[ie/Vbox7] Fix extractor (#9100)

Closes #1098, Closes #5661
Authored by: seproDev
This commit is contained in:
sepro 2024-01-29 21:16:46 +01:00 committed by GitHub
parent 9b5efaf86b
commit 67bb70cd70
No known key found for this signature in database
GPG key ID: B5690EEEBB952194

View file

@ -1,5 +1,6 @@
from .common import InfoExtractor from .common import InfoExtractor
from ..utils import ExtractorError from ..utils import ExtractorError, base_url, int_or_none, url_basename
from ..utils.traversal import traverse_obj
class Vbox7IE(InfoExtractor): class Vbox7IE(InfoExtractor):
@ -19,7 +20,7 @@ class Vbox7IE(InfoExtractor):
_GEO_COUNTRIES = ['BG'] _GEO_COUNTRIES = ['BG']
_TESTS = [{ _TESTS = [{
'url': 'http://vbox7.com/play:0946fff23c', 'url': 'http://vbox7.com/play:0946fff23c',
'md5': 'a60f9ab3a3a2f013ef9a967d5f7be5bf', 'md5': '50ca1f78345a9c15391af47d8062d074',
'info_dict': { 'info_dict': {
'id': '0946fff23c', 'id': '0946fff23c',
'ext': 'mp4', 'ext': 'mp4',
@ -29,19 +30,25 @@ class Vbox7IE(InfoExtractor):
'timestamp': 1470982814, 'timestamp': 1470982814,
'upload_date': '20160812', 'upload_date': '20160812',
'uploader': 'zdraveibulgaria', 'uploader': 'zdraveibulgaria',
}, 'view_count': int,
'params': { 'duration': 2640,
'proxy': '127.0.0.1:8118',
}, },
}, { }, {
'url': 'http://vbox7.com/play:249bb972c2', 'url': 'http://vbox7.com/play:249bb972c2',
'md5': '99f65c0c9ef9b682b97313e052734c3f', 'md5': 'da1dd2eb245200cb86e6d09d43232116',
'info_dict': { 'info_dict': {
'id': '249bb972c2', 'id': '249bb972c2',
'ext': 'mp4', 'ext': 'mp4',
'title': 'Смях! Чудо - чист за секунди - Скрита камера', 'title': 'Смях! Чудо - чист за секунди - Скрита камера',
'uploader': 'svideteliat_ot_varshava',
'view_count': int,
'timestamp': 1360215023,
'thumbnail': 'https://i49.vbox7.com/design/iconci/png/noimg6.png',
'description': 'Смях! Чудо - чист за секунди - Скрита камера',
'upload_date': '20130207',
'duration': 83,
}, },
'skip': 'georestricted', 'expected_warnings': ['Failed to download m3u8 information'],
}, { }, {
'url': 'http://vbox7.com/emb/external.php?vid=a240d20f9c&autoplay=1', 'url': 'http://vbox7.com/emb/external.php?vid=a240d20f9c&autoplay=1',
'only_matching': True, 'only_matching': True,
@ -53,41 +60,38 @@ class Vbox7IE(InfoExtractor):
def _real_extract(self, url): def _real_extract(self, url):
video_id = self._match_id(url) video_id = self._match_id(url)
response = self._download_json( data = self._download_json(
'https://www.vbox7.com/ajax/video/nextvideo.php?vid=%s' % video_id, 'https://www.vbox7.com/aj/player/item/options', video_id,
video_id) query={'vid': video_id})['options']
if 'error' in response: src_url = data.get('src')
raise ExtractorError( if src_url in (None, '', 'blank'):
'%s said: %s' % (self.IE_NAME, response['error']), expected=True) raise ExtractorError('Video is unavailable', expected=True)
video = response['options'] fmt_base = url_basename(src_url).rsplit('.', 1)[0].rsplit('_', 1)[0]
if fmt_base == 'vn':
self.raise_geo_restricted()
title = video['title'] fmt_base = base_url(src_url) + fmt_base
video_url = video['src']
if '/na.mp4' in video_url: formats = self._extract_m3u8_formats(
self.raise_geo_restricted(countries=self._GEO_COUNTRIES) f'{fmt_base}.m3u8', video_id, m3u8_id='hls', fatal=False)
# TODO: Add MPD formats, when dash range support is added
for res in traverse_obj(data, ('resolutions', lambda _, v: v != 0, {int})):
formats.append({
'url': f'{fmt_base}_{res}.mp4',
'format_id': f'http-{res}',
'height': res,
})
uploader = video.get('uploader') return {
webpage = self._download_webpage(
'http://vbox7.com/play:%s' % video_id, video_id, fatal=None)
info = {}
if webpage:
info = self._search_json_ld(
webpage.replace('"/*@context"', '"@context"'), video_id,
fatal=False)
info.update({
'id': video_id, 'id': video_id,
'title': title, 'formats': formats,
'url': video_url, **self._search_json_ld(self._download_webpage(
'uploader': uploader, f'https://www.vbox7.com/play:{video_id}', video_id, fatal=False) or '', video_id, fatal=False),
'thumbnail': self._proto_relative_url( **traverse_obj(data, {
info.get('thumbnail') or self._og_search_thumbnail(webpage), 'title': ('title', {str}),
'http:'), 'uploader': ('uploader', {str}),
}) 'duration': ('duration', {int_or_none}),
return info }),
}