_ 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.get('artwork_url')
 147         if isinstance(thumbnail, compat_str):
 148             thumbnail = thumbnail.replace('-large', '-t500x500')
 149         ext = 'mp3'
 150         result = {
 151             'id': track_id,
 152             'uploader': info.get('user', {}).get('username'),
 153             'upload_date': unified_strdate(info.get('created_at')),
 154             'title': info['title'],
 155             'description': info.get('description'),
 156             'thumbnail': thumbnail,
 157             'duration': int_or_none(info.get('duration'), 1000),
 158             'webpage_url': info.get('permalink_url'),
 159             'license': info.get('license'),
 160         }
 161         formats = []
 162         if info.get('downloadable', False):
 163             # We can build a direct link to the song
 164             format_url = (
 165                 'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(
 166                     track_id, self._CLIENT_ID))
 167             formats.append({
 168                 'format_id': 'download',
 169                 'ext': info.get('original_format', 'mp3'),
 170                 'url': format_url,
 171                 'vcodec': 'none',
 172                 'preference': 10,
 173             })
 174
 175         # We have to retrieve the url
 176         streams_url = ('http://api.soundcloud.com/i1/tracks/{0}/streams?'
 177                        'client_id={1}&secret_token={2}'.format(track_id, self._IPHONE_CLIENT_ID, secret_token))
 178         format_dict = self._download_json(
 179             streams_url,
 180             track_id, 'Downloading track url')
 181
 182         for key, stream_url in format_dict.items():
 183             if key.startswith('http'):
 184                 formats.append({
 185                     'format_id': key,
 186                     'ext': ext,
 187                     'url': stream_url,
 188                     'vcodec': 'none',
 189                 })
 190             elif key.startswith('rtmp'):
 191                 # The url doesn't have an rtmp app, we have to extract the playpath
 192                 url, path = stream_url.split('mp3:', 1)
 193                 formats.append({
 194                     'format_id': key,
 195                     'url': url,
 196                     'play_path': 'mp3:' + path,
 197                     'ext': 'flv',
 198                     'vcodec': 'none',
 199                 })
 200
 201             if not formats:
 202                 # We fallback to the stream_url in the original info, this
 203                 # cannot be always used, sometimes it can give an HTTP 404 error
 204                 formats.append({
 205                     'format_id': 'fallback',
 206                     'url': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
 207                     'ext': ext,
 208                     'vcodec': 'none',
 209                 })
 210
 211             for f in formats:
 212                 if f['format_id'].startswith('http'):
 213                     f['protocol'] = 'http'
 214                 if f['format_id'].startswith('rtmp'):
 215                     f['protocol'] = 'rtmp'
 216
 217         self._check_formats(formats, track_id)
 218         self._sort_formats(formats)
 219         result['formats'] = formats
 220
 221         return result
 222
 223     def _real_extract(self, url):
 224         mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
 225         if mobj is None:
 226             raise ExtractorError('Invalid URL: %s' % url)
 227
 228         track_id = mobj.group('track_id')
 229
 230         if track_id is not None:
 231             info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 232             full_title = track_id
 233             token = mobj.group('secret_token')
 234             if token:
 235                 info_json_url += '&secret_token=' + token
 236         elif mobj.group('player'):
 237             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 238             real_url = query['url'][0]
 239             # If the token is in the query of the original url we have to
 240             # manually add it
 241             if 'secret_token' in query:
 242                 real_url += '?secret_token=' + query['secret_token'][0]
 243             return self.url_result(real_url)
 244         else:
 245             # extract uploader (which is in the url)
 246             uploader = mobj.group('uploader')
 247             # extract simple title (uploader + slug of song title)
 248             slug_title = mobj.group('title')
 249             token = mobj.group('token')
 250             full_title = resolve_title = '%s/%s' % (uploader, slug_title)
 251             if token:
 252                 resolve_title += '/%s' % token
 253
 254             self.report_resolve(full_title)
 255
 256             url = 'http://soundcloud.com/%s' % resolve_title
 257             info_json_url = self._resolv_url(url)
 258         info = self._download_json(info_json_url, full_title, 'Downloading info JSON')
 259
 260         return self._extract_info_dict(info, full_title, secret_token=token)
 261
 262
 263 class SoundcloudPlaylistBaseIE(SoundcloudIE):
 264     @staticmethod
 265     def _extract_id(e):
 266         return compat_str(e['id']) if e.get('id') else None
 267
 268     def _extract_track_entries(self, tracks):
 269         return [
 270             self.url_result(
 271                 track['permalink_url'], SoundcloudIE.ie_key(),
 272                 video_id=self._extract_id(track))
 273             for track in tracks if track.get('permalink_url')]
 274
 275
 276 class SoundcloudSetIE(SoundcloudPlaylistBaseIE):
 277     _VALID_URL = r'https?://(?:(?:www|m)\.)?soundcloud\.com/(?P<uploader>[\w\d-]+)/sets/(?P<slug_title>[\w\d-]+)(?:/(?P<token>[^?/]+))?'
