mirror of
https://codeberg.org/polarisfm/youtube-dl
synced 2024-11-25 09:54:32 +01:00
[common] fix dash codec information for mixed videos and fragment url construction(#11490)
This commit is contained in:
parent
8ab7e6c4cc
commit
7fe1592073
@ -461,6 +461,7 @@ class TestUtil(unittest.TestCase):
|
||||
self.assertEqual(urljoin('http://foo.de/', None), None)
|
||||
self.assertEqual(urljoin('http://foo.de/', ''), None)
|
||||
self.assertEqual(urljoin('http://foo.de/', ['foobar']), None)
|
||||
self.assertEqual(urljoin('http://foo.de/a/b/c.txt', '.././../d.txt'), 'http://foo.de/d.txt')
|
||||
|
||||
def test_parse_age_limit(self):
|
||||
self.assertEqual(parse_age_limit(None), None)
|
||||
|
@ -59,6 +59,7 @@ from ..utils import (
|
||||
parse_m3u8_attributes,
|
||||
extract_attributes,
|
||||
parse_codecs,
|
||||
urljoin,
|
||||
)
|
||||
|
||||
|
||||
@ -1631,11 +1632,6 @@ class InfoExtractor(object):
|
||||
extract_Initialization(segment_template)
|
||||
return ms_info
|
||||
|
||||
def combine_url(base_url, target_url):
|
||||
if re.match(r'^https?://', target_url):
|
||||
return target_url
|
||||
return '%s%s%s' % (base_url, '' if base_url.endswith('/') else '/', target_url)
|
||||
|
||||
mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
|
||||
formats = []
|
||||
for period in mpd_doc.findall(_add_ns('Period')):
|
||||
@ -1685,12 +1681,11 @@ class InfoExtractor(object):
|
||||
'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),
|
||||
'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
|
||||
'fps': int_or_none(representation_attrib.get('frameRate')),
|
||||
'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'),
|
||||
'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'),
|
||||
'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
|
||||
'format_note': 'DASH %s' % content_type,
|
||||
'filesize': filesize,
|
||||
}
|
||||
f.update(parse_codecs(representation_attrib.get('codecs')))
|
||||
representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
|
||||
if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info:
|
||||
|
||||
@ -1774,7 +1769,7 @@ class InfoExtractor(object):
|
||||
f['fragments'].append({'url': initialization_url})
|
||||
f['fragments'].extend(representation_ms_info['fragments'])
|
||||
for fragment in f['fragments']:
|
||||
fragment['url'] = combine_url(base_url, fragment['url'])
|
||||
fragment['url'] = urljoin(base_url, fragment['url'])
|
||||
try:
|
||||
existing_format = next(
|
||||
fo for fo in formats
|
||||
|
Loading…
Reference in New Issue
Block a user