From edd83104b432a971feb86c69c4963dc67cc3dfc0 Mon Sep 17 00:00:00 2001 From: George Schizas Date: Tue, 11 Aug 2020 12:20:47 +0300 Subject: [PATCH] Use initial data from JS instead to get chapters There are probably a lot more useful data in there. --- youtube_dl/extractor/youtube.py | 14 ++++---------- 1 file changed, 4 insertions(+), 10 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index b35bf03aa..03b726942 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1661,21 +1661,15 @@ def extract_id(cls, url): def _extract_chapters_from_json(self, webpage, video_id, duration): if not webpage: return - player = self._parse_json( + initial_data = self._parse_json( self._search_regex( - r'RELATED_PLAYER_ARGS["\']\s*:\s*({.+})\s*,?\s*\n', webpage, + r'window\["ytInitialData"\] = (.+);\n', webpage, 'player args', default='{}'), video_id, fatal=False) - if not player or not isinstance(player, dict): - return - watch_next_response = player.get('watch_next_response') - if not isinstance(watch_next_response, compat_str): - return - response = self._parse_json(watch_next_response, video_id, fatal=False) - if not response or not isinstance(response, dict): + if not initial_data or not isinstance(initial_data, dict): return chapters_list = try_get( - response, + initial_data, lambda x: x['playerOverlays'] ['playerOverlayRenderer'] ['decoratedPlayerBarRenderer']