Merge branch 'atomicparsley' (closes #2436)

[youtube-dl] / youtube_dl / extractor / ted.py

diff --git a/youtube_dl/extractor/ted.py b/youtube_dl/extractor/ted.py
index ad1a46c3385713056b94d2f00e38558e8ea69b1c..be7a0f866cb0ce37d977e6ce504ff3d612af5bbb 100644 (file)
--- a/youtube_dl/extractor/ted.py
+++ b/youtube_dl/extractor/ted.py
@@ -18,12 +18,14 @@ class TEDIE(SubtitlesInfoExtractor):
             (?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"
+        /(?P<name>[\w-]+) # Here goes the name and then ".html"
         .*)$
         '''
-    _TEST = {
+    _TESTS = [{
         'url': 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html',
         'md5': '4ea1dada91e4174b53dac2bb8ace429d',
         'info_dict': {
@@ -35,13 +37,24 @@ class TEDIE(SubtitlesInfoExtractor):
                 '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.',
+        }
+    }]
 
-    _FORMATS_PREFERENCE = {
-        'low': 1,
-        'medium': 2,
-        'high': 3,
+    _NATIVE_FORMATS = {
+        'low': {'preference': 1, 'width': 320, 'height': 180},
+        'medium': {'preference': 2, 'width': 512, 'height': 288},
+        'high': {'preference': 3, 'width': 854, 'height': 480},
     }
 
     def _extract_info(self, webpage):
@@ -57,6 +70,8 @@ class TEDIE(SubtitlesInfoExtractor):
         name = m.group('name')
         if m.group('type_talk'):
             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)
 
@@ -84,12 +99,14 @@ class TEDIE(SubtitlesInfoExtractor):
         talk_info = self._extract_info(webpage)['talks'][0]
 
         formats = [{
-            'ext': 'mp4',
             'url': format_url,
             'format_id': format_id,
             'format': format_id,
-            'preference': self._FORMATS_PREFERENCE.get(format_id, -1),
         } for (format_id, format_url) in talk_info['nativeDownloads'].items()]
+        for f in formats:
+            finfo = self._NATIVE_FORMATS.get(f['format_id'])
+            if finfo:
+                f.update(finfo)
         self._sort_formats(formats)
 
         video_id = compat_str(talk_info['id'])
@@ -123,3 +140,26 @@ class TEDIE(SubtitlesInfoExtractor):
         else:
             self._downloader.report_warning(u'video doesn\'t have subtitles')
             return {}
+
+    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>',
+            webpage, 'description', fatal=False)
+
+        return {
+            'id': name,
+            'url': video_url,
+            'title': title,
+            'thumbnail': thumbnail,
+            'description': description,
+        }