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

   1 # encoding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5 import itertools
   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 )
  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                             (?P<uploader>[\w\d-]+)/
  35                             (?!(?:tracks|sets(?:/.+?)?|reposts|likes|spotlight)/?(?:$|[?#]))
  36                             (?P<title>[\w\d-]+)/?
  37                             (?P<token>[^?]+?)?(?:[?].*)?$)
  38                        |(?:api\.soundcloud\.com/tracks/(?P<track_id>\d+)
  39                           (?:/?\?secret_token=(?P<secret_token>[^&]+))?)
  40                        |(?P<player>(?:w|player|p.)\.soundcloud\.com/player/?.*?url=.*)
  41                     )
  42                     '''
  43     IE_NAME = 'soundcloud'
  44     _TESTS = [
  45         {
  46             'url': 'http://soundcloud.com/ethmusic/lostin-powers-she-so-heavy',
  47             'md5': 'ebef0a451b909710ed1d7787dddbf0d7',
  48             'info_dict': {
  49                 'id': '62986583',
  50                 'ext': 'mp3',
  51                 'upload_date': '20121011',
  52                 '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',
  53                 'uploader': 'E.T. ExTerrestrial Music',
  54                 'title': 'Lostin Powers - She so Heavy (SneakPreview) Adrian Ackers Blueprint 1',
  55                 'duration': 143,
  56                 'license': 'all-rights-reserved',
  57             }
  58         },
  59         # not streamable song
  60         {
  61             'url': 'https://soundcloud.com/the-concept-band/goldrushed-mastered?in=the-concept-band/sets/the-royal-concept-ep',
  62             'info_dict': {
  63                 'id': '47127627',
  64                 'ext': 'mp3',
  65                 'title': 'Goldrushed',
  66                 'description': 'From Stockholm Sweden\r\nPovel / Magnus / Filip / David\r\nwww.theroyalconcept.com',
  67                 'uploader': 'The Royal Concept',
  68                 'upload_date': '20120521',
  69                 'duration': 227,
  70                 'license': 'all-rights-reserved',
  71             },
  72             'params': {
  73                 # rtmp
  74                 'skip_download': True,
  75             },
  76         },
  77         # private link
  78         {
  79             'url': 'https://soundcloud.com/jaimemf/youtube-dl-test-video-a-y-baw/s-8Pjrp',
  80             'md5': 'aa0dd32bfea9b0c5ef4f02aacd080604',
  81             'info_dict': {
  82                 'id': '123998367',
  83                 'ext': 'mp3',
  84                 'title': 'Youtube - Dl Test Video \'\' Ä↭',
  85                 'uploader': 'jaimeMF',
  86                 'description': 'test chars:  \"\'/\\ä↭',
  87                 'upload_date': '20131209',
  88                 'duration': 9,
  89                 'license': 'all-rights-reserved',
  90             },
  91         },
  92         # private link (alt format)
  93         {
  94             'url': 'https://api.soundcloud.com/tracks/123998367?secret_token=s-8Pjrp',
  95             'md5': 'aa0dd32bfea9b0c5ef4f02aacd080604',
  96             'info_dict': {
  97                 'id': '123998367',
  98                 'ext': 'mp3',
  99                 'title': 'Youtube - Dl Test Video \'\' Ä↭',
 100                 'uploader': 'jaimeMF',
 101                 'description': 'test chars:  \"\'/\\ä↭',
 102                 'upload_date': '20131209',
 103                 'duration': 9,
 104                 'license': 'all-rights-reserved',
 105             },
 106         },
 107         # downloadable song
 108         {
 109             'url': 'https://soundcloud.com/oddsamples/bus-brakes',
 110             'md5': '7624f2351f8a3b2e7cd51522496e7631',
 111             'info_dict': {
 112                 'id': '128590877',
 113                 'ext': 'mp3',
 114                 'title': 'Bus Brakes',
 115                 'description': 'md5:0053ca6396e8d2fd7b7e1595ef12ab66',
 116                 'uploader': 'oddsamples',
 117                 'upload_date': '20140109',
 118                 'duration': 17,
 119                 'license': 'cc-by-sa',
 120             },
 121         },
 122     ]
 123
 124     _CLIENT_ID = '02gUJC0hH2ct1EGOcYXQIzRFU91c72Ea'
 125     _IPHONE_CLIENT_ID = '376f225bf427445fc4bfb6b99b72e0bf'
 126
 127     @staticmethod
 128     def _extract_urls(webpage):
 129         return [m.group('url') for m in re.finditer(
 130             r'<iframe[^>]+src=(["\'])(?P<url>(?:https?://)?(?:w\.)?soundcloud\.com/player.+?)\1',
 131             webpage)]
 132
 133     def report_resolve(self, video_id):
 134         """Report information extraction."""
 135         self.to_screen('%s: Resolving id' % video_id)
 136
 137     @classmethod
 138     def _resolv_url(cls, url):
 139         return 'http://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID
 140
 141     def _extract_info_dict(self, info, full_title=None, quiet=False, secret_token=None):
 142         track_id = compat_str(info['id'])
 143         name = full_title or track_id
 144         if quiet:
 145             self.report_extraction(name)
 146         thumbnail = info['artwork_url']
 147         track_license = info['license']
 148         if thumbnail is not None:
 149             thumbnail = thumbnail.replace('-large', '-t500x500')
 150         ext = 'mp3'
 151         result = {
 152             'id': track_id,
 153             'uploader': info['user']['username'],
 154             'upload_date': unified_strdate(info['created_at']),
 155             'title': info['title'],
 156             'description': info['description'],
 157             'thumbnail': thumbnail,
 158             'duration': int_or_none(info.get('duration'), 1000),
 159             'webpage_url': info.get('permalink_url'),
 160             'license': track_license,
 161         }
 162         formats = []
 163         if info.get('downloadable', False):
 164             # We can build a direct link to the song
 165             format_url = (
 166                 'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(
 167                     track_id, self._CLIENT_ID))
 168             formats.append({
 169                 'format_id': 'download',
 170                 'ext': info.get('original_format', 'mp3'),
 171                 'url': format_url,
 172                 'vcodec': 'none',
 173                 'preference': 10,
 174             })
 175
 176         # We have to retrieve the url
 177         streams_url = ('http://api.soundcloud.com/i1/tracks/{0}/streams?'
 178                        'client_id={1}&secret_token={2}'.format(track_id, self._IPHONE_CLIENT_ID, secret_token))
 179         format_dict = self._download_json(
 180             streams_url,
 181             track_id, 'Downloading track url')
 182
 183         for key, stream_url in format_dict.items():
 184             if key.startswith('http'):
 185                 formats.append({
 186                     'format_id': key,
 187                     'ext': ext,
 188                     'url': stream_url,
 189                     'vcodec': 'none',
 190                 })
 191             elif key.startswith('rtmp'):
 192                 # The url doesn't have an rtmp app, we have to extract the playpath
 193                 url, path = stream_url.split('mp3:', 1)
 194                 formats.append({
 195                     'format_id': key,
 196                     'url': url,
 197                     'play_path': 'mp3:' + path,
 198                     'ext': 'flv',
 199                     'vcodec': 'none',
 200                 })
 201
 202             if not formats:
 203                 # We fallback to the stream_url in the original info, this
 204                 # cannot be always used, sometimes it can give an HTTP 404 error
 205                 formats.append({
 206                     'format_id': 'fallback',
 207                     'url': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
 208                     'ext': ext,
 209                     'vcodec': 'none',
 210                 })
 211
 212             for f in formats:
 213                 if f['format_id'].startswith('http'):
 214                     f['protocol'] = 'http'
 215                 if f['format_id'].startswith('rtmp'):
 216                     f['protocol'] = 'rtmp'
 217
 218         self._check_formats(formats, track_id)
 219         self._sort_formats(formats)
 220         result['formats'] = formats
 221
 222         return result
 223
 224     def _real_extract(self, url):
 225         mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
 226         if mobj is None:
