yt_dlp/extractor/facebook.py

@@ -19,7 +19,6 @@ from ..utils import (
     get_element_by_id,
     get_first,
     int_or_none,
-    join_nonempty,
     js_to_json,
     merge_dicts,
     parse_count,
@@ -422,25 +421,29 @@ class FacebookIE(InfoExtractor):
             post = traverse_obj(post_data, (
                 ..., 'require', ..., ..., ..., '__bbox', 'require', ..., ..., ..., '__bbox', 'result', 'data'), expected_type=dict) or []
             snippet = traverse_obj(post, (..., 'video', ..., 'attachments', ..., lambda k, v: (
-                k == 'media' and str(v['id']) == video_id and v['__typename'] == 'Video')))
+                k == 'media' and str(v['id']) == video_id and v['__typename'] == 'Video')), expected_type=dict) or {}
             locale = self._html_search_meta(['og:locale', 'twitter:locale'], webpage, 'locale', default='en_US')
             captions = get_first(snippet, 'video_available_captions_locales', 'captions_url')
-            is_video_broadcast = get_first(snippet, 'is_video_broadcast', expected_type=bool)
+            useIsVideoBroadcast = get_first(snippet, ('is_video_broadcast')) or False
             automatic_captions = {}
             subtitles = {}
-            if url_or_none(captions):  # snippet only had 'captions_url'
+            if isinstance(captions, str):
-                subtitles[locale] = [{'url': captions}]
+                subtitles[locale] = [{'ext': determine_ext(captions, default_ext='srt'), 'url': captions}]
-            else:
+            elif isinstance(captions, list):
-                captions = sorted(captions, key=lambda c: c['locale'])
+                if len(captions) > 1:
-            for caption in traverse_obj(captions, lambda _, v: v['captions_url'] and v['locale']):
+                    captions = sorted(captions, key=lambda c: (c['locale'] != locale, c['locale']))
-                subs = {
+                for c in captions:
-                    'url': caption['captions_url'],
+                    s = {
-                    'name': join_nonempty('localized_language', 'localized_country', from_dict=caption),
+                        'ext': determine_ext(c['captions_url'], default_ext='srt'),
+                        'url': c['captions_url'],
+                        'name': (c['localized_language']
+                                 + (' (' + c['localized_country'] + ')' if c['localized_country'] else '')
+                                 + (' (' + c['localized_creation_method'] + ')' if c['localized_creation_method'] else '')),
                     }
-                if caption.get('localized_creation_method') or is_video_broadcast:
+                    if c['localized_creation_method'] or useIsVideoBroadcast:
-                    automatic_captions.setdefault(caption['locale'], []).append(subs)
+                        automatic_captions.setdefault(c['locale'], []).append(s)
                     else:
-                    subtitles.setdefault(caption['locale'], []).append(subs)
+                        subtitles.setdefault(c['locale'], []).append(s)
             media = traverse_obj(post, (..., 'attachments', ..., lambda k, v: (
                 k == 'media' and str(v['id']) == video_id and v['__typename'] == 'Video')), expected_type=dict)
             title = get_first(media, ('title', 'text'))
@@ -731,7 +734,6 @@ class FacebookIE(InfoExtractor):
         video_data = video_data[0]
 
         formats = []
-        subtitles = {}
         for f in video_data:
             format_id = f['stream_type']
             if f and isinstance(f, dict):
@@ -754,14 +756,10 @@ class FacebookIE(InfoExtractor):
                             'height': 720 if quality == 'hd' else None
                         })
             extract_dash_manifest(f[0], formats)
-            subtitles_src = f[0].get('subtitles_src')
-            if subtitles_src:
-                subtitles.setdefault('en', []).append({'url': subtitles_src})
 
         info_dict = {
             'id': video_id,
             'formats': formats,
-            'subtitles': subtitles,
         }
         process_formats(info_dict)
         info_dict.update(extract_metadata(webpage))