[common] recognize hls manifests that contain video only formats(#11394)

pull/8/head
Remita Amine 8 years ago
parent 0d7d9f9404
commit 8821a718cf

@ -283,11 +283,6 @@ class CBCWatchVideoIE(CBCWatchBaseIE):
formats = self._extract_m3u8_formats(re.sub(r'/([^/]+)/[^/?]+\.m3u8', r'/\1/\1.m3u8', m3u8_url), video_id, 'mp4', fatal=False) formats = self._extract_m3u8_formats(re.sub(r'/([^/]+)/[^/?]+\.m3u8', r'/\1/\1.m3u8', m3u8_url), video_id, 'mp4', fatal=False)
if len(formats) < 2: if len(formats) < 2:
formats = self._extract_m3u8_formats(m3u8_url, video_id, 'mp4') formats = self._extract_m3u8_formats(m3u8_url, video_id, 'mp4')
# Despite metadata in m3u8 all video+audio formats are
# actually video-only (no audio)
for f in formats:
if f.get('acodec') != 'none' and f.get('vcodec') != 'none':
f['acodec'] = 'none'
self._sort_formats(formats) self._sort_formats(formats)
info = { info = {

@ -1224,6 +1224,7 @@ class InfoExtractor(object):
'protocol': entry_protocol, 'protocol': entry_protocol,
'preference': preference, 'preference': preference,
}] }]
audio_groups = set()
last_info = {} last_info = {}
last_media = {} last_media = {}
for line in m3u8_doc.splitlines(): for line in m3u8_doc.splitlines():
@ -1239,15 +1240,18 @@ class InfoExtractor(object):
for v in (media.get('GROUP-ID'), media.get('NAME')): for v in (media.get('GROUP-ID'), media.get('NAME')):
if v: if v:
format_id.append(v) format_id.append(v)
formats.append({ f = {
'format_id': '-'.join(format_id), 'format_id': '-'.join(format_id),
'url': format_url(media_url), 'url': format_url(media_url),
'language': media.get('LANGUAGE'), 'language': media.get('LANGUAGE'),
'vcodec': 'none' if media_type == 'AUDIO' else None,
'ext': ext, 'ext': ext,
'protocol': entry_protocol, 'protocol': entry_protocol,
'preference': preference, 'preference': preference,
}) }
if media_type == 'AUDIO':
f['vcodec'] = 'none'
audio_groups.add(media['GROUP-ID'])
formats.append(f)
else: else:
# When there is no URI in EXT-X-MEDIA let this tag's # When there is no URI in EXT-X-MEDIA let this tag's
# data be used by regular URI lines below # data be used by regular URI lines below
@ -1295,6 +1299,9 @@ class InfoExtractor(object):
'abr': abr, 'abr': abr,
}) })
f.update(parse_codecs(last_info.get('CODECS'))) f.update(parse_codecs(last_info.get('CODECS')))
if last_info.get('AUDIO') in audio_groups:
# TODO: update acodec for for audio only formats with the same GROUP-ID
f['acodec'] = 'none'
formats.append(f) formats.append(f)
last_info = {} last_info = {}
last_media = {} last_media = {}

@ -78,11 +78,6 @@ class MSNIE(InfoExtractor):
m3u8_formats = self._extract_m3u8_formats( m3u8_formats = self._extract_m3u8_formats(
format_url, display_id, 'mp4', format_url, display_id, 'mp4',
m3u8_id='hls', fatal=False) m3u8_id='hls', fatal=False)
# Despite metadata in m3u8 all video+audio formats are
# actually video-only (no audio)
for f in m3u8_formats:
if f.get('acodec') != 'none' and f.get('vcodec') != 'none':
f['acodec'] = 'none'
formats.extend(m3u8_formats) formats.extend(m3u8_formats)
else: else:
formats.append({ formats.append({

Loading…
Cancel
Save