administration/yt-dlp
1
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-17 14:41:44 +00:00
Code Issues Projects Releases Wiki Activity

[extractor/gmanetwork] Add extractor (#5945)

Browse Source 
Authored by: HobbyistDev
Partially fixes #5770
This commit is contained in:
HobbyistDev 2023-04-13 03:19:08 +09:00 committed by GitHub
parent c3f624ef0a
commit 2d97d154fe
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 84 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
yt_dlp/extractor/_extractors.py
View File

@ -681,6 +681,7 @@ from .globo import (
GloboIE,
GloboArticleIE,
)
from .gmanetwork import GMANetworkVideoIE
from .go import GoIE
from .godtube import GodTubeIE
from .gofile import GofileIE

83
yt_dlp/extractor/gmanetwork.py Normal file
View File

@ -0,0 +1,83 @@
from .common import InfoExtractor
from .dailymotion import DailymotionIE
from .youtube import YoutubeIE
class GMANetworkVideoIE(InfoExtractor):
_VALID_URL = r'https?://(?:www)\.gmanetwork\.com/(?:\w+/){3}(?P<id>\d+)/(?P<display_id>[\w-]+)/video'
_TESTS = [{
'url': 'https://www.gmanetwork.com/fullepisodes/home/running_man_philippines/168677/running-man-philippines-catch-the-thief-full-chapter-2/video?section=home',
'info_dict': {
'id': '28BqW0AXPe0',
'ext': 'mp4',
'upload_date': '20220919',
'uploader_url': 'http://www.youtube.com/channel/UChsoPNR5x-wdSO2GrOSIWqQ',
'like_count': int,
'view_count': int,
'uploader': 'YoüLOL',
'channel_id': 'UChsoPNR5x-wdSO2GrOSIWqQ',
'duration': 5313,
'comment_count': int,
'tags': 'count:22',
'uploader_id': 'UChsoPNR5x-wdSO2GrOSIWqQ',
'title': 'Running Man Philippines: Catch the Thief (FULL CHAPTER 2)',
'channel_url': 'https://www.youtube.com/channel/UChsoPNR5x-wdSO2GrOSIWqQ',
'thumbnail': 'https://i.ytimg.com/vi/28BqW0AXPe0/maxresdefault.jpg',
'release_timestamp': 1663594212,
'age_limit': 0,
'channel_follower_count': int,
'categories': ['Entertainment'],
'description': 'md5:811bdcea74f9c48051824e494756e926',
'live_status': 'not_live',
'playable_in_embed': True,
'channel': 'YoüLOL',
'availability': 'public',
'release_date': '20220919',
}
}, {
'url': 'https://www.gmanetwork.com/fullepisodes/home/more_than_words/87059/more-than-words-full-episode-80/video?section=home',
'info_dict': {
'id': 'yiDOExw2aSA',
'ext': 'mp4',
'live_status': 'not_live',
'channel': 'GMANetwork',
'like_count': int,
'channel_follower_count': int,
'description': 'md5:6d00cd658394fa1a5071200d3ed4be05',
'duration': 1419,
'age_limit': 0,
'comment_count': int,
'upload_date': '20181003',
'thumbnail': 'https://i.ytimg.com/vi_webp/yiDOExw2aSA/maxresdefault.webp',
'availability': 'public',
'playable_in_embed': True,
'channel_id': 'UCKL5hAuzgFQsyrsQKgU0Qng',
'title': 'More Than Words: Full Episode 80 (Finale)',
'uploader_id': 'GMANETWORK',
'categories': ['Entertainment'],
'uploader': 'GMANetwork',
'channel_url': 'https://www.youtube.com/channel/UCKL5hAuzgFQsyrsQKgU0Qng',
'tags': 'count:29',
'view_count': int,
'uploader_url': 'http://www.youtube.com/user/GMANETWORK',
}
}]
def _real_extract(self, url):
content_id, display_id = self._match_valid_url(url).group('id', 'display_id')
webpage = self._download_webpage(url, display_id)
# webpage route
youtube_id = self._search_regex(
r'var\s*YOUTUBE_VIDEO\s*=\s*[\'"]+(?P<yt_id>[\w-]+)', webpage, 'youtube_id', fatal=False)
if youtube_id:
return self.url_result(youtube_id, YoutubeIE, youtube_id)
# api call route
# more info at https://aphrodite.gmanetwork.com/fullepisodes/assets/fullepisodes/js/dist/fullepisodes_video.js?v=1.1.11
network_url = self._search_regex(
r'NETWORK_URL\s*=\s*[\'"](?P<url>[^\'"]+)', webpage, 'network_url')
json_data = self._download_json(f'{network_url}api/data/content/video/{content_id}', display_id)
if json_data.get('video_file'):
return self.url_result(json_data['video_file'], YoutubeIE, json_data['video_file'])
else:
return self.url_result(json_data['dailymotion_file'], DailymotionIE, json_data['dailymotion_file'])