]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/ted.py
Prepare to upload
[youtubedl] / youtube_dl / extractor / ted.py
index a48d77c309dcd1f9984cd0a6c71b7af574ca5498..06a27fd0428b469abb64e6428faf263fc080cc33 100644 (file)
@@ -6,7 +6,10 @@ import re
 from .common import InfoExtractor
 
 from ..compat import compat_str
-from ..utils import int_or_none
+from ..utils import (
+    int_or_none,
+    try_get,
+)
 
 
 class TEDIE(InfoExtractor):
@@ -27,7 +30,7 @@ class TEDIE(InfoExtractor):
         '''
     _TESTS = [{
         'url': 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html',
-        'md5': 'fc94ac279feebbce69f21c0c6ee82810',
+        'md5': '0de43ac406aa3e4ea74b66c9c7789b13',
         'info_dict': {
             'id': '102',
             'ext': 'mp4',
@@ -37,21 +40,26 @@ class TEDIE(InfoExtractor):
                             'consciousness, but that half the time our brains are '
                             'actively fooling us.'),
             'uploader': 'Dan Dennett',
-            'width': 854,
+            'width': 853,
             'duration': 1308,
         }
     }, {
         'url': 'http://www.ted.com/watch/ted-institute/ted-bcg/vishal-sikka-the-beauty-and-power-of-algorithms',
-        'md5': '226f4fb9c62380d11b7995efa4c87994',
+        'md5': 'b899ac15e345fb39534d913f7606082b',
         'info_dict': {
-            'id': 'vishal-sikka-the-beauty-and-power-of-algorithms',
+            'id': 'tSVI8ta_P4w',
             '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.',
-        }
+            'thumbnail': r're:^https?://.+\.jpg',
+            'description': 'md5:6261fdfe3e02f4f579cbbfc00aff73f4',
+            'upload_date': '20140122',
+            'uploader_id': 'TEDInstitute',
+            'uploader': 'TED Institute',
+        },
+        'add_ie': ['Youtube'],
     }, {
         'url': 'http://www.ted.com/talks/gabby_giffords_and_mark_kelly_be_passionate_be_courageous_be_your_best',
+        'md5': '71b3ab2f4233012dce09d515c9c39ce2',
         'info_dict': {
             'id': '1972',
             'ext': 'mp4',
@@ -73,7 +81,7 @@ class TEDIE(InfoExtractor):
         'add_ie': ['Youtube'],
         'info_dict': {
             'id': '_ZG8HBuDjgc',
-            'ext': 'mp4',
+            'ext': 'webm',
             'title': 'Douglas Adams: Parrots the Universe and Everything',
             'description': 'md5:01ad1e199c49ac640cb1196c0e9016af',
             'uploader': 'University of California Television (UCTV)',
@@ -102,14 +110,15 @@ class TEDIE(InfoExtractor):
     }]
 
     _NATIVE_FORMATS = {
-        'low': {'preference': 1, 'width': 320, 'height': 180},
-        'medium': {'preference': 2, 'width': 512, 'height': 288},
-        'high': {'preference': 3, 'width': 854, 'height': 480},
+        'low': {'width': 320, 'height': 180},
+        'medium': {'width': 512, 'height': 288},
+        'high': {'width': 854, 'height': 480},
     }
 
     def _extract_info(self, webpage):
-        info_json = self._search_regex(r'q\("\w+.init",({.+})\)</script>',
-                                       webpage, 'info json')
+        info_json = self._search_regex(
+            r'(?s)q\(\s*"\w+.init"\s*,\s*({.+})\)\s*</script>',
+            webpage, 'info json')
         return json.loads(info_json)
 
     def _real_extract(self, url):
@@ -131,11 +140,16 @@ class TEDIE(InfoExtractor):
         webpage = self._download_webpage(url, name,
                                          'Downloading playlist webpage')
         info = self._extract_info(webpage)
-        playlist_info = info['playlist']
+
+        playlist_info = try_get(
+            info, lambda x: x['__INITIAL_DATA__']['playlist'],
+            dict) or info['playlist']
 
         playlist_entries = [
             self.url_result('http://www.ted.com/talks/' + talk['slug'], self.ie_key())
-            for talk in info['talks']
+            for talk in try_get(
+                info, lambda x: x['__INITIAL_DATA__']['talks'],
+                dict) or info['talks']
         ]
         return self.playlist_result(
             playlist_entries,
@@ -144,9 +158,14 @@ class TEDIE(InfoExtractor):
 
     def _talk_info(self, url, video_name):
         webpage = self._download_webpage(url, video_name)
-        self.report_extraction(video_name)
 
-        talk_info = self._extract_info(webpage)['talks'][0]
+        info = self._extract_info(webpage)
+
+        talk_info = try_get(
+            info, lambda x: x['__INITIAL_DATA__']['talks'][0],
+            dict) or info['talks'][0]
+
+        title = talk_info['title'].strip()
 
         external = talk_info.get('external')
         if external:
@@ -160,26 +179,40 @@ class TEDIE(InfoExtractor):
                 'url': ext_url or external['uri'],
             }
 
