mirrors/yt-dlp
1
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-27 10:31:29 +00:00
Code Issues Packages Projects Releases Wiki Activity

[ie/N1Info:article] Fix extractor (#7373)

Browse source 
Authored by: u-spec-png
This commit is contained in:
u-spec-png 2023-09-19 01:36:10 +02:00 committed by GitHub
parent 69b03f84f8
commit 8ac5b6d96a
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 39 additions and 13 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

52
yt_dlp/extractor/n1.py
View file

@ -33,7 +33,7 @@ def _real_extract(self, url):
class N1InfoIIE(InfoExtractor):
IE_NAME = 'N1Info:article'
_VALID_URL = r'https?://(?:(?:(?:ba|rs|hr)\.)?n1info\.(?:com|si)|nova\.rs)/(?:[^/]+/){1,2}(?P<id>[^/]+)'
_VALID_URL = r'https?://(?:(?:\w+\.)?n1info\.\w+|nova\.rs)/(?:[^/?#]+/){1,2}(?P<id>[^/?#]+)'
_TESTS = [{
# Youtube embedded
'url': 'https://rs.n1info.com/sport-klub/tenis/kako-je-djokovic-propustio-istorijsku-priliku-video/',
@ -94,6 +94,16 @@ class N1InfoIIE(InfoExtractor):
'upload_date': '20211102',
'timestamp': 1635861677,
},
}, {
'url': 'https://n1info.rs/vesti/cuta-biti-u-kosovskoj-mitrovici-znaci-da-te-docekaju-eksplozivnim-napravama/',
'info_dict': {
'id': '1332368',
'ext': 'mp4',
'title': 'Ćuta: Biti u Kosovskoj Mitrovici znači da te dočekaju eksplozivnim napravama',
'upload_date': '20230620',
'timestamp': 1687290536,
'thumbnail': 'https://cdn.brid.tv/live/partners/26827/snapshot/1332368_th_6492013a8356f_1687290170.jpg'
},
}, {
'url': 'https://hr.n1info.com/vijesti/pravobraniteljica-o-ubojstvu-u-zagrebu-radi-se-o-doista-nezapamcenoj-situaciji/',
'only_matching': True,
@ -105,19 +115,35 @@ def _real_extract(self, url):
title = self._html_search_regex(r'<h1[^>]+>(.+?)</h1>', webpage, 'title')
timestamp = unified_timestamp(self._html_search_meta('article:published_time', webpage))
videos = re.findall(r'(?m)(<video[^>]+>)', webpage)
plugin_data = self._html_search_meta('BridPlugin', webpage)
entries = []
for video in videos:
video_data = extract_attributes(video)
entries.append({
'_type': 'url_transparent',
'url': video_data.get('data-url'),
'id': video_data.get('id'),
'title': title,
'thumbnail': video_data.get('data-thumbnail'),
'timestamp': timestamp,
'ie_key': 'N1InfoAsset'})
if plugin_data:
site_id = self._html_search_regex(r'site:(\d+)', webpage, 'site id')
for video_data in re.findall(r'\$bp\("Brid_\d+", (.+)\);', webpage):
video_id = self._parse_json(video_data, title)['video']
entries.append({
'id': video_id,
'title': title,
'timestamp': timestamp,
'thumbnail': self._html_search_meta('thumbnailURL', webpage),
'formats': self._extract_m3u8_formats(
f'https://cdn-uc.brid.tv/live/partners/{site_id}/streaming/{video_id}/{video_id}.m3u8',
video_id, fatal=False),
})
else:
# Old player still present in older articles
videos = re.findall(r'(?m)(<video[^>]+>)', webpage)
for video in videos:
video_data = extract_attributes(video)
entries.append({
'_type': 'url_transparent',
'url': video_data.get('data-url'),
'id': video_data.get('id'),
'title': title,
'thumbnail': video_data.get('data-thumbnail'),
'timestamp': timestamp,
'ie_key': 'N1InfoAsset',
})
embedded_videos = re.findall(r'(<iframe[^>]+>)', webpage)
for embedded_video in embedded_videos: