From cf6413e840476c15e5b166dc2f7cc2a90a4a9aad Mon Sep 17 00:00:00 2001
From: HobbyistDev <105957301+HobbyistDev@users.noreply.github.com>
Date: Fri, 19 Jan 2024 08:27:25 +0900
Subject: [PATCH] [ie/BiliIntl] Fix and improve subtitles extraction (#7077)

Closes #7075, Closes #6664
Authored by: HobbyistDev, itachi-19, dirkf, seproDev

Co-authored-by: itachi-19 <16500619+itachi-19@users.noreply.github.com>
Co-authored-by: dirkf <fieldhouse@gmx.net>
Co-authored-by: sepro <4618135+seproDev@users.noreply.github.com>
---
 yt_dlp/extractor/bilibili.py | 42 +++++++++++++++++++++++++-----------
 1 file changed, 29 insertions(+), 13 deletions(-)

diff --git a/yt_dlp/extractor/bilibili.py b/yt_dlp/extractor/bilibili.py
index bc25dc75e..5475b3650 100644
--- a/yt_dlp/extractor/bilibili.py
+++ b/yt_dlp/extractor/bilibili.py
@@ -18,6 +18,7 @@
     OnDemandPagedList,
     bool_or_none,
     clean_html,
+    determine_ext,
     filter_dict,
     float_or_none,
     format_field,
@@ -1658,19 +1659,34 @@ def _get_subtitles(self, *, ep_id=None, aid=None):
                 'aid': aid,
             })) or {}
         subtitles = {}
-        for sub in sub_json.get('subtitles') or []:
-            sub_url = sub.get('url')
-            if not sub_url:
-                continue
-            sub_data = self._download_json(
-                sub_url, ep_id or aid, errnote='Unable to download subtitles', fatal=False,
-                note='Downloading subtitles%s' % f' for {sub["lang"]}' if sub.get('lang') else '')
-            if not sub_data:
-                continue
-            subtitles.setdefault(sub.get('lang_key', 'en'), []).append({
-                'ext': 'srt',
-                'data': self.json2srt(sub_data)
-            })
+        fetched_urls = set()
+        for sub in traverse_obj(sub_json, (('subtitles', 'video_subtitle'), ..., {dict})):
+            for url in traverse_obj(sub, ((None, 'ass', 'srt'), 'url', {url_or_none})):
+                if url in fetched_urls:
+                    continue
+                fetched_urls.add(url)
+                sub_ext = determine_ext(url)
+                sub_lang = sub.get('lang_key') or 'en'
+
+                if sub_ext == 'ass':
+                    subtitles.setdefault(sub_lang, []).append({
+                        'ext': 'ass',
+                        'url': url,
+                    })
+                elif sub_ext == 'json':
+                    sub_data = self._download_json(
+                        url, ep_id or aid, fatal=False,
+                        note=f'Downloading subtitles{format_field(sub, "lang", " for %s")} ({sub_lang})',
+                        errnote='Unable to download subtitles')
+
+                    if sub_data:
+                        subtitles.setdefault(sub_lang, []).append({
+                            'ext': 'srt',
+                            'data': self.json2srt(sub_data),
+                        })
+                else:
+                    self.report_warning('Unexpected subtitle extension', ep_id or aid)
+
         return subtitles
 
     def _get_formats(self, *, ep_id=None, aid=None):