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 int_or_none
  10
  11
  12 class TEDIE(InfoExtractor):
  13     IE_NAME = 'ted'
  14     _VALID_URL = r'''(?x)
  15         (?P<proto>https?://)
  16         (?P<type>www|embed(?:-ssl)?)(?P<urlmain>\.ted\.com/
  17         (
  18             (?P<type_playlist>playlists(?:/\d+)?) # We have a playlist
  19             |
  20             ((?P<type_talk>talks)) # We have a simple talk
  21             |
  22             (?P<type_watch>watch)/[^/]+/[^/]+
  23         )
  24         (/lang/(.*?))? # The url may contain the language
  25         /(?P<name>[\w-]+) # Here goes the name and then ".html"
  26         .*)$
  27         '''
  28     _TESTS = [{
  29         'url': 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html',
  30         'md5': 'fc94ac279feebbce69f21c0c6ee82810',
  31         'info_dict': {
  32             'id': '102',
  33             'ext': 'mp4',
  34             'title': 'The illusion of consciousness',
  35             'description': ('Philosopher Dan Dennett makes a compelling '
  36                             'argument that not only don\'t we understand our own '
  37                             'consciousness, but that half the time our brains are '
  38                             'actively fooling us.'),
  39             'uploader': 'Dan Dennett',
  40             'width': 854,
  41             'duration': 1308,
  42         }
  43     }, {
  44         'url': 'http://www.ted.com/watch/ted-institute/ted-bcg/vishal-sikka-the-beauty-and-power-of-algorithms',
  45         'md5': '226f4fb9c62380d11b7995efa4c87994',
  46         'info_dict': {
  47             'id': 'vishal-sikka-the-beauty-and-power-of-algorithms',
  48             'ext': 'mp4',
  49             'title': 'Vishal Sikka: The beauty and power of algorithms',
  50             'thumbnail': 're:^https?://.+\.jpg',
  51             'description': 'Adaptive, intelligent, and consistent, algorithms are emerging as the ultimate app for everything from matching consumers to products to assessing medical diagnoses. Vishal Sikka shares his appreciation for the algorithm, charting both its inherent beauty and its growing power.',
  52         }
  53     }, {
  54         'url': 'http://www.ted.com/talks/gabby_giffords_and_mark_kelly_be_passionate_be_courageous_be_your_best',
  55         'info_dict': {
  56             'id': '1972',
  57             'ext': 'mp4',
  58             'title': 'Be passionate. Be courageous. Be your best.',
  59             'uploader': 'Gabby Giffords and Mark Kelly',
  60             'description': 'md5:5174aed4d0f16021b704120360f72b92',
  61             'duration': 1128,
  62         },
  63     }, {
  64         'url': 'http://www.ted.com/playlists/who_are_the_hackers',
  65         'info_dict': {
  66             'id': '10',
  67             'title': 'Who are the hackers?',
  68         },
  69         'playlist_mincount': 6,
  70     }, {
  71         # contains a youtube video
  72         'url': 'https://www.ted.com/talks/douglas_adams_parrots_the_universe_and_everything',
  73         'add_ie': ['Youtube'],
  74         'info_dict': {
  75             'id': '_ZG8HBuDjgc',
  76             'ext': 'mp4',
  77             'title': 'Douglas Adams: Parrots the Universe and Everything',
  78             'description': 'md5:01ad1e199c49ac640cb1196c0e9016af',
  79             'uploader': 'University of California Television (UCTV)',
  80             'uploader_id': 'UCtelevision',
  81             'upload_date': '20080522',
  82         },
  83         'params': {
  84             'skip_download': True,
  85         },
  86     }, {
  87         # YouTube video
  88         'url': 'http://www.ted.com/talks/jeffrey_kluger_the_sibling_bond',
  89         'add_ie': ['Youtube'],
  90         'info_dict': {
  91             'id': 'aFBIPO-P7LM',
  92             'ext': 'mp4',
  93             'title': 'The hidden power of siblings: Jeff Kluger at TEDxAsheville',
  94             'description': 'md5:3d7a4f50d95ca5dd67104e2a20f43fe1',
  95             'uploader': 'TEDx Talks',
  96             'uploader_id': 'TEDxTalks',
  97             'upload_date': '20111216',
  98         },
  99         'params': {
 100             'skip_download': True,
 101         },
 102     }]
 103
 104     _NATIVE_FORMATS = {
 105         'low': {'preference': 1, 'width': 320, 'height': 180},
 106         'medium': {'preference': 2, 'width': 512, 'height': 288},
 107         'high': {'preference': 3, 'width': 854, 'height': 480},
 108     }
 109
 110     def _extract_info(self, webpage):
 111         info_json = self._search_regex(r'q\("\w+.init",({.+})\)</script>',
 112                                        webpage, 'info json')
 113         return json.loads(info_json)
 114
 115     def _real_extract(self, url):
 116         m = re.match(self._VALID_URL, url, re.VERBOSE)
 117         if m.group('type').startswith('embed'):
 118             desktop_url = m.group('proto') + 'www' + m.group('urlmain')
 119             return self.url_result(desktop_url, 'TED')
 120         name = m.group('name')
 121         if m.group('type_talk'):
 122             return self._talk_info(url, name)
 123         elif m.group('type_watch'):
 124             return self._watch_info(url, name)
 125         else:
 126             return self._playlist_videos_info(url, name)
 127
 128     def _playlist_videos_info(self, url, name):
 129         '''Returns the videos of the playlist'''
 130
 131         webpage = self._download_webpage(url, name,
 132                                          'Downloading playlist webpage')
 133         info = self._extract_info(webpage)
 134         playlist_info = info['playlist']
 135
 136         playlist_entries = [
 137             self.url_result('http://www.ted.com/talks/' + talk['slug'], self.ie_key())
 138             for talk in info['talks']
 139         ]
