_ Git - youtube-dl/blob - youtube_dl/extractor/soundcloud.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import itertools
   5 import re
   6
   7 from .common import (
   8     InfoExtractor,
   9     SearchInfoExtractor
  10 )
  11 from ..compat import (
  12     compat_str,
  13     compat_urlparse,
  14     compat_urllib_parse_urlencode,
  15 )
  16 from ..utils import (
  17     ExtractorError,
  18     int_or_none,
  19     unified_strdate,
  20     update_url_query)
  21
  22
  23 class SoundcloudIE(InfoExtractor):
  24     """Information extractor for soundcloud.com
  25        To access the media, the uid of the song and a stream token
  26        must be extracted from the page source and the script must make
  27        a request to media.soundcloud.com/crossdomain.xml. Then
  28        the media can be grabbed by requesting from an url composed
  29        of the stream token and uid
  30      """
  31
  32     _VALID_URL = r'''(?x)^(?:https?://)?
  33                     (?:(?:(?:www\.|m\.)?soundcloud\.com/
  34                             (?!stations/track)
  35                             (?P<uploader>[\w\d-]+)/
  36                             (?!(?:tracks|sets(?:/.+?)?|reposts|likes|spotlight)/?(?:$|[?#]))
  37                             (?P<title>[\w\d-]+)/?
  38                             (?P<token>[^?]+?)?(?:[?].*)?$)
  39                        |(?:api\.soundcloud\.com/tracks/(?P<track_id>\d+)
  40                           (?:/?\?secret_token=(?P<secret_token>[^&]+))?)
  41                        |(?P<player>(?:w|player|p.)\.soundcloud\.com/player/?.*?url=.*)
  42                     )
  43                     '''
  44     IE_NAME = 'soundcloud'
  45     _TESTS = [
  46         {
  47             'url': 'http://soundcloud.com/ethmusic/lostin-powers-she-so-heavy',
  48             'md5': 'ebef0a451b909710ed1d7787dddbf0d7',
  49             'info_dict': {
  50                 'id': '62986583',
  51                 'ext': 'mp3',
  52                 'upload_date': '20121011',
  53                 'description': 'No Downloads untill we record the finished version this weekend, i was too pumped n i had to post it , earl is prolly gonna b hella p.o\'d',
  54                 'uploader': 'E.T. ExTerrestrial Music',
  55                 'title': 'Lostin Powers - She so Heavy (SneakPreview) Adrian Ackers Blueprint 1',
  56                 'duration': 143,
  57                 'license': 'all-rights-reserved',
  58             }
  59         },
  60         # not streamable song
  61         {
  62             'url': 'https://soundcloud.com/the-concept-band/goldrushed-mastered?in=the-concept-band/sets/the-royal-concept-ep',
  63             'info_dict': {
  64                 'id': '47127627',
  65                 'ext': 'mp3',
  66                 'title': 'Goldrushed',
  67                 'description': 'From Stockholm Sweden\r\nPovel / Magnus / Filip / David\r\nwww.theroyalconcept.com',
  68                 'uploader': 'The Royal Concept',
  69                 'upload_date': '20120521',
  70                 'duration': 227,
  71                 'license': 'all-rights-reserved',
  72             },
  73             'params': {
  74                 # rtmp
  75                 'skip_download': True,
  76             },
  77         },
  78         # private link
  79         {
  80             'url': 'https://soundcloud.com/jaimemf/youtube-dl-test-video-a-y-baw/s-8Pjrp',
  81             'md5': 'aa0dd32bfea9b0c5ef4f02aacd080604',
  82             'info_dict': {
  83                 'id': '123998367',
  84                 'ext': 'mp3',
  85                 'title': 'Youtube - Dl Test Video \'\' Ä↭',
  86                 'uploader': 'jaimeMF',
  87                 'description': 'test chars:  \"\'/\\ä↭',
  88                 'upload_date': '20131209',
  89                 'duration': 9,
  90                 'license': 'all-rights-reserved',
  91             },
  92         },
  93         # private link (alt format)
  94         {
  95             'url': 'https://api.soundcloud.com/tracks/123998367?secret_token=s-8Pjrp',
