_ Git - youtube-dl/blob - 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                 'description': 'From Stockholm Sweden\r\nPovel / Magnus / Filip / David\r\nwww.theroyalconcept.com',
  58                 'uploader': 'The Royal Concept',
  59                 'upload_date': '20120521',
  60             },
  61             'params': {
  62                 # rtmp
  63                 'skip_download': True,
  64             },
  65         },
  66         # private link
  67         {
  68             'url': 'https://soundcloud.com/jaimemf/youtube-dl-test-video-a-y-baw/s-8Pjrp',
  69             'md5': 'aa0dd32bfea9b0c5ef4f02aacd080604',
  70             'info_dict': {
  71                 'id': '123998367',
  72                 'ext': 'mp3',
  73                 'title': 'Youtube - Dl Test Video \'\' Ä↭',
  74                 'uploader': 'jaimeMF',
  75                 'description': 'test chars:  \"\'/\\ä↭',
  76                 'upload_date': '20131209',
  77             },
  78         },
  79         # downloadable song
  80         {
  81             'url': 'https://soundcloud.com/simgretina/just-your-problem-baby-1',
  82             'md5': '56a8b69568acaa967b4c49f9d1d52d19',
  83             'info_dict': {
  84                 'id': '105614606',
  85                 'ext': 'wav',
  86                 'title': 'Just Your Problem Baby (Acapella)',
  87                 'description': 'Vocals',
  88                 'uploader': 'Sim Gretina',
  89                 'upload_date': '20130815',
  90             },
  91         },
  92     ]
  93
  94     _CLIENT_ID = 'b45b1aa10f1ac2941910a7f0d10f8e28'
  95     _IPHONE_CLIENT_ID = '376f225bf427445fc4bfb6b99b72e0bf'
  96
  97     @classmethod
  98     def suitable(cls, url):
  99         return re.match(cls._VALID_URL, url, flags=re.VERBOSE) is not None
 100
 101     def report_resolve(self, video_id):
 102         """Report information extraction."""
 103         self.to_screen('%s: Resolving id' % video_id)
 104
 105     @classmethod
 106     def _resolv_url(cls, url):
 107         return 'http://api.soundcloud.com/resolve.json?url=' + url + '&client_id=' + cls._CLIENT_ID
 108
 109     def _extract_info_dict(self, info, full_title=None, quiet=False, secret_token=None):
 110         track_id = compat_str(info['id'])
 111         name = full_title or track_id
 112         if quiet:
 113             self.report_extraction(name)
 114
 115         thumbnail = info['artwork_url']
 116         if thumbnail is not None:
 117             thumbnail = thumbnail.replace('-large', '-t500x500')
 118         ext = 'mp3'
 119         result = {
 120             'id': track_id,
 121             'uploader': info['user']['username'],
 122             'upload_date': unified_strdate(info['created_at']),
 123             'title': info['title'],
 124             'description': info['description'],
 125             'thumbnail': thumbnail,
 126         }
 127         formats = []
 128         if info.get('downloadable', False):
 129             # We can build a direct link to the song
 130             format_url = (
 131                 'https://api.soundcloud.com/tracks/{0}/download?client_id={1}'.format(
 132                     track_id, self._CLIENT_ID))
 133             formats.append({
 134                 'format_id': 'download',
 135                 'ext': info.get('original_format', 'mp3'),
 136                 'url': format_url,
 137                 'vcodec': 'none',
 138                 'preference': 10,
 139             })
 140
 141         # We have to retrieve the url
 142         streams_url = ('http://api.soundcloud.com/i1/tracks/{0}/streams?'