+        native_downloads = try_get(
+            talk_info, lambda x: x['downloads']['nativeDownloads'],
+            dict) or talk_info['nativeDownloads']
+
         formats = [{
             '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]
+        } for (format_id, format_url) in native_downloads.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)
 
-        for format_id, resources in talk_info['resources'].items():
+        player_talk = talk_info['player_talks'][0]
+
+        resources_ = player_talk.get('resources') or talk_info.get('resources')
+
+        http_url = None
+        for format_id, resources in resources_.items():
             if format_id == 'h264':
                 for resource in resources:
+                    h264_url = resource.get('file')
+                    if not h264_url:
+                        continue
                     bitrate = int_or_none(resource.get('bitrate'))
                     formats.append({
-                        'url': resource['file'],
+                        'url': h264_url,
                         'format_id': '%s-%sk' % (format_id, bitrate),
                         'tbr': bitrate,
                     })
+                    if re.search(r'\d+k', h264_url):
+                        http_url = h264_url
             elif format_id == 'rtmp':
                 streamer = talk_info.get('streamer')
                 if not streamer:
@@ -195,16 +228,24 @@ class TEDIE(InfoExtractor):
                         'tbr': int_or_none(resource.get('bitrate')),
                     })
             elif format_id == 'hls':
-                hls_formats = self._extract_m3u8_formats(
-                    resources.get('stream'), video_name, 'mp4', m3u8_id=format_id)
-                for f in hls_formats:
-                    if f.get('format_id') == 'hls-meta':
-                        continue
-                    if not f.get('height'):
-                        f['vcodec'] = 'none'
-                    else:
-                        f['acodec'] = 'none'
-                formats.extend(hls_formats)
+                formats.extend(self._extract_m3u8_formats(
+                    resources.get('stream'), video_name, 'mp4', m3u8_id=format_id, fatal=False))
+
+        m3u8_formats = list(filter(
+            lambda f: f.get('protocol') == 'm3u8' and f.get('vcodec') != 'none',
+            formats))
+        if http_url:
+            for m3u8_format in m3u8_formats:
+                bitrate = self._search_regex(r'(\d+k)', m3u8_format['url'], 'bitrate', default=None)
+                if not bitrate:
+                    continue
+                f = m3u8_format.copy()
+                f.update({
+                    'url': re.sub(r'\d+k', bitrate, http_url),
+                    'format_id': m3u8_format['format_id'].replace('hls', 'http'),
+                    'protocol': 'http',
+                })
+                formats.append(f)
 
         audio_download = talk_info.get('audioDownload')
         if audio_download:
@@ -212,21 +253,17 @@ class TEDIE(InfoExtractor):
                 'url': audio_download,
                 'format_id': 'audio',
                 'vcodec': 'none',
-                'preference': -0.5,
             })
 
         self._sort_formats(formats)
 
         video_id = compat_str(talk_info['id'])
 
-        thumbnail = talk_info['thumb']
-        if not thumbnail.startswith('http'):
-            thumbnail = 'http://' + thumbnail
         return {
             'id': video_id,
-            'title': talk_info['title'].strip(),
-            'uploader': talk_info['speaker'],
-            'thumbnail': thumbnail,
+            'title': title,
+            'uploader': player_talk.get('speaker') or talk_info.get('speaker'),
+            'thumbnail': player_talk.get('thumb') or talk_info.get('thumb'),
             'description': self._og_search_description(webpage),
             'subtitles': self._get_subtitles(video_id, talk_info),
             'formats': formats,
@@ -234,27 +271,33 @@ class TEDIE(InfoExtractor):
         }
 
     def _get_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:
-                sub_lang_list[l] = [
-                    {
-                        'url': 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/%s' % (video_id, l, ext),
-                        'ext': ext,
-                    }
-                    for ext in ['ted', 'srt']
-                ]
-            return sub_lang_list
-        else:
-            return {}
+        sub_lang_list = {}
+        for language in try_get(
+                talk_info,
+                (lambda x: x['downloads']['languages'],
+                 lambda x: x['languages']), list):
+            lang_code = language.get('languageCode') or language.get('ianaCode')
+            if not lang_code:
+                continue
+            sub_lang_list[lang_code] = [
+                {
+                    'url': 'http://www.ted.com/talks/subtitles/id/%s/lang/%s/format/%s' % (video_id, lang_code, ext),
+                    'ext': ext,
+                }
+                for ext in ['ted', 'srt']
+            ]
+        return sub_lang_list
 
     def _watch_info(self, url, name):
         webpage = self._download_webpage(url, name)
 
         config_json = self._html_search_regex(
             r'"pages\.jwplayer"\s*,\s*({.+?})\s*\)\s*</script>',
-            webpage, 'config')
+            webpage, 'config', default=None)
+        if not config_json:
+            embed_url = self._search_regex(
+                r"<iframe[^>]+class='pages-video-embed__video__object'[^>]+src='([^']+)'", webpage, 'embed url')
+            return self.url_result(self._proto_relative_url(embed_url))
         config = json.loads(config_json)['config']
         video_url = config['video']['url']
         thumbnail = config.get('image', {}).get('url')