[soundcloud] Add support for new rendition and improve extraction (closes #20699)

This commit is contained in:
Sergey M․ 2019-04-23 00:39:16 +07:00
parent 3fd86cfe13
commit e09965d550
No known key found for this signature in database
GPG key ID: 2C393E0F18A9236D

View file

@ -15,7 +15,12 @@ from ..compat import (
) )
from ..utils import ( from ..utils import (
ExtractorError, ExtractorError,
float_or_none,
int_or_none, int_or_none,
KNOWN_EXTENSIONS,
merge_dicts,
mimetype2ext,
str_or_none,
try_get, try_get,
unified_timestamp, unified_timestamp,
update_url_query, update_url_query,
@ -57,7 +62,7 @@ class SoundcloudIE(InfoExtractor):
'uploader': 'E.T. ExTerrestrial Music', 'uploader': 'E.T. ExTerrestrial Music',
'timestamp': 1349920598, 'timestamp': 1349920598,
'upload_date': '20121011', 'upload_date': '20121011',
'duration': 143, 'duration': 143.216,
'license': 'all-rights-reserved', 'license': 'all-rights-reserved',
'view_count': int, 'view_count': int,
'like_count': int, 'like_count': int,
@ -100,7 +105,7 @@ class SoundcloudIE(InfoExtractor):
'uploader': 'jaimeMF', 'uploader': 'jaimeMF',
'timestamp': 1386604920, 'timestamp': 1386604920,
'upload_date': '20131209', 'upload_date': '20131209',
'duration': 9, 'duration': 9.927,
'license': 'all-rights-reserved', 'license': 'all-rights-reserved',
'view_count': int, 'view_count': int,
'like_count': int, 'like_count': int,
@ -120,7 +125,7 @@ class SoundcloudIE(InfoExtractor):
'uploader': 'jaimeMF', 'uploader': 'jaimeMF',
'timestamp': 1386604920, 'timestamp': 1386604920,
'upload_date': '20131209', 'upload_date': '20131209',
'duration': 9, 'duration': 9.927,
'license': 'all-rights-reserved', 'license': 'all-rights-reserved',
'view_count': int, 'view_count': int,
'like_count': int, 'like_count': int,
@ -140,7 +145,7 @@ class SoundcloudIE(InfoExtractor):
'uploader': 'oddsamples', 'uploader': 'oddsamples',
'timestamp': 1389232924, 'timestamp': 1389232924,
'upload_date': '20140109', 'upload_date': '20140109',
'duration': 17, 'duration': 17.346,
'license': 'cc-by-sa', 'license': 'cc-by-sa',
'view_count': int, 'view_count': int,
'like_count': int, 'like_count': int,
@ -160,7 +165,7 @@ class SoundcloudIE(InfoExtractor):
'uploader': 'Ori Uplift Music', 'uploader': 'Ori Uplift Music',
'timestamp': 1504206263, 'timestamp': 1504206263,
'upload_date': '20170831', 'upload_date': '20170831',
'duration': 7449, 'duration': 7449.096,
'license': 'all-rights-reserved', 'license': 'all-rights-reserved',
'view_count': int, 'view_count': int,
'like_count': int, 'like_count': int,
@ -180,7 +185,7 @@ class SoundcloudIE(InfoExtractor):
'uploader': 'garyvee', 'uploader': 'garyvee',
'timestamp': 1488152409, 'timestamp': 1488152409,
'upload_date': '20170226', 'upload_date': '20170226',
'duration': 207, 'duration': 207.012,
'thumbnail': r're:https?://.*\.jpg', 'thumbnail': r're:https?://.*\.jpg',
'license': 'all-rights-reserved', 'license': 'all-rights-reserved',
'view_count': int, 'view_count': int,
@ -192,9 +197,31 @@ class SoundcloudIE(InfoExtractor):
'skip_download': True, 'skip_download': True,
}, },
}, },
# not avaialble via api.soundcloud.com/i1/tracks/id/streams
{
'url': 'https://soundcloud.com/giovannisarani/mezzo-valzer',
'md5': 'e22aecd2bc88e0e4e432d7dcc0a1abf7',
'info_dict': {
'id': '583011102',
'ext': 'mp3',
'title': 'Mezzo Valzer',
'description': 'md5:4138d582f81866a530317bae316e8b61',
'uploader': 'Giovanni Sarani',
'timestamp': 1551394171,
'upload_date': '20190228',
'duration': 180.157,
'thumbnail': r're:https?://.*\.jpg',
'license': 'all-rights-reserved',
'view_count': int,
'like_count': int,
'comment_count': int,
'repost_count': int,
},
'expected_warnings': ['Unable to download JSON metadata'],
}
] ]
_CLIENT_ID = 'NmW1FlPaiL94ueEu7oziOWjYEzZzQDcK' _CLIENT_ID = 'FweeGBOOEOYJWLJN3oEyToGLKhmSz0I7'
@staticmethod @staticmethod
def _extract_urls(webpage): def _extract_urls(webpage):
@ -202,10 +229,6 @@ class SoundcloudIE(InfoExtractor):
r'<iframe[^>]+src=(["\'])(?P<url>(?:https?://)?(?:w\.)?soundcloud\.com/player.+?)\1', r'<iframe[^>]+src=(["\'])(?P<url>(?:https?://)?(?:w\.)?soundcloud\.com/player.+?)\1',
webpage)] webpage)]
def report_resolve(self, video_id):
"""Report information extraction."""
