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

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5
   6 from .common import InfoExtractor
   7 from ..compat import compat_urlparse
   8 from ..utils import (
   9     determine_ext,
  10     float_or_none,
  11     int_or_none,
  12     js_to_json,
  13     mimetype2ext,
  14 )
  15
  16
  17 class JWPlatformBaseIE(InfoExtractor):
  18     @staticmethod
  19     def _find_jwplayer_data(webpage):
  20         # TODO: Merge this with JWPlayer-related codes in generic.py
  21
  22         mobj = re.search(
  23             r'jwplayer\((?P<quote>[\'"])[^\'" ]+(?P=quote)\)\.setup\s*\((?P<options>[^)]+)\)',
  24             webpage)
  25         if mobj:
  26             return mobj.group('options')
  27
  28     def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
  29         jwplayer_data = self._parse_json(
  30             self._find_jwplayer_data(webpage), video_id,
  31             transform_source=js_to_json)
  32         return self._parse_jwplayer_data(
  33             jwplayer_data, video_id, *args, **kwargs)
  34
  35     def _parse_jwplayer_data(self, jwplayer_data, video_id=None, require_title=True,
  36                              m3u8_id=None, mpd_id=None, rtmp_params=None, base_url=None):
  37         # JWPlayer backward compatibility: flattened playlists
  38         # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/api/config.js#L81-L96
  39         if 'playlist' not in jwplayer_data:
  40             jwplayer_data = {'playlist': [jwplayer_data]}
  41
  42         entries = []
  43
  44         # JWPlayer backward compatibility: single playlist item
  45         # https://github.com/jwplayer/jwplayer/blob/v7.7.0/src/js/playlist/playlist.js#L10
  46         if not isinstance(jwplayer_data['playlist'], list):
  47             jwplayer_data['playlist'] = [jwplayer_data['playlist']]
  48
  49         for video_data in jwplayer_data['playlist']:
  50             # JWPlayer backward compatibility: flattened sources
  51             # https://github.com/jwplayer/jwplayer/blob/v7.4.3/src/js/playlist/item.js#L29-L35
  52             if 'sources' not in video_data:
  53                 video_data['sources'] = [video_data]
  54
  55             this_video_id = video_id or video_data['mediaid']
  56
  57             formats = []
  58             for source in video_data['sources']:
  59                 source_url = self._proto_relative_url(source['file'])
  60                 if base_url:
  61                     source_url = compat_urlparse.urljoin(base_url, source_url)
  62                 source_type = source.get('type') or ''
  63                 ext = mimetype2ext(source_type) or determine_ext(source_url)
  64                 if source_type == 'hls' or ext == 'm3u8':
  65                     formats.extend(self._extract_m3u8_formats(
  66                         source_url, this_video_id, 'mp4', 'm3u8_native', m3u8_id=m3u8_id, fatal=False))
  67                 elif ext == 'mpd':
  68                     formats.extend(self._extract_mpd_formats(
  69                         source_url, this_video_id, mpd_id=mpd_id, fatal=False))
  70                 # https://github.com/jwplayer/jwplayer/blob/master/src/js/providers/default.js#L67
  71                 elif source_type.startswith('audio') or ext in ('oga', 'aac', 'mp3', 'mpeg', 'vorbis'):
  72                     formats.append({
  73                         'url': source_url,
  74                         'vcodec': 'none',
  75                         'ext': ext,
  76                     })
  77                 else:
  78                     height = int_or_none(source.get('height'))
  79                     if height is None:
  80                         # Often no height is provided but there is a label in
  81                         # format like 1080p.
  82                         height = int_or_none(self._search_regex(
  83                             r'^(\d{3,})[pP]$', source.get('label') or '',
  84                             'height', default=None))
  85                     a_format = {
  86                         'url': source_url,
  87                         'width': int_or_none(source.get('width')),
  88                         'height': height,
  89                         'ext': ext,
  90                     }
  91                     if source_url.startswith('rtmp'):
  92                         a_format['ext'] = 'flv'
  93
  94                         # See com/longtailvideo/jwplayer/media/RTMPMediaProvider.as
  95                         # of jwplayer.flash.swf
  96                         rtmp_url_parts = re.split(
  97                             r'((?:mp4|mp3|flv):)', source_url, 1)
  98                         if len(rtmp_url_parts) == 3:
  99                             rtmp_url, prefix, play_path = rtmp_url_parts
 100                             a_format.update({
 101                                 'url': rtmp_url,
 102                                 'play_path': prefix + play_path,
 103                             })
 104                         if rtmp_params:
 105                             a_format.update(rtmp_params)
 106                     formats.append(a_format)
 107             self._sort_formats(formats)
 108
 109             subtitles = {}
 110             tracks = video_data.get('tracks')
 111             if tracks and isinstance(tracks, list):
 112                 for track in tracks:
 113                     if track.get('file') and track.get('kind') == 'captions':
 114                         subtitles.setdefault(track.get('label') or 'en', []).append({
 115                             'url': self._proto_relative_url(track['file'])
 116                         })
 117
 118             entries.append({
 119                 'id': this_video_id,
 120                 'title': video_data['title'] if require_title else video_data.get('title'),
 121                 'description': video_data.get('description'),
 122                 'thumbnail': self._proto_relative_url(video_data.get('image')),
 123                 'timestamp': int_or_none(video_data.get('pubdate')),
 124                 'duration': float_or_none(jwplayer_data.get('duration')),
 125                 'subtitles': subtitles,
 126                 'formats': formats,
 127             })
 128         if len(entries) == 1:
 129             return entries[0]
 130         else:
 131             return self.playlist_result(entries)
 132
 133
 134 class JWPlatformIE(JWPlatformBaseIE):
 135     _VALID_URL = r'(?:https?://content\.jwplatform\.com/(?:feeds|players|jw6)/|jwplatform:)(?P<id>[a-zA-Z0-9]{8})'
 136     _TEST = {
 137         'url': 'http://content.jwplatform.com/players/nPripu9l-ALJ3XQCI.js',
 138         'md5': 'fa8899fa601eb7c83a64e9d568bdf325',
 139         'info_dict': {
 140             'id': 'nPripu9l',
 141             'ext': 'mov',
 142             'title': 'Big Buck Bunny Trailer',
 143             'description': 'Big Buck Bunny is a short animated film by the Blender Institute. It is made using free and open source software.',
 144             'upload_date': '20081127',
 145             'timestamp': 1227796140,
 146         }
 147     }
 148
 149     @staticmethod
 150     def _extract_url(webpage):
 151         mobj = re.search(
 152             r'<script[^>]+?src=["\'](?P<url>(?:https?:)?//content.jwplatform.com/players/[a-zA-Z0-9]{8})',
 153             webpage)
 154         if mobj:
 155             return mobj.group('url')
 156
 157     def _real_extract(self, url):
 158         video_id = self._match_id(url)
 159         json_data = self._download_json('http://content.jwplatform.com/feeds/%s.json' % video_id, video_id)
 160         return self._parse_jwplayer_data(json_data, video_id)