youtube_dl/extractor/soundcloud.py

   1 # encoding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import json
   5 import re
   6 import itertools
   7
   8 from .common import InfoExtractor
   9 from ..utils import (
  10     compat_str,
  11     compat_urlparse,
  12     compat_urllib_parse,
  13
  14     ExtractorError,
  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'''^(?: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             }
  49         },
  50         # not streamable song
  51         {
  52             'url': 'https://soundcloud.com/the-concept-band/goldrushed-mastered?in=the-concept-band/sets/the-royal-concept-ep',
  53             'info_dict': {
  54                 'id': '47127627',
  55                 'ext': 'mp3',
  56                 'title': 'Goldrushed',
  57                 'description': 'From Stockholm Sweden\r\nPovel / Magnus / Filip / David\r\nwww.theroyalconcept.com',
  58                 'uploader': 'The Royal Concept',
  59                 'upload_date': '20120521',
  60             },
  61             'params': {
  62                 # rtmp
  63                 'skip_download': True,
  64             },
  65         },
  66         # private link
  67         {
  68             'url': 'https://soundcloud.com/jaimemf/youtube-dl-test-video-a-y-baw/s-8Pjrp',
  69             'md5': 'aa0dd32bfea9b0c5ef4f02aacd080604',
  70             'info_dict': {
  71                 'id': '123998367',
  72                 'ext': 'mp3',
  73                 'title': 'Youtube - Dl Test Video \'\' Ä↭',
  74                 'uploader': 'jaimeMF',
  75                 'description': 'test chars:  \"\'/\\ä↭',
  76                 'upload_date': '20131209',
  77             },
  78         },
  79         # downloadable song
  80         {
  81             'url': 'https://soundcloud.com/simgretina/just-your-problem-baby-1',
  82             'md5': '56a8b69568acaa967b4c49f9d1d52d19',
  83             'info_dict': {
  84                 'id': '105614606',
  85                 'ext': 'wav',
  86                 'title': 'Just Your Problem Baby (Acapella)',
  87                 'description': 'Vocals',
  88                 'uploader': 'Sim Gretina',
  89                 'upload_date': '20130815',
  90             },
  91         },
  92     ]
  93
  94     _CLIENT_ID = 'b45b1aa10f1ac2941910a7f0d10f8e28'
  95     _IPHONE_CLIENT_ID = '376f225bf427445fc4bfb6b99b72e0bf'
  96
  97     @classmethod
  98     def suitable(cls, url):
  99         return re.match(cls._VALID_URL, url, flags=re.VERBOSE) is not None
 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         }
 127         if info.get('downloadable', False):
 128             # We can build a direct link to the song
 129             format_url = (
 130                 'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(
 131                     track_id, self._CLIENT_ID))
 132             result['formats'] = [{
 133                 'format_id': 'download',
 134                 'ext': info.get('original_format', 'mp3'),
 135                 'url': format_url,
 136                 'vcodec': 'none',
 137             }]
 138         else:
 139             # We have to retrieve the url
 140             streams_url = ('http://api.soundcloud.com/i1/tracks/{0}/streams?'