self.to_screen('%s: Resolving id' % video_id)
@classmethod @classmethod
def _resolv_url(cls, url): def _resolv_url(cls, url):
return 'https://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID return 'https://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID
@ -224,6 +247,10 @@ class SoundcloudIE(InfoExtractor):
def extract_count(key): def extract_count(key):
return int_or_none(info.get('%s_count' % key)) return int_or_none(info.get('%s_count' % key))
like_count = extract_count('favoritings')
if like_count is None:
like_count = extract_count('likes')
result = { result = {
'id': track_id, 'id': track_id,
'uploader': username, 'uploader': username,
@ -231,15 +258,17 @@ class SoundcloudIE(InfoExtractor):
'title': title, 'title': title,
'description': info.get('description'), 'description': info.get('description'),
'thumbnail': thumbnail, 'thumbnail': thumbnail,
'duration': int_or_none(info.get('duration'), 1000), 'duration': float_or_none(info.get('duration'), 1000),
'webpage_url': info.get('permalink_url'), 'webpage_url': info.get('permalink_url'),
'license': info.get('license'), 'license': info.get('license'),
'view_count': extract_count('playback'), 'view_count': extract_count('playback'),
'like_count': extract_count('favoritings'), 'like_count': like_count,
'comment_count': extract_count('comment'), 'comment_count': extract_count('comment'),
'repost_count': extract_count('reposts'), 'repost_count': extract_count('reposts'),
'genre': info.get('genre'), 'genre': info.get('genre'),
} }
format_urls = set()
formats = [] formats = []
query = {'client_id': self._CLIENT_ID} query = {'client_id': self._CLIENT_ID}
if secret_token is not None: if secret_token is not None:
@ -248,6 +277,7 @@ class SoundcloudIE(InfoExtractor):
# We can build a direct link to the song # We can build a direct link to the song
format_url = update_url_query( format_url = update_url_query(
'https://api.soundcloud.com/tracks/%s/download' % track_id, query) 'https://api.soundcloud.com/tracks/%s/download' % track_id, query)
format_urls.add(format_url)
formats.append({ formats.append({
'format_id': 'download', 'format_id': 'download',
'ext': info.get('original_format', 'mp3'), 'ext': info.get('original_format', 'mp3'),
@ -256,12 +286,17 @@ class SoundcloudIE(InfoExtractor):
'preference': 10, 'preference': 10,
}) })
