[extractor/common] Improved support for HTML5 subtitles

Ref: #10625

In a strict sense, <track>s with kind=captions are not subtitles. [1]
openload misuses this attribute, and I guess there will be more
examples, so I add it to common.py.

Also allow extracting information for subtitles-only <video> or <audio>
tags, which is the case of openload.

[1] https://www.w3.org/TR/html5/embedded-content-0.html#attr-track-kind
totalwebcasting
Yen Chi Hsuan 8 years ago
parent e6332059ac
commit 5968d7d2fe
No known key found for this signature in database
GPG Key ID: 3FDDD575826C5C30

@ -1,3 +1,9 @@
vesion <unreleased>
Core
+ Improved support for HTML5 subtitles
version 2016.09.24
Core

@ -1828,7 +1828,7 @@ class InfoExtractor(object):
for track_tag in re.findall(r'<track[^>]+>', media_content):
track_attributes = extract_attributes(track_tag)
kind = track_attributes.get('kind')
if not kind or kind == 'subtitles':
if not kind or kind in ('subtitles', 'captions'):
src = track_attributes.get('src')
if not src:
continue
@ -1836,7 +1836,7 @@ class InfoExtractor(object):
media_info['subtitles'].setdefault(lang, []).append({
'url': absolute_url(src),
})
if media_info['formats']:
if media_info['formats'] or media_info['subtitles']:
entries.append(media_info)
return entries

Loading…
Cancel
Save