 141                 'client_id={1}&secret_token={2}'.format(track_id, self._IPHONE_CLIENT_ID, secret_token))
 142             stream_json = self._download_webpage(
 143                 streams_url,
 144                 track_id, 'Downloading track url')
 145
 146             formats = []
 147             format_dict = json.loads(stream_json)
 148             for key, stream_url in format_dict.items():
 149                 if key.startswith('http'):
 150                     formats.append({
 151                         'format_id': key,
 152                         'ext': ext,
 153                         'url': stream_url,
 154                         'vcodec': 'none',
 155                     })
 156                 elif key.startswith('rtmp'):
 157                     # The url doesn't have an rtmp app, we have to extract the playpath
 158                     url, path = stream_url.split('mp3:', 1)
 159                     formats.append({
 160                         'format_id': key,
 161                         'url': url,
 162                         'play_path': 'mp3:' + path,
 163                         'ext': ext,
 164                         'vcodec': 'none',
 165                     })
 166
 167             if not formats:
 168                 # We fallback to the stream_url in the original info, this
 169                 # cannot be always used, sometimes it can give an HTTP 404 error
 170                 formats.append({
 171                     'format_id': 'fallback',
 172                     'url': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
 173                     'ext': ext,
 174                     'vcodec': 'none',
 175                 })
 176
 177             for f in formats:
 178                 if f['format_id'].startswith('http'):
 179                     f['protocol'] = 'http'
 180                 if f['format_id'].startswith('rtmp'):
 181                     f['protocol'] = 'rtmp'
 182
 183             self._sort_formats(formats)
 184             result['formats'] = formats
 185
 186         return result
 187
 188     def _real_extract(self, url):
 189         mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
 190         if mobj is None:
 191             raise ExtractorError('Invalid URL: %s' % url)
 192
 193         track_id = mobj.group('track_id')
 194         token = None
 195         if track_id is not None:
 196             info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 197             full_title = track_id
 198         elif mobj.group('player'):
 199             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 200             return self.url_result(query['url'][0], ie='Soundcloud')
 201         else:
 202             # extract uploader (which is in the url)
 203             uploader = mobj.group('uploader')
 204             # extract simple title (uploader + slug of song title)
 205             slug_title =  mobj.group('title')
 206             token = mobj.group('token')
 207             full_title = resolve_title = '%s/%s' % (uploader, slug_title)
 208             if token:
 209                 resolve_title += '/%s' % token
 210
 211             self.report_resolve(full_title)
 212
 213             url = 'http://soundcloud.com/%s' % resolve_title
 214             info_json_url = self._resolv_url(url)
 215         info_json = self._download_webpage(info_json_url, full_title, 'Downloading info JSON')
 216
 217         info = json.loads(info_json)
 218         return self._extract_info_dict(info, full_title, secret_token=token)
 219
 220 class SoundcloudSetIE(SoundcloudIE):
 221     _VALID_URL = r'https?://(?:www\.)?soundcloud\.com/([\w\d-]+)/sets/([\w\d-]+)'
 222     IE_NAME = 'soundcloud:set'
 223     # it's in tests/test_playlists.py
 224     _TESTS = []
 225
 226     def _real_extract(self, url):
 227         mobj = re.match(self._VALID_URL, url)
 228         if mobj is None:
 229             raise ExtractorError('Invalid URL: %s' % url)
 230
 231         # extract uploader (which is in the url)
 232         uploader = mobj.group(1)
 233         # extract simple title (uploader + slug of song title)
 234         slug_title =  mobj.group(2)
 235         full_title = '%s/sets/%s' % (uploader, slug_title)
 236
 237         self.report_resolve(full_title)
 238
 239         url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 240         resolv_url = self._resolv_url(url)
 241         info_json = self._download_webpage(resolv_url, full_title)
 242
 243         info = json.loads(info_json)
 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>[^/]+)(/?(tracks/)?)?(\?.*)?$'
 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
 268         url = 'http://soundcloud.com/%s/' % uploader
 269         resolv_url = self._resolv_url(url)
 270         user_json = self._download_webpage(resolv_url, uploader,
 271             'Downloading user info')
 272         user = json.loads(user_json)
 273
 274         tracks = []
 275         for i in itertools.count():
 276             data = compat_urllib_parse.urlencode({'offset': i*50,
 277                                                   'client_id': self._CLIENT_ID,
 278                                                   })
 279             tracks_url = 'http://api.soundcloud.com/users/%s/tracks.json?' % user['id'] + data
 280             response = self._download_webpage(tracks_url, uploader,
 281                 'Downloading tracks page %s' % (i+1))
 282             new_tracks = json.loads(response)
 283             tracks.extend(self._extract_info_dict(track, quiet=True) for track in new_tracks)
 284             if len(new_tracks) < 50:
 285                 break
 286
 287         return {
 288             '_type': 'playlist',
 289             'id': compat_str(user['id']),
 290             'title': user['username'],
 291             'entries': tracks,
 292         }