 143             'client_id={1}&secret_token={2}'.format(track_id, self._IPHONE_CLIENT_ID, secret_token))
 144         stream_json = self._download_webpage(
 145             streams_url,
 146             track_id, 'Downloading track url')
 147
 148         format_dict = json.loads(stream_json)
 149         for key, stream_url in format_dict.items():
 150             if key.startswith('http'):
 151                 formats.append({
 152                     'format_id': key,
 153                     'ext': ext,
 154                     'url': stream_url,
 155                     'vcodec': 'none',
 156                 })
 157             elif key.startswith('rtmp'):
 158                 # The url doesn't have an rtmp app, we have to extract the playpath
 159                 url, path = stream_url.split('mp3:', 1)
 160                 formats.append({
 161                     'format_id': key,
 162                     'url': url,
 163                     'play_path': 'mp3:' + path,
 164                     'ext': ext,
 165                     'vcodec': 'none',
 166                 })
 167
 168             if not formats:
 169                 # We fallback to the stream_url in the original info, this
 170                 # cannot be always used, sometimes it can give an HTTP 404 error
 171                 formats.append({
 172                     'format_id': 'fallback',
 173                     'url': info['stream_url'] + '?client_id=' + self._CLIENT_ID,
 174                     'ext': ext,
 175                     'vcodec': 'none',
 176                 })
 177
 178             for f in formats:
 179                 if f['format_id'].startswith('http'):
 180                     f['protocol'] = 'http'
 181                 if f['format_id'].startswith('rtmp'):
 182                     f['protocol'] = 'rtmp'
 183
 184             self._sort_formats(formats)
 185             result['formats'] = formats
 186
 187         return result
 188
 189     def _real_extract(self, url):
 190         mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
 191         if mobj is None:
 192             raise ExtractorError('Invalid URL: %s' % url)
 193
 194         track_id = mobj.group('track_id')
 195         token = None
 196         if track_id is not None:
 197             info_json_url = 'http://api.soundcloud.com/tracks/' + track_id + '.json?client_id=' + self._CLIENT_ID
 198             full_title = track_id
 199         elif mobj.group('player'):
 200             query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 201             return self.url_result(query['url'][0], ie='Soundcloud')
 202         else:
 203             # extract uploader (which is in the url)
 204             uploader = mobj.group('uploader')
 205             # extract simple title (uploader + slug of song title)
 206             slug_title =  mobj.group('title')
 207             token = mobj.group('token')
 208             full_title = resolve_title = '%s/%s' % (uploader, slug_title)
 209             if token:
 210                 resolve_title += '/%s' % token
 211
 212             self.report_resolve(full_title)
 213
 214             url = 'http://soundcloud.com/%s' % resolve_title
 215             info_json_url = self._resolv_url(url)
 216         info_json = self._download_webpage(info_json_url, full_title, 'Downloading info JSON')
 217
 218         info = json.loads(info_json)
 219         return self._extract_info_dict(info, full_title, secret_token=token)
 220
 221 class SoundcloudSetIE(SoundcloudIE):
 222     _VALID_URL = r'https?://(?:www\.)?soundcloud\.com/([\w\d-]+)/sets/([\w\d-]+)'
 223     IE_NAME = 'soundcloud:set'
 224     # it's in tests/test_playlists.py
 225     _TESTS = []
 226
 227     def _real_extract(self, url):
 228         mobj = re.match(self._VALID_URL, url)
 229         if mobj is None:
 230             raise ExtractorError('Invalid URL: %s' % url)
 231
 232         # extract uploader (which is in the url)
 233         uploader = mobj.group(1)
 234         # extract simple title (uploader + slug of song title)
 235         slug_title =  mobj.group(2)
 236         full_title = '%s/sets/%s' % (uploader, slug_title)
 237
 238         self.report_resolve(full_title)
 239
 240         url = 'http://soundcloud.com/%s/sets/%s' % (uploader, slug_title)
 241         resolv_url = self._resolv_url(url)
 242         info_json = self._download_webpage(resolv_url, full_title)
 243
 244         info = json.loads(info_json)
 245         if 'errors' in info:
 246             for err in info['errors']:
 247                 self._downloader.report_error('unable to download video webpage: %s' % compat_str(err['error_message']))
 248             return
 249
 250         self.report_extraction(full_title)
 251         return {'_type': 'playlist',
 252                 'entries': [self._extract_info_dict(track) for track in info['tracks']],
 253                 'id': info['id'],
 254                 'title': info['title'],
 255                 }
 256
 257
 258 class SoundcloudUserIE(SoundcloudIE):
 259     _VALID_URL = r'https?://(www\.)?soundcloud\.com/(?P<user>[^/]+)(/?(tracks/)?)?(\?.*)?$'
 260     IE_NAME = 'soundcloud:user'
 261
 262     # it's in tests/test_playlists.py
 263     _TESTS = []
 264
 265     def _real_extract(self, url):
 266         mobj = re.match(self._VALID_URL, url)
 267         uploader = mobj.group('user')
 268
 269         url = 'http://soundcloud.com/%s/' % uploader
 270         resolv_url = self._resolv_url(url)
 271         user_json = self._download_webpage(resolv_url, uploader,
 272             'Downloading user info')
 273         user = json.loads(user_json)
 274
 275         tracks = []
 276         for i in itertools.count():
 277             data = compat_urllib_parse.urlencode({'offset': i*50,
 278                                                   'client_id': self._CLIENT_ID,
 279                                                   })
 280             tracks_url = 'http://api.soundcloud.com/users/%s/tracks.json?' % user['id'] + data
 281             response = self._download_webpage(tracks_url, uploader,
 282                 'Downloading tracks page %s' % (i+1))
 283             new_tracks = json.loads(response)
 284             tracks.extend(self._extract_info_dict(track, quiet=True) for track in new_tracks)
 285             if len(new_tracks) < 50:
 286                 break
 287
 288         return {
 289             '_type': 'playlist',
 290             'id': compat_str(user['id']),
 291             'title': user['username'],
 292             'entries': tracks,
 293         }