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                 'uploader': 'The Royal Concept',
  58                 'upload_date': '20120521',
  59             },
  60             'params': {
  61                 # rtmp
  62                 'skip_download': True,
  63             },
  64         },
  65         # private link
  66         {
  67             'url': 'https://soundcloud.com/jaimemf/youtube-dl-test-video-a-y-baw/s-8Pjrp',
  68             'md5': 'aa0dd32bfea9b0c5ef4f02aacd080604',
  69             'info_dict': {
  70                 'id': '123998367',
  71                 'ext': 'mp3',
  72                 'title': 'Youtube - Dl Test Video \'\' Ä↭',
  73                 'uploader': 'jaimeMF',
  74                 'description': 'test chars:  \"\'/\\ä↭',
  75                 'upload_date': '20131209',
  76             },
  77         },
  78         # downloadable song
  79         {
  80             'url': 'https://soundcloud.com/simgretina/just-your-problem-baby-1',
  81             'md5': '56a8b69568acaa967b4c49f9d1d52d19',
  82             'info_dict': {
  83                 'id': '105614606',
  84                 'ext': 'wav',
  85                 'title': 'Just Your Problem Baby (Acapella)',
  86                 'description': 'Vocals',
  87                 'uploader': 'Sim Gretina',
  88                 'upload_date': '20130815',
  89             },
  90         },
  91     ]
  92
  93     _CLIENT_ID = 'b45b1aa10f1ac2941910a7f0d10f8e28'
  94     _IPHONE_CLIENT_ID = '376f225bf427445fc4bfb6b99b72e0bf'
  95
  96     @classmethod
  97     def suitable(cls, url):
  98         return re.match(cls._VALID_URL, url, flags=re.VERBOSE) is not None
  99
 100     def report_resolve(self, video_id):
 101         """Report information extraction."""
 102         self.to_screen(u'%s: Resolving id' % video_id)
 103
 104     @classmethod
 105     def _resolv_url(cls, url):
 106         return 'http://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID
 107
 108     def _extract_info_dict(self, info, full_title=None, quiet=False, secret_token=None):
 109         track_id = compat_str(info['id'])
 110         name = full_title or track_id
 111         if quiet:
 112             self.report_extraction(name)
 113
 114         thumbnail = info['artwork_url']
 115         if thumbnail is not None:
 116             thumbnail = thumbnail.replace('-large', '-t500x500')
 117         ext = 'mp3'
 118         result = {
 119             'id': track_id,
 120             'uploader': info['user']['username'],
 121             'upload_date': unified_strdate(info['created_at']),
 122             'title': info['title'],
 123             'description': info['description'],
 124             'thumbnail': thumbnail,
 125         }
 126         if info.get('downloadable', False):
 127             # We can build a direct link to the song
 128             format_url = (
 129                 'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(
 130                     track_id, self._CLIENT_ID))
 131             result['formats'] = [{
 132                 'format_id': 'download',
 133                 'ext': info.get('original_format', 'mp3'),
 134                 'url': format_url,
 135                 'vcodec': 'none',
 136             }]
 137         else:
 138             # We have to retrieve the url
 139             streams_url = ('http://api.soundcloud.com/i1/tracks/{0}/streams?'
