youtube_dl/extractor/soundcloud.py

   1 import json
   2 import re
   3 import itertools
   4
   5 from .common import InfoExtractor
   6 from ..utils import (
   7     compat_str,
   8     compat_urlparse,
   9     compat_urllib_parse,
  10
  11     ExtractorError,
  12     unified_strdate,
  13 )
  14
  15
  16 class SoundcloudIE(InfoExtractor):
  17     """Information extractor for soundcloud.com
  18        To access the media, the uid of the song and a stream token
  19        must be extracted from the page source and the script must make
  20        a request to media.soundcloud.com/crossdomain.xml. Then
  21        the media can be grabbed by requesting from an url composed
  22        of the stream token and uid
  23      """
  24
  25     _VALID_URL = r'''^(?:https?://)?
  26                     (?:(?:(?:www\.)?soundcloud\.com/([\w\d-]+)/([\w\d-]+)/?(?:[?].*)?$)
  27                        |(?:api\.soundcloud\.com/tracks/(?P<track_id>\d+))
  28                        |(?P<widget>w.soundcloud.com/player/?.*?url=.*)
  29                     )
  30                     '''
  31     IE_NAME = u'soundcloud'
  32     _TESTS = [
  33         {
  34             u'url': u'http://soundcloud.com/ethmusic/lostin-powers-she-so-heavy',
  35             u'file': u'62986583.mp3',
  36             u'md5': u'ebef0a451b909710ed1d7787dddbf0d7',
  37             u'info_dict': {
  38                 u"upload_date": u"20121011",
  39                 u"description": u"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",
  40                 u"uploader": u"E.T. ExTerrestrial Music",
  41                 u"title": u"Lostin Powers - She so Heavy (SneakPreview) Adrian Ackers Blueprint 1"
  42             }
  43         },
  44         # not streamable song
  45         {
  46             u'url': u'https://soundcloud.com/the-concept-band/goldrushed-mastered?in=the-concept-band/sets/the-royal-concept-ep',
  47             u'info_dict': {
  48                 u'id': u'47127627',
  49                 u'ext': u'mp3',
  50                 u'title': u'Goldrushed',
  51                 u'uploader': u'The Royal Concept',
  52                 u'upload_date': u'20120521',
  53             },
  54             u'params': {
  55                 # rtmp
  56                 u'skip_download': True,
  57             },
  58         },
  59     ]
  60
  61     _CLIENT_ID = 'b45b1aa10f1ac2941910a7f0d10f8e28'
  62
  63     @classmethod
  64     def suitable(cls, url):
  65         return re.match(cls._VALID_URL, url, flags=re.VERBOSE) is not None
  66
  67     def report_resolve(self, video_id):
  68         """Report information extraction."""
  69         self.to_screen(u'%s: Resolving id' % video_id)
  70
  71     @classmethod
  72     def _resolv_url(cls, url):
  73         return 'http://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID
  74
  75     def _extract_info_dict(self, info, full_title=None, quiet=False):
  76         track_id = compat_str(info['id'])
  77         name = full_title or track_id
  78         if quiet == False:
  79             self.report_extraction(name)
  80
  81         thumbnail = info['artwork_url']
  82         if thumbnail is not None:
  83             thumbnail = thumbnail.replace('-large', '-t500x500')
  84         result = {
  85             'id':       track_id,
  86             'url':      info['stream_url'] + '?client_id=' + self._CLIENT_ID,
  87             'uploader': info['user']['username'],
  88             'upload_date': unified_strdate(info['created_at']),
  89             'title':    info['title'],
  90             'ext':      info.get('original_format', u'mp3'),
  91             'description': info['description'],
  92             'thumbnail': thumbnail,
  93         }
  94         if info.get('downloadable', False):
  95             result['url'] = 'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(track_id, self._CLIENT_ID)
  96         if not info.get('streamable', False):
  97             # We have to get the rtmp url
  98             stream_json = self._download_webpage(
  99                 'http://api.soundcloud.com/i1/tracks/{0}/streams?client_id={1}'.format(track_id, self._CLIENT_ID),
 100                 track_id, u'Downloading track url')
 101             rtmp_url = json.loads(stream_json)['rtmp_mp3_128_url']
 102             # The url doesn't have an rtmp app, we have to extract the playpath
 103             url, path = rtmp_url.split('mp3:', 1)
 104             result.update({
 105                 'url': url,
