itdominator/SolarFM
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Upgrade yt_dlp and download script

Browse Source
This commit is contained in:
itdominator
2025-05-02 16:11:08 -05:00
parent 3a2e8eeb08
commit d68d9ce4f9
1194 changed files with 60099 additions and 44436 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
plugins/youtube_download/yt_dlp/extractor/wsj.py
View File

@@ -1,7 +1,8 @@
from .common import InfoExtractor
from ..utils import (
int_or_none,
float_or_none,
int_or_none,
join_nonempty,
unified_strdate,
)
@@ -76,7 +77,7 @@ class WSJIE(InfoExtractor):
tbr = int_or_none(v.get('bitrate'))
formats.append({
'url': mp4_url,
'format_id': 'http' + ('-%d' % tbr if tbr else ''),
'format_id': join_nonempty('http', tbr),
'tbr': tbr,
'width': int_or_none(v.get('width')),
'height': int_or_none(v.get('height')),
@@ -99,8 +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',
@@ -108,13 +109,22 @@ class WSJArticleIE(InfoExtractor):
'upload_date': '20170221',
'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')
return self.url_result('wsj:%s' % video_id, WSJIE.ie_key(), video_id)
return self.url_result(f'wsj:{video_id}', WSJIE.ie_key(), video_id)