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

   1 from __future__ import unicode_literals
   2
   3 import json
   4 import re
   5
   6 from .common import InfoExtractor
   7
   8 from ..compat import compat_str
   9 from ..utils import (
  10     float_or_none,
  11     int_or_none,
  12     try_get,
  13     url_or_none,
  14 )
  15
  16
  17 class TEDIE(InfoExtractor):
  18     IE_NAME = 'ted'
  19     _VALID_URL = r'''(?x)
  20         (?P<proto>https?://)
  21         (?P<type>www|embed(?:-ssl)?)(?P<urlmain>\.ted\.com/
  22         (
  23             (?P<type_playlist>playlists(?:/\d+)?) # We have a playlist
  24             |
  25             ((?P<type_talk>talks)) # We have a simple talk
  26             |
  27             (?P<type_watch>watch)/[^/]+/[^/]+
  28         )
  29         (/lang/(.*?))? # The url may contain the language
  30         /(?P<name>[\w-]+) # Here goes the name and then ".html"
  31         .*)$
  32         '''
  33     _TESTS = [{
  34         'url': 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html',
  35         'md5': 'b0ce2b05ca215042124fbc9e3886493a',
  36         'info_dict': {
  37             'id': '102',
  38             'ext': 'mp4',
  39             'title': 'The illusion of consciousness',
  40             'description': ('Philosopher Dan Dennett makes a compelling '
  41                             'argument that not only don\'t we understand our own '
  42                             'consciousness, but that half the time our brains are '
  43                             'actively fooling us.'),
  44             'uploader': 'Dan Dennett',
  45             'width': 853,
  46             'duration': 1308,
  47             'view_count': int,
  48             'comment_count': int,
  49             'tags': list,
  50         },
  51         'params': {
  52             'skip_download': True,
  53         },
  54     }, {
  55         # missing HTTP bitrates
  56         'url': 'https://www.ted.com/talks/vishal_sikka_the_beauty_and_power_of_algorithms',
  57         'info_dict': {
  58             'id': '6069',
  59             'ext': 'mp4',
  60             'title': 'The beauty and power of algorithms',
  61             'thumbnail': r're:^https?://.+\.jpg',
  62             'description': 'md5:734e352710fb00d840ab87ae31aaf688',
  63             'uploader': 'Vishal Sikka',
  64         },
  65         'params': {
  66             'skip_download': True,
  67         },
  68     }, {
  69         'url': 'http://www.ted.com/talks/gabby_giffords_and_mark_kelly_be_passionate_be_courageous_be_your_best',
  70         'md5': 'e6b9617c01a7970ceac8bb2c92c346c0',
  71         'info_dict': {
  72             'id': '1972',
  73             'ext': 'mp4',
  74             'title': 'Be passionate. Be courageous. Be your best.',
  75             'uploader': 'Gabby Giffords and Mark Kelly',
  76             'description': 'md5:5174aed4d0f16021b704120360f72b92',
  77             'duration': 1128,
  78         },
  79         'params': {
  80             'skip_download': True,
  81         },
  82     }, {
  83         'url': 'http://www.ted.com/playlists/who_are_the_hackers',
  84         'info_dict': {
  85             'id': '10',
  86             'title': 'Who are the hackers?',
  87         },
  88         'playlist_mincount': 6,
  89     }, {
  90         # contains a youtube video
  91         'url': 'https://www.ted.com/talks/douglas_adams_parrots_the_universe_and_everything',
  92         'add_ie': ['Youtube'],
  93         'info_dict': {
  94             'id': '_ZG8HBuDjgc',
  95             'ext': 'webm',
  96             'title': 'Douglas Adams: Parrots the Universe and Everything',
  97             'description': 'md5:01ad1e199c49ac640cb1196c0e9016af',
  98             'uploader': 'University of California Television (UCTV)',
  99             'uploader_id': 'UCtelevision',
 100             'upload_date': '20080522',
 101         },
 102         'params': {
 103             'skip_download': True,
 104         },
 105     }, {
 106         # no nativeDownloads
 107         'url': 'https://www.ted.com/talks/tom_thum_the_orchestra_in_my_mouth',
 108         'info_dict': {
 109             'id': '1792',
 110             'ext': 'mp4',
 111             'title': 'The orchestra in my mouth',
 112             'description': 'md5:5d1d78650e2f8dfcbb8ebee2951ac29a',
 113             'uploader': 'Tom Thum',
 114             'view_count': int,
 115             'comment_count': int,
 116             'tags': list,
 117         },
 118         'params': {
 119             'skip_download': True,
 120         },
 121     }]
 122
 123     _NATIVE_FORMATS = {
 124         'low': {'width': 320, 'height': 180},
 125         'medium': {'width': 512, 'height': 288},
 126         'high': {'width': 854, 'height': 480},
