From 741fd809bc4d301c19b53877692ae510334a6750 Mon Sep 17 00:00:00 2001 From: "Sergey B (Troex Nevelin)" <436912+troex@users.noreply.github.com> Date: Wed, 23 Apr 2025 02:14:42 +0200 Subject: [PATCH] [ie/GetCourseRu] Fix extractors (#12943) Closes #12941 Authored by: troex --- yt_dlp/extractor/getcourseru.py | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/yt_dlp/extractor/getcourseru.py b/yt_dlp/extractor/getcourseru.py index b7581d77e2..2d923cf540 100644 --- a/yt_dlp/extractor/getcourseru.py +++ b/yt_dlp/extractor/getcourseru.py @@ -8,7 +8,7 @@ from ..utils.traversal import traverse_obj class GetCourseRuPlayerIE(InfoExtractor): - _VALID_URL = r'https?://player02\.getcourse\.ru/sign-player/?\?(?:[^#]+&)?json=[^#&]+' + _VALID_URL = r'https?://(?:player02\.getcourse\.ru|cf-api-2\.vhcdn\.com)/sign-player/?\?(?:[^#]+&)?json=[^#&]+' _EMBED_REGEX = [rf']+\bsrc=[\'"](?P{_VALID_URL}[^\'"]*)'] _TESTS = [{ 'url': 'http://player02.getcourse.ru/sign-player/?json=eyJ2aWRlb19oYXNoIjoiMTkwYmRmOTNmMWIyOTczNTMwOTg1M2E3YTE5ZTI0YjMiLCJ1c2VyX2lkIjozNTk1MjUxODMsInN1Yl9sb2dpbl91c2VyX2lkIjpudWxsLCJsZXNzb25faWQiOm51bGwsImlwIjoiNDYuMTQyLjE4Mi4yNDciLCJnY19ob3N0IjoiYWNhZGVteW1lbC5vbmxpbmUiLCJ0aW1lIjoxNzA1NDQ5NjQyLCJwYXlsb2FkIjoidV8zNTk1MjUxODMiLCJ1aV9sYW5ndWFnZSI6InJ1IiwiaXNfaGF2ZV9jdXN0b21fc3R5bGUiOnRydWV9&s=354ad2c993d95d5ac629e3133d6cefea&vh-static-feature=zigzag', @@ -20,6 +20,16 @@ class GetCourseRuPlayerIE(InfoExtractor): 'duration': 1693, }, 'skip': 'JWT expired', + }, { + 'url': 'https://cf-api-2.vhcdn.com/sign-player/?json=example', + 'info_dict': { + 'id': '435735291', + 'title': '8afd7c489952108e00f019590f3711f3', + 'ext': 'mp4', + 'thumbnail': 'https://preview-htz.vhcdn.com/preview/8afd7c489952108e00f019590f3711f3/preview.jpg?version=1682170973&host=vh-72', + 'duration': 777, + }, + 'skip': 'JWT expired', }] def _real_extract(self, url): @@ -168,7 +178,7 @@ class GetCourseRuIE(InfoExtractor): playlist_id = self._search_regex( r'window\.(?:lessonId|gcsObjectId)\s*=\s*(\d+)', webpage, 'playlist id', default=display_id) - title = self._og_search_title(webpage) or self._html_extract_title(webpage) + title = self._og_search_title(webpage, default=None) or self._html_extract_title(webpage) return self.playlist_from_matches( re.findall(GetCourseRuPlayerIE._EMBED_REGEX[0], webpage),