 227             raise ExtractorError('Invalid URL: %s' % url)
 228
 229         track_id = mobj.group('track_id')
 230         token = None
 231
 232         if track_id is not None:
 233             info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 234             full_title = track_id
 235             token = mobj.group('secret_token')
 236             if token:
 237                 info_json_url += '&secret_token=' + token
 238         elif mobj.group('player'):
 239             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 240             real_url = query['url'][0]
 241             # If the token is in the query of the original url we have to
 242             # manually add it
 243             if 'secret_token' in query:
 244                 real_url += '?secret_token=' + query['secret_token'][0]
 245             return self.url_result(real_url)
 246         else:
 247             # extract uploader (which is in the url)
 248             uploader = mobj.group('uploader')
 249             # extract simple title (uploader + slug of song title)
 250             slug_title = mobj.group('title')
 251             token = mobj.group('token')
 252             full_title = resolve_title = '%s/%s' % (uploader, slug_title)
 253             if token:
 254                 resolve_title += '/%s' % token
 255
 256             self.report_resolve(full_title)
 257
 258             url = 'http://soundcloud.com/%s' % resolve_title
 259             info_json_url = self._resolv_url(url)
 260         info = self._download_json(info_json_url, full_title, 'Downloading info JSON')
 261
 262         return self._extract_info_dict(info, full_title, secret_token=token)
 263
 264
 265 class SoundcloudSetIE(SoundcloudIE):
 266     _VALID_URL = r'https?://(?:(?:www|m)\.)?soundcloud\.com/(?P<uploader>[\w\d-]+)/sets/(?P<slug_title>[\w\d-]+)(?:/(?P<token>[^?/]+))?'
 267     IE_NAME = 'soundcloud:set'
 268     _TESTS = [{
 269         'url': 'https://soundcloud.com/the-concept-band/sets/the-royal-concept-ep',
 270         'info_dict': {
 271             'id': '2284613',
 272             'title': 'The Royal Concept EP',
 273         },
 274         'playlist_mincount': 6,
 275     }, {
 276         'url': 'https://soundcloud.com/the-concept-band/sets/the-royal-concept-ep/token',
 277         'only_matching': True,
 278     }]
 279
 280     def _real_extract(self, url):
 281         mobj = re.match(self._VALID_URL, url)
 282
 283         # extract uploader (which is in the url)
 284         uploader = mobj.group('uploader')
 285         # extract simple title (uploader + slug of song title)
 286         slug_title = mobj.group('slug_title')
 287         full_title = '%s/sets/%s' % (uploader, slug_title)
 288         url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 289
 290         token = mobj.group('token')
 291         if token:
 292             full_title += '/' + token
 293             url += '/' + token
 294
 295         self.report_resolve(full_title)
 296
 297         resolv_url = self._resolv_url(url)
 298         info = self._download_json(resolv_url, full_title)
 299
 300         if 'errors' in info:
 301             msgs = (compat_str(err['error_message']) for err in info['errors'])
 302             raise ExtractorError('unable to download video webpage: %s' % ','.join(msgs))
 303
 304         entries = [self.url_result(track['permalink_url'], 'Soundcloud') for track in info['tracks']]
 305
 306         return {
 307             '_type': 'playlist',
 308             'entries': entries,
 309             'id': '%s' % info['id'],
 310             'title': info['title'],
 311         }
 312
 313
 314 class SoundcloudUserIE(SoundcloudIE):
 315     _VALID_URL = r'''(?x)
 316                         https?://
 317                             (?:(?:www|m)\.)?soundcloud\.com/
 318                             (?P<user>[^/]+)
 319                             (?:/
 320                                 (?P<rsrc>tracks|sets|reposts|likes|spotlight)
 321                             )?