  96             'md5': 'aa0dd32bfea9b0c5ef4f02aacd080604',
  97             'info_dict': {
  98                 'id': '123998367',
  99                 'ext': 'mp3',
 100                 'title': 'Youtube - Dl Test Video \'\' Ä↭',
 101                 'uploader': 'jaimeMF',
 102                 'description': 'test chars:  \"\'/\\ä↭',
 103                 'upload_date': '20131209',
 104                 'duration': 9,
 105                 'license': 'all-rights-reserved',
 106             },
 107         },
 108         # downloadable song
 109         {
 110             'url': 'https://soundcloud.com/oddsamples/bus-brakes',
 111             'md5': '7624f2351f8a3b2e7cd51522496e7631',
 112             'info_dict': {
 113                 'id': '128590877',
 114                 'ext': 'mp3',
 115                 'title': 'Bus Brakes',
 116                 'description': 'md5:0053ca6396e8d2fd7b7e1595ef12ab66',
 117                 'uploader': 'oddsamples',
 118                 'upload_date': '20140109',
 119                 'duration': 17,
 120                 'license': 'cc-by-sa',
 121             },
 122         },
 123     ]
 124
 125     _CLIENT_ID = 'JlZIsxg2hY5WnBgtn3jfS0UYCl0K8DOg'
 126     _IPHONE_CLIENT_ID = '376f225bf427445fc4bfb6b99b72e0bf'
 127
 128     @staticmethod
 129     def _extract_urls(webpage):
 130         return [m.group('url') for m in re.finditer(
 131             r'<iframe[^>]+src=(["\'])(?P<url>(?:https?://)?(?:w\.)?soundcloud\.com/player.+?)\1',
 132             webpage)]
 133
 134     def report_resolve(self, video_id):
 135         """Report information extraction."""
 136         self.to_screen('%s: Resolving id' % video_id)
 137
 138     @classmethod
 139     def _resolv_url(cls, url):
 140         return 'https://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID
 141
 142     def _extract_info_dict(self, info, full_title=None, quiet=False, secret_token=None):
 143         track_id = compat_str(info['id'])
 144         name = full_title or track_id
 145         if quiet:
 146             self.report_extraction(name)
 147         thumbnail = info.get('artwork_url')
 148         if isinstance(thumbnail, compat_str):
 149             thumbnail = thumbnail.replace('-large', '-t500x500')
 150         ext = 'mp3'
 151         result = {
 152             'id': track_id,
 153             'uploader': info.get('user', {}).get('username'),
 154             'upload_date': unified_strdate(info.get('created_at')),
 155             'title': info['title'],
 156             'description': info.get('description'),
 157             'thumbnail': thumbnail,
 158             'duration': int_or_none(info.get('duration'), 1000),
 159             'webpage_url': info.get('permalink_url'),
 160             'license': info.get('license'),
 161         }
 162         formats = []
 163         query = {'client_id': self._CLIENT_ID}
 164         if secret_token is not None:
 165             query['secret_token'] = secret_token
 166         if info.get('downloadable', False):
 167             # We can build a direct link to the song
 168             format_url = update_url_query(
 169                 'https://api.soundcloud.com/tracks/{0}/download'.format(track_id), query)
 170             formats.append({
 171                 'format_id': 'download',
 172                 'ext': info.get('original_format', 'mp3'),
 173                 'url': format_url,
 174                 'vcodec': 'none',
 175                 'preference': 10,
 176             })
 177
 178         # We have to retrieve the url
 179         format_dict = self._download_json(
 180             'https://api.soundcloud.com/i1/tracks/%s/streams' % track_id,
 181             track_id, 'Downloading track url', query=query)
 182
 183         for key, stream_url in format_dict.items():
 184             abr = int_or_none(self._search_regex(
 185                 r'_(\d+)_url', key, 'audio bitrate', default=None))
 186             if key.startswith('http'):
 187                 stream_formats = [{
 188                     'format_id': key,
 189                     'ext': ext,
