Raphaël G. Git Repositories - youtubedl/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 (
   9     compat_str,
  10     compat_urlparse
  11 )
  12 from ..utils import (
  13     extract_attributes,
  14     float_or_none,
  15     int_or_none,
  16     try_get,
  17     url_or_none,
  18 )
  19
  20
  21 class TEDIE(InfoExtractor):
  22     IE_NAME = 'ted'
  23     _VALID_URL = r'''(?x)
  24         (?P<proto>https?://)
  25         (?P<type>www|embed(?:-ssl)?)(?P<urlmain>\.ted\.com/
  26         (
  27             (?P<type_playlist>playlists(?:/(?P<playlist_id>\d+))?) # We have a playlist
  28             |
  29             ((?P<type_talk>talks)) # We have a simple talk
  30             |
  31             (?P<type_watch>watch)/[^/]+/[^/]+
  32         )
  33         (/lang/(.*?))? # The url may contain the language
  34         /(?P<name>[\w-]+) # Here goes the name and then ".html"
  35         .*)$
  36         '''
  37     _TESTS = [{
  38         'url': 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html',
  39         'md5': 'b0ce2b05ca215042124fbc9e3886493a',
  40         'info_dict': {
  41             'id': '102',
  42             'ext': 'mp4',
  43             'title': 'The illusion of consciousness',
  44             'description': ('Philosopher Dan Dennett makes a compelling '
  45                             'argument that not only don\'t we understand our own '
  46                             'consciousness, but that half the time our brains are '
  47                             'actively fooling us.'),
  48             'uploader': 'Dan Dennett',
  49             'width': 853,
  50             'duration': 1308,
  51             'view_count': int,
  52             'comment_count': int,
  53             'tags': list,
  54         },
  55         'params': {
  56             'skip_download': True,
  57         },
  58     }, {
  59         # missing HTTP bitrates
  60         'url': 'https://www.ted.com/talks/vishal_sikka_the_beauty_and_power_of_algorithms',
  61         'info_dict': {
  62             'id': '6069',
  63             'ext': 'mp4',
  64             'title': 'The beauty and power of algorithms',
  65             'thumbnail': r're:^https?://.+\.jpg',
  66             'description': 'md5:734e352710fb00d840ab87ae31aaf688',
  67             'uploader': 'Vishal Sikka',
  68         },
  69         'params': {
  70             'skip_download': True,
  71         },
  72     }, {
  73         'url': 'http://www.ted.com/talks/gabby_giffords_and_mark_kelly_be_passionate_be_courageous_be_your_best',
  74         'md5': 'e6b9617c01a7970ceac8bb2c92c346c0',
  75         'info_dict': {
  76             'id': '1972',
  77             'ext': 'mp4',
  78             'title': 'Be passionate. Be courageous. Be your best.',
  79             'uploader': 'Gabby Giffords and Mark Kelly',
  80             'description': 'md5:5174aed4d0f16021b704120360f72b92',
  81             'duration': 1128,
  82         },
  83         'params': {
  84             'skip_download': True,
  85         },
  86     }, {
  87         'url': 'http://www.ted.com/playlists/who_are_the_hackers',
  88         'info_dict': {
  89             'id': '10',
  90             'title': 'Who are the hackers?',
  91             'description': 'md5:49a0dbe8fb76d81a0e64b4a80af7f15a'
  92         },
  93         'playlist_mincount': 6,
  94     }, {
  95         # contains a youtube video
  96         'url': 'https://www.ted.com/talks/douglas_adams_parrots_the_universe_and_everything',
  97         'add_ie': ['Youtube'],
  98         'info_dict': {
  99             'id': '_ZG8HBuDjgc',
 100             'ext': 'webm',
 101             'title': 'Douglas Adams: Parrots the Universe and Everything',
 102             'description': 'md5:01ad1e199c49ac640cb1196c0e9016af',
 103             'uploader': 'University of California Television (UCTV)',
 104             'uploader_id': 'UCtelevision',
 105             'upload_date': '20080522',
 106         },
 107         'params': {
 108             'skip_download': True,
 109         },
 110     }, {
 111         # no nativeDownloads
 112         'url': 'https://www.ted.com/talks/tom_thum_the_orchestra_in_my_mouth',
 113         'info_dict': {
 114             'id': '1792',
 115             'ext': 'mp4',
 116             'title': 'The orchestra in my mouth',
 117             'description': 'md5:5d1d78650e2f8dfcbb8ebee2951ac29a',
 118             'uploader': 'Tom Thum',
 119             'view_count': int,
 120             'comment_count': int,
 121             'tags': list,
 122         },
 123         'params': {
 124             'skip_download': True,
 125         },
 126     }]
 127
 128     _NATIVE_FORMATS = {
 129         'low': {'width': 320, 'height': 180},
 130         'medium': {'width': 512, 'height': 288},
 131         'high': {'width': 854, 'height': 480},
 132     }
 133
