[extractor/premiershiprugby] Add extractor (#4129)

Closes #2980
Authored by: HobbyistDev
This commit is contained in:
HobbyistDev 2022-06-25 23:43:32 +09:00 committed by GitHub
parent c043c24625
commit 30d22d775b
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 40 additions and 0 deletions

View file

@ -1330,6 +1330,7 @@
PuhuTVIE, PuhuTVIE,
PuhuTVSerieIE, PuhuTVSerieIE,
) )
from .premiershiprugby import PremiershipRugbyIE
from .presstv import PressTVIE from .presstv import PressTVIE
from .projectveritas import ProjectVeritasIE from .projectveritas import ProjectVeritasIE
from .prosiebensat1 import ProSiebenSat1IE from .prosiebensat1 import ProSiebenSat1IE

View file

@ -0,0 +1,39 @@
from .common import InfoExtractor
from ..utils import int_or_none, traverse_obj
class PremiershipRugbyIE(InfoExtractor):
_VALID_URL = r'https?://(?:\w+\.)premiershiprugby\.(?:com)/watch/(?P<id>[\w-]+)'
_TESTS = [{
'url': 'https://www.premiershiprugby.com/watch/full-match-harlequins-v-newcastle-falcons',
'info_dict': {
'id': '0_mbkb7ldt',
'title': 'Full Match: Harlequins v Newcastle Falcons',
'ext': 'mp4',
'thumbnail': 'https://open.http.mp.streamamg.com/p/3000914/sp/300091400/thumbnail/entry_id/0_mbkb7ldt//width/960/height/540/type/1/quality/75',
'duration': 6093.0,
'tags': ['video'],
'categories': ['Full Match', 'Harlequins', 'Newcastle Falcons', 'gallaher premiership'],
}
}]
def _real_extract(self, url):
display_id = self._match_id(url)
json_data = self._download_json(
f'https://article-cms-api.incrowdsports.com/v2/articles/slug/{display_id}',
display_id, query={'clientId': 'PRL'})['data']['article']
formats, subs = self._extract_m3u8_formats_and_subtitles(
json_data['heroMedia']['content']['videoLink'], display_id)
return {
'id': json_data['heroMedia']['content']['sourceSystemId'],
'display_id': display_id,
'title': traverse_obj(json_data, ('heroMedia', 'title')),
'formats': formats,
'subtitles': subs,
'thumbnail': traverse_obj(json_data, ('heroMedia', 'content', 'videoThumbnail')),
'duration': int_or_none(traverse_obj(json_data, ('heroMedia', 'content', 'metadata', 'msDuration')), scale=1000),
'tags': json_data.get('tags'),
'categories': traverse_obj(json_data, ('categories', ..., 'text')),
}