 190                     'url': stream_url,
 191                 }]
 192             elif key.startswith('rtmp'):
 193                 # The url doesn't have an rtmp app, we have to extract the playpath
 194                 url, path = stream_url.split('mp3:', 1)
 195                 stream_formats = [{
 196                     'format_id': key,
 197                     'url': url,
 198                     'play_path': 'mp3:' + path,
 199                     'ext': 'flv',
 200                 }]
 201             elif key.startswith('hls'):
 202                 stream_formats = self._extract_m3u8_formats(
 203                     stream_url, track_id, 'mp3', entry_protocol='m3u8_native',
 204                     m3u8_id=key, fatal=False)
 205             else:
 206                 continue
 207
 208             for f in stream_formats:
 209                 f['abr'] = abr
 210
 211             formats.extend(stream_formats)
 212
 213         if not formats:
 214             # We fallback to the stream_url in the original info, this
 215             # cannot be always used, sometimes it can give an HTTP 404 error
 216             formats.append({
 217                 'format_id': 'fallback',
 218                 'url': update_url_query(info['stream_url'], query),
 219                 'ext': ext,
 220             })
 221
 222         for f in formats:
 223             f['vcodec'] = 'none'
 224
 225         self._check_formats(formats, track_id)
 226         self._sort_formats(formats)
 227         result['formats'] = formats
 228
 229         return result
 230
 231     def _real_extract(self, url):
 232         mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
 233         if mobj is None:
 234             raise ExtractorError('Invalid URL: %s' % url)
 235
 236         track_id = mobj.group('track_id')
 237
 238         if track_id is not None:
 239             info_json_url = 'https://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 240             full_title = track_id
 241             token = mobj.group('secret_token')
 242             if token:
 243                 info_json_url += '&secret_token=' + token
 244         elif mobj.group('player'):
 245             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 246             real_url = query['url'][0]
 247             # If the token is in the query of the original url we have to
 248             # manually add it
 249             if 'secret_token' in query:
 250                 real_url += '?secret_token=' + query['secret_token'][0]
 251             return self.url_result(real_url)
 252         else:
 253             # extract uploader (which is in the url)
 254             uploader = mobj.group('uploader')
 255             # extract simple title (uploader + slug of song title)
 256             slug_title = mobj.group('title')
 257             token = mobj.group('token')
 258             full_title = resolve_title = '%s/%s' % (uploader, slug_title)
 259             if token:
 260                 resolve_title += '/%s' % token
 261
 262             self.report_resolve(full_title)
 263
 264             url = 'https://soundcloud.com/%s' % resolve_title
 265             info_json_url = self._resolv_url(url)
 266         info = self._download_json(info_json_url, full_title, 'Downloading info JSON')
 267
 268         return self._extract_info_dict(info, full_title, secret_token=token)
 269
 270
 271 class SoundcloudPlaylistBaseIE(SoundcloudIE):
 272     @staticmethod
 273     def _extract_id(e):
 274         return compat_str(e['id']) if e.get('id') else None
 275
 276     def _extract_track_entries(self, tracks):
 277         return [
 278             self.url_result(
 279                 track['permalink_url'], SoundcloudIE.ie_key(),
 280                 video_id=self._extract_id(track))
 281             for track in tracks if track.get('permalink_url')]
 282
 283
 284 class SoundcloudSetIE(SoundcloudPlaylistBaseIE):
 285     _VALID_URL = r'https?://(?:(?:www|m)\.)?soundcloud\.com/(?P<uploader>[\w\d-]+)/sets/(?P<slug_title>[\w\d-]+)(?:/(?P<token>[^?/]+))?'
