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