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     _IPHONE_CLIENT_ID = '376f225bf427445fc4bfb6b99b72e0bf'
  63
  64     @classmethod
  65     def suitable(cls, url):
  66         return re.match(cls._VALID_URL, url, flags=re.VERBOSE) is not None
  67
  68     def report_resolve(self, video_id):
  69         """Report information extraction."""
  70         self.to_screen(u'%s: Resolving id' % video_id)
  71
  72     @classmethod
  73     def _resolv_url(cls, url):
  74         return 'http://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID
  75
  76     def _extract_info_dict(self, info, full_title=None, quiet=False):
  77         track_id = compat_str(info['id'])
  78         name = full_title or track_id
  79         if quiet:
  80             self.report_extraction(name)
  81
  82         thumbnail = info['artwork_url']
  83         if thumbnail is not None:
  84             thumbnail = thumbnail.replace('-large', '-t500x500')
  85         ext = info.get('original_format', u'mp3')
  86         result = {
  87             'id': track_id,
  88             'uploader': info['user']['username'],
  89             'upload_date': unified_strdate(info['created_at']),
  90             'title': info['title'],
  91             'description': info['description'],
  92             'thumbnail': thumbnail,
  93         }
  94         if info.get('downloadable', False):
  95             # We can build a direct link to the song
  96             format_url = (
  97                 u'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(
  98                     track_id, self._CLIENT_ID))
  99             result['formats'] = [{
 100                 'format_id': 'download',
 101                 'ext': ext,
 102                 'url': format_url,
 103             }]
 104         else:
 105             # We have to retrieve the url
 106             stream_json = self._download_webpage(
 107                 'http://api.soundcloud.com/i1/tracks/{0}/streams?client_id={1}'.format(track_id, self._IPHONE_CLIENT_ID),
 108                 track_id, u'Downloading track url')
 109
 110             formats = []
 111             format_dict = json.loads(stream_json)
 112             for key, stream_url in format_dict.items():
 113                 if key.startswith(u'http'):
 114                     formats.append({
 115                         'format_id': key,
 116                         'ext': ext,
 117                         'url': stream_url,
 118                     })
 119                 elif key.startswith(u'rtmp'):
 120                     # The url doesn't have an rtmp app, we have to extract the playpath
 121                     url, path = stream_url.split('mp3:', 1)
 122                     formats.append({
 123                         'format_id': key,
 124                         'url': url,
 125                         'play_path': 'mp3:' + path,
 126                         'ext': ext,
 127                     })
 128
 129             if not formats:
 130                 # We fallback to the stream_url in the original info, this
 131                 # cannot be always used, sometimes it can give an HTTP 404 error
 132                 formats.append({
 133                     'format_id': u'fallback',
 134                     'url': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
 135                     'ext': ext,
 136                 })
 137
 138             def format_pref(f):
 139                 if f['format_id'].startswith('http'):
 140                     return 2
 141                 if f['format_id'].startswith('rtmp'):
 142                     return 1
 143                 return 0
 144
 145             formats.sort(key=format_pref)
 146             result['formats'] = formats
 147
 148         return result
 149
 150     def _real_extract(self, url):
 151         mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
 152         if mobj is None:
 153             raise ExtractorError(u'Invalid URL: %s' % url)
 154
 155         track_id = mobj.group('track_id')
 156         if track_id is not None:
 157             info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 158             full_title = track_id
 159         elif mobj.group('widget'):
 160             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 161             return self.url_result(query['url'][0], ie='Soundcloud')
 162         else:
 163             # extract uploader (which is in the url)
 164             uploader = mobj.group(1)
 165             # extract simple title (uploader + slug of song title)
 166             slug_title =  mobj.group(2)
 167             full_title = '%s/%s' % (uploader, slug_title)
 168
 169             self.report_resolve(full_title)
 170
 171             url = 'http://soundcloud.com/%s/%s' % (uploader, slug_title)
 172             info_json_url = self._resolv_url(url)
 173         info_json = self._download_webpage(info_json_url, full_title, u'Downloading info JSON')
 174
 175         info = json.loads(info_json)
 176         return self._extract_info_dict(info, full_title)
 177
 178 class SoundcloudSetIE(SoundcloudIE):
 179     _VALID_URL = r'^(?:https?://)?(?:www\.)?soundcloud\.com/([\w\d-]+)/sets/([\w\d-]+)(?:[?].*)?$'
 180     IE_NAME = u'soundcloud:set'
 181     # it's in tests/test_playlists.py
 182     _TESTS = []
 183
 184     def _real_extract(self, url):
 185         mobj = re.match(self._VALID_URL, url)
 186         if mobj is None:
 187             raise ExtractorError(u'Invalid URL: %s' % url)
 188
 189         # extract uploader (which is in the url)
 190         uploader = mobj.group(1)
 191         # extract simple title (uploader + slug of song title)
 192         slug_title =  mobj.group(2)
 193         full_title = '%s/sets/%s' % (uploader, slug_title)
 194
 195         self.report_resolve(full_title)
 196
 197         url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 198         resolv_url = self._resolv_url(url)
 199         info_json = self._download_webpage(resolv_url, full_title)
 200
 201         info = json.loads(info_json)
 202         if 'errors' in info:
 203             for err in info['errors']:
 204                 self._downloader.report_error(u'unable to download video webpage: %s' % compat_str(err['error_message']))
 205             return
 206
 207         self.report_extraction(full_title)
 208         return {'_type': 'playlist',
 209                 'entries': [self._extract_info_dict(track) for track in info['tracks']],
 210                 'id': info['id'],
 211                 'title': info['title'],
 212                 }
 213
 214
 215 class SoundcloudUserIE(SoundcloudIE):
 216     _VALID_URL = r'https?://(www\.)?soundcloud.com/(?P<user>[^/]+)(/?(tracks/)?)?(\?.*)?$'
 217     IE_NAME = u'soundcloud:user'
 218
 219     # it's in tests/test_playlists.py
 220     _TESTS = []
 221
 222     def _real_extract(self, url):
 223         mobj = re.match(self._VALID_URL, url)
 224         uploader = mobj.group('user')
 225
 226         url = 'http://soundcloud.com/%s/' % uploader
 227         resolv_url = self._resolv_url(url)
 228         user_json = self._download_webpage(resolv_url, uploader,
 229             u'Downloading user info')
 230         user = json.loads(user_json)
 231
 232         tracks = []
 233         for i in itertools.count():
 234             data = compat_urllib_parse.urlencode({'offset': i*50,
 235                                                   'client_id': self._CLIENT_ID,
 236                                                   })
 237             tracks_url = 'http://api.soundcloud.com/users/%s/tracks.json?' % user['id'] + data
 238             response = self._download_webpage(tracks_url, uploader,
 239                 u'Downloading tracks page %s' % (i+1))
 240             new_tracks = json.loads(response)
 241             tracks.extend(self._extract_info_dict(track, quiet=True) for track in new_tracks)
 242             if len(new_tracks) < 50:
 243                 break
 244
 245         return {
 246             '_type': 'playlist',
 247             'id': compat_str(user['id']),
 248             'title': user['username'],
 249             'entries': tracks,
 250         }