[extractor/europarl] Add EuroParlWebstream Extractor (#5547)

Authored by: HobbyistDev
Closes #4933
This commit is contained in:
HobbyistDev 2022-12-10 17:44:43 +09:00 committed by GitHub
parent 3ac5476430
commit 22697a84f6
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 85 additions and 1 deletions

View File

@ -537,7 +537,7 @@ from .espn import (
ESPNCricInfoIE, ESPNCricInfoIE,
) )
from .esri import EsriVideoIE from .esri import EsriVideoIE
from .europa import EuropaIE from .europa import EuropaIE, EuroParlWebstreamIE
from .europeantour import EuropeanTourIE from .europeantour import EuropeanTourIE
from .eurosport import EurosportIE from .eurosport import EurosportIE
from .euscreen import EUScreenIE from .euscreen import EUScreenIE

View File

@ -3,6 +3,7 @@ from ..utils import (
int_or_none, int_or_none,
orderedSet, orderedSet,
parse_duration, parse_duration,
parse_iso8601,
parse_qs, parse_qs,
qualities, qualities,
unified_strdate, unified_strdate,
@ -87,3 +88,86 @@ class EuropaIE(InfoExtractor):
'view_count': view_count, 'view_count': view_count,
'formats': formats 'formats': formats
} }
class EuroParlWebstreamIE(InfoExtractor):
_VALID_URL = r'''(?x)
https?://(?:multimedia|webstreaming)\.europarl\.europa\.eu/[^/#?]+/
(?:embed/embed\.html\?event=|(?!video)[^/#?]+/[\w-]+_)(?P<id>[\w-]+)
'''
_TESTS = [{
'url': 'https://multimedia.europarl.europa.eu/pl/webstreaming/plenary-session_20220914-0900-PLENARY',
'info_dict': {
'id': 'bcaa1db4-76ef-7e06-8da7-839bd0ad1dbe',
'ext': 'mp4',
'release_timestamp': 1663137900,
'title': 'Plenary session',
'release_date': '20220914',
},
'params': {
'skip_download': True,
}
}, {
'url': 'https://multimedia.europarl.europa.eu/pl/webstreaming/eu-cop27-un-climate-change-conference-in-sharm-el-sheikh-egypt-ep-delegation-meets-with-ngo-represen_20221114-1600-SPECIAL-OTHER',
'info_dict': {
'id': 'a8428de8-b9cd-6a2e-11e4-3805d9c9ff5c',
'ext': 'mp4',
'release_timestamp': 1668434400,
'release_date': '20221114',
'title': 'md5:d3550280c33cc70e0678652e3d52c028',
},
'params': {
'skip_download': True,
}
}, {
# embed webpage
'url': 'https://webstreaming.europarl.europa.eu/ep/embed/embed.html?event=20220914-0900-PLENARY&language=en&autoplay=true&logo=true',
'info_dict': {
'id': 'bcaa1db4-76ef-7e06-8da7-839bd0ad1dbe',
'ext': 'mp4',
'title': 'Plenary session',
'release_date': '20220914',
'release_timestamp': 1663137900,
},
'params': {
'skip_download': True,
}
}, {
# live webstream
'url': 'https://multimedia.europarl.europa.eu/en/webstreaming/euroscola_20221115-1000-SPECIAL-EUROSCOLA',
'info_dict': {
'ext': 'mp4',
'id': '510eda7f-ba72-161b-7ee7-0e836cd2e715',
'release_timestamp': 1668502800,
'title': 'Euroscola 2022-11-15 19:21',
'release_date': '20221115',
'live_status': 'is_live',
},
'skip': 'not live anymore'
}]
def _real_extract(self, url):
display_id = self._match_id(url)
json_info = self._download_json(
'https://vis-api.vuplay.co.uk/event/external', display_id,
query={
'player_key': 'europarl|718f822c-a48c-4841-9947-c9cb9bb1743c',
'external_id': display_id,
})
formats, subtitles = self._extract_mpd_formats_and_subtitles(json_info['streaming_url'], display_id)
fmts, subs = self._extract_m3u8_formats_and_subtitles(
json_info['streaming_url'].replace('.mpd', '.m3u8'), display_id)
formats.extend(fmts)
self._merge_subtitles(subs, target=subtitles)
return {
'id': json_info['id'],
'title': json_info.get('title'),
'formats': formats,
'subtitles': subtitles,
'release_timestamp': parse_iso8601(json_info.get('published_start')),
'is_live': 'LIVE' in json_info.get('state', '')
}