]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/ted.py
debian/control: Annotate with bug numbers closed by this release.
[youtubedl] / youtube_dl / extractor / ted.py
index dfa1176a3e4e4eef333dcb829773c189bf9916ba..d260c91c2172deabb697c3ad242cfda24f395d41 100644 (file)
+from __future__ import unicode_literals
+
 import json
 import re
 
 import json
 import re
 
-from .common import InfoExtractor
-
-
-class TEDIE(InfoExtractor):
-    _VALID_URL=r'''http://www\.ted\.com/
-                   (
-                        ((?P<type_playlist>playlists)/(?P<playlist_id>\d+)) # We have a playlist
-                        |
-                        ((?P<type_talk>talks)) # We have a simple talk
-                   )
-                   (/lang/(.*?))? # The url may contain the language
-                   /(?P<name>\w+) # Here goes the name and then ".html"
-                   '''
-    _TEST = {
-        u'url': u'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html',
-        u'file': u'102.mp4',
-        u'md5': u'2d76ee1576672e0bd8f187513267adf6',
-        u'info_dict': {
-            u"description": u"md5:c6fa72e6eedbd938c9caf6b2702f5922", 
-            u"title": u"Dan Dennett: The illusion of consciousness"
+from .subtitles import SubtitlesInfoExtractor
+
+from ..utils import (
+    compat_str,
+)
+
+
+class TEDIE(SubtitlesInfoExtractor):
+    _VALID_URL = r'''(?x)
+        (?P<proto>https?://)
+        (?P<type>www|embed)(?P<urlmain>\.ted\.com/
+        (
+            (?P<type_playlist>playlists(?:/\d+)?) # We have a playlist
+            |
+            ((?P<type_talk>talks)) # We have a simple talk
+            |
+            (?P<type_watch>watch)/[^/]+/[^/]+
+        )
+        (/lang/(.*?))? # The url may contain the language
+        /(?P<name>[\w-]+) # Here goes the name and then ".html"
+        .*)$
+        '''
+    _TESTS = [{
+        'url': 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html',
+        'md5': '4ea1dada91e4174b53dac2bb8ace429d',
+        'info_dict': {
+            'id': '102',
+            'ext': 'mp4',
+            'title': 'The illusion of consciousness',
+            'description': ('Philosopher Dan Dennett makes a compelling '
+                'argument that not only don\'t we understand our own '
+                'consciousness, but that half the time our brains are '
+                'actively fooling us.'),
+            'uploader': 'Dan Dennett',
+            'width': 854,
+        }
+    }, {
+        'url': 'http://www.ted.com/watch/ted-institute/ted-bcg/vishal-sikka-the-beauty-and-power-of-algorithms',
+        'md5': '226f4fb9c62380d11b7995efa4c87994',
+        'info_dict': {
+            'id': 'vishal-sikka-the-beauty-and-power-of-algorithms',
+            'ext': 'mp4',
+            'title': 'Vishal Sikka: The beauty and power of algorithms',
+            'thumbnail': 're:^https?://.+\.jpg',
+            '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.',
         }
         }
+    }, {
+        'url': 'http://www.ted.com/talks/gabby_giffords_and_mark_kelly_be_passionate_be_courageous_be_your_best',
+        'md5': '49144e345a899b8cb34d315f3b9cfeeb',
+        'info_dict': {
+            'id': '1972',
+            'ext': 'mp4',
+            'title': 'Be passionate. Be courageous. Be your best.',
+            'uploader': 'Gabby Giffords and Mark Kelly',
+            'description': 'md5:5174aed4d0f16021b704120360f72b92',
+        },
+    }]
+
+    _NATIVE_FORMATS = {
+        'low': {'preference': 1, 'width': 320, 'height': 180},
+        'medium': {'preference': 2, 'width': 512, 'height': 288},
+        'high': {'preference': 3, 'width': 854, 'height': 480},
     }
 
