWM' in ytie_results)
@@ -66,7 +66,7 @@ class TestYoutubeLists(unittest.TestCase):
     def test_youtube_playlist_empty(self):
         dl = FakeYDL()
         ie = YoutubePlaylistIE(dl)
-        result = ie.extract('https://www.youtube.com/playlist?list=PLtPgu7CB4gbZDA7i_euNxn75ISqxwZPYx')[0]
+        result = ie.extract('https://www.youtube.com/playlist?list=PLtPgu7CB4gbZDA7i_euNxn75ISqxwZPYx')
         self.assertIsPlaylist(result)
         self.assertEqual(len(result['entries']), 0)
 
@@ -74,7 +74,7 @@ class TestYoutubeLists(unittest.TestCase):
         dl = FakeYDL()
         ie = YoutubePlaylistIE(dl)
         # TODO find a > 100 (paginating?) videos course
-        result = ie.extract('https://www.youtube.com/course?list=ECUl4u3cNGP61MdtwGTqZA0MreSaDybji8')[0]
+        result = ie.extract('https://www.youtube.com/course?list=ECUl4u3cNGP61MdtwGTqZA0MreSaDybji8')
         entries = result['entries']
         self.assertEqual(YoutubeIE()._extract_id(entries[0]['url']), 'j9WZyLZCBzs')
         self.assertEqual(len(entries), 25)
@@ -99,7 +99,7 @@ class TestYoutubeLists(unittest.TestCase):
     def test_youtube_safe_search(self):
         dl = FakeYDL()
         ie = YoutubePlaylistIE(dl)
-        result = ie.extract('PLtPgu7CB4gbY9oDN3drwC3cMbJggS7dKl')[0]
+        result = ie.extract('PLtPgu7CB4gbY9oDN3drwC3cMbJggS7dKl')
         self.assertEqual(len(result['entries']), 2)
 
     def test_youtube_show(self):
index c992cba978441081a50e5f134a4a694c83a6685c..d97ea8c8368dda3af8bd3c23244501af5c12f326 100644 (file)
@@ -1506,8 +1506,9 @@ class YoutubePlaylistIE(InfoExtractor):
                      |
                         ((?:PL|EC|UU|FL)[0-9A-Za-z-_]{10,})
                      )"""
-    _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/playlists/%s?max-results=%i&start-index=%i&v=2&alt=json&safeSearch=none'
-    _MAX_RESULTS = 50
+    _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s&page=%s'
+    _MORE_PAGES_INDICATOR = r'data-link-type="next"'
+    _VIDEO_RE = r'href="/watch\?v=([0-9A-Za-z_-]{11})&amp;'
     IE_NAME = u'youtube:playlist'
 
     @classmethod
@@ -1532,41 +1533,23 @@ class YoutubePlaylistIE(InfoExtractor):
             else:
                 self.to_screen(u'Downloading playlist PL%s - add --no-playlist to just download video %s' % (playlist_id, video_id))
 
-        # Download playlist videos from API
-        videos = []
+        # Extract the video ids from the playlist pages
+        ids = []
 
         for page_num in itertools.count(1):
-            start_index = self._MAX_RESULTS * (page_num - 1) + 1
-            if start_index >= 1000:
-                self._downloader.report_warning(u'Max number of results reached')
-                break
-            url = self._TEMPLATE_URL % (playlist_id, self._MAX_RESULTS, start_index)
+            url = self._TEMPLATE_URL % (playlist_id, page_num)
             page = self._download_webpage(url, playlist_id, u'Downloading page #%s' % page_num)
+            # The ids are duplicated
+            new_ids = orderedSet(re.findall(self._VIDEO_RE, page))
+            ids.extend(new_ids)
 
-            try:
-                response = json.loads(page)
-            except ValueError as err:
-                raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
-
-            if 'feed' not in response:
-                raise ExtractorError(u'Got a malformed response from YouTube API')
-            playlist_title = response['feed']['title']['$t']
-            if 'entry' not in response['feed']:
-                # Number of videos is a multiple of self._MAX_RESULTS
+            if re.search(self._MORE_PAGES_INDICATOR, page) is None:
                 break
 
-            for entry in response['feed']['entry']:
-                index = entry['yt$position']['$t']
-                if 'media$group' in entry and 'yt$videoid' in entry['media$group']:
-                    videos.append((
-                        index,
-                        'https://www.youtube.com/watch?v=' + entry['media$group']['yt$videoid']['$t']
-                    ))
-
-        videos = [v[1] for v in sorted(videos)]
+        playlist_title = self._og_search_title(page)
 
-        url_results = [self.url_result(vurl, 'Youtube') for vurl in videos]
-        return [self.playlist_result(url_results, playlist_id, playlist_title)]
+        url_results = [self.url_result(vid, 'Youtube') for vid in ids]
+        return self.playlist_result(url_results, playlist_id, playlist_title)
 
 
 class YoutubeChannelIE(InfoExtractor):