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