     }
 
-    @classmethod
-    def suitable(cls, url):
-        """Receives a URL and returns True if suitable for this IE."""
-        return re.match(cls._VALID_URL, url, re.VERBOSE) is not None
+    def _extract_info(self, webpage):
+        info_json = self._search_regex(r'q\("\w+.init",({.+})\)</script>',
+            webpage, 'info json')
+        return json.loads(info_json)
 
     def _real_extract(self, url):
 
     def _real_extract(self, url):
-        m=re.match(self._VALID_URL, url, re.VERBOSE)
+        m = re.match(self._VALID_URL, url, re.VERBOSE)
+        if m.group('type') == 'embed':
+            desktop_url = m.group('proto') + 'www' + m.group('urlmain')
+            return self.url_result(desktop_url, 'TED')
+        name = m.group('name')
         if m.group('type_talk'):
         if m.group('type_talk'):
-            return [self._talk_info(url)]
-        else :
-            playlist_id=m.group('playlist_id')
-            name=m.group('name')
-            self.to_screen(u'Getting info of playlist %s: "%s"' % (playlist_id,name))
-            return [self._playlist_videos_info(url,name,playlist_id)]
-
-    def _playlist_videos_info(self,url,name,playlist_id=0):
+            return self._talk_info(url, name)
+        elif m.group('type_watch'):
+            return self._watch_info(url, name)
+        else:
+            return self._playlist_videos_info(url, name)
+
+    def _playlist_videos_info(self, url, name):
         '''Returns the videos of the playlist'''
         '''Returns the videos of the playlist'''
-        video_RE=r'''
-                     <li\ id="talk_(\d+)"([.\s]*?)data-id="(?P<video_id>\d+)"
-                     ([.\s]*?)data-playlist_item_id="(\d+)"
-                     ([.\s]*?)data-mediaslug="(?P<mediaSlug>.+?)"
-                     '''
-        video_name_RE=r'<p\ class="talk-title"><a href="(?P<talk_url>/talks/(.+).html)">(?P<fullname>.+?)</a></p>'
-        webpage=self._download_webpage(url, playlist_id, 'Downloading playlist webpage')
-        m_videos=re.finditer(video_RE,webpage,re.VERBOSE)
-        m_names=re.finditer(video_name_RE,webpage)
-
-        playlist_title = self._html_search_regex(r'div class="headline">\s*?<h1>\s*?<span>(.*?)</span>',
-                                                 webpage, 'playlist title')
-
-        playlist_entries = []
-        for m_video, m_name in zip(m_videos,m_names):
-            talk_url='http://www.ted.com%s' % m_name.group('talk_url')
-            playlist_entries.append(self.url_result(talk_url, 'TED'))
-        return self.playlist_result(playlist_entries, playlist_id = playlist_id, playlist_title = playlist_title)
-
-    def _talk_info(self, url, video_id=0):
-        """Return the video for the talk in the url"""
-        m = re.match(self._VALID_URL, url,re.VERBOSE)
-        video_name = m.group('name')
-        webpage = self._download_webpage(url, video_id, 'Downloading \"%s\" page' % video_name)
+
+        webpage = self._download_webpage(url, name,
+            'Downloading playlist webpage')
+        info = self._extract_info(webpage)
+        playlist_info = info['playlist']
+
+        playlist_entries = [
+            self.url_result('http://www.ted.com/talks/' + talk['slug'], self.ie_key())
+            for talk in info['talks']
+        ]
+        return self.playlist_result(
+            playlist_entries,
+            playlist_id=compat_str(playlist_info['id']),
+            playlist_title=playlist_info['title'])
+
+    def _talk_info(self, url, video_name):
+        webpage = self._download_webpage(url, video_name)
         self.report_extraction(video_name)
         self.report_extraction(video_name)
-        # If the url includes the language we get the title translated
-        title = self._html_search_regex(r'<span .*?id="altHeadline".+?>(?P<title>.*)</span>',
-                                        webpage, 'title')
-        json_data = self._search_regex(r'<script.*?>var talkDetails = ({.*?})</script>',
-                                    webpage, 'json data')
-        info = json.loads(json_data)
-        desc = self._html_search_regex(r'<div class="talk-intro">.*?<p.*?>(.*?)</p>',
-                                       webpage, 'description', flags = re.DOTALL)
-        
-        thumbnail = self._search_regex(r'</span>[\s.]*</div>[\s.]*<img src="(.*?)"',
-                                       webpage, 'thumbnail')
+
+        talk_info = self._extract_info(webpage)['talks'][0]
+
         formats = [{
         formats = [{
-            'ext': 'mp4',
-            'url': stream['file'],
-            'format': stream['id']
-            } for stream in info['htmlStreams']]
-        info = {
-            'id': info['id'],
-            'title': title,
+            'url': format_url,
+            'format_id': format_id,
+            'format': format_id,
+        } for (format_id, format_url) in talk_info['nativeDownloads'].items() if format_url is not None]
+        if formats:
+            for f in formats:
+                finfo = self._NATIVE_FORMATS.get(f['format_id'])
+                if finfo:
+                    f.update(finfo)
+        else:
+            # Use rtmp downloads
+            formats = [{
+                'format_id': f['name'],
+                'url': talk_info['streamer'],
+                'play_path': f['file'],
+                'ext': 'flv',
+                'width': f['width'],
+                'height': f['height'],
+                'tbr': f['bitrate'],
+            } for f in talk_info['resources']['rtmp']]
+        self._sort_formats(formats)
+
+        video_id = compat_str(talk_info['id'])
+        # subtitles
+        video_subtitles = self.extract_subtitles(video_id, talk_info)
+        if self._downloader.params.get('listsubtitles', False):
+            self._list_available_subtitles(video_id, talk_info)
+            return
+
+        thumbnail = talk_info['thumb']
+        if not thumbnail.startswith('http'):
+            thumbnail = 'http://' + thumbnail
+        return {
+            'id': video_id,
+            'title': talk_info['title'],
+            'uploader': talk_info['speaker'],
             'thumbnail': thumbnail,
             'thumbnail': thumbnail,
-            'description': desc,
+            'description': self._og_search_description(webpage),
+            'subtitles': video_subtitles,
             'formats': formats,
         }
 