 140         return self.playlist_result(
 141             playlist_entries,
 142             playlist_id=compat_str(playlist_info['id']),
 143             playlist_title=playlist_info['title'])
 144
 145     def _talk_info(self, url, video_name):
 146         webpage = self._download_webpage(url, video_name)
 147         self.report_extraction(video_name)
 148
 149         talk_info = self._extract_info(webpage)['talks'][0]
 150
 151         external = talk_info.get('external')
 152         if external:
 153             service = external['service']
 154             self.to_screen('Found video from %s' % service)
 155             ext_url = None
 156             if service.lower() == 'youtube':
 157                 ext_url = external.get('code')
 158             return {
 159                 '_type': 'url',
 160                 'url': ext_url or external['uri'],
 161             }
 162
 163         formats = [{
 164             'url': format_url,
 165             'format_id': format_id,
 166             'format': format_id,
 167         } for (format_id, format_url) in talk_info['nativeDownloads'].items() if format_url is not None]
 168         if formats:
 169             for f in formats:
 170                 finfo = self._NATIVE_FORMATS.get(f['format_id'])
 171                 if finfo:
 172                     f.update(finfo)
 173
 174         for format_id, resources in talk_info['resources'].items():
 175             if format_id == 'h264':
 176                 for resource in resources:
 177                     bitrate = int_or_none(resource.get('bitrate'))
 178                     formats.append({
 179                         'url': resource['file'],
 180                         'format_id': '%s-%sk' % (format_id, bitrate),
 181                         'tbr': bitrate,
 182                     })
 183             elif format_id == 'rtmp':
 184                 streamer = talk_info.get('streamer')
 185                 if not streamer:
 186                     continue
 187                 for resource in resources:
 188                     formats.append({
 189                         'format_id': '%s-%s' % (format_id, resource.get('name')),
 190                         'url': streamer,
 191                         'play_path': resource['file'],
 192                         'ext': 'flv',
 193                         'width': int_or_none(resource.get('width')),
 194                         'height': int_or_none(resource.get('height')),
 195                         'tbr': int_or_none(resource.get('bitrate')),
 196                     })
 197             elif format_id == 'hls':
 198                 hls_formats = self._extract_m3u8_formats(
 199                     resources.get('stream'), video_name, 'mp4', m3u8_id=format_id)
 200                 for f in hls_formats:
 201                     if f.get('format_id') == 'hls-meta':
 202                         continue
 203                     if not f.get('height'):
 204                         f['vcodec'] = 'none'
 205                     else:
 206                         f['acodec'] = 'none'
 207                 formats.extend(hls_formats)
 208
 209         audio_download = talk_info.get('audioDownload')
 210         if audio_download:
 211             formats.append({
 212                 'url': audio_download,
 213                 'format_id': 'audio',
 214                 'vcodec': 'none',
 215                 'preference': -0.5,
 216             })
 217
 218         self._sort_formats(formats)
 219
 220         video_id = compat_str(talk_info['id'])
 221
 222         thumbnail = talk_info['thumb']
 223         if not thumbnail.startswith('http'):
 224             thumbnail = 'http://' + thumbnail
 225         return {
 226             'id': video_id,
 227             'title': talk_info['title'].strip(),
 228             'uploader': talk_info['speaker'],
 229             'thumbnail': thumbnail,
 230             'description': self._og_search_description(webpage),
 231             'subtitles': self._get_subtitles(video_id, talk_info),
 232             'formats': formats,
 233             'duration': talk_info.get('duration'),
 234         }
 235
 236     def _get_subtitles(self, video_id, talk_info):
 237         languages = [lang['languageCode'] for lang in talk_info.get('languages', [])]
 238         if languages:
 239             sub_lang_list = {}
 240             for l in languages:
 241                 sub_lang_list[l] = [
 242                     {
 243                         'url': 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/%s' % (video_id, l, ext),
 244                         'ext': ext,
 245                     }
 246                     for ext in ['ted', 'srt']
 247                 ]
 248             return sub_lang_list
 249         else:
 250             return {}
 251
 252     def _watch_info(self, url, name):
 253         webpage = self._download_webpage(url, name)
 254
 255         config_json = self._html_search_regex(
 256             r'"pages\.jwplayer"\s*,\s*({.+?})\s*\)\s*</script>',
 257             webpage, 'config')
 258         config = json.loads(config_json)['config']
 259         video_url = config['video']['url']
 260         thumbnail = config.get('image', {}).get('url')
 261
 262         title = self._html_search_regex(
 263             r"(?s)<h1(?:\s+class='[^']+')?>(.+?)</h1>", webpage, 'title')
 264         description = self._html_search_regex(
 265             [
 266                 r'(?s)<h4 class="[^"]+" id="h3--about-this-talk">.*?</h4>(.*?)</div>',
 267                 r'(?s)<p><strong>About this talk:</strong>\s+(.*?)</p>',
 268             ],
 269             webpage, 'description', fatal=False)
 270
 271         return {
 272             'id': name,
 273             'url': video_url,
 274             'title': title,
 275             'thumbnail': thumbnail,
 276             'description': description,
 277         }