 286     IE_NAME = 'soundcloud:set'
 287     _TESTS = [{
 288         'url': 'https://soundcloud.com/the-concept-band/sets/the-royal-concept-ep',
 289         'info_dict': {
 290             'id': '2284613',
 291             'title': 'The Royal Concept EP',
 292         },
 293         'playlist_mincount': 5,
 294     }, {
 295         'url': 'https://soundcloud.com/the-concept-band/sets/the-royal-concept-ep/token',
 296         'only_matching': True,
 297     }]
 298
 299     def _real_extract(self, url):
 300         mobj = re.match(self._VALID_URL, url)
 301
 302         # extract uploader (which is in the url)
 303         uploader = mobj.group('uploader')
 304         # extract simple title (uploader + slug of song title)
 305         slug_title = mobj.group('slug_title')
 306         full_title = '%s/sets/%s' % (uploader, slug_title)
 307         url = 'https://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 308
 309         token = mobj.group('token')
 310         if token:
 311             full_title += '/' + token
 312             url += '/' + token
 313
 314         self.report_resolve(full_title)
 315
 316         resolv_url = self._resolv_url(url)
 317         info = self._download_json(resolv_url, full_title)
 318
 319         if 'errors' in info:
 320             msgs = (compat_str(err['error_message']) for err in info['errors'])
 321             raise ExtractorError('unable to download video webpage: %s' % ','.join(msgs))
 322
 323         entries = self._extract_track_entries(info['tracks'])
 324
 325         return {
 326             '_type': 'playlist',
 327             'entries': entries,
 328             'id': '%s' % info['id'],
 329             'title': info['title'],
 330         }
 331
 332
 333 class SoundcloudPagedPlaylistBaseIE(SoundcloudPlaylistBaseIE):
 334     _API_BASE = 'https://api.soundcloud.com'
 335     _API_V2_BASE = 'https://api-v2.soundcloud.com'
 336
 337     def _extract_playlist(self, base_url, playlist_id, playlist_title):
 338         COMMON_QUERY = {
 339             'limit': 50,
 340             'client_id': self._CLIENT_ID,
 341             'linked_partitioning': '1',
 342         }
 343
 344         query = COMMON_QUERY.copy()
 345         query['offset'] = 0
 346
 347         next_href = base_url + '?' + compat_urllib_parse_urlencode(query)
 348
 349         entries = []
 350         for i in itertools.count():
 351             response = self._download_json(
 352                 next_href, playlist_id, 'Downloading track page %s' % (i + 1))
 353
 354             collection = response['collection']
 355             if not collection:
 356                 break
 357
 358             def resolve_permalink_url(candidates):
 359                 for cand in candidates:
 360                     if isinstance(cand, dict):
 361                         permalink_url = cand.get('permalink_url')
 362                         entry_id = self._extract_id(cand)
 363                         if permalink_url and permalink_url.startswith('http'):
 364                             return permalink_url, entry_id
 365
 366             for e in collection:
 367                 permalink_url, entry_id = resolve_permalink_url((e, e.get('track'), e.get('playlist')))
 368                 if permalink_url:
 369                     entries.append(self.url_result(permalink_url, video_id=entry_id))
 370
 371             next_href = response.get('next_href')
 372             if not next_href:
 373                 break
 374
 375             parsed_next_href = compat_urlparse.urlparse(response['next_href'])
 376             qs = compat_urlparse.parse_qs(parsed_next_href.query)
 377             qs.update(COMMON_QUERY)
 378             next_href = compat_urlparse.urlunparse(
 379                 parsed_next_href._replace(query=compat_urllib_parse_urlencode(qs, True)))
 380
 381         return {
 382             '_type': 'playlist',
 383             'id': playlist_id,
 384             'title': playlist_title,
 385             'entries': entries,
 386         }
 387
 388
 389 class SoundcloudUserIE(SoundcloudPagedPlaylistBaseIE):
 390     _VALID_URL = r'''(?x)
 391                         https?://
 392                             (?:(?:www|m)\.)?soundcloud\.com/
 393                             (?P<user>[^/]+)
 394                             (?:/
 395                                 (?P<rsrc>tracks|sets|reposts|likes|spotlight)
 396                             )?
