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