paradizelost/youtube-dl
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[ARD] Back-port subtitle extraction from yt-dlp PR 2409

Browse Source 
Authored by: fstirlitz
Fixes #30543
Closes #17766 (thanks ngdio)
This commit is contained in:
dirkf 2022-02-07 13:28:21 +00:00
parent 734dfbb4e3
commit 47b0c8697a
1 changed files with 15 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
youtube_dl/extractor/ard.py
View File

@ -332,9 +332,24 @@ class ARDIE(InfoExtractor):
formats.append(f) formats.append(f)
self._sort_formats(formats) self._sort_formats(formats)
_SUB_FORMATS = (
('./dataTimedText', 'ttml'),
('./dataTimedTextNoOffset', 'ttml'),
('./dataTimedTextVtt', 'vtt'),
)
subtitles = {}
for subsel, subext in _SUB_FORMATS:
for node in video_node.findall(subsel):
subtitles.setdefault('de', []).append({
'url': node.attrib['url'],
'ext': subext,
})
return { return {
'id': xpath_text(video_node, './videoId', default=display_id), 'id': xpath_text(video_node, './videoId', default=display_id),
'formats': formats, 'formats': formats,
'subtitles': subtitles,
'display_id': display_id, 'display_id': display_id,
'title': video_node.find('./title').text, 'title': video_node.find('./title').text,
'duration': parse_duration(video_node.find('./duration').text), 'duration': parse_duration(video_node.find('./duration').text),