 397                             /?(?:[?#].*)?$
 398                     '''
 399     IE_NAME = 'soundcloud:user'
 400     _TESTS = [{
 401         'url': 'https://soundcloud.com/the-akashic-chronicler',
 402         'info_dict': {
 403             'id': '114582580',
 404             'title': 'The Akashic Chronicler (All)',
 405         },
 406         'playlist_mincount': 74,
 407     }, {
 408         'url': 'https://soundcloud.com/the-akashic-chronicler/tracks',
 409         'info_dict': {
 410             'id': '114582580',
 411             'title': 'The Akashic Chronicler (Tracks)',
 412         },
 413         'playlist_mincount': 37,
 414     }, {
 415         'url': 'https://soundcloud.com/the-akashic-chronicler/sets',
 416         'info_dict': {
 417             'id': '114582580',
 418             'title': 'The Akashic Chronicler (Playlists)',
 419         },
 420         'playlist_mincount': 2,
 421     }, {
 422         'url': 'https://soundcloud.com/the-akashic-chronicler/reposts',
 423         'info_dict': {
 424             'id': '114582580',
 425             'title': 'The Akashic Chronicler (Reposts)',
 426         },
 427         'playlist_mincount': 7,
 428     }, {
 429         'url': 'https://soundcloud.com/the-akashic-chronicler/likes',
 430         'info_dict': {
 431             'id': '114582580',
 432             'title': 'The Akashic Chronicler (Likes)',
 433         },
 434         'playlist_mincount': 321,
 435     }, {
 436         'url': 'https://soundcloud.com/grynpyret/spotlight',
 437         'info_dict': {
 438             'id': '7098329',
 439             'title': 'Grynpyret (Spotlight)',
 440         },
 441         'playlist_mincount': 1,
 442     }]
 443
 444     _BASE_URL_MAP = {
 445         'all': '%s/profile/soundcloud:users:%%s' % SoundcloudPagedPlaylistBaseIE._API_V2_BASE,
 446         'tracks': '%s/users/%%s/tracks' % SoundcloudPagedPlaylistBaseIE._API_BASE,
 447         'sets': '%s/users/%%s/playlists' % SoundcloudPagedPlaylistBaseIE._API_V2_BASE,
 448         'reposts': '%s/profile/soundcloud:users:%%s/reposts' % SoundcloudPagedPlaylistBaseIE._API_V2_BASE,
 449         'likes': '%s/users/%%s/likes' % SoundcloudPagedPlaylistBaseIE._API_V2_BASE,
 450         'spotlight': '%s/users/%%s/spotlight' % SoundcloudPagedPlaylistBaseIE._API_V2_BASE,
 451     }
 452
 453     _TITLE_MAP = {
 454         'all': 'All',
 455         'tracks': 'Tracks',
 456         'sets': 'Playlists',
 457         'reposts': 'Reposts',
 458         'likes': 'Likes',
 459         'spotlight': 'Spotlight',
 460     }
 461
 462     def _real_extract(self, url):
 463         mobj = re.match(self._VALID_URL, url)
 464         uploader = mobj.group('user')
 465
 466         url = 'https://soundcloud.com/%s/' % uploader
 467         resolv_url = self._resolv_url(url)
 468         user = self._download_json(
 469             resolv_url, uploader, 'Downloading user info')
 470
 471         resource = mobj.group('rsrc') or 'all'
 472
 473         return self._extract_playlist(
 474             self._BASE_URL_MAP[resource] % user['id'], compat_str(user['id']),
 475             '%s (%s)' % (user['username'], self._TITLE_MAP[resource]))
 476
 477
 478 class SoundcloudTrackStationIE(SoundcloudPagedPlaylistBaseIE):
 479     _VALID_URL = r'https?://(?:(?:www|m)\.)?soundcloud\.com/stations/track/[^/]+/(?P<id>[^/?#&]+)'
 480     IE_NAME = 'soundcloud:trackstation'
 481     _TESTS = [{
 482         'url': 'https://soundcloud.com/stations/track/officialsundial/your-text',
 483         'info_dict': {
 484             'id': '286017854',
 485             'title': 'Track station: your-text',
 486         },
 487         'playlist_mincount': 47,
 488     }]
 489
 490     def _real_extract(self, url):
 491         track_name = self._match_id(url)
 492
 493         webpage = self._download_webpage(url, track_name)
 494
 495         track_id = self._search_regex(
 496             r'soundcloud:track-stations:(\d+)', webpage, 'track id')
