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

   1 # coding: 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 = 'fDoItMDbsbZz8dY16ZzARCZmzgHBPotA'
 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         format_dict = self._download_json(
 177             'http://api.soundcloud.com/i1/tracks/%s/streams' % track_id,
 178             track_id, 'Downloading track url', query={
 179                 'client_id': self._CLIENT_ID,
 180                 'secret_token': secret_token,
 181             })
 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
 231         if track_id is not None:
 232             info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 233             full_title = track_id
 234             token = mobj.group('secret_token')
 235             if token:
 236                 info_json_url += '&secret_token=' + token
 237         elif mobj.group('player'):
 238             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 239             real_url = query['url'][0]
 240             # If the token is in the query of the original url we have to
 241             # manually add it
 242             if 'secret_token' in query:
 243                 real_url += '?secret_token=' + query['secret_token'][0]
 244             return self.url_result(real_url)
 245         else:
 246             # extract uploader (which is in the url)
 247             uploader = mobj.group('uploader')
 248             # extract simple title (uploader + slug of song title)
 249             slug_title = mobj.group('title')
 250             token = mobj.group('token')
 251             full_title = resolve_title = '%s/%s' % (uploader, slug_title)
 252             if token:
 253                 resolve_title += '/%s' % token
 254
 255             self.report_resolve(full_title)
 256
 257             url = 'http://soundcloud.com/%s' % resolve_title
 258             info_json_url = self._resolv_url(url)
 259         info = self._download_json(info_json_url, full_title, 'Downloading info JSON')
 260
 261         return self._extract_info_dict(info, full_title, secret_token=token)
 262
 263
 264 class SoundcloudPlaylistBaseIE(SoundcloudIE):
 265     @staticmethod
 266     def _extract_id(e):
 267         return compat_str(e['id']) if e.get('id') else None
 268
 269     def _extract_track_entries(self, tracks):
 270         return [
 271             self.url_result(
 272                 track['permalink_url'], SoundcloudIE.ie_key(),
 273                 video_id=self._extract_id(track))
 274             for track in tracks if track.get('permalink_url')]
 275
 276
 277 class SoundcloudSetIE(SoundcloudPlaylistBaseIE):
 278     _VALID_URL = r'https?://(?:(?:www|m)\.)?soundcloud\.com/(?P<uploader>[\w\d-]+)/sets/(?P<slug_title>[\w\d-]+)(?:/(?P<token>[^?/]+))?'
 279     IE_NAME = 'soundcloud:set'
 280     _TESTS = [{
 281         'url': 'https://soundcloud.com/the-concept-band/sets/the-royal-concept-ep',
 282         'info_dict': {
 283             'id': '2284613',
 284             'title': 'The Royal Concept EP',
 285         },
 286         'playlist_mincount': 6,
 287     }, {
 288         'url': 'https://soundcloud.com/the-concept-band/sets/the-royal-concept-ep/token',
 289         'only_matching': True,
 290     }]
 291
 292     def _real_extract(self, url):
 293         mobj = re.match(self._VALID_URL, url)
 294
 295         # extract uploader (which is in the url)
 296         uploader = mobj.group('uploader')
 297         # extract simple title (uploader + slug of song title)
 298         slug_title = mobj.group('slug_title')
 299         full_title = '%s/sets/%s' % (uploader, slug_title)
 300         url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 301
 302         token = mobj.group('token')
 303         if token:
 304             full_title += '/' + token
 305             url += '/' + token
 306
 307         self.report_resolve(full_title)
 308
 309         resolv_url = self._resolv_url(url)
 310         info = self._download_json(resolv_url, full_title)
 311
 312         if 'errors' in info:
 313             msgs = (compat_str(err['error_message']) for err in info['errors'])
 314             raise ExtractorError('unable to download video webpage: %s' % ','.join(msgs))
 315
 316         entries = self._extract_track_entries(info['tracks'])
 317
 318         return {
 319             '_type': 'playlist',
 320             'entries': entries,
 321             'id': '%s' % info['id'],
 322             'title': info['title'],
 323         }
 324
 325
 326 class SoundcloudUserIE(SoundcloudPlaylistBaseIE):
 327     _VALID_URL = r'''(?x)
 328                         https?://
 329                             (?:(?:www|m)\.)?soundcloud\.com/
 330                             (?P<user>[^/]+)
 331                             (?:/
 332                                 (?P<rsrc>tracks|sets|reposts|likes|spotlight)
 333                             )?
