[ie/tiktok] Fix and deprioritize JSON subtitles (#10516)

Fixes regression caused by 5ce5824 Closes #10514 Authored by: bashonly
yt-dlp · Jul 23, 2024 · 2f97779 · 2f97779
1 parent 713b4cd
commit 2f97779
Showing 1 changed file with 16 additions and 3 deletions.
diff --git a/yt_dlp/extractor/tiktok.py b/yt_dlp/extractor/tiktok.py
@@ -23,7 +23,6 @@
     mimetype2ext,
     parse_qs,
     qualities,
-    remove_start,
     srt_subtitles_timecode,
     str_or_none,
     traverse_obj,
@@ -254,7 +253,16 @@ def _extract_web_data_and_status(self, url, video_id, fatal=True):
 
     def _get_subtitles(self, aweme_detail, aweme_id, user_name):
         # TODO: Extract text positioning info
+
+        EXT_MAP = {  # From lowest to highest preference
+            'creator_caption': 'json',
+            'srt': 'srt',
+            'webvtt': 'vtt',
+        }
+        preference = qualities(tuple(EXT_MAP.values()))
+
         subtitles = {}
+
         # aweme/detail endpoint subs
         captions_info = traverse_obj(
             aweme_detail, ('interaction_stickers', ..., 'auto_video_caption_info', 'auto_captions', ...), expected_type=dict)
@@ -278,8 +286,8 @@ def _get_subtitles(self, aweme_detail, aweme_id, user_name):
                 if not caption.get('url'):
                     continue
                 subtitles.setdefault(caption.get('lang') or 'en', []).append({
-                    'ext': remove_start(caption.get('caption_format'), 'web'),
                     'url': caption['url'],
+                    'ext': EXT_MAP.get(caption.get('Format')),
                 })
         # webpage subs
         if not subtitles:
@@ -288,9 +296,14 @@ def _get_subtitles(self, aweme_detail, aweme_id, user_name):
                     self._create_url(user_name, aweme_id), aweme_id, fatal=False)
             for caption in traverse_obj(aweme_detail, ('video', 'subtitleInfos', lambda _, v: v['Url'])):
                 subtitles.setdefault(caption.get('LanguageCodeName') or 'en', []).append({
-                    'ext': remove_start(caption.get('Format'), 'web'),
                     'url': caption['Url'],
+                    'ext': EXT_MAP.get(caption.get('Format')),
                 })
+
+        # Deprioritize creator_caption json since it can't be embedded or used by media players
+        for lang, subs_list in subtitles.items():
+            subtitles[lang] = sorted(subs_list, key=lambda x: preference(x['ext']))
+
         return subtitles
 
     def _parse_url_key(self, url_key):