[youtube] Improve format filesize extraction (#16453)

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

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 1f29e8a4ea940418ad335d27459c07082161a9d7..7f4298c088c7514534e66f6a8d1921fe57196f03 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -1697,9 +1697,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         self.report_information_extraction(video_id)
 
         # uploader
-        if 'author' not in video_info:
-            raise ExtractorError('Unable to extract uploader name')
-        video_uploader = compat_urllib_parse_unquote_plus(video_info['author'][0])
+        video_uploader = try_get(video_info, lambda x: x['author'][0], compat_str)
+        if video_uploader:
+            video_uploader = compat_urllib_parse_unquote_plus(video_uploader)
+        else:
+            self._downloader.report_warning('unable to extract uploader name')
 
         # uploader_id
         video_uploader_id = None
@@ -1813,6 +1815,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
         chapters = self._extract_chapters(description_original, video_duration)
 
+        def _extract_filesize(media_url):
+            return int_or_none(self._search_regex(
+                r'\bclen[=/](\d+)', media_url, 'filesize', default=None))
+
         if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
             self.report_rtmp_download()
             formats = [{
@@ -1917,8 +1923,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 mobj = re.search(r'^(?P<width>\d+)[xX](?P<height>\d+)$', url_data.get('size', [''])[0])
                 width, height = (int(mobj.group('width')), int(mobj.group('height'))) if mobj else (None, None)
 
+                filesize = int_or_none(url_data.get(
+                    'clen', [None])[0]) or _extract_filesize(url)
+
                 more_fields = {
-                    'filesize': int_or_none(url_data.get('clen', [None])[0]),
+                    'filesize': filesize,
                     'tbr': float_or_none(url_data.get('bitrate', [None])[0], 1000),
                     'width': width,
                     'height': height,
@@ -1992,6 +2001,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     for df in self._extract_mpd_formats(
                             mpd_url, video_id, fatal=dash_mpd_fatal,
                             formats_dict=self._formats):
+                        if not df.get('filesize'):
+                            df['filesize'] = _extract_filesize(df['url'])
                         # 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