[youtube] Clarify itag 36 height and abr (Closes #8457)

[yt-dlp.git] / youtube_dl / extractor / youtube.py

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 8d3fcfa5f7ebace20248ee85c9d9dfb74ce910c4..b9a91dea2bd1d0f0aa21abcb211766c2ab15304b 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -181,7 +181,7 @@ def _real_initialize(self):
             return
 
 
-class YoutubeEntryListBaseInfoExtractor(InfoExtractor):
+class YoutubeEntryListBaseInfoExtractor(YoutubeBaseInfoExtractor):
     # Extract entries from page with "Load more" button
     def _entries(self, page, playlist_id):
         more_widget_html = content_html = page
@@ -233,7 +233,7 @@ def extract_videos_from_page(self, page):
 
 class YoutubePlaylistsBaseInfoExtractor(YoutubeEntryListBaseInfoExtractor):
     def _process_page(self, content):
-        for playlist_id in re.findall(r'href="/?playlist\?list=(.+?)"', content):
+        for playlist_id in orderedSet(re.findall(r'href="/?playlist\?list=([0-9A-Za-z-_]{10,})"', content)):
             yield self.url_result(
                 'https://www.youtube.com/playlist?list=%s' % playlist_id, 'YoutubePlaylist')
 
@@ -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'},
@@ -369,6 +370,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         # RTMP (unnamed)
         '_rtmp': {'protocol': 'rtmp'},
     }
+    _SUBTITLE_FORMATS = ('ttml', 'vtt')
 
     IE_NAME = 'youtube'
     _TESTS = [
@@ -918,7 +920,7 @@ def _get_subtitles(self, video_id, webpage):
             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 @@ def _get_automatic_captions(self, video_id, webpage):
             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,
@@ -1463,7 +1465,7 @@ def _map_to_format_list(urlmap):
         # Look for the DASH manifest
         if self._downloader.params.get('youtube_include_dash_manifest', True):
             dash_mpd_fatal = True
-            for dash_manifest_url in dash_mpds:
+            for mpd_url in dash_mpds:
                 dash_formats = {}
                 try:
                     def decrypt_sig(mobj):
@@ -1471,15 +1473,11 @@ def decrypt_sig(mobj):
                         dec_s = self._decrypt_signature(s, video_id, player_url, age_gate)
                         return '/signature/%s' % dec_s
 
-                    dash_manifest_url = re.sub(r'/s/([a-fA-F0-9\.]+)', decrypt_sig, dash_manifest_url)
-                    dash_doc = self._download_xml(
-                        dash_manifest_url, video_id,
-                        note='Downloading DASH manifest',
-                        errnote='Could not download DASH manifest',
-                        fatal=dash_mpd_fatal)
+                    mpd_url = re.sub(r'/s/([a-fA-F0-9\.]+)', decrypt_sig, mpd_url)
 
-                    for df in self._parse_dash_manifest(
-                            video_id, dash_doc, namespace='urn:mpeg:DASH:schema:MPD:2011', formats_dict=self._formats, fatal=dash_mpd_fatal):
+                    for df in self._extract_mpd_formats(
+                            mpd_url, video_id, fatal=dash_mpd_fatal,
+                            formats_dict=self._formats):
                         # Do not overwrite DASH format found in some previous DASH manifest
                         if df['format_id'] not in dash_formats:
                             dash_formats[df['format_id']] = df
@@ -1547,7 +1545,7 @@ def decrypt_sig(mobj):
         }
 
 
-class YoutubePlaylistIE(YoutubeBaseInfoExtractor, YoutubePlaylistBaseInfoExtractor):
+class YoutubePlaylistIE(YoutubePlaylistBaseInfoExtractor):
     IE_DESC = 'YouTube.com playlists'
     _VALID_URL = r"""(?x)(?:
                         (?:https?://)?