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