             'formats': formats,
         }
 
-        # TODO: Remove when #980 has been merged
-        info.update(info['formats'][-1])
+    def _get_available_subtitles(self, video_id, talk_info):
+        languages = [lang['languageCode'] for lang in talk_info.get('languages', [])]
+        if languages:
+            sub_lang_list = {}
+            for l in languages:
+                url = 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/srt' % (video_id, l)
+                sub_lang_list[l] = url
+            return sub_lang_list
+        else:
+            self._downloader.report_warning('video doesn\'t have subtitles')
+            return {}
 
 
-        return info
+    def _watch_info(self, url, name):
+        webpage = self._download_webpage(url, name)
+
+        config_json = self._html_search_regex(
+            r"data-config='([^']+)", webpage, 'config')
+        config = json.loads(config_json)
+        video_url = config['video']['url']
+        thumbnail = config.get('image', {}).get('url')
+
+        title = self._html_search_regex(
+            r"(?s)<h1(?:\s+class='[^']+')?>(.+?)</h1>", webpage, 'title')
+        description = self._html_search_regex(
+            [
+                r'(?s)<h4 class="[^"]+" id="h3--about-this-talk">.*?</h4>(.*?)</div>',
+                r'(?s)<p><strong>About this talk:</strong>\s+(.*?)</p>',
+            ],
+            webpage, 'description', fatal=False)
+
+        return {
+            'id': name,
+            'url': video_url,
+            'title': title,
+            'thumbnail': thumbnail,
+            'description': description,
+        }