 127     }
 128
 129     def _extract_info(self, webpage):
 130         info_json = self._search_regex(
 131             r'(?s)q\(\s*"\w+.init"\s*,\s*({.+})\)\s*</script>',
 132             webpage, 'info json')
 133         return json.loads(info_json)
 134
 135     def _real_extract(self, url):
 136         m = re.match(self._VALID_URL, url, re.VERBOSE)
 137         if m.group('type').startswith('embed'):
 138             desktop_url = m.group('proto') + 'www' + m.group('urlmain')
 139             return self.url_result(desktop_url, 'TED')
 140         name = m.group('name')
 141         if m.group('type_talk'):
 142             return self._talk_info(url, name)
 143         elif m.group('type_watch'):
 144             return self._watch_info(url, name)
 145         else:
 146             return self._playlist_videos_info(url, name)
 147
 148     def _playlist_videos_info(self, url, name):
 149         '''Returns the videos of the playlist'''
 150
 151         webpage = self._download_webpage(url, name,
 152                                          'Downloading playlist webpage')
 153         info = self._extract_info(webpage)
 154
 155         playlist_info = try_get(
 156             info, lambda x: x['__INITIAL_DATA__']['playlist'],
 157             dict) or info['playlist']
 158
 159         playlist_entries = [
 160             self.url_result('http://www.ted.com/talks/' + talk['slug'], self.ie_key())
 161             for talk in try_get(
 162                 info, lambda x: x['__INITIAL_DATA__']['talks'],
 163                 dict) or info['talks']
 164         ]
 165         return self.playlist_result(
 166             playlist_entries,
 167             playlist_id=compat_str(playlist_info['id']),
 168             playlist_title=playlist_info['title'])
 169
 170     def _talk_info(self, url, video_name):
 171         webpage = self._download_webpage(url, video_name)
 172
 173         info = self._extract_info(webpage)
 174
 175         data = try_get(info, lambda x: x['__INITIAL_DATA__'], dict) or info
 176         talk_info = data['talks'][0]
 177
 178         title = talk_info['title'].strip()
 179
 180         native_downloads = try_get(
 181             talk_info,
 182             (lambda x: x['downloads']['nativeDownloads'],
 183              lambda x: x['nativeDownloads']),
 184             dict) or {}
 185
 186         formats = [{
 187             'url': format_url,
 188             'format_id': format_id,
 189             'format': format_id,
 190         } for (format_id, format_url) in native_downloads.items() if format_url is not None]
 191         if formats:
 192             for f in formats:
 193                 finfo = self._NATIVE_FORMATS.get(f['format_id'])
 194                 if finfo:
 195                     f.update(finfo)
 196
 197         player_talk = talk_info['player_talks'][0]
 198
 199         external = player_talk.get('external')
 200         if isinstance(external, dict):
 201             service = external.get('service')
 202             if isinstance(service, compat_str):
 203                 ext_url = None
 204                 if service.lower() == 'youtube':
 205                     ext_url = external.get('code')
 206
 207                 return self.url_result(ext_url or external['uri'])
 208
 209         resources_ = player_talk.get('resources') or talk_info.get('resources')
 210
 211         http_url = None
 212         for format_id, resources in resources_.items():
 213             if format_id == 'h264':
 214                 for resource in resources:
 215                     h264_url = resource.get('file')
 216                     if not h264_url:
 217                         continue
 218                     bitrate = int_or_none(resource.get('bitrate'))
 219                     formats.append({
 220                         'url': h264_url,
 221                         'format_id': '%s-%sk' % (format_id, bitrate),
 222                         'tbr': bitrate,
 223                     })
 224                     if re.search(r'\d+k', h264_url):
 225                         http_url = h264_url
 226             elif format_id == 'rtmp':
 227                 streamer = talk_info.get('streamer')
 228                 if not streamer:
 229                     continue
 230                 for resource in resources:
 231                     formats.append({
 232                         'format_id': '%s-%s' % (format_id, resource.get('name')),
 233                         'url': streamer,
 234                         'play_path': resource['file'],
 235                         'ext': 'flv',
 236                         'width': int_or_none(resource.get('width')),
 237                         'height': int_or_none(resource.get('height')),