 322                             /?(?:[?#].*)?$
 323                     '''
 324     IE_NAME = 'soundcloud:user'
 325     _TESTS = [{
 326         'url': 'https://soundcloud.com/the-akashic-chronicler',
 327         'info_dict': {
 328             'id': '114582580',
 329             'title': 'The Akashic Chronicler (All)',
 330         },
 331         'playlist_mincount': 111,
 332     }, {
 333         'url': 'https://soundcloud.com/the-akashic-chronicler/tracks',
 334         'info_dict': {
 335             'id': '114582580',
 336             'title': 'The Akashic Chronicler (Tracks)',
 337         },
 338         'playlist_mincount': 50,
 339     }, {
 340         'url': 'https://soundcloud.com/the-akashic-chronicler/sets',
 341         'info_dict': {
 342             'id': '114582580',
 343             'title': 'The Akashic Chronicler (Playlists)',
 344         },
 345         'playlist_mincount': 3,
 346     }, {
 347         'url': 'https://soundcloud.com/the-akashic-chronicler/reposts',
 348         'info_dict': {
 349             'id': '114582580',
 350             'title': 'The Akashic Chronicler (Reposts)',
 351         },
 352         'playlist_mincount': 7,
 353     }, {
 354         'url': 'https://soundcloud.com/the-akashic-chronicler/likes',
 355         'info_dict': {
 356             'id': '114582580',
 357             'title': 'The Akashic Chronicler (Likes)',
 358         },
 359         'playlist_mincount': 321,
 360     }, {
 361         'url': 'https://soundcloud.com/grynpyret/spotlight',
 362         'info_dict': {
 363             'id': '7098329',
 364             'title': 'Grynpyret (Spotlight)',
 365         },
 366         'playlist_mincount': 1,
 367     }]
 368
 369     _API_BASE = 'https://api.soundcloud.com'
 370     _API_V2_BASE = 'https://api-v2.soundcloud.com'
 371
 372     _BASE_URL_MAP = {
 373         'all': '%s/profile/soundcloud:users:%%s' % _API_V2_BASE,
 374         'tracks': '%s/users/%%s/tracks' % _API_BASE,
 375         'sets': '%s/users/%%s/playlists' % _API_V2_BASE,
 376         'reposts': '%s/profile/soundcloud:users:%%s/reposts' % _API_V2_BASE,
 377         'likes': '%s/users/%%s/likes' % _API_V2_BASE,
 378         'spotlight': '%s/users/%%s/spotlight' % _API_V2_BASE,
 379     }
 380
 381     _TITLE_MAP = {
 382         'all': 'All',
 383         'tracks': 'Tracks',
 384         'sets': 'Playlists',
 385         'reposts': 'Reposts',
 386         'likes': 'Likes',
 387         'spotlight': 'Spotlight',
 388     }
 389
 390     def _real_extract(self, url):
 391         mobj = re.match(self._VALID_URL, url)
 392         uploader = mobj.group('user')
 393
 394         url = 'http://soundcloud.com/%s/' % uploader
 395         resolv_url = self._resolv_url(url)
 396         user = self._download_json(
 397             resolv_url, uploader, 'Downloading user info')
 398
 399         resource = mobj.group('rsrc') or 'all'
 400         base_url = self._BASE_URL_MAP[resource] % user['id']
 401
 402         COMMON_QUERY = {
 403             'limit': 50,
 404             'client_id': self._CLIENT_ID,
 405             'linked_partitioning': '1',
 406         }
 407
 408         query = COMMON_QUERY.copy()
 409         query['offset'] = 0
 410
 411         next_href = base_url + '?' + compat_urllib_parse_urlencode(query)
 412
 413         entries = []
 414         for i in itertools.count():
 415             response = self._download_json(
 416                 next_href, uploader, 'Downloading track page %s' % (i + 1))
 417
 418             collection = response['collection']
 419             if not collection:
 420                 break
 421
 422             def resolve_permalink_url(candidates):
 423                 for cand in candidates:
 424                     if isinstance(cand, dict):
 425                         permalink_url = cand.get('permalink_url')
 426                         if permalink_url and permalink_url.startswith('http'):
 427                             return permalink_url
 428
 429             for e in collection:
 430                 permalink_url = resolve_permalink_url((e, e.get('track'), e.get('playlist')))
 431                 if permalink_url:
 432                     entries.append(self.url_result(permalink_url))
 433
 434             next_href = response.get('next_href')
 435             if not next_href:
 436                 break
