[ie/mediasite] Provide untimed transcript in ttml format

kclauhk · Dec 26, 2024 · a9df4a4 · a9df4a4
1 parent 363863e
commit a9df4a4
Showing 1 changed file with 34 additions and 7 deletions.
diff --git a/yt_dlp/extractor/mediasite.py b/yt_dlp/extractor/mediasite.py
@@ -233,6 +233,25 @@ def __extract_slides(self, *, stream_id, snum, stream, duration, images):
             'fragment_base_url': slide_base_url,
         }
 
+    def _get_transcript_txt(self, transcript_url, resource_id, lang_code='und', lang_name=None):
+        ts = {
+            'name': join_nonempty(lang_name, '(Untimed)', delim=' '),
+            'ext': 'ttml',
+        }
+        if ((self.get_param('writesubtitles') or self.get_param('writeautomaticsub'))
+                and 'ttml' in self.get_param('subtitlesformat')):
+            if transcript := self._download_webpage(
+                    transcript_url, resource_id, note='Downloading transcript', fatal=False):
+                d = ('<?xml version="1.0" encoding="utf-8" ?>\n'
+                    f'<tt xml:lang="{lang_code}" xmlns="http://www.w3.org/ns/ttml" />\n'
+                    '<head>\n</head>\n<body>\n<div>\n<p xml:id="transcript">\n'
+                    + transcript.strip()
+                    + '\n</p>\n</div>\n</body>\n</tt>')
+                return {'data': d, **ts}
+        else:
+            return {'url': transcript_url, **ts}
+        return {}
+
     def _real_extract(self, url):
         url, data = unsmuggle_url(url, {})
         mobj = self._match_valid_url(url)
@@ -370,14 +389,22 @@ def _real_extract(self, url):
             else:
                 subtitles.setdefault(lang_code, []).append(t)
         if transcript_url := presentation.get('TranscriptUrl'):
+            if 'playbackTicket' not in transcript_url:
+                transcript_url = join_nonempty(
+                    transcript_url, traverse_obj(presentation, ('Streams', 0, 'SlidePlaybackTicketId', {str_or_none})),
+                    delim='?playbackTicket=')
             if determine_ext(transcript_url) != 'txt':
-                if len(transcripts) == 1:
-                    (captions or subtitles).setdefault(lang_code, []).append({
-                        'url': transcript_url,
-                        'name': lang_name,
-                    })
-                else:
-                    subtitles.setdefault('und', []).append({'url': transcript_url})
+                ts = {'url': transcript_url}
+            else:
+                ts = self._get_transcript_txt(
+                    transcript_url, resource_id, *([lang_code, lang_name] if len(transcripts) == 1 else ['und']))
+            if len(transcripts) == 1:
+                (captions or subtitles).setdefault(lang_code, []).insert(0, {
+                    'name': lang_name,
+                    **ts,
+                })
+            else:
+                subtitles.setdefault('und', []).insert(0, ts)
 
         return {
             'id': resource_id,