 497
 498         return self._extract_playlist(
 499             '%s/stations/soundcloud:track-stations:%s/tracks'
 500             % (self._API_V2_BASE, track_id),
 501             track_id, 'Track station: %s' % track_name)
 502
 503
 504 class SoundcloudPlaylistIE(SoundcloudPlaylistBaseIE):
 505     _VALID_URL = r'https?://api\.soundcloud\.com/playlists/(?P<id>[0-9]+)(?:/?\?secret_token=(?P<token>[^&]+?))?$'
 506     IE_NAME = 'soundcloud:playlist'
 507     _TESTS = [{
 508         'url': 'https://api.soundcloud.com/playlists/4110309',
 509         'info_dict': {
 510             'id': '4110309',
 511             'title': 'TILT Brass - Bowery Poetry Club, August \'03 [Non-Site SCR 02]',
 512             'description': 're:.*?TILT Brass - Bowery Poetry Club',
 513         },
 514         'playlist_count': 6,
 515     }]
 516
 517     def _real_extract(self, url):
 518         mobj = re.match(self._VALID_URL, url)
 519         playlist_id = mobj.group('id')
 520         base_url = '%s//api.soundcloud.com/playlists/%s.json?' % (self.http_scheme(), playlist_id)
 521
 522         data_dict = {
 523             'client_id': self._CLIENT_ID,
 524         }
 525         token = mobj.group('token')
 526
 527         if token:
 528             data_dict['secret_token'] = token
 529
 530         data = compat_urllib_parse_urlencode(data_dict)
 531         data = self._download_json(
 532             base_url + data, playlist_id, 'Downloading playlist')
 533
 534         entries = self._extract_track_entries(data['tracks'])
 535
 536         return {
 537             '_type': 'playlist',
 538             'id': playlist_id,
 539             'title': data.get('title'),
 540             'description': data.get('description'),
 541             'entries': entries,
 542         }
 543
 544
 545 class SoundcloudSearchIE(SearchInfoExtractor, SoundcloudIE):
 546     IE_NAME = 'soundcloud:search'
 547     IE_DESC = 'Soundcloud search'
 548     _MAX_RESULTS = float('inf')
 549     _TESTS = [{
 550         'url': 'scsearch15:post-avant jazzcore',
 551         'info_dict': {
 552             'title': 'post-avant jazzcore',
 553         },
 554         'playlist_count': 15,
 555     }]
 556
 557     _SEARCH_KEY = 'scsearch'
 558     _MAX_RESULTS_PER_PAGE = 200
 559     _DEFAULT_RESULTS_PER_PAGE = 50
 560     _API_V2_BASE = 'https://api-v2.soundcloud.com'
 561
 562     def _get_collection(self, endpoint, collection_id, **query):
 563         limit = min(
 564             query.get('limit', self._DEFAULT_RESULTS_PER_PAGE),
 565             self._MAX_RESULTS_PER_PAGE)
 566         query['limit'] = limit
 567         query['client_id'] = self._CLIENT_ID
 568         query['linked_partitioning'] = '1'
 569         query['offset'] = 0
 570         data = compat_urllib_parse_urlencode(query)
 571         next_url = '{0}{1}?{2}'.format(self._API_V2_BASE, endpoint, data)
 572
 573         collected_results = 0
 574
 575         for i in itertools.count(1):
 576             response = self._download_json(
 577                 next_url, collection_id, 'Downloading page {0}'.format(i),
 578                 'Unable to download API page')
 579
 580             collection = response.get('collection', [])
 581             if not collection:
 582                 break
 583
 584             collection = list(filter(bool, collection))
 585             collected_results += len(collection)
 586
 587             for item in collection:
 588                 yield self.url_result(item['uri'], SoundcloudIE.ie_key())
 589
 590             if not collection or collected_results >= limit:
 591                 break
 592
 593             next_url = response.get('next_href')
 594             if not next_url:
 595                 break
 596
 597     def _get_n_results(self, query, n):
 598         tracks = self._get_collection('/search/tracks', query, limit=n, q=query)
 599         return self.playlist_result(tracks, playlist_title=query)