youtube_dl/extractor/ted.py

   1 from __future__ import unicode_literals
   2
   3 import json
   4 import re
   5
   6 from .subtitles import SubtitlesInfoExtractor
   7
   8 from ..utils import (
   9     compat_str,
  10 )
  11
  12
  13 class TEDIE(SubtitlesInfoExtractor):
  14     _VALID_URL = r'''(?x)
  15         (?P<proto>https?://)
  16         (?P<type>www|embed)(?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': '4ea1dada91e4174b53dac2bb8ace429d',
  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         }
  42     }, {
  43         'url': 'http://www.ted.com/watch/ted-institute/ted-bcg/vishal-sikka-the-beauty-and-power-of-algorithms',
  44         'md5': '226f4fb9c62380d11b7995efa4c87994',
  45         'info_dict': {
  46             'id': 'vishal-sikka-the-beauty-and-power-of-algorithms',
  47             'ext': 'mp4',
  48             'title': 'Vishal Sikka: The beauty and power of algorithms',
  49             'thumbnail': 're:^https?://.+\.jpg',
  50             '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.',
  51         }
  52     }]
  53
  54     _NATIVE_FORMATS = {
  55         'low': {'preference': 1, 'width': 320, 'height': 180},
  56         'medium': {'preference': 2, 'width': 512, 'height': 288},
  57         'high': {'preference': 3, 'width': 854, 'height': 480},
  58     }
  59
  60     def _extract_info(self, webpage):
  61         info_json = self._search_regex(r'q\("\w+.init",({.+})\)</script>',
  62             webpage, 'info json')
  63         return json.loads(info_json)
  64
  65     def _real_extract(self, url):
  66         m = re.match(self._VALID_URL, url, re.VERBOSE)
  67         if m.group('type') == 'embed':
  68             desktop_url = m.group('proto') + 'www' + m.group('urlmain')
  69             return self.url_result(desktop_url, 'TED')
  70         name = m.group('name')
  71         if m.group('type_talk'):
  72             return self._talk_info(url, name)
  73         elif m.group('type_watch'):
  74             return self._watch_info(url, name)
  75         else:
  76             return self._playlist_videos_info(url, name)
  77
  78     def _playlist_videos_info(self, url, name):
  79         '''Returns the videos of the playlist'''
  80
  81         webpage = self._download_webpage(url, name,
  82             'Downloading playlist webpage')
  83         info = self._extract_info(webpage)
  84         playlist_info = info['playlist']
  85
  86         playlist_entries = [
  87             self.url_result(u'http://www.ted.com/talks/' + talk['slug'], self.ie_key())
  88             for talk in info['talks']
  89         ]
  90         return self.playlist_result(
  91             playlist_entries,
  92             playlist_id=compat_str(playlist_info['id']),
  93             playlist_title=playlist_info['title'])
  94
  95     def _talk_info(self, url, video_name):
  96         webpage = self._download_webpage(url, video_name)
  97         self.report_extraction(video_name)
  98
  99         talk_info = self._extract_info(webpage)['talks'][0]
 100
 101         formats = [{
 102             'url': format_url,
 103             'format_id': format_id,
 104             'format': format_id,
 105         } for (format_id, format_url) in talk_info['nativeDownloads'].items()]
 106         for f in formats:
 107             finfo = self._NATIVE_FORMATS.get(f['format_id'])
 108             if finfo:
 109                 f.update(finfo)
 110         self._sort_formats(formats)
 111
 112         video_id = compat_str(talk_info['id'])
 113         # subtitles
 114         video_subtitles = self.extract_subtitles(video_id, talk_info)
 115         if self._downloader.params.get('listsubtitles', False):
 116             self._list_available_subtitles(video_id, talk_info)
 117             return
 118
 119         thumbnail = talk_info['thumb']
 120         if not thumbnail.startswith('http'):
 121             thumbnail = 'http://' + thumbnail
 122         return {
 123             'id': video_id,
 124             'title': talk_info['title'],
 125             'uploader': talk_info['speaker'],
 126             'thumbnail': thumbnail,
 127             'description': self._og_search_description(webpage),
 128             'subtitles': video_subtitles,
 129             'formats': formats,
 130         }
 131
 132     def _get_available_subtitles(self, video_id, talk_info):
 133         languages = [lang['languageCode'] for lang in talk_info.get('languages', [])]
 134         if languages:
 135             sub_lang_list = {}
 136             for l in languages:
 137                 url = 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/srt' % (video_id, l)
 138                 sub_lang_list[l] = url
 139             return sub_lang_list
 140         else:
 141             self._downloader.report_warning(u'video doesn\'t have subtitles')
 142             return {}
 143
 144     def _watch_info(self, url, name):
 145         webpage = self._download_webpage(url, name)
 146
 147         config_json = self._html_search_regex(
 148             r"data-config='([^']+)", webpage, 'config')
 149         config = json.loads(config_json)
 150         video_url = config['video']['url']
 151         thumbnail = config.get('image', {}).get('url')
 152
 153         title = self._html_search_regex(
 154             r"(?s)<h1(?:\s+class='[^']+')?>(.+?)</h1>", webpage, 'title')
 155         description = self._html_search_regex(
 156             r'(?s)<h4 class="[^"]+" id="h3--about-this-talk">.*?</h4>(.*?)</div>',
 157             webpage, 'description', fatal=False)
 158
 159         return {
 160             'id': name,
 161             'url': video_url,
 162             'title': title,
 163             'thumbnail': thumbnail,
 164             'description': description,
 165         }