_ 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 = '2t9loNQH90kzJcsFCODdigxfp325aq4z'
 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 'https://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             'https://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             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': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
 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 SoundcloudUserIE(SoundcloudPlaylistBaseIE):
 334     _VALID_URL = r'''(?x)
 335                         https?://
 336                             (?:(?:www|m)\.)?soundcloud\.com/
 337                             (?P<user>[^/]+)
 338                             (?:/
 339                                 (?P<rsrc>tracks|sets|reposts|likes|spotlight)
 340                             )?
 341                             /?(?:[?#].*)?$
 342                     '''
 343     IE_NAME = 'soundcloud:user'
 344     _TESTS = [{
 345         'url': 'https://soundcloud.com/the-akashic-chronicler',
 346         'info_dict': {
 347             'id': '114582580',
 348             'title': 'The Akashic Chronicler (All)',
 349         },
 350         'playlist_mincount': 74,
 351     }, {
 352         'url': 'https://soundcloud.com/the-akashic-chronicler/tracks',
 353         'info_dict': {
 354             'id': '114582580',
 355             'title': 'The Akashic Chronicler (Tracks)',
 356         },
 357         'playlist_mincount': 37,
 358     }, {
 359         'url': 'https://soundcloud.com/the-akashic-chronicler/sets',
 360         'info_dict': {
 361             'id': '114582580',
 362             'title': 'The Akashic Chronicler (Playlists)',
 363         },
 364         'playlist_mincount': 2,
 365     }, {
 366         'url': 'https://soundcloud.com/the-akashic-chronicler/reposts',
 367         'info_dict': {
 368             'id': '114582580',
 369             'title': 'The Akashic Chronicler (Reposts)',
 370         },
 371         'playlist_mincount': 7,
 372     }, {
 373         'url': 'https://soundcloud.com/the-akashic-chronicler/likes',
 374         'info_dict': {
 375             'id': '114582580',
 376             'title': 'The Akashic Chronicler (Likes)',
 377         },
 378         'playlist_mincount': 321,
 379     }, {
 380         'url': 'https://soundcloud.com/grynpyret/spotlight',
 381         'info_dict': {
 382             'id': '7098329',
 383             'title': 'Grynpyret (Spotlight)',
 384         },
 385         'playlist_mincount': 1,
 386     }]
 387
 388     _API_BASE = 'https://api.soundcloud.com'
 389     _API_V2_BASE = 'https://api-v2.soundcloud.com'
 390
 391     _BASE_URL_MAP = {
 392         'all': '%s/profile/soundcloud:users:%%s' % _API_V2_BASE,
 393         'tracks': '%s/users/%%s/tracks' % _API_BASE,
 394         'sets': '%s/users/%%s/playlists' % _API_V2_BASE,
 395         'reposts': '%s/profile/soundcloud:users:%%s/reposts' % _API_V2_BASE,
 396         'likes': '%s/users/%%s/likes' % _API_V2_BASE,
 397         'spotlight': '%s/users/%%s/spotlight' % _API_V2_BASE,
 398     }
 399
 400     _TITLE_MAP = {
 401         'all': 'All',
 402         'tracks': 'Tracks',
 403         'sets': 'Playlists',
 404         'reposts': 'Reposts',
 405         'likes': 'Likes',
 406         'spotlight': 'Spotlight',
 407     }
 408
 409     def _real_extract(self, url):
 410         mobj = re.match(self._VALID_URL, url)
 411         uploader = mobj.group('user')
 412
 413         url = 'https://soundcloud.com/%s/' % uploader
 414         resolv_url = self._resolv_url(url)
 415         user = self._download_json(
 416             resolv_url, uploader, 'Downloading user info')
 417
 418         resource = mobj.group('rsrc') or 'all'
 419         base_url = self._BASE_URL_MAP[resource] % user['id']
 420
 421         COMMON_QUERY = {
 422             'limit': 50,
 423             'client_id': self._CLIENT_ID,
 424             'linked_partitioning': '1',
 425         }
 426
 427         query = COMMON_QUERY.copy()
 428         query['offset'] = 0
 429
 430         next_href = base_url + '?' + compat_urllib_parse_urlencode(query)
 431
 432         entries = []
 433         for i in itertools.count():
 434             response = self._download_json(
 435                 next_href, uploader, 'Downloading track page %s' % (i + 1))
 436
 437             collection = response['collection']
 438             if not collection:
 439                 break
 440
 441             def resolve_permalink_url(candidates):
 442                 for cand in candidates:
 443                     if isinstance(cand, dict):
 444                         permalink_url = cand.get('permalink_url')
 445                         entry_id = self._extract_id(cand)
 446                         if permalink_url and permalink_url.startswith('http'):
