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 InfoExtractor
   8 from ..utils import (
   9     compat_str,
  10     compat_urlparse,
  11     compat_urllib_parse,
  12
  13     ExtractorError,
  14     int_or_none,
  15     unified_strdate,
  16 )
  17
  18
  19 class SoundcloudIE(InfoExtractor):
  20     """Information extractor for soundcloud.com
  21        To access the media, the uid of the song and a stream token
  22        must be extracted from the page source and the script must make
  23        a request to media.soundcloud.com/crossdomain.xml. Then
  24        the media can be grabbed by requesting from an url composed
  25        of the stream token and uid
  26      """
  27
  28     _VALID_URL = r'''(?x)^(?:https?://)?
  29                     (?:(?:(?:www\.|m\.)?soundcloud\.com/
  30                             (?P<uploader>[\w\d-]+)/
  31                             (?!sets/|likes/?(?:$|[?#]))
  32                             (?P<title>[\w\d-]+)/?
  33                             (?P<token>[^?]+?)?(?:[?].*)?$)
  34                        |(?:api\.soundcloud\.com/tracks/(?P<track_id>\d+)
  35                           (?:/?\?secret_token=(?P<secret_token>[^&]+?))?$)
  36                        |(?P<player>(?:w|player|p.)\.soundcloud\.com/player/?.*?url=.*)
  37                     )
  38                     '''
  39     IE_NAME = 'soundcloud'
  40     _TESTS = [
  41         {
  42             'url': 'http://soundcloud.com/ethmusic/lostin-powers-she-so-heavy',
  43             'file': '62986583.mp3',
  44             'md5': 'ebef0a451b909710ed1d7787dddbf0d7',
  45             'info_dict': {
  46                 "upload_date": "20121011",
  47                 "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",
  48                 "uploader": "E.T. ExTerrestrial Music",
  49                 "title": "Lostin Powers - She so Heavy (SneakPreview) Adrian Ackers Blueprint 1",
  50                 "duration": 143,
  51             }
  52         },
  53         # not streamable song
  54         {
  55             'url': 'https://soundcloud.com/the-concept-band/goldrushed-mastered?in=the-concept-band/sets/the-royal-concept-ep',
  56             'info_dict': {
  57                 'id': '47127627',
  58                 'ext': 'mp3',
  59                 'title': 'Goldrushed',
  60                 'description': 'From Stockholm Sweden\r\nPovel / Magnus / Filip / David\r\nwww.theroyalconcept.com',
  61                 'uploader': 'The Royal Concept',
  62                 'upload_date': '20120521',
  63                 'duration': 227,
  64             },
  65             'params': {
  66                 # rtmp
  67                 'skip_download': True,
  68             },
  69         },
  70         # private link
  71         {
  72             'url': 'https://soundcloud.com/jaimemf/youtube-dl-test-video-a-y-baw/s-8Pjrp',
  73             'md5': 'aa0dd32bfea9b0c5ef4f02aacd080604',
  74             'info_dict': {
  75                 'id': '123998367',
  76                 'ext': 'mp3',
  77                 'title': 'Youtube - Dl Test Video \'\' Ä↭',
  78                 'uploader': 'jaimeMF',
  79                 'description': 'test chars:  \"\'/\\ä↭',
  80                 'upload_date': '20131209',
  81                 'duration': 9,
  82             },
  83         },
  84         # private link (alt format)
  85         {
  86             'url': 'https://api.soundcloud.com/tracks/123998367?secret_token=s-8Pjrp',
  87             'md5': 'aa0dd32bfea9b0c5ef4f02aacd080604',
  88             'info_dict': {
  89                 'id': '123998367',
  90                 'ext': 'mp3',
  91                 'title': 'Youtube - Dl Test Video \'\' Ä↭',
  92                 'uploader': 'jaimeMF',
  93                 'description': 'test chars:  \"\'/\\ä↭',
  94                 'upload_date': '20131209',
  95                 'duration': 9,
  96             },
  97         },
  98         # downloadable song
  99         {
 100             'url': 'https://soundcloud.com/oddsamples/bus-brakes',
 101             'md5': '7624f2351f8a3b2e7cd51522496e7631',
 102             'info_dict': {
 103                 'id': '128590877',
 104                 'ext': 'mp3',
 105                 'title': 'Bus Brakes',
 106                 'description': 'md5:0170be75dd395c96025d210d261c784e',
 107                 'uploader': 'oddsamples',
 108                 'upload_date': '20140109',
 109                 'duration': 17,
 110             },
 111         },
 112     ]
 113
 114     _CLIENT_ID = 'b45b1aa10f1ac2941910a7f0d10f8e28'
 115     _IPHONE_CLIENT_ID = '376f225bf427445fc4bfb6b99b72e0bf'
 116
 117     def report_resolve(self, video_id):
 118         """Report information extraction."""