 437
 438             parsed_next_href = compat_urlparse.urlparse(response['next_href'])
 439             qs = compat_urlparse.parse_qs(parsed_next_href.query)
 440             qs.update(COMMON_QUERY)
 441             next_href = compat_urlparse.urlunparse(
 442                 parsed_next_href._replace(query=compat_urllib_parse_urlencode(qs, True)))
 443
 444         return {
 445             '_type': 'playlist',
 446             'id': compat_str(user['id']),
 447             'title': '%s (%s)' % (user['username'], self._TITLE_MAP[resource]),
 448             'entries': entries,
 449         }
 450
 451
 452 class SoundcloudPlaylistIE(SoundcloudIE):
 453     _VALID_URL = r'https?://api\.soundcloud\.com/playlists/(?P<id>[0-9]+)(?:/?\?secret_token=(?P<token>[^&]+?))?$'
 454     IE_NAME = 'soundcloud:playlist'
 455     _TESTS = [{
 456         'url': 'http://api.soundcloud.com/playlists/4110309',
 457         'info_dict': {
 458             'id': '4110309',
 459             'title': 'TILT Brass - Bowery Poetry Club, August \'03 [Non-Site SCR 02]',
 460             'description': 're:.*?TILT Brass - Bowery Poetry Club',
 461         },
 462         'playlist_count': 6,
 463     }]
 464
 465     def _real_extract(self, url):
 466         mobj = re.match(self._VALID_URL, url)
 467         playlist_id = mobj.group('id')
 468         base_url = '%s//api.soundcloud.com/playlists/%s.json?' % (self.http_scheme(), playlist_id)
 469
 470         data_dict = {
 471             'client_id': self._CLIENT_ID,
 472         }
 473         token = mobj.group('token')
 474
 475         if token:
 476             data_dict['secret_token'] = token
 477
 478         data = compat_urllib_parse_urlencode(data_dict)
 479         data = self._download_json(
 480             base_url + data, playlist_id, 'Downloading playlist')
 481
 482         entries = [self.url_result(track['permalink_url'], 'Soundcloud') for track in data['tracks']]
 483
 484         return {
 485             '_type': 'playlist',
 486             'id': playlist_id,
 487             'title': data.get('title'),
 488             'description': data.get('description'),
 489             'entries': entries,
 490         }
 491
 492
 493 class SoundcloudSearchIE(SearchInfoExtractor, SoundcloudIE):
 494     IE_NAME = 'soundcloud:search'
 495     IE_DESC = 'Soundcloud search'
 496     _MAX_RESULTS = float('inf')
 497     _TESTS = [{
 498         'url': 'scsearch15:post-avant jazzcore',
 499         'info_dict': {
 500             'title': 'post-avant jazzcore',
 501         },
 502         'playlist_count': 15,
 503     }]
 504
 505     _SEARCH_KEY = 'scsearch'
 506     _MAX_RESULTS_PER_PAGE = 200
 507     _DEFAULT_RESULTS_PER_PAGE = 50
 508     _API_V2_BASE = 'https://api-v2.soundcloud.com'
 509
 510     def _get_collection(self, endpoint, collection_id, **query):
 511         limit = min(
 512             query.get('limit', self._DEFAULT_RESULTS_PER_PAGE),
 513             self._MAX_RESULTS_PER_PAGE)
 514         query['limit'] = limit
 515         query['client_id'] = self._CLIENT_ID
 516         query['linked_partitioning'] = '1'
 517         query['offset'] = 0
 518         data = compat_urllib_parse_urlencode(query)
 519         next_url = '{0}{1}?{2}'.format(self._API_V2_BASE, endpoint, data)
 520
 521         collected_results = 0
 522
 523         for i in itertools.count(1):
 524             response = self._download_json(
 525                 next_url, collection_id, 'Downloading page {0}'.format(i),
 526                 'Unable to download API page')
 527
 528             collection = response.get('collection', [])
 529             if not collection:
 530                 break
 531
 532             collection = list(filter(bool, collection))
 533             collected_results += len(collection)
 534
 535             for item in collection:
 536                 yield self.url_result(item['uri'], SoundcloudIE.ie_key())
 537
 538             if not collection or collected_results >= limit:
 539                 break
 540
 541             next_url = response.get('next_href')
 542             if not next_url:
 543                 break
 544
 545     def _get_n_results(self, query, n):
 546         tracks = self._get_collection('/search/tracks', query, limit=n, q=query)
 547         return self.playlist_result(tracks, playlist_title=query)