 278     IE_NAME = 'soundcloud:set'
 279     _TESTS = [{
 280         'url': 'https://soundcloud.com/the-concept-band/sets/the-royal-concept-ep',
 281         'info_dict': {
 282             'id': '2284613',
 283             'title': 'The Royal Concept EP',
 284         },
 285         'playlist_mincount': 6,
 286     }, {
 287         'url': 'https://soundcloud.com/the-concept-band/sets/the-royal-concept-ep/token',
 288         'only_matching': True,
 289     }]
 290
 291     def _real_extract(self, url):
 292         mobj = re.match(self._VALID_URL, url)
 293
 294         # extract uploader (which is in the url)
 295         uploader = mobj.group('uploader')
 296         # extract simple title (uploader + slug of song title)
 297         slug_title = mobj.group('slug_title')
 298         full_title = '%s/sets/%s' % (uploader, slug_title)
 299         url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 300
 301         token = mobj.group('token')
 302         if token:
 303             full_title += '/' + token
 304             url += '/' + token
 305
 306         self.report_resolve(full_title)
 307
 308         resolv_url = self._resolv_url(url)
 309         info = self._download_json(resolv_url, full_title)
 310
 311         if 'errors' in info:
 312             msgs = (compat_str(err['error_message']) for err in info['errors'])
 313             raise ExtractorError('unable to download video webpage: %s' % ','.join(msgs))
 314
 315         entries = self._extract_track_entries(info['tracks'])
 316
 317         return {
 318             '_type': 'playlist',
 319             'entries': entries,
 320             'id': '%s' % info['id'],
 321             'title': info['title'],
 322         }
 323
 324
 325 class SoundcloudUserIE(SoundcloudPlaylistBaseIE):
 326     _VALID_URL = r'''(?x)
 327                         https?://
 328                             (?:(?:www|m)\.)?soundcloud\.com/
 329                             (?P<user>[^/]+)
 330                             (?:/
 331                                 (?P<rsrc>tracks|sets|reposts|likes|spotlight)
 332                             )?