 119         self.to_screen('%s: Resolving id' % video_id)
 120
 121     @classmethod
 122     def _resolv_url(cls, url):
 123         return 'http://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID
 124
 125     def _extract_info_dict(self, info, full_title=None, quiet=False, secret_token=None):
 126         track_id = compat_str(info['id'])
 127         name = full_title or track_id
 128         if quiet:
 129             self.report_extraction(name)
 130
 131         thumbnail = info['artwork_url']
 132         if thumbnail is not None:
 133             thumbnail = thumbnail.replace('-large', '-t500x500')
 134         ext = 'mp3'
 135         result = {
 136             'id': track_id,
 137             'uploader': info['user']['username'],
 138             'upload_date': unified_strdate(info['created_at']),
 139             'title': info['title'],
 140             'description': info['description'],
 141             'thumbnail': thumbnail,
 142             'duration': int_or_none(info.get('duration'), 1000),
 143         }
 144         formats = []
 145         if info.get('downloadable', False):
 146             # We can build a direct link to the song
 147             format_url = (
 148                 'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(
 149                     track_id, self._CLIENT_ID))
 150             formats.append({
 151                 'format_id': 'download',
 152                 'ext': info.get('original_format', 'mp3'),
 153                 'url': format_url,
 154                 'vcodec': 'none',
 155                 'preference': 10,
 156             })
 157
 158         # We have to retrieve the url
 159         streams_url = ('http://api.soundcloud.com/i1/tracks/{0}/streams?'
 160             'client_id={1}&secret_token={2}'.format(track_id, self._IPHONE_CLIENT_ID, secret_token))
 161         format_dict = self._download_json(
 162             streams_url,
 163             track_id, 'Downloading track url')
 164
 165         for key, stream_url in format_dict.items():
 166             if key.startswith('http'):
 167                 formats.append({
 168                     'format_id': key,
 169                     'ext': ext,
 170                     'url': stream_url,
 171                     'vcodec': 'none',
 172                 })
 173             elif key.startswith('rtmp'):
 174                 # The url doesn't have an rtmp app, we have to extract the playpath
 175                 url, path = stream_url.split('mp3:', 1)
 176                 formats.append({
 177                     'format_id': key,
 178                     'url': url,
 179                     'play_path': 'mp3:' + path,
 180                     'ext': ext,
 181                     'vcodec': 'none',
 182                 })
 183
 184             if not formats:
 185                 # We fallback to the stream_url in the original info, this
 186                 # cannot be always used, sometimes it can give an HTTP 404 error
 187                 formats.append({
 188                     'format_id': 'fallback',
 189                     'url': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
 190                     'ext': ext,
 191                     'vcodec': 'none',
 192                 })
 193
 194             for f in formats:
 195                 if f['format_id'].startswith('http'):
 196                     f['protocol'] = 'http'
 197                 if f['format_id'].startswith('rtmp'):
 198                     f['protocol'] = 'rtmp'
 199
 200             self._sort_formats(formats)
 201             result['formats'] = formats
 202
 203         return result
 204
 205     def _real_extract(self, url):
 206         mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
 207         if mobj is None:
 208             raise ExtractorError('Invalid URL: %s' % url)
 209
 210         track_id = mobj.group('track_id')
 211         token = None
 212         if track_id is not None:
 213             info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 214             full_title = track_id
 215             token = mobj.group('secret_token')
 216             if token:
 217                 info_json_url += "&secret_token=" + token
 218         elif mobj.group('player'):
 219             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 220             return self.url_result(query['url'][0])
 221         else:
 222             # extract uploader (which is in the url)
 223             uploader = mobj.group('uploader')
 224             # extract simple title (uploader + slug of song title)
 225             slug_title =  mobj.group('title')
 226             token = mobj.group('token')
 227             full_title = resolve_title = '%s/%s' % (uploader, slug_title)
 228             if token:
 229                 resolve_title += '/%s' % token
 230
 231             self.report_resolve(full_title)
 232
 233             url = 'http://soundcloud.com/%s' % resolve_title
 234             info_json_url = self._resolv_url(url)
 235         info = self._download_json(info_json_url, full_title, 'Downloading info JSON')
 236
 237         return self._extract_info_dict(info, full_title, secret_token=token)
 238
 239
 240 class SoundcloudSetIE(SoundcloudIE):
 241     _VALID_URL = r'https?://(?:www\.)?soundcloud\.com/(?P<uploader>[\w\d-]+)/sets/(?P<slug_title>[\w\d-]+)(?:/(?P<token>[^?/]+))?'
