[youtube] Improve multifeed videos extraction (Closes #8536)

[youtube-dl] / youtube_dl / extractor / youtube.py

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 828f5d1f4b0dbe3245a9fff6076553cd5dfbb87c..97d6219072d988e5ed16ad0277e18e69144fc7d6 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -286,7 +286,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         '22': {'ext': 'mp4', 'width': 1280, 'height': 720, 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264'},
         '34': {'ext': 'flv', 'width': 640, 'height': 360, 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264'},
         '35': {'ext': 'flv', 'width': 854, 'height': 480, 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264'},
-        '36': {'ext': '3gp', 'width': 320, 'height': 240, 'acodec': 'aac', 'abr': 32, 'vcodec': 'mp4v'},
+        # itag 36 videos are either 320x180 (BaW_jenozKc) or 320x240 (__2ABJjxzNo), abr varies as well
+        '36': {'ext': '3gp', 'width': 320, 'acodec': 'aac', 'vcodec': 'mp4v'},
         '37': {'ext': 'mp4', 'width': 1920, 'height': 1080, 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264'},
         '38': {'ext': 'mp4', 'width': 4096, 'height': 3072, 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264'},
         '43': {'ext': 'webm', 'width': 640, 'height': 360, 'acodec': 'vorbis', 'abr': 128, 'vcodec': 'vp8'},
@@ -369,11 +370,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         # RTMP (unnamed)
         '_rtmp': {'protocol': 'rtmp'},
     }
+    _SUBTITLE_FORMATS = ('ttml', 'vtt')
 
     IE_NAME = 'youtube'
     _TESTS = [
         {
-            'url': 'http://www.youtube.com/watch?v=BaW_jenozKcj&t=1s&end=9',
+            'url': 'http://www.youtube.com/watch?v=BaW_jenozKc&t=1s&end=9',
             'info_dict': {
                 'id': 'BaW_jenozKc',
                 'ext': 'mp4',
@@ -439,7 +441,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             }
         },
         {
-            'url': 'http://www.youtube.com/watch?v=BaW_jenozKcj&v=UxxajLWwzqY',
+            'url': 'http://www.youtube.com/watch?v=BaW_jenozKc&v=UxxajLWwzqY',
             'note': 'Use the first video ID in the URL',
             'info_dict': {
                 'id': 'BaW_jenozKc',
@@ -918,7 +920,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             if lang in sub_lang_list:
                 continue
             sub_formats = []
-            for ext in ['sbv', 'vtt', 'srt']:
+            for ext in self._SUBTITLE_FORMATS:
                 params = compat_urllib_parse.urlencode({
                     'lang': lang,
                     'v': video_id,
@@ -988,7 +990,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             for lang_node in caption_list.findall('target'):
                 sub_lang = lang_node.attrib['lang_code']
                 sub_formats = []
-                for ext in ['sbv', 'vtt', 'srt']:
+                for ext in self._SUBTITLE_FORMATS:
                     params = compat_urllib_parse.urlencode({
                         'lang': original_lang,
                         'tlang': sub_lang,
@@ -1194,9 +1196,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             if not self._downloader.params.get('noplaylist'):
                 entries = []
                 feed_ids = []
-                multifeed_metadata_list = compat_urllib_parse_unquote_plus(video_info['multifeed_metadata_list'][0])
+                multifeed_metadata_list = video_info['multifeed_metadata_list'][0]
                 for feed in multifeed_metadata_list.split(','):
-                    feed_data = compat_parse_qs(feed)
+                    # Unquote should take place before split on comma (,) since textual
+                    # fields may contain comma as well (see
+                    # https://github.com/rg3/youtube-dl/issues/8536)
+                    feed_data = compat_parse_qs(compat_urllib_parse_unquote_plus(feed))
                     entries.append({
                         '_type': 'url_transparent',
                         'ie_key': 'Youtube',