Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/soundcloud.py

   1 # encoding: utf-8
   2 import json
   3 import re
   4 import itertools
   5
   6 from .common import InfoExtractor
   7 from ..utils import (
   8     compat_str,
   9     compat_urlparse,
  10     compat_urllib_parse,
  11
  12     ExtractorError,
  13     unified_strdate,
  14 )
  15
  16
  17 class SoundcloudIE(InfoExtractor):
  18     """Information extractor for soundcloud.com
  19        To access the media, the uid of the song and a stream token
  20        must be extracted from the page source and the script must make
  21        a request to media.soundcloud.com/crossdomain.xml. Then
  22        the media can be grabbed by requesting from an url composed
  23        of the stream token and uid
  24      """
  25
  26     _VALID_URL = r'''^(?:https?://)?
  27                     (?:(?:(?:www\.|m\.)?soundcloud\.com/
  28                             (?P<uploader>[\w\d-]+)/
  29                             (?!sets/)(?P<title>[\w\d-]+)/?
  30                             (?P<token>[^?]+?)?(?:[?].*)?$)
  31                        |(?:api\.soundcloud\.com/tracks/(?P<track_id>\d+))
  32                        |(?P<widget>w\.soundcloud\.com/player/?.*?url=.*)
  33                     )
  34                     '''
  35     IE_NAME = u'soundcloud'
  36     _TESTS = [
  37         {
  38             u'url': u'http://soundcloud.com/ethmusic/lostin-powers-she-so-heavy',
  39             u'file': u'62986583.mp3',
  40             u'md5': u'ebef0a451b909710ed1d7787dddbf0d7',
  41             u'info_dict': {
  42                 u"upload_date": u"20121011",
  43                 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",
  44                 u"uploader": u"E.T. ExTerrestrial Music",
  45                 u"title": u"Lostin Powers - She so Heavy (SneakPreview) Adrian Ackers Blueprint 1"
  46             }
  47         },
  48         # not streamable song
  49         {
  50             u'url': u'https://soundcloud.com/the-concept-band/goldrushed-mastered?in=the-concept-band/sets/the-royal-concept-ep',
  51             u'info_dict': {
  52                 u'id': u'47127627',
  53                 u'ext': u'mp3',
  54                 u'title': u'Goldrushed',
  55                 u'uploader': u'The Royal Concept',
  56                 u'upload_date': u'20120521',
  57             },
  58             u'params': {
  59                 # rtmp
  60                 u'skip_download': True,
  61             },
  62         },
  63         # private link
  64         {
  65             u'url': u'https://soundcloud.com/jaimemf/youtube-dl-test-video-a-y-baw/s-8Pjrp',
  66             u'md5': u'aa0dd32bfea9b0c5ef4f02aacd080604',
  67             u'info_dict': {
  68                 u'id': u'123998367',
  69                 u'ext': u'mp3',
  70                 u'title': u'Youtube - Dl Test Video \'\' Ä↭',
  71                 u'uploader': u'jaimeMF',
  72                 u'description': u'test chars:  \"\'/\\ä↭',
  73                 u'upload_date': u'20131209',
  74             },
  75         },
  76         # downloadable song
  77         {
  78             u'url': u'https://soundcloud.com/simgretina/just-your-problem-baby-1',
  79             u'md5': u'56a8b69568acaa967b4c49f9d1d52d19',
  80             u'info_dict': {
  81                 u'id': u'105614606',
  82                 u'ext': u'wav',
  83                 u'title': u'Just Your Problem Baby (Acapella)',
  84                 u'description': u'Vocals',
  85                 u'uploader': u'Sim Gretina',
  86                 u'upload_date': u'20130815',
  87             },
  88         },
  89     ]
  90
  91     _CLIENT_ID = 'b45b1aa10f1ac2941910a7f0d10f8e28'
  92     _IPHONE_CLIENT_ID = '376f225bf427445fc4bfb6b99b72e0bf'
  93
  94     @classmethod
  95     def suitable(cls, url):
  96         return re.match(cls._VALID_URL, url, flags=re.VERBOSE) is not None
  97
  98     def report_resolve(self, video_id):
  99         """Report information extraction."""
 100         self.to_screen(u'%s: Resolving id' % video_id)
 101
 102     @classmethod
 103     def _resolv_url(cls, url):
 104         return 'http://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID
 105
 106     def _extract_info_dict(self, info, full_title=None, quiet=False, secret_token=None):
 107         track_id = compat_str(info['id'])
 108         name = full_title or track_id
 109         if quiet:
 110             self.report_extraction(name)
 111
 112         thumbnail = info['artwork_url']
 113         if thumbnail is not None:
 114             thumbnail = thumbnail.replace('-large', '-t500x500')
 115         ext = u'mp3'
 116         result = {
 117             'id': track_id,
 118             'uploader': info['user']['username'],
 119             'upload_date': unified_strdate(info['created_at']),
 120             'title': info['title'],
 121             'description': info['description'],
 122             'thumbnail': thumbnail,
 123         }
 124         if info.get('downloadable', False):
 125             # We can build a direct link to the song
 126             format_url = (
 127                 u'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(
 128                     track_id, self._CLIENT_ID))
 129             result['formats'] = [{
 130                 'format_id': 'download',
 131                 'ext': info.get('original_format', u'mp3'),
 132                 'url': format_url,
 133                 'vcodec': 'none',
 134             }]
 135         else:
 136             # We have to retrieve the url
 137             streams_url = ('http://api.soundcloud.com/i1/tracks/{0}/streams?'