 333                             /?(?:[?#].*)?$
 334                     '''
 335     IE_NAME = 'soundcloud:user'
 336     _TESTS = [{
 337         'url': 'https://soundcloud.com/the-akashic-chronicler',
 338         'info_dict': {
 339             'id': '114582580',
 340             'title': 'The Akashic Chronicler (All)',
 341         },
 342         'playlist_mincount': 74,
 343     }, {
 344         'url': 'https://soundcloud.com/the-akashic-chronicler/tracks',
 345         'info_dict': {
 346             'id': '114582580',
 347             'title': 'The Akashic Chronicler (Tracks)',
 348         },
 349         'playlist_mincount': 37,
 350     }, {
 351         'url': 'https://soundcloud.com/the-akashic-chronicler/sets',
 352         'info_dict': {
 353             'id': '114582580',
 354             'title': 'The Akashic Chronicler (Playlists)',
 355         },
 356         'playlist_mincount': 2,
 357     }, {
 358         'url': 'https://soundcloud.com/the-akashic-chronicler/reposts',
 359         'info_dict': {
 360             'id': '114582580',
 361             'title': 'The Akashic Chronicler (Reposts)',
 362         },
 363         'playlist_mincount': 7,
 364     }, {
 365         'url': 'https://soundcloud.com/the-akashic-chronicler/likes',
 366         'info_dict': {
 367             'id': '114582580',
 368             'title': 'The Akashic Chronicler (Likes)',
 369         },
 370         'playlist_mincount': 321,
 371     }, {
 372         'url': 'https://soundcloud.com/grynpyret/spotlight',
 373         'info_dict': {
 374             'id': '7098329',
 375             'title': 'GRYNPYRET (Spotlight)',
 376         },
 377         'playlist_mincount': 1,
 378     }]
 379
 380     _API_BASE = 'https://api.soundcloud.com'
 381     _API_V2_BASE = 'https://api-v2.soundcloud.com'
 382
 383     _BASE_URL_MAP = {
 384         'all': '%s/profile/soundcloud:users:%%s' % _API_V2_BASE,
 385         'tracks': '%s/users/%%s/tracks' % _API_BASE,
 386         'sets': '%s/users/%%s/playlists' % _API_V2_BASE,
 387         'reposts': '%s/profile/soundcloud:users:%%s/reposts' % _API_V2_BASE,
 388         'likes': '%s/users/%%s/likes' % _API_V2_BASE,
 389         'spotlight': '%s/users/%%s/spotlight' % _API_V2_BASE,
 390     }
 391
 392     _TITLE_MAP = {
 393         'all': 'All',
 394         'tracks': 'Tracks',
 395         'sets': 'Playlists',
 396         'reposts': 'Reposts',
 397         'likes': 'Likes',
 398         'spotlight': 'Spotlight',
 399     }
 400
 401     def _real_extract(self, url):
 402         mobj = re.match(self._VALID_URL, url)
 403         uploader = mobj.group('user')
 404
 405         url = 'http://soundcloud.com/%s/' % uploader
 406         resolv_url = self._resolv_url(url)
 407         user = self._download_json(
 408             resolv_url, uploader, 'Downloading user info')
 409
 410         resource = mobj.group('rsrc') or 'all'
 411         base_url = self._BASE_URL_MAP[resource] % user['id']
 412
 413         COMMON_QUERY = {
 414             'limit': 50,
 415             'client_id': self._CLIENT_ID,
 416             'linked_partitioning': '1',
 417         }
 418
 419         query = COMMON_QUERY.copy()
 420         query['offset'] = 0
 421
 422         next_href = base_url + '?' + compat_urllib_parse_urlencode(query)
 423
 424         entries = []
 425         for i in itertools.count():
 426             response = self._download_json(
 427                 next_href, uploader, 'Downloading track page %s' % (i + 1))
 428
 429             collection = response['collection']
 430             if not collection:
 431                 break
 432
 433             def resolve_permalink_url(candidates):
 434                 for cand in candidates:
 435                     if isinstance(cand, dict):
 436                         permalink_url = cand.get('permalink_url')
 437                         entry_id = self._extract_id(cand)
 438                         if permalink_url and permalink_url.startswith('http'):
 439                             return permalink_url, entry_id
 440
 441             for e in collection:
 442                 permalink_url, entry_id = resolve_permalink_url((e, e.get('track'), e.get('playlist')))
 443                 if permalink_url:
 444                     entries.append(self.url_result(permalink_url, video_id=entry_id))
 445
 446             next_href = response.get('next_href')