 134     def _extract_info(self, webpage):
 135         info_json = self._search_regex(
 136             r'(?s)q\(\s*"\w+.init"\s*,\s*({.+?})\)\s*</script>',
 137             webpage, 'info json')
 138         return json.loads(info_json)
 139
 140     def _real_extract(self, url):
 141         m = re.match(self._VALID_URL, url, re.VERBOSE)
 142         if m.group('type').startswith('embed'):
 143             desktop_url = m.group('proto') + 'www' + m.group('urlmain')
 144             return self.url_result(desktop_url, 'TED')
 145         name = m.group('name')
 146         if m.group('type_talk'):
 147             return self._talk_info(url, name)
 148         elif m.group('type_watch'):
 149             return self._watch_info(url, name)
 150         else:
 151             return self._playlist_videos_info(url, name)
 152
 153     def _playlist_videos_info(self, url, name):
 154         '''Returns the videos of the playlist'''
 155
 156         webpage = self._download_webpage(url, name,
 157                                          'Downloading playlist webpage')
 158
 159         playlist_entries = []
 160         for entry in re.findall(r'(?s)<[^>]+data-ga-context=["\']playlist["\'][^>]*>', webpage):
 161             attrs = extract_attributes(entry)
 162             entry_url = compat_urlparse.urljoin(url, attrs['href'])
 163             playlist_entries.append(self.url_result(entry_url, self.ie_key()))
 164
 165         final_url = self._og_search_url(webpage, fatal=False)
 166         playlist_id = (
 167             re.match(self._VALID_URL, final_url).group('playlist_id')
 168             if final_url else None)
 169
 170         return self.playlist_result(
 171             playlist_entries, playlist_id=playlist_id,
 172             playlist_title=self._og_search_title(webpage, fatal=False),
 173             playlist_description=self._og_search_description(webpage))
 174
 175     def _talk_info(self, url, video_name):
 176         webpage = self._download_webpage(url, video_name)
 177
 178         info = self._extract_info(webpage)
 179
 180         data = try_get(info, lambda x: x['__INITIAL_DATA__'], dict) or info
 181         talk_info = data['talks'][0]
 182
 183         title = talk_info['title'].strip()
 184
 185         native_downloads = try_get(
 186             talk_info,
 187             (lambda x: x['downloads']['nativeDownloads'],
 188              lambda x: x['nativeDownloads']),
 189             dict) or {}
 190
 191         formats = [{
 192             'url': format_url,
 193             'format_id': format_id,
 194             'format': format_id,
 195         } for (format_id, format_url) in native_downloads.items() if format_url is not None]
 196         if formats:
 197             for f in formats:
 198                 finfo = self._NATIVE_FORMATS.get(f['format_id'])
 199                 if finfo:
 200                     f.update(finfo)
 201
 202         player_talk = talk_info['player_talks'][0]
 203
 204         external = player_talk.get('external')
 205         if isinstance(external, dict):
 206             service = external.get('service')
 207             if isinstance(service, compat_str):
 208                 ext_url = None
 209                 if service.lower() == 'youtube':
 210                     ext_url = external.get('code')
 211
 212                 return self.url_result(ext_url or external['uri'])
 213
 214         resources_ = player_talk.get('resources') or talk_info.get('resources')
 215
 216         http_url = None
 217         for format_id, resources in resources_.items():
 218             if format_id == 'h264':
 219                 for resource in resources:
 220                     h264_url = resource.get('file')
 221                     if not h264_url:
 222                         continue
 223                     bitrate = int_or_none(resource.get('bitrate'))
 224                     formats.append({
 225                         'url': h264_url,
 226                         'format_id': '%s-%sk' % (format_id, bitrate),
 227                         'tbr': bitrate,
 228                     })
 229                     if re.search(r'\d+k', h264_url):
 230                         http_url = h264_url
 231             elif format_id == 'rtmp':
 232                 streamer = talk_info.get('streamer')
 233                 if not streamer:
 234                     continue
 235                 for resource in resources:
 236                     formats.append({
 237                         'format_id': '%s-%s' % (format_id, resource.get('name')),
 238                         'url': streamer,
 239                         'play_path': resource['file'],
 240                         'ext': 'flv',
 241                         'width': int_or_none(resource.get('width')),
 242                         'height': int_or_none(resource.get('height')),
 243                         'tbr': int_or_none(resource.get('bitrate')),