 447                             return permalink_url, entry_id
 448
 449             for e in collection:
 450                 permalink_url, entry_id = resolve_permalink_url((e, e.get('track'), e.get('playlist')))
 451                 if permalink_url:
 452                     entries.append(self.url_result(permalink_url, video_id=entry_id))
 453
 454             next_href = response.get('next_href')
 455             if not next_href:
 456                 break
 457
 458             parsed_next_href = compat_urlparse.urlparse(response['next_href'])
 459             qs = compat_urlparse.parse_qs(parsed_next_href.query)
 460             qs.update(COMMON_QUERY)
 461             next_href = compat_urlparse.urlunparse(
 462                 parsed_next_href._replace(query=compat_urllib_parse_urlencode(qs, True)))
 463
 464         return {
 465             '_type': 'playlist',
 466             'id': compat_str(user['id']),
 467             'title': '%s (%s)' % (user['username'], self._TITLE_MAP[resource]),
 468             'entries': entries,
 469         }
 470
 471
 472 class SoundcloudPlaylistIE(SoundcloudPlaylistBaseIE):
 473     _VALID_URL = r'https?://api\.soundcloud\.com/playlists/(?P<id>[0-9]+)(?:/?\?secret_token=(?P<token>[^&]+?))?$'
 474     IE_NAME = 'soundcloud:playlist'
 475     _TESTS = [{
 476         'url': 'https://api.soundcloud.com/playlists/4110309',
 477         'info_dict': {
 478             'id': '4110309',
 479             'title': 'TILT Brass - Bowery Poetry Club, August \'03 [Non-Site SCR 02]',
 480             'description': 're:.*?TILT Brass - Bowery Poetry Club',
 481         },
 482         'playlist_count': 6,
 483     }]
 484
 485     def _real_extract(self, url):
 486         mobj = re.match(self._VALID_URL, url)
 487         playlist_id = mobj.group('id')
 488         base_url = '%s//api.soundcloud.com/playlists/%s.json?' % (self.http_scheme(), playlist_id)
 489
 490         data_dict = {
 491             'client_id': self._CLIENT_ID,
 492         }
 493         token = mobj.group('token')
 494
 495         if token:
 496             data_dict['secret_token'] = token
 497
 498         data = compat_urllib_parse_urlencode(data_dict)
 499         data = self._download_json(
 500             base_url + data, playlist_id, 'Downloading playlist')
 501
 502         entries = self._extract_track_entries(data['tracks'])
 503
 504         return {
 505             '_type': 'playlist',
 506             'id': playlist_id,
 507             'title': data.get('title'),
 508             'description': data.get('description'),
 509             'entries': entries,
 510         }
 511
 512
 513 class SoundcloudSearchIE(SearchInfoExtractor, SoundcloudIE):
 514     IE_NAME = 'soundcloud:search'
 515     IE_DESC = 'Soundcloud search'
 516     _MAX_RESULTS = float('inf')
 517     _TESTS = [{
 518         'url': 'scsearch15:post-avant jazzcore',
 519         'info_dict': {
 520             'title': 'post-avant jazzcore',
 521         },
 522         'playlist_count': 15,
 523     }]
 524
 525     _SEARCH_KEY = 'scsearch'
 526     _MAX_RESULTS_PER_PAGE = 200
 527     _DEFAULT_RESULTS_PER_PAGE = 50
 528     _API_V2_BASE = 'https://api-v2.soundcloud.com'
 529
 530     def _get_collection(self, endpoint, collection_id, **query):
 531         limit = min(
 532             query.get('limit', self._DEFAULT_RESULTS_PER_PAGE),
 533             self._MAX_RESULTS_PER_PAGE)
 534         query['limit'] = limit
 535         query['client_id'] = self._CLIENT_ID
 536         query['linked_partitioning'] = '1'
 537         query['offset'] = 0
 538         data = compat_urllib_parse_urlencode(query)
 539         next_url = '{0}{1}?{2}'.format(self._API_V2_BASE, endpoint, data)
 540
 541         collected_results = 0
 542
 543         for i in itertools.count(1):
 544             response = self._download_json(
 545                 next_url, collection_id, 'Downloading page {0}'.format(i),
 546                 'Unable to download API page')
 547
 548             collection = response.get('collection', [])
 549             if not collection:
 550                 break
 551
 552             collection = list(filter(bool, collection))
 553             collected_results += len(collection)
 554
 555             for item in collection:
 556                 yield self.url_result(item['uri'], SoundcloudIE.ie_key())
 557
 558             if not collection or collected_results >= limit:
 559                 break
 560
 561             next_url = response.get('next_href')
 562             if not next_url:
 563                 break
 564
 565     def _get_n_results(self, query, n):
 566         tracks = self._get_collection('/search/tracks', query, limit=n, q=query)
 567         return self.playlist_result(tracks, playlist_title=query)