 106                 'play_path': 'mp3:' + path,
 107             })
 108         return result
 109
 110     def _real_extract(self, url):
 111         mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
 112         if mobj is None:
 113             raise ExtractorError(u'Invalid URL: %s' % url)
 114
 115         track_id = mobj.group('track_id')
 116         if track_id is not None:
 117             info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 118             full_title = track_id
 119         elif mobj.group('widget'):
 120             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 121             return self.url_result(query['url'][0], ie='Soundcloud')
 122         else:
 123             # extract uploader (which is in the url)
 124             uploader = mobj.group(1)
 125             # extract simple title (uploader + slug of song title)
 126             slug_title =  mobj.group(2)
 127             full_title = '%s/%s' % (uploader, slug_title)
 128
 129             self.report_resolve(full_title)
 130
 131             url = 'http://soundcloud.com/%s/%s' % (uploader, slug_title)
 132             info_json_url = self._resolv_url(url)
 133         info_json = self._download_webpage(info_json_url, full_title, u'Downloading info JSON')
 134
 135         info = json.loads(info_json)
 136         return self._extract_info_dict(info, full_title)
 137
 138 class SoundcloudSetIE(SoundcloudIE):
 139     _VALID_URL = r'^(?:https?://)?(?:www\.)?soundcloud\.com/([\w\d-]+)/sets/([\w\d-]+)(?:[?].*)?$'
 140     IE_NAME = u'soundcloud:set'
 141     # it's in tests/test_playlists.py
 142     _TESTS = []
 143
 144     def _real_extract(self, url):
 145         mobj = re.match(self._VALID_URL, url)
 146         if mobj is None:
 147             raise ExtractorError(u'Invalid URL: %s' % url)
 148
 149         # extract uploader (which is in the url)
 150         uploader = mobj.group(1)
 151         # extract simple title (uploader + slug of song title)
 152         slug_title =  mobj.group(2)
 153         full_title = '%s/sets/%s' % (uploader, slug_title)
 154
 155         self.report_resolve(full_title)
 156
 157         url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 158         resolv_url = self._resolv_url(url)
 159         info_json = self._download_webpage(resolv_url, full_title)
 160
 161         info = json.loads(info_json)
 162         if 'errors' in info:
 163             for err in info['errors']:
 164                 self._downloader.report_error(u'unable to download video webpage: %s' % compat_str(err['error_message']))
 165             return
 166
 167         self.report_extraction(full_title)
 168         return {'_type': 'playlist',
 169                 'entries': [self._extract_info_dict(track) for track in info['tracks']],
 170                 'id': info['id'],
 171                 'title': info['title'],
 172                 }
 173
 174
 175 class SoundcloudUserIE(SoundcloudIE):
 176     _VALID_URL = r'https?://(www\.)?soundcloud.com/(?P<user>[^/]+)(/?(tracks/)?)?(\?.*)?$'
 177     IE_NAME = u'soundcloud:user'
 178
 179     # it's in tests/test_playlists.py
 180     _TESTS = []
 181
 182     def _real_extract(self, url):
 183         mobj = re.match(self._VALID_URL, url)
 184         uploader = mobj.group('user')
 185
 186         url = 'http://soundcloud.com/%s/' % uploader
 187         resolv_url = self._resolv_url(url)
 188         user_json = self._download_webpage(resolv_url, uploader,
 189             u'Downloading user info')
 190         user = json.loads(user_json)
 191
 192         tracks = []
 193         for i in itertools.count():
 194             data = compat_urllib_parse.urlencode({'offset': i*50,
 195                                                   'client_id': self._CLIENT_ID,
 196                                                   })
 197             tracks_url = 'http://api.soundcloud.com/users/%s/tracks.json?' % user['id'] + data
 198             response = self._download_webpage(tracks_url, uploader,
 199                 u'Downloading tracks page %s' % (i+1))
 200             new_tracks = json.loads(response)
 201             tracks.extend(self._extract_info_dict(track, quiet=True) for track in new_tracks)
 202             if len(new_tracks) < 50:
 203                 break
 204
 205         return {
 206             '_type': 'playlist',
 207             'id': compat_str(user['id']),
 208             'title': user['username'],
 209             'entries': tracks,
 210         }