 140                 'client_id={1}&secret_token={2}'.format(track_id, self._IPHONE_CLIENT_ID, secret_token))
 141             stream_json = self._download_webpage(
 142                 streams_url,
 143                 track_id, 'Downloading track url')
 144
 145             formats = []
 146             format_dict = json.loads(stream_json)
 147             for key, stream_url in format_dict.items():
 148                 if key.startswith(u'http'):
 149                     formats.append({
 150                         'format_id': key,
 151                         'ext': ext,
 152                         'url': stream_url,
 153                         'vcodec': 'none',
 154                     })
 155                 elif key.startswith(u'rtmp'):
 156                     # The url doesn't have an rtmp app, we have to extract the playpath
 157                     url, path = stream_url.split('mp3:', 1)
 158                     formats.append({
 159                         'format_id': key,
 160                         'url': url,
 161                         'play_path': 'mp3:' + path,
 162                         'ext': ext,
 163                         'vcodec': 'none',
 164                     })
 165
 166             if not formats:
 167                 # We fallback to the stream_url in the original info, this
 168                 # cannot be always used, sometimes it can give an HTTP 404 error
 169                 formats.append({
 170                     'format_id': 'fallback',
 171                     'url': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
 172                     'ext': ext,
 173                     'vcodec': 'none',
 174                 })
 175
 176             for f in formats:
 177                 if f['format_id'].startswith('http'):
 178                     f['protocol'] = 'http'
 179                 if f['format_id'].startswith('rtmp'):
 180                     f['protocol'] = 'rtmp'
 181
 182             self._sort_formats(formats)
 183             result['formats'] = formats
 184
 185         return result
 186
 187     def _real_extract(self, url):
 188         mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
 189         if mobj is None:
 190             raise ExtractorError(u'Invalid URL: %s' % url)
 191
 192         track_id = mobj.group('track_id')
 193         token = None
 194         if track_id is not None:
 195             info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 196             full_title = track_id
 197         elif mobj.group('player'):
 198             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 199             return self.url_result(query['url'][0], ie='Soundcloud')
 200         else:
 201             # extract uploader (which is in the url)
 202             uploader = mobj.group('uploader')
 203             # extract simple title (uploader + slug of song title)
 204             slug_title =  mobj.group('title')
 205             token = mobj.group('token')
 206             full_title = resolve_title = '%s/%s' % (uploader, slug_title)
 207             if token:
 208                 resolve_title += '/%s' % token
 209
 210             self.report_resolve(full_title)
 211
 212             url = 'http://soundcloud.com/%s' % resolve_title
 213             info_json_url = self._resolv_url(url)
 214         info_json = self._download_webpage(info_json_url, full_title, 'Downloading info JSON')
 215
 216         info = json.loads(info_json)
 217         return self._extract_info_dict(info, full_title, secret_token=token)
 218
 219 class SoundcloudSetIE(SoundcloudIE):
 220     _VALID_URL = r'^(?:https?://)?(?:www\.)?soundcloud\.com/([\w\d-]+)/sets/([\w\d-]+)(?:[?].*)?$'
 221     IE_NAME = 'soundcloud:set'
 222     # it's in tests/test_playlists.py
 223     _TESTS = []
 224
 225     def _real_extract(self, url):
 226         mobj = re.match(self._VALID_URL, url)
 227         if mobj is None:
 228             raise ExtractorError(u'Invalid URL: %s' % url)
 229
 230         # extract uploader (which is in the url)
 231         uploader = mobj.group(1)
 232         # extract simple title (uploader + slug of song title)
 233         slug_title =  mobj.group(2)
 234         full_title = '%s/sets/%s' % (uploader, slug_title)
 235
 236         self.report_resolve(full_title)
 237
 238         url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 239         resolv_url = self._resolv_url(url)
 240         info_json = self._download_webpage(resolv_url, full_title)
 241
 242         info = json.loads(info_json)
 243         if 'errors' in info:
 244             for err in info['errors']:
 245                 self._downloader.report_error(u'unable to download video webpage: %s' % compat_str(err['error_message']))
 246             return
 247
 248         self.report_extraction(full_title)
 249         return {'_type': 'playlist',
 250                 'entries': [self._extract_info_dict(track) for track in info['tracks']],
 251                 'id': info['id'],
 252                 'title': info['title'],
 253                 }
 254
 255
 256 class SoundcloudUserIE(SoundcloudIE):
 257     _VALID_URL = r'https?://(www\.)?soundcloud\.com/(?P<user>[^/]+)(/?(tracks/)?)?(\?.*)?$'
 258     IE_NAME = 'soundcloud:user'
 259
 260     # it's in tests/test_playlists.py
 261     _TESTS = []
 262
 263     def _real_extract(self, url):
 264         mobj = re.match(self._VALID_URL, url)
 265         uploader = mobj.group('user')
 266
 267         url = 'http://soundcloud.com/%s/' % uploader
 268         resolv_url = self._resolv_url(url)
 269         user_json = self._download_webpage(resolv_url, uploader,
 270             'Downloading user info')
 271         user = json.loads(user_json)
 272
 273         tracks = []
 274         for i in itertools.count():
 275             data = compat_urllib_parse.urlencode({'offset': i*50,
 276                                                   'client_id': self._CLIENT_ID,
 277                                                   })
 278             tracks_url = 'http://api.soundcloud.com/users/%s/tracks.json?' % user['id'] + data
 279             response = self._download_webpage(tracks_url, uploader,
 280                 'Downloading tracks page %s' % (i+1))
 281             new_tracks = json.loads(response)
 282             tracks.extend(self._extract_info_dict(track, quiet=True) for track in new_tracks)
 283             if len(new_tracks) < 50:
 284                 break
 285
 286         return {
 287             '_type': 'playlist',
 288             'id': compat_str(user['id']),
 289             'title': user['username'],
 290             'entries': tracks,
 291         }