From 3ad0b7f422d547204df687b6d0b2d9110fff3990 Mon Sep 17 00:00:00 2001
From: bashonly <88596187+bashonly@users.noreply.github.com>
Date: Tue, 24 Sep 2024 17:10:42 -0500
Subject: [PATCH] [ie/tiktok] Fix web formats extraction (#11074)

Closes #11034
Authored by: bashonly
---
 yt_dlp/extractor/tiktok.py | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

diff --git a/yt_dlp/extractor/tiktok.py b/yt_dlp/extractor/tiktok.py
index 9d823a3154..f7e103fe9f 100644
--- a/yt_dlp/extractor/tiktok.py
+++ b/yt_dlp/extractor/tiktok.py
@@ -542,16 +542,12 @@ def _extract_web_formats(self, aweme_detail):
                 **COMMON_FORMAT_INFO,
                 'format_id': 'download',
                 'url': self._proto_relative_url(download_url),
+                'format_note': 'watermarked',
+                'preference': -2,
             })
 
         self._remove_duplicate_formats(formats)
 
-        for f in traverse_obj(formats, lambda _, v: 'unwatermarked' not in v['url']):
-            f.update({
-                'format_note': join_nonempty(f.get('format_note'), 'watermarked', delim=', '),
-                'preference': f.get('preference') or -2,
-            })
-
         # Is it a slideshow with only audio for download?
         if not formats and traverse_obj(aweme_detail, ('music', 'playUrl', {url_or_none})):
             audio_url = aweme_detail['music']['playUrl']
@@ -565,7 +561,8 @@ def _extract_web_formats(self, aweme_detail):
                 'vcodec': 'none',
             })
 
-        return formats
+        # Filter out broken formats, see https://github.com/yt-dlp/yt-dlp/issues/11034
+        return [f for f in formats if urllib.parse.urlparse(f['url']).hostname != 'www.tiktok.com']
 
     def _parse_aweme_video_web(self, aweme_detail, webpage_url, video_id, extract_flat=False):
         author_info = traverse_obj(aweme_detail, (('authorInfo', 'author', None), {