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

   1 from __future__ import unicode_literals
   2
   3 import json
   4 import time
   5
   6 from .common import InfoExtractor
   7 from ..compat import compat_urllib_parse
   8 from ..utils import (
   9     ExtractorError,
  10     sanitized_Request,
  11 )
  12
  13
  14 class HypemIE(InfoExtractor):
  15     _VALID_URL = r'http://(?:www\.)?hypem\.com/track/(?P<id>[^/]+)/'
  16     _TEST = {
  17         'url': 'http://hypem.com/track/1v6ga/BODYWORK+-+TAME',
  18         'md5': 'b9cc91b5af8995e9f0c1cee04c575828',
  19         'info_dict': {
  20             'id': '1v6ga',
  21             'ext': 'mp3',
  22             'title': 'Tame',
  23             'uploader': 'BODYWORK',
  24         }
  25     }
  26
  27     def _real_extract(self, url):
  28         track_id = self._match_id(url)
  29
  30         data = {'ax': 1, 'ts': time.time()}
  31         request = sanitized_Request(url + '?' + compat_urllib_parse.urlencode(data))
  32         response, urlh = self._download_webpage_handle(
  33             request, track_id, 'Downloading webpage with the url')
  34
  35         html_tracks = self._html_search_regex(
  36             r'(?ms)<script type="application/json" id="displayList-data">(.+?)</script>',
  37             response, 'tracks')
  38         try:
  39             track_list = json.loads(html_tracks)
  40             track = track_list['tracks'][0]
  41         except ValueError:
  42             raise ExtractorError('Hypemachine contained invalid JSON.')
  43
  44         key = track['key']
  45         track_id = track['id']
  46         title = track['song']
  47
  48         request = sanitized_Request(
  49             'http://hypem.com/serve/source/%s/%s' % (track_id, key),
  50             '', {'Content-Type': 'application/json'})
  51         song_data = self._download_json(request, track_id, 'Downloading metadata')
  52         final_url = song_data['url']
  53         artist = track.get('artist')
  54
  55         return {
  56             'id': track_id,
  57             'url': final_url,
  58             'ext': 'mp3',
  59             'title': title,
  60             'uploader': artist,
  61         }