 138                 'client_id={1}&secret_token={2}'.format(track_id, self._IPHONE_CLIENT_ID, secret_token))
 139             stream_json = self._download_webpage(
 140                 streams_url,
 141                 track_id, u'Downloading track url')
 142
 143             formats = []
 144             format_dict = json.loads(stream_json)
 145             for key, stream_url in format_dict.items():
 146                 if key.startswith(u'http'):
 147                     formats.append({
 148                         'format_id': key,
 149                         'ext': ext,
 150                         'url': stream_url,
 151                         'vcodec': 'none',
 152                     })
 153                 elif key.startswith(u'rtmp'):
 154                     # The url doesn't have an rtmp app, we have to extract the playpath
 155                     url, path = stream_url.split('mp3:', 1)
 156                     formats.append({
 157                         'format_id': key,
 158                         'url': url,
 159                         'play_path': 'mp3:' + path,
 160                         'ext': ext,
 161                         'vcodec': 'none',
 162                     })
 163
 164             if not formats:
 165                 # We fallback to the stream_url in the original info, this
 166                 # cannot be always used, sometimes it can give an HTTP 404 error
 167                 formats.append({
 168                     'format_id': u'fallback',
 169                     'url': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
 170                     'ext': ext,
 171                     'vcodec': 'none',
 172                 })
 173
 174             def format_pref(f):
 175                 if f['format_id'].startswith('http'):
 176                     return 2
 177                 if f['format_id'].startswith('rtmp'):
 178                     return 1
 179                 return 0
 180
 181             formats.sort(key=format_pref)
 182             result['formats'] = formats
 183
 184         return result
 185
 186     def _real_extract(self, url):
 187         mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
 188         if mobj is None:
 189             raise ExtractorError(u'Invalid URL: %s' % url)
 190
 191         track_id = mobj.group('track_id')
 192         token = None
 193         if track_id is not None:
 194             info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 195             full_title = track_id
 196         elif mobj.group('widget'):
 197             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 198             return self.url_result(query['url'][0], ie='Soundcloud')
 199         else:
 200             # extract uploader (which is in the url)
 201             uploader = mobj.group('uploader')
 202             # extract simple title (uploader + slug of song title)
 203             slug_title =  mobj.group('title')
 204             token = mobj.group('token')
 205             full_title = resolve_title = '%s/%s' % (uploader, slug_title)
 206             if token:
 207                 resolve_title += '/%s' % token
 208
 209             self.report_resolve(full_title)
 210
 211             url = 'http://soundcloud.com/%s' % resolve_title
 212             info_json_url = self._resolv_url(url)
 213         info_json = self._download_webpage(info_json_url, full_title, u'Downloading info JSON')
 214
 215         info = json.loads(info_json)
 216         return self._extract_info_dict(info, full_title, secret_token=token)
 217
 218 class SoundcloudSetIE(SoundcloudIE):
 219     _VALID_URL = r'^(?:https?://)?(?:www\.)?soundcloud\.com/([\w\d-]+)/sets/([\w\d-]+)(?:[?].*)?$'
 220     IE_NAME = u'soundcloud:set'
 221     # it's in tests/test_playlists.py
 222     _TESTS = []
 223
 224     def _real_extract(self, url):
 225         mobj = re.match(self._VALID_URL, url)
 226         if mobj is None:
 227             raise ExtractorError(u'Invalid URL: %s' % url)
 228
 229         # extract uploader (which is in the url)
 230         uploader = mobj.group(1)
 231         # extract simple title (uploader + slug of song title)
 232         slug_title =  mobj.group(2)
 233         full_title = '%s/sets/%s' % (uploader, slug_title)
 234
 235         self.report_resolve(full_title)
 236
 237         url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 238         resolv_url = self._resolv_url(url)
 239         info_json = self._download_webpage(resolv_url, full_title)
 240
 241         info = json.loads(info_json)
 242         if 'errors' in info:
 243             for err in info['errors']:
 244                 self._downloader.report_error(u'unable to download video webpage: %s' % compat_str(err['error_message']))
 245             return
 246
 247         self.report_extraction(full_title)
 248         return {'_type': 'playlist',
 249                 'entries': [self._extract_info_dict(track) for track in info['tracks']],
 250                 'id': info['id'],
 251                 'title': info['title'],
 252                 }
 253
 254
 255 class SoundcloudUserIE(SoundcloudIE):
 256     _VALID_URL = r'https?://(www\.)?soundcloud\.com/(?P<user>[^/]+)(/?(tracks/)?)?(\?.*)?$'
 257     IE_NAME = u'soundcloud:user'
 258
 259     # it's in tests/test_playlists.py
 260     _TESTS = []
 261
 262     def _real_extract(self, url):
 263         mobj = re.match(self._VALID_URL, url)
 264         uploader = mobj.group('user')
 265
 266         url = 'http://soundcloud.com/%s/' % uploader
 267         resolv_url = self._resolv_url(url)
 268         user_json = self._download_webpage(resolv_url, uploader,
 269             u'Downloading user info')
 270         user = json.loads(user_json)
 271
 272         tracks = []
 273         for i in itertools.count():
 274             data = compat_urllib_parse.urlencode({'offset': i*50,
 275                                                   'client_id': self._CLIENT_ID,
 276                                                   })
 277             tracks_url = 'http://api.soundcloud.com/users/%s/tracks.json?' % user['id'] + data
 278             response = self._download_webpage(tracks_url, uploader,
 279                 u'Downloading tracks page %s' % (i+1))
 280             new_tracks = json.loads(response)
 281             tracks.extend(self._extract_info_dict(track, quiet=True) for track in new_tracks)
 282             if len(new_tracks) < 50:
 283                 break
 284
 285         return {
 286             '_type': 'playlist',
 287             'id': compat_str(user['id']),
 288             'title': user['username'],
 289             'entries': tracks,
 290         }