 334                             /?(?:[?#].*)?$
 335                     '''
 336     IE_NAME = 'soundcloud:user'
 337     _TESTS = [{
 338         'url': 'https://soundcloud.com/the-akashic-chronicler',
 339         'info_dict': {
 340             'id': '114582580',
 341             'title': 'The Akashic Chronicler (All)',
 342         },
 343         'playlist_mincount': 74,
 344     }, {
 345         'url': 'https://soundcloud.com/the-akashic-chronicler/tracks',
 346         'info_dict': {
 347             'id': '114582580',
 348             'title': 'The Akashic Chronicler (Tracks)',
 349         },
 350         'playlist_mincount': 37,
 351     }, {
 352         'url': 'https://soundcloud.com/the-akashic-chronicler/sets',
 353         'info_dict': {
 354             'id': '114582580',
 355             'title': 'The Akashic Chronicler (Playlists)',
 356         },
 357         'playlist_mincount': 2,
 358     }, {
 359         'url': 'https://soundcloud.com/the-akashic-chronicler/reposts',
 360         'info_dict': {
 361             'id': '114582580',
 362             'title': 'The Akashic Chronicler (Reposts)',
 363         },
 364         'playlist_mincount': 7,
 365     }, {
 366         'url': 'https://soundcloud.com/the-akashic-chronicler/likes',
 367         'info_dict': {
 368             'id': '114582580',
 369             'title': 'The Akashic Chronicler (Likes)',
 370         },
 371         'playlist_mincount': 321,
 372     }, {
 373         'url': 'https://soundcloud.com/grynpyret/spotlight',
 374         'info_dict': {
 375             'id': '7098329',
 376             'title': 'GRYNPYRET (Spotlight)',
 377         },
 378         'playlist_mincount': 1,
 379     }]
 380
 381     _API_BASE = 'https://api.soundcloud.com'
 382     _API_V2_BASE = 'https://api-v2.soundcloud.com'
 383
 384     _BASE_URL_MAP = {
 385         'all': '%s/profile/soundcloud:users:%%s' % _API_V2_BASE,
 386         'tracks': '%s/users/%%s/tracks' % _API_BASE,
 387         'sets': '%s/users/%%s/playlists' % _API_V2_BASE,
 388         'reposts': '%s/profile/soundcloud:users:%%s/reposts' % _API_V2_BASE,
 389         'likes': '%s/users/%%s/likes' % _API_V2_BASE,
 390         'spotlight': '%s/users/%%s/spotlight' % _API_V2_BASE,
 391     }
 392
 393     _TITLE_MAP = {
 394         'all': 'All',
 395         'tracks': 'Tracks',
 396         'sets': 'Playlists',
 397         'reposts': 'Reposts',
 398         'likes': 'Likes',
 399         'spotlight': 'Spotlight',
 400     }
 401
 402     def _real_extract(self, url):
 403         mobj = re.match(self._VALID_URL, url)
 404         uploader = mobj.group('user')
 405
 406         url = 'http://soundcloud.com/%s/' % uploader
 407         resolv_url = self._resolv_url(url)
 408         user = self._download_json(
 409             resolv_url, uploader, 'Downloading user info')
 410
 411         resource = mobj.group('rsrc') or 'all'
 412         base_url = self._BASE_URL_MAP[resource] % user['id']
 413
 414         COMMON_QUERY = {
 415             'limit': 50,
 416             'client_id': self._CLIENT_ID,
 417             'linked_partitioning': '1',
 418         }
 419
 420         query = COMMON_QUERY.copy()
 421         query['offset'] = 0
 422
 423         next_href = base_url + '?' + compat_urllib_parse_urlencode(query)
 424
 425         entries = []
 426         for i in itertools.count():
 427             response = self._download_json(
 428                 next_href, uploader, 'Downloading track page %s' % (i + 1))
 429
 430             collection = response['collection']
 431             if not collection:
 432                 break
 433
 434             def resolve_permalink_url(candidates):
 435                 for cand in candidates:
 436                     if isinstance(cand, dict):
 437                         permalink_url = cand.get('permalink_url')
 438                         entry_id = self._extract_id(cand)
 439                         if permalink_url and permalink_url.startswith('http'):
 440                             return permalink_url, entry_id
 441
 442             for e in collection:
 443                 permalink_url, entry_id = resolve_permalink_url((e, e.get('track'), e.get('playlist')))
 444                 if permalink_url:
 445                     entries.append(self.url_result(permalink_url, video_id=entry_id))
 446
 447             next_href = response.get('next_href')