 242     IE_NAME = 'soundcloud:set'
 243     _TESTS = [{
 244         'url': 'https://soundcloud.com/the-concept-band/sets/the-royal-concept-ep',
 245         'info_dict': {
 246             'title': 'The Royal Concept EP',
 247         },
 248         'playlist_mincount': 6,
 249     }]
 250
 251     def _real_extract(self, url):
 252         mobj = re.match(self._VALID_URL, url)
 253
 254         # extract uploader (which is in the url)
 255         uploader = mobj.group('uploader')
 256         # extract simple title (uploader + slug of song title)
 257         slug_title = mobj.group('slug_title')
 258         full_title = '%s/sets/%s' % (uploader, slug_title)
 259         url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 260
 261         token = mobj.group('token')
 262         if token:
 263             full_title += '/' + token
 264             url += '/' + token
 265
 266         self.report_resolve(full_title)
 267
 268         resolv_url = self._resolv_url(url)
 269         info = self._download_json(resolv_url, full_title)
 270
 271         if 'errors' in info:
 272             for err in info['errors']:
 273                 self._downloader.report_error('unable to download video webpage: %s' % compat_str(err['error_message']))
 274             return
 275
 276         return {
 277             '_type': 'playlist',
 278             'entries': [self._extract_info_dict(track, secret_token=token) for track in info['tracks']],
 279             'id': info['id'],
 280             'title': info['title'],
 281         }
 282
 283
 284 class SoundcloudUserIE(SoundcloudIE):
 285     _VALID_URL = r'https?://(www\.)?soundcloud\.com/(?P<user>[^/]+)/?((?P<rsrc>tracks|likes)/?)?(\?.*)?$'
 286     IE_NAME = 'soundcloud:user'
 287     _TESTS = [{
 288         'url': 'https://soundcloud.com/the-concept-band',
 289         'info_dict': {
 290             'id': '9615865',
 291             'title': 'The Royal Concept',
 292         },
 293         'playlist_mincount': 12
 294     }, {
 295         'url': 'https://soundcloud.com/the-concept-band/likes',
 296         'info_dict': {
 297             'id': '9615865',
 298             'title': 'The Royal Concept',
 299         },
 300         'playlist_mincount': 1,
 301     }]
 302
 303     def _real_extract(self, url):
 304         mobj = re.match(self._VALID_URL, url)
 305         uploader = mobj.group('user')
 306         resource = mobj.group('rsrc')
 307         if resource is None:
 308             resource = 'tracks'
 309         elif resource == 'likes':
 310             resource = 'favorites'
 311
 312         url = 'http://soundcloud.com/%s/' % uploader
 313         resolv_url = self._resolv_url(url)
 314         user = self._download_json(
 315             resolv_url, uploader, 'Downloading user info')
 316         base_url = 'http://api.soundcloud.com/users/%s/%s.json?' % (uploader, resource)
 317
 318         entries = []
 319         for i in itertools.count():
 320             data = compat_urllib_parse.urlencode({
 321                 'offset': i * 50,
 322                 'limit': 50,
 323                 'client_id': self._CLIENT_ID,
 324             })
 325             new_entries = self._download_json(
 326                 base_url + data, uploader, 'Downloading track page %s' % (i + 1))
 327             if len(new_entries) == 0:
 328                 self.to_screen('%s: End page received' % uploader)
 329                 break
 330             entries.extend(self._extract_info_dict(e, quiet=True) for e in new_entries)
 331
 332         return {
 333             '_type': 'playlist',
 334             'id': compat_str(user['id']),
 335             'title': user['username'],
 336             'entries': entries,
 337         }
 338
 339
 340 class SoundcloudPlaylistIE(SoundcloudIE):
 341     _VALID_URL = r'https?://api\.soundcloud\.com/playlists/(?P<id>[0-9]+)(?:/?\?secret_token=(?P<token>[^&]+?))$'
 342     IE_NAME = 'soundcloud:playlist'
 343     _TESTS = [
 344
 345         {
 346             'url': 'http://api.soundcloud.com/playlists/4110309',
 347             'info_dict': {
 348                 'id': '4110309',
 349                 'title': 'TILT Brass - Bowery Poetry Club, August \'03 [Non-Site SCR 02]',
 350                 'description': 're:.*?TILT Brass - Bowery Poetry Club',
 351             },
 352             'playlist_count': 6,
 353         }
 354     ]
 355
 356     def _real_extract(self, url):
 357         mobj = re.match(self._VALID_URL, url)
 358         playlist_id = mobj.group('id')
 359         base_url = '%s//api.soundcloud.com/playlists/%s.json?' % (self.http_scheme(), playlist_id)
 360
 361         data_dict = {
 362             'client_id': self._CLIENT_ID,
 363         }
 364         token = mobj.group('token')
 365
 366         if token:
 367             data_dict['secret_token'] = token
 368
 369         data = compat_urllib_parse.urlencode(data_dict)
 370         data = self._download_json(
 371             base_url + data, playlist_id, 'Downloading playlist')
 372
 373         entries = [
 374             self._extract_info_dict(t, quiet=True, secret_token=token)
 375                 for t in data['tracks']]
 376
 377         return {
 378             '_type': 'playlist',
 379             'id': playlist_id,
 380             'title': data.get('title'),
 381             'description': data.get('description'),
 382             'entries': entries,
 383         }