# We have to retrieve the url # Old API, does not work for some tracks (e.g.
# https://soundcloud.com/giovannisarani/mezzo-valzer)
format_dict = self._download_json( format_dict = self._download_json(
'https://api.soundcloud.com/i1/tracks/%s/streams' % track_id, 'https://api.soundcloud.com/i1/tracks/%s/streams' % track_id,
track_id, 'Downloading track url', query=query) track_id, 'Downloading track url', query=query, fatal=False)
if format_dict:
for key, stream_url in format_dict.items(): for key, stream_url in format_dict.items():
if stream_url in format_urls:
continue
format_urls.add(stream_url)
ext, abr = 'mp3', None ext, abr = 'mp3', None
mobj = re.search(r'_([^_]+)_(\d+)_url', key) mobj = re.search(r'_([^_]+)_(\d+)_url', key)
if mobj: if mobj:
@ -295,6 +330,48 @@ class SoundcloudIE(InfoExtractor):
formats.extend(stream_formats) formats.extend(stream_formats)
# New API
transcodings = try_get(
info, lambda x: x['media']['transcodings'], list) or []
for t in transcodings:
if not isinstance(t, dict):
continue
format_url = url_or_none(t.get('url'))
if not format_url:
continue
stream = self._download_json(
update_url_query(format_url, query), track_id, fatal=False)
if not isinstance(stream, dict):
continue
stream_url = url_or_none(stream.get('url'))
if not stream_url:
continue
if stream_url in format_urls:
continue
format_urls.add(stream_url)
protocol = try_get(t, lambda x: x['format']['protocol'], compat_str)
if protocol != 'hls' and '/hls' in format_url:
protocol = 'hls'
ext = None
preset = str_or_none(t.get('preset'))
if preset:
ext = preset.split('_')[0]
if ext not in KNOWN_EXTENSIONS:
mimetype = try_get(
t, lambda x: x['format']['mime_type'], compat_str)
ext = mimetype2ext(mimetype) or 'mp3'
format_id_list = []
if protocol:
format_id_list.append(protocol)
format_id_list.append(ext)
format_id = '_'.join(format_id_list)
formats.append({
'url': stream_url,
'format_id': format_id,
'ext': ext,
'protocol': 'm3u8_native' if protocol == 'hls' else 'http',
})
if not formats: if not formats:
# We fallback to the stream_url in the original info, this # We fallback to the stream_url in the original info, this
# cannot be always used, sometimes it can give an HTTP 404 error # cannot be always used, sometimes it can give an HTTP 404 error
@ -303,11 +380,11 @@ class SoundcloudIE(InfoExtractor):
'url': update_url_query(info['stream_url'], query), 'url': update_url_query(info['stream_url'], query),
'ext': 'mp3', 'ext': 'mp3',
}) })
self._check_formats(formats, track_id)
for f in formats: for f in formats:
f['vcodec'] = 'none' f['vcodec'] = 'none'
self._check_formats(formats, track_id)
self._sort_formats(formats) self._sort_formats(formats)
result['formats'] = formats result['formats'] = formats
@ -319,6 +396,7 @@ class SoundcloudIE(InfoExtractor):
raise ExtractorError('Invalid URL: %s' % url) raise ExtractorError('Invalid URL: %s' % url)
track_id = mobj.group('track_id') track_id = mobj.group('track_id')
new_info = {}
if track_id is not None: if track_id is not None:
info_json_url = 'https://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID info_json_url = 'https://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
@ -344,13 +422,31 @@ class SoundcloudIE(InfoExtractor):
if token: if token:
resolve_title += '/%s' % token resolve_title += '/%s' % token
self.report_resolve(full_title) webpage = self._download_webpage(url, full_title, fatal=False)
if webpage:
entries = self._parse_json(
self._search_regex(
r'var\s+c\s*=\s*(\[.+?\])\s*,\s*o\s*=Date\b', webpage,
'data', default='[]'), full_title, fatal=False)
if entries:
for e in entries:
if not isinstance(e, dict):
continue
if e.get('id') != 67:
continue
data = try_get(e, lambda x: x['data'][0], dict)
if data:
new_info = data
break
info_json_url = self._resolv_url(
'https://soundcloud.com/%s' % resolve_title)
url = 'https://soundcloud.com/%s' % resolve_title # Contains some additional info missing from new_info
info_json_url = self._resolv_url(url) info = self._download_json(
info = self._download_json(info_json_url, full_title, 'Downloading info JSON') info_json_url, full_title, 'Downloading info JSON')
return self._extract_info_dict(info, full_title, secret_token=token) return self._extract_info_dict(
merge_dicts(info, new_info), full_title, secret_token=token)
class SoundcloudPlaylistBaseIE(SoundcloudIE): class SoundcloudPlaylistBaseIE(SoundcloudIE):
@ -396,8 +492,6 @@ class SoundcloudSetIE(SoundcloudPlaylistBaseIE):
full_title += '/' + token full_title += '/' + token
url += '/' + token url += '/' + token
self.report_resolve(full_title)
resolv_url = self._resolv_url(url) resolv_url = self._resolv_url(url)
info = self._download_json(resolv_url, full_title) info = self._download_json(resolv_url, full_title)