 447             if not next_href:
 448                 break
 449
 450             parsed_next_href = compat_urlparse.urlparse(response['next_href'])
 451             qs = compat_urlparse.parse_qs(parsed_next_href.query)
 452             qs.update(COMMON_QUERY)
 453             next_href = compat_urlparse.urlunparse(
 454                 parsed_next_href._replace(query=compat_urllib_parse_urlencode(qs, True)))
 455
 456         return {
 457             '_type': 'playlist',
 458             'id': compat_str(user['id']),
 459             'title': '%s (%s)' % (user['username'], self._TITLE_MAP[resource]),
 460             'entries': entries,
 461         }
 462
 463
 464 class SoundcloudPlaylistIE(SoundcloudPlaylistBaseIE):
 465     _VALID_URL = r'https?://api\.soundcloud\.com/playlists/(?P<id>[0-9]+)(?:/?\?secret_token=(?P<token>[^&]+?))?$'
 466     IE_NAME = 'soundcloud:playlist'
 467     _TESTS = [{
 468         'url': 'http://api.soundcloud.com/playlists/4110309',
 469         'info_dict': {
 470             'id': '4110309',
 471             'title': 'TILT Brass - Bowery Poetry Club, August \'03 [Non-Site SCR 02]',
 472             'description': 're:.*?TILT Brass - Bowery Poetry Club',
 473         },
 474         'playlist_count': 6,
 475     }]
 476
 477     def _real_extract(self, url):
 478         mobj = re.match(self._VALID_URL, url)
 479         playlist_id = mobj.group('id')
 480         base_url = '%s//api.soundcloud.com/playlists/%s.json?' % (self.http_scheme(), playlist_id)
 481
 482         data_dict = {
 483             'client_id': self._CLIENT_ID,
 484         }
 485         token = mobj.group('token')
 486
 487         if token:
 488             data_dict['secret_token'] = token
 489
 490         data = compat_urllib_parse_urlencode(data_dict)
 491         data = self._download_json(
 492             base_url + data, playlist_id, 'Downloading playlist')
 493
 494         entries = self._extract_track_entries(data['tracks'])
 495
 496         return {
 497             '_type': 'playlist',
 498             'id': playlist_id,
 499             'title': data.get('title'),
 500             'description': data.get('description'),
 501             'entries': entries,
 502         }
 503
 504
 505 class SoundcloudSearchIE(SearchInfoExtractor, SoundcloudIE):
 506     IE_NAME = 'soundcloud:search'
 507     IE_DESC = 'Soundcloud search'
 508     _MAX_RESULTS = float('inf')
 509     _TESTS = [{
 510         'url': 'scsearch15:post-avant jazzcore',
 511         'info_dict': {
 512             'title': 'post-avant jazzcore',
 513         },
 514         'playlist_count': 15,
 515     }]
 516
 517     _SEARCH_KEY = 'scsearch'
 518     _MAX_RESULTS_PER_PAGE = 200
 519     _DEFAULT_RESULTS_PER_PAGE = 50
 520     _API_V2_BASE = 'https://api-v2.soundcloud.com'
 521
 522     def _get_collection(self, endpoint, collection_id, **query):
 523         limit = min(
 524             query.get('limit', self._DEFAULT_RESULTS_PER_PAGE),
 525             self._MAX_RESULTS_PER_PAGE)
 526         query['limit'] = limit
 527         query['client_id'] = self._CLIENT_ID
 528         query['linked_partitioning'] = '1'
 529         query['offset'] = 0
 530         data = compat_urllib_parse_urlencode(query)
 531         next_url = '{0}{1}?{2}'.format(self._API_V2_BASE, endpoint, data)
 532
 533         collected_results = 0
 534
 535         for i in itertools.count(1):
 536             response = self._download_json(
 537                 next_url, collection_id, 'Downloading page {0}'.format(i),
 538                 'Unable to download API page')
 539
 540             collection = response.get('collection', [])
 541             if not collection:
 542                 break
 543
 544             collection = list(filter(bool, collection))
 545             collected_results += len(collection)
 546
 547             for item in collection:
 548                 yield self.url_result(item['uri'], SoundcloudIE.ie_key())
 549
 550             if not collection or collected_results >= limit:
 551                 break
 552
 553             next_url = response.get('next_href')
 554             if not next_url:
 555                 break
 556
 557     def _get_n_results(self, query, n):
 558         tracks = self._get_collection('/search/tracks', query, limit=n, q=query)
 559         return self.playlist_result(tracks, playlist_title=query)