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