diff options
| author | Refael Ackermann <refack@gmail.com> | 2025-02-22 19:40:53 -0500 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-02-23 00:40:53 +0000 |
| commit | 7f3006eb0c0659982bb956d71b0bc806bcb0a5f2 (patch) | |
| tree | 0afec822a50727c650072b968c78fc328205772b | |
| parent | 4445f37a7a66b248dbd8376c43137e6e441f138e (diff) | |
[ie/wsj] Support opinion URLs and impersonation (#12431)
Authored by: refack
| -rw-r--r-- | yt_dlp/extractor/wsj.py | 17 |
1 files changed, 13 insertions, 4 deletions
diff --git a/yt_dlp/extractor/wsj.py b/yt_dlp/extractor/wsj.py index b6b656f7d..7cf46141c 100644 --- a/yt_dlp/extractor/wsj.py +++ b/yt_dlp/extractor/wsj.py @@ -100,8 +100,8 @@ class WSJIE(InfoExtractor): class WSJArticleIE(InfoExtractor): - _VALID_URL = r'(?i)https?://(?:www\.)?wsj\.com/articles/(?P<id>[^/?#&]+)' - _TEST = { + _VALID_URL = r'(?i)https?://(?:www\.)?wsj\.com/(?:articles|opinion)/(?P<id>[^/?#&]+)' + _TESTS = [{ 'url': 'https://www.wsj.com/articles/dont-like-china-no-pandas-for-you-1490366939?', 'info_dict': { 'id': '4B13FA62-1D8C-45DB-8EA1-4105CB20B362', @@ -110,11 +110,20 @@ class WSJArticleIE(InfoExtractor): 'uploader_id': 'ralcaraz', 'title': 'Bao Bao the Panda Leaves for China', }, - } + }, { + 'url': 'https://www.wsj.com/opinion/hamas-hostages-caskets-bibas-family-israel-gaza-29da083b', + 'info_dict': { + 'id': 'CE68D629-8DB8-4CD3-B30A-92112C102054', + 'ext': 'mp4', + 'upload_date': '20241007', + 'uploader_id': 'Tinnes, David', + 'title': 'WSJ Opinion: "Get the Jew": The Crown Heights Riot Revisited', + }, + }] def _real_extract(self, url): article_id = self._match_id(url) - webpage = self._download_webpage(url, article_id) + webpage = self._download_webpage(url, article_id, impersonate=True) video_id = self._search_regex( r'(?:id=["\']video|video-|iframe\.html\?guid=|data-src=["\'])([a-fA-F0-9-]{36})', webpage, 'video id') |