[wistia] improve format extraction and extract subtitles(closes #22590)

youtube_dl/extractor/wistia.py

@@ -13,8 +13,7 @@
 
 class WistiaIE(InfoExtractor):
     _VALID_URL = r'(?:wistia:|https?://(?:fast\.)?wistia\.(?:net|com)/embed/(?:iframe|medias)/)(?P<id>[a-z0-9]{10})'
-    _API_URL = 'http://fast.wistia.com/embed/medias/%s.json'
+    _EMBED_BASE_URL = 'http://fast.wistia.com/embed/'
-    _IFRAME_URL = 'http://fast.wistia.net/embed/iframe/%s'
 
     _TESTS = [{
         'url': 'http://fast.wistia.net/embed/iframe/sh7fpupwlt',
@@ -67,10 +66,10 @@ def _real_extract(self, url):
         video_id = self._match_id(url)
 
         data_json = self._download_json(
-            self._API_URL % video_id, video_id,
+            self._EMBED_BASE_URL + 'medias/%s.json' % video_id, video_id,
             # Some videos require this.
             headers={
-                'Referer': url if url.startswith('http') else self._IFRAME_URL % video_id,
+                'Referer': url if url.startswith('http') else self._EMBED_BASE_URL + 'iframe/' + video_id,
             })
 
         if data_json.get('error'):
@@ -95,27 +94,61 @@ def _real_extract(self, url):
                     'url': aurl,
                     'width': int_or_none(a.get('width')),
                     'height': int_or_none(a.get('height')),
+                    'filesize': int_or_none(a.get('size')),
                 })
             else:
                 aext = a.get('ext')
-                is_m3u8 = a.get('container') == 'm3u8' or aext == 'm3u8'
+                display_name = a.get('display_name')
-                formats.append({
+                format_id = atype
-                    'format_id': atype,
+                if atype and atype.endswith('_video') and display_name:
+                    format_id = '%s-%s' % (atype[:-6], display_name)
+                f = {
+                    'format_id': format_id,
                     'url': aurl,
-                    'tbr': int_or_none(a.get('bitrate')),
+                    'tbr': int_or_none(a.get('bitrate')) or None,
-                    'vbr': int_or_none(a.get('opt_vbitrate')),
-                    'width': int_or_none(a.get('width')),
-                    'height': int_or_none(a.get('height')),
-                    'filesize': int_or_none(a.get('size')),
-                    'vcodec': a.get('codec'),
-                    'container': a.get('container'),
-                    'ext': 'mp4' if is_m3u8 else aext,
-                    'protocol': 'm3u8' if is_m3u8 else None,
                     'preference': 1 if atype == 'original' else None,
-                })
+                }
+                if display_name == 'Audio':
+                    f.update({
+                        'vcodec': 'none',
+                    })
+                else:
+                    f.update({
+                        'width': int_or_none(a.get('width')),
+                        'height': int_or_none(a.get('height')),
+                        'vcodec': a.get('codec'),
+                    })
+                if a.get('container') == 'm3u8' or aext == 'm3u8':
+                    ts_f = f.copy()
+                    ts_f.update({
+                        'ext': 'ts',
+                        'format_id': f['format_id'].replace('hls-', 'ts-'),
+                        'url': f['url'].replace('.bin', '.ts'),
+                    })
+                    formats.append(ts_f)
+                    f.update({
+                        'ext': 'mp4',
+                        'protocol': 'm3u8_native',
+                    })
+                else:
+                    f.update({
+                        'container': a.get('container'),
+                        'ext': aext,
+                        'filesize': int_or_none(a.get('size')),
+                    })
+                formats.append(f)
 
         self._sort_formats(formats)
 
+        subtitles = {}
+        for caption in data.get('captions', []):
+            language = caption.get('language')
+            if not language:
+                continue
+            subtitles[language] = [{
+                'url': self._EMBED_BASE_URL + 'captions/' + video_id + '.vtt?language=' + language,
+            }]
+
         return {
             'id': video_id,
             'title': title,
@@ -124,4 +157,5 @@ def _real_extract(self, url):
             'thumbnails': thumbnails,
             'duration': float_or_none(data.get('duration')),
             'timestamp': int_or_none(data.get('createdAt')),
+            'subtitles': subtitles,
         }