 448             if not next_href:
 449                 break
 450
 451             parsed_next_href = compat_urlparse.urlparse(response['next_href'])
 452             qs = compat_urlparse.parse_qs(parsed_next_href.query)
 453             qs.update(COMMON_QUERY)
 454             next_href = compat_urlparse.urlunparse(
 455                 parsed_next_href._replace(query=compat_urllib_parse_urlencode(qs, True)))
 456
 457         return {
 458             '_type': 'playlist',
 459             'id': compat_str(user['id']),
 460             'title': '%s (%s)' % (user['username'], self._TITLE_MAP[resource]),
 461             'entries': entries,
 462         }
 463
 464
 465 class SoundcloudPlaylistIE(SoundcloudPlaylistBaseIE):
 466     _VALID_URL = r'https?://api\.soundcloud\.com/playlists/(?P<id>[0-9]+)(?:/?\?secret_token=(?P<token>[^&]+?))?$'
 467     IE_NAME = 'soundcloud:playlist'
 468     _TESTS = [{
 469         'url': 'http://api.soundcloud.com/playlists/4110309',
 470         'info_dict': {
 471             'id': '4110309',
 472             'title': 'TILT Brass - Bowery Poetry Club, August \'03 [Non-Site SCR 02]',
 473             'description': 're:.*?TILT Brass - Bowery Poetry Club',
 474         },
 475         'playlist_count': 6,
 476     }]
 477
 478     def _real_extract(self, url):
 479         mobj = re.match(self._VALID_URL, url)
 480         playlist_id = mobj.group('id')
 481         base_url = '%s//api.soundcloud.com/playlists/%s.json?' % (self.http_scheme(), playlist_id)
 482
 483         data_dict = {
 484             'client_id': self._CLIENT_ID,
 485         }
 486         token = mobj.group('token')
 487
 488         if token:
 489             data_dict['secret_token'] = token
 490
 491         data = compat_urllib_parse_urlencode(data_dict)
 492         data = self._download_json(
 493             base_url + data, playlist_id, 'Downloading playlist')
 494
 495         entries = self._extract_track_entries(data['tracks'])
 496
 497         return {
 498             '_type': 'playlist',
 499             'id': playlist_id,
 500             'title': data.get('title'),
 501             'description': data.get('description'),
 502             'entries': entries,
 503         }
 504
 505
 506 class SoundcloudSearchIE(SearchInfoExtractor, SoundcloudIE):
 507     IE_NAME = 'soundcloud:search'
 508     IE_DESC = 'Soundcloud search'
 509     _MAX_RESULTS = float('inf')
 510     _TESTS = [{
 511         'url': 'scsearch15:post-avant jazzcore',
 512         'info_dict': {
 513             'title': 'post-avant jazzcore',
 514         },
 515         'playlist_count': 15,
 516     }]
 517
 518     _SEARCH_KEY = 'scsearch'
 519     _MAX_RESULTS_PER_PAGE = 200
 520     _DEFAULT_RESULTS_PER_PAGE = 50
 521     _API_V2_BASE = 'https://api-v2.soundcloud.com'
 522
 523     def _get_collection(self, endpoint, collection_id, **query):
 524         limit = min(
 525             query.get('limit', self._DEFAULT_RESULTS_PER_PAGE),
 526             self._MAX_RESULTS_PER_PAGE)
 527         query['limit'] = limit
 528         query['client_id'] = self._CLIENT_ID
 529         query['linked_partitioning'] = '1'
 530         query['offset'] = 0
 531         data = compat_urllib_parse_urlencode(query)
 532         next_url = '{0}{1}?{2}'.format(self._API_V2_BASE, endpoint, data)
 533
 534         collected_results = 0
 535
 536         for i in itertools.count(1):
 537             response = self._download_json(
 538                 next_url, collection_id, 'Downloading page {0}'.format(i),
 539                 'Unable to download API page')
 540
 541             collection = response.get('collection', [])
 542             if not collection:
 543                 break
 544
 545             collection = list(filter(bool, collection))
 546             collected_results += len(collection)
 547
 548             for item in collection:
 549                 yield self.url_result(item['uri'], SoundcloudIE.ie_key())
 550
 551             if not collection or collected_results >= limit:
 552                 break
 553
 554             next_url = response.get('next_href')
 555             if not next_url:
 556                 break
 557
 558     def _get_n_results(self, query, n):
 559         tracks = self._get_collection('/search/tracks', query, limit=n, q=query)
 560         return self.playlist_result(tracks, playlist_title=query)