_ Git - youtube-dl/blob - youtube_dl/extractor/periscope.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5
   6 from .common import InfoExtractor
   7 from ..utils import (
   8     parse_iso8601,
   9     unescapeHTML,
  10 )
  11
  12
  13 class PeriscopeBaseIE(InfoExtractor):
  14     def _call_api(self, method, query, item_id):
  15         return self._download_json(
  16             'https://api.periscope.tv/api/v2/%s' % method,
  17             item_id, query=query)
  18
  19
  20 class PeriscopeIE(PeriscopeBaseIE):
  21     IE_DESC = 'Periscope'
  22     IE_NAME = 'periscope'
  23     _VALID_URL = r'https?://(?:www\.)?(?:periscope|pscp)\.tv/[^/]+/(?P<id>[^/?#]+)'
  24     # Alive example URLs can be found here http://onperiscope.com/
  25     _TESTS = [{
  26         'url': 'https://www.periscope.tv/w/aJUQnjY3MjA3ODF8NTYxMDIyMDl2zCg2pECBgwTqRpQuQD352EMPTKQjT4uqlM3cgWFA-g==',
  27         'md5': '65b57957972e503fcbbaeed8f4fa04ca',
  28         'info_dict': {
  29             'id': '56102209',
  30             'ext': 'mp4',
  31             'title': 'Bec Boop - 🚠✈️🇬🇧 Fly above #London in Emirates Air Line cable car at night 🇬🇧✈️🚠 #BoopScope 🎀💗',
  32             'timestamp': 1438978559,
  33             'upload_date': '20150807',
  34             'uploader': 'Bec Boop',
  35             'uploader_id': '1465763',
  36         },
  37         'skip': 'Expires in 24 hours',
  38     }, {
  39         'url': 'https://www.periscope.tv/w/1ZkKzPbMVggJv',
  40         'only_matching': True,
  41     }, {
  42         'url': 'https://www.periscope.tv/bastaakanoggano/1OdKrlkZZjOJX',
  43         'only_matching': True,
  44     }, {
  45         'url': 'https://www.periscope.tv/w/1ZkKzPbMVggJv',
  46         'only_matching': True,
  47     }]
  48
  49     @staticmethod
  50     def _extract_url(webpage):
  51         mobj = re.search(
  52             r'<iframe[^>]+src=([\'"])(?P<url>(?:https?:)?//(?:www\.)?(?:periscope|pscp)\.tv/(?:(?!\1).)+)\1', webpage)
  53         if mobj:
  54             return mobj.group('url')
  55
  56     def _real_extract(self, url):
  57         token = self._match_id(url)
  58
  59         broadcast_data = self._call_api(
  60             'getBroadcastPublic', {'broadcast_id': token}, token)
  61         broadcast = broadcast_data['broadcast']
  62         status = broadcast['status']
  63
  64         user = broadcast_data.get('user', {})
  65
  66         uploader = broadcast.get('user_display_name') or user.get('display_name')
  67         uploader_id = (broadcast.get('username') or user.get('username') or
  68                        broadcast.get('user_id') or user.get('id'))
  69
  70         title = '%s - %s' % (uploader, status) if uploader else status
  71         state = broadcast.get('state').lower()
  72         if state == 'running':
  73             title = self._live_title(title)
  74         timestamp = parse_iso8601(broadcast.get('created_at'))
  75
  76         thumbnails = [{
  77             'url': broadcast[image],
  78         } for image in ('image_url', 'image_url_small') if broadcast.get(image)]
  79
  80         stream = self._call_api(
  81             'getAccessPublic', {'broadcast_id': token}, token)
  82
  83         video_urls = set()
  84         formats = []
  85         for format_id in ('replay', 'rtmp', 'hls', 'https_hls', 'lhls', 'lhlsweb'):
  86             video_url = stream.get(format_id + '_url')
  87             if not video_url or video_url in video_urls:
  88                 continue
  89             video_urls.add(video_url)
  90             if format_id != 'rtmp':
  91                 formats.extend(self._extract_m3u8_formats(
  92                     video_url, token, 'mp4',
  93                     entry_protocol='m3u8_native'
  94                     if state in ('ended', 'timed_out') else 'm3u8',
  95                     m3u8_id=format_id, fatal=False))
  96                 continue
  97             formats.append({
  98                 'url': video_url,
  99                 'ext': 'flv' if format_id == 'rtmp' else 'mp4',
 100             })
 101         self._sort_formats(formats)
 102
 103         return {
 104             'id': broadcast.get('id') or token,
 105             'title': title,
 106             'timestamp': timestamp,
 107             'uploader': uploader,
 108             'uploader_id': uploader_id,
 109             'thumbnails': thumbnails,
 110             'formats': formats,
 111         }
 112
 113
 114 class PeriscopeUserIE(PeriscopeBaseIE):
 115     _VALID_URL = r'https?://(?:www\.)?(?:periscope|pscp)\.tv/(?P<id>[^/]+)/?$'
 116     IE_DESC = 'Periscope user videos'
 117     IE_NAME = 'periscope:user'
 118
 119     _TEST = {
 120         'url': 'https://www.periscope.tv/LularoeHusbandMike/',
 121         'info_dict': {
 122             'id': 'LularoeHusbandMike',
 123             'title': 'LULAROE HUSBAND MIKE',
 124             'description': 'md5:6cf4ec8047768098da58e446e82c82f0',
 125         },
 126         # Periscope only shows videos in the last 24 hours, so it's possible to
 127         # get 0 videos
 128         'playlist_mincount': 0,
 129     }
 130
 131     def _real_extract(self, url):
 132         user_name = self._match_id(url)
 133
 134         webpage = self._download_webpage(url, user_name)
 135
 136         data_store = self._parse_json(
 137             unescapeHTML(self._search_regex(
 138                 r'data-store=(["\'])(?P<data>.+?)\1',
 139                 webpage, 'data store', default='{}', group='data')),
 140             user_name)
 141
 142         user = list(data_store['UserCache']['users'].values())[0]['user']
 143         user_id = user['id']
 144         session_id = data_store['SessionToken']['public']['broadcastHistory']['token']['session_id']
 145
 146         broadcasts = self._call_api(
 147             'getUserBroadcastsPublic',
 148             {'user_id': user_id, 'session_id': session_id},
 149             user_name)['broadcasts']
 150
 151         broadcast_ids = [
 152             broadcast['id'] for broadcast in broadcasts if broadcast.get('id')]
 153
 154         title = user.get('display_name') or user.get('username') or user_name
 155         description = user.get('description')
 156
 157         entries = [
 158             self.url_result(
 159                 'https://www.periscope.tv/%s/%s' % (user_name, broadcast_id))
 160             for broadcast_id in broadcast_ids]
 161
 162         return self.playlist_result(entries, user_id, title, description)