 244                     })
 245             elif format_id == 'hls':
 246                 if not isinstance(resources, dict):
 247                     continue
 248                 stream_url = url_or_none(resources.get('stream'))
 249                 if not stream_url:
 250                     continue
 251                 formats.extend(self._extract_m3u8_formats(
 252                     stream_url, video_name, 'mp4', m3u8_id=format_id,
 253                     fatal=False))
 254
 255         m3u8_formats = list(filter(
 256             lambda f: f.get('protocol') == 'm3u8' and f.get('vcodec') != 'none',
 257             formats))
 258         if http_url:
 259             for m3u8_format in m3u8_formats:
 260                 bitrate = self._search_regex(r'(\d+k)', m3u8_format['url'], 'bitrate', default=None)
 261                 if not bitrate:
 262                     continue
 263                 bitrate_url = re.sub(r'\d+k', bitrate, http_url)
 264                 if not self._is_valid_url(
 265                         bitrate_url, video_name, '%s bitrate' % bitrate):
 266                     continue
 267                 f = m3u8_format.copy()
 268                 f.update({
 269                     'url': bitrate_url,
 270                     'format_id': m3u8_format['format_id'].replace('hls', 'http'),
 271                     'protocol': 'http',
 272                 })
 273                 if f.get('acodec') == 'none':
 274                     del f['acodec']
 275                 formats.append(f)
 276
 277         audio_download = talk_info.get('audioDownload')
 278         if audio_download:
 279             formats.append({
 280                 'url': audio_download,
 281                 'format_id': 'audio',
 282                 'vcodec': 'none',
 283             })
 284
 285         self._sort_formats(formats)
 286
 287         video_id = compat_str(talk_info['id'])
 288
 289         return {
 290             'id': video_id,
 291             'title': title,
 292             'uploader': player_talk.get('speaker') or talk_info.get('speaker'),
 293             'thumbnail': player_talk.get('thumb') or talk_info.get('thumb'),
 294             'description': self._og_search_description(webpage),
 295             'subtitles': self._get_subtitles(video_id, talk_info),
 296             'formats': formats,
 297             'duration': float_or_none(talk_info.get('duration')),
 298             'view_count': int_or_none(data.get('viewed_count')),
 299             'comment_count': int_or_none(
 300                 try_get(data, lambda x: x['comments']['count'])),
 301             'tags': try_get(talk_info, lambda x: x['tags'], list),
 302         }
 303
 304     def _get_subtitles(self, video_id, talk_info):
 305         sub_lang_list = {}
 306         for language in try_get(
 307                 talk_info,
 308                 (lambda x: x['downloads']['languages'],
 309                  lambda x: x['languages']), list):
 310             lang_code = language.get('languageCode') or language.get('ianaCode')
 311             if not lang_code:
 312                 continue
 313             sub_lang_list[lang_code] = [
 314                 {
 315                     'url': 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/%s' % (video_id, lang_code, ext),
 316                     'ext': ext,
 317                 }
 318                 for ext in ['ted', 'srt']
 319             ]
 320         return sub_lang_list
 321
 322     def _watch_info(self, url, name):
 323         webpage = self._download_webpage(url, name)
 324
 325         config_json = self._html_search_regex(
 326             r'"pages\.jwplayer"\s*,\s*({.+?})\s*\)\s*</script>',
 327             webpage, 'config', default=None)
 328         if not config_json:
 329             embed_url = self._search_regex(
 330                 r"<iframe[^>]+class='pages-video-embed__video__object'[^>]+src='([^']+)'", webpage, 'embed url')
 331             return self.url_result(self._proto_relative_url(embed_url))
 332         config = json.loads(config_json)['config']
 333         video_url = config['video']['url']
 334         thumbnail = config.get('image', {}).get('url')
 335
 336         title = self._html_search_regex(
 337             r"(?s)<h1(?:\s+class='[^']+')?>(.+?)</h1>", webpage, 'title')
 338         description = self._html_search_regex(
 339             [
 340                 r'(?s)<h4 class="[^"]+" id="h3--about-this-talk">.*?</h4>(.*?)</div>',
 341                 r'(?s)<p><strong>About this talk:</strong>\s+(.*?)</p>',
 342             ],
 343             webpage, 'description', fatal=False)
 344
 345         return {
 346             'id': name,
 347             'url': video_url,
 348             'title': title,
 349             'thumbnail': thumbnail,
 350             'description': description,
 351         }