 238                         'tbr': int_or_none(resource.get('bitrate')),
 239                     })
 240             elif format_id == 'hls':
 241                 if not isinstance(resources, dict):
 242                     continue
 243                 stream_url = url_or_none(resources.get('stream'))
 244                 if not stream_url:
 245                     continue
 246                 formats.extend(self._extract_m3u8_formats(
 247                     stream_url, video_name, 'mp4', m3u8_id=format_id,
 248                     fatal=False))
 249
 250         m3u8_formats = list(filter(
 251             lambda f: f.get('protocol') == 'm3u8' and f.get('vcodec') != 'none',
 252             formats))
 253         if http_url:
 254             for m3u8_format in m3u8_formats:
 255                 bitrate = self._search_regex(r'(\d+k)', m3u8_format['url'], 'bitrate', default=None)
 256                 if not bitrate:
 257                     continue
 258                 bitrate_url = re.sub(r'\d+k', bitrate, http_url)
 259                 if not self._is_valid_url(
 260                         bitrate_url, video_name, '%s bitrate' % bitrate):
 261                     continue
 262                 f = m3u8_format.copy()
 263                 f.update({
 264                     'url': bitrate_url,
 265                     'format_id': m3u8_format['format_id'].replace('hls', 'http'),
 266                     'protocol': 'http',
 267                 })
 268                 formats.append(f)
 269
 270         audio_download = talk_info.get('audioDownload')
 271         if audio_download:
 272             formats.append({
 273                 'url': audio_download,
 274                 'format_id': 'audio',
 275                 'vcodec': 'none',
 276             })
 277
 278         self._sort_formats(formats)
 279
 280         video_id = compat_str(talk_info['id'])
 281
 282         return {
 283             'id': video_id,
 284             'title': title,
 285             'uploader': player_talk.get('speaker') or talk_info.get('speaker'),
 286             'thumbnail': player_talk.get('thumb') or talk_info.get('thumb'),
 287             'description': self._og_search_description(webpage),
 288             'subtitles': self._get_subtitles(video_id, talk_info),
 289             'formats': formats,
 290             'duration': float_or_none(talk_info.get('duration')),
 291             'view_count': int_or_none(data.get('viewed_count')),
 292             'comment_count': int_or_none(
 293                 try_get(data, lambda x: x['comments']['count'])),
 294             'tags': try_get(talk_info, lambda x: x['tags'], list),
 295         }
 296
 297     def _get_subtitles(self, video_id, talk_info):
 298         sub_lang_list = {}
 299         for language in try_get(
 300                 talk_info,
 301                 (lambda x: x['downloads']['languages'],
 302                  lambda x: x['languages']), list):
 303             lang_code = language.get('languageCode') or language.get('ianaCode')
 304             if not lang_code:
 305                 continue
 306             sub_lang_list[lang_code] = [
 307                 {
 308                     'url': 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/%s' % (video_id, lang_code, ext),
 309                     'ext': ext,
 310                 }
 311                 for ext in ['ted', 'srt']
 312             ]
 313         return sub_lang_list
 314
 315     def _watch_info(self, url, name):
 316         webpage = self._download_webpage(url, name)
 317
 318         config_json = self._html_search_regex(
 319             r'"pages\.jwplayer"\s*,\s*({.+?})\s*\)\s*</script>',
 320             webpage, 'config', default=None)
 321         if not config_json:
 322             embed_url = self._search_regex(
 323                 r"<iframe[^>]+class='pages-video-embed__video__object'[^>]+src='([^']+)'", webpage, 'embed url')
 324             return self.url_result(self._proto_relative_url(embed_url))
 325         config = json.loads(config_json)['config']
 326         video_url = config['video']['url']
 327         thumbnail = config.get('image', {}).get('url')
 328
 329         title = self._html_search_regex(
 330             r"(?s)<h1(?:\s+class='[^']+')?>(.+?)</h1>", webpage, 'title')
 331         description = self._html_search_regex(
 332             [
 333                 r'(?s)<h4 class="[^"]+" id="h3--about-this-talk">.*?</h4>(.*?)</div>',
 334                 r'(?s)<p><strong>About this talk:</strong>\s+(.*?)</p>',
 335             ],
 336             webpage, 'description', fatal=False)
 337
 338         return {
 339             'id': name,
 340             'url': video_url,
 341             'title': title,
 342             'thumbnail': thumbnail,
 343             'description': description,
 344         }