0
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-18 01:35:12 +00:00
yt-dlp/youtube_dlc/extractor/tv2.py

193 lines
7.3 KiB
Python
Raw Normal View History

2016-10-02 11:39:18 +00:00
# coding: utf-8
2015-05-16 21:01:52 +00:00
from __future__ import unicode_literals
import re
2015-05-16 21:01:52 +00:00
from .common import InfoExtractor
from ..compat import compat_HTTPError
2015-05-16 21:01:52 +00:00
from ..utils import (
determine_ext,
ExtractorError,
2015-05-16 21:01:52 +00:00
int_or_none,
float_or_none,
js_to_json,
2015-05-16 21:01:52 +00:00
parse_iso8601,
remove_end,
strip_or_none,
try_get,
2015-05-16 21:01:52 +00:00
)
class TV2IE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?tv2\.no/v/(?P<id>\d+)'
2015-05-16 21:01:52 +00:00
_TEST = {
'url': 'http://www.tv2.no/v/916509/',
'info_dict': {
'id': '916509',
'ext': 'flv',
2016-01-30 19:42:34 +00:00
'title': 'Se Frode Gryttens hyllest av Steven Gerrard',
2015-05-16 21:01:52 +00:00
'description': 'TV 2 Sportens huspoet tar avskjed med Liverpools kaptein Steven Gerrard.',
'timestamp': 1431715610,
'upload_date': '20150515',
'duration': 156.967,
'view_count': int,
'categories': list,
2016-01-30 19:42:34 +00:00
},
2015-05-16 21:01:52 +00:00
}
_API_DOMAIN = 'sumo.tv2.no'
_PROTOCOLS = ('HDS', 'HLS', 'DASH')
_GEO_COUNTRIES = ['NO']
2015-05-16 21:01:52 +00:00
def _real_extract(self, url):
video_id = self._match_id(url)
api_base = 'http://%s/api/web/asset/%s' % (self._API_DOMAIN, video_id)
2015-05-16 21:01:52 +00:00
formats = []
format_urls = []
for protocol in self._PROTOCOLS:
try:
data = self._download_json(
api_base + '/play.json?protocol=%s&videoFormat=SMIL+ISMUSP' % protocol,
video_id, 'Downloading play JSON')['playback']
except ExtractorError as e:
if isinstance(e.cause, compat_HTTPError) and e.cause.code == 401:
error = self._parse_json(e.cause.read().decode(), video_id)['error']
error_code = error.get('code')
if error_code == 'ASSET_PLAYBACK_INVALID_GEO_LOCATION':
self.raise_geo_restricted(countries=self._GEO_COUNTRIES)
elif error_code == 'SESSION_NOT_AUTHENTICATED':
self.raise_login_required()
raise ExtractorError(error['description'])
raise
items = try_get(data, lambda x: x['items']['item'])
if not items:
continue
if not isinstance(items, list):
items = [items]
for item in items:
if not isinstance(item, dict):
continue
2015-05-16 21:01:52 +00:00
video_url = item.get('url')
if not video_url or video_url in format_urls:
continue
format_id = '%s-%s' % (protocol.lower(), item.get('mediaFormat'))
if not self._is_valid_url(video_url, video_id, format_id):
continue
format_urls.append(video_url)
ext = determine_ext(video_url)
if ext == 'f4m':
formats.extend(self._extract_f4m_formats(
2016-07-29 15:01:34 +00:00
video_url, video_id, f4m_id=format_id, fatal=False))
2015-05-16 21:01:52 +00:00
elif ext == 'm3u8':
2019-11-30 14:50:17 +00:00
if not data.get('drmProtected'):
formats.extend(self._extract_m3u8_formats(
video_url, video_id, 'mp4', entry_protocol='m3u8_native',
m3u8_id=format_id, fatal=False))
elif ext == 'mpd':
formats.extend(self._extract_mpd_formats(
video_url, video_id, format_id, fatal=False))
2015-05-16 21:01:52 +00:00
elif ext == 'ism' or video_url.endswith('.ism/Manifest'):
pass
else:
formats.append({
'url': video_url,
'format_id': format_id,
'tbr': int_or_none(item.get('bitrate')),
'filesize': int_or_none(item.get('fileSize')),
})
2019-11-30 14:50:17 +00:00
if not formats and data.get('drmProtected'):
raise ExtractorError('This video is DRM protected.', expected=True)
2015-05-16 21:01:52 +00:00
self._sort_formats(formats)
asset = self._download_json(
api_base + '.json', video_id,
'Downloading metadata JSON')['asset']
2015-05-16 21:01:52 +00:00
title = asset['title']
thumbnails = [{
'id': thumbnail.get('@type'),
'url': thumbnail.get('url'),
2019-11-30 14:50:17 +00:00
} for _, thumbnail in (asset.get('imageVersions') or {}).items()]
2015-05-16 21:01:52 +00:00
return {
'id': video_id,
'url': video_url,
'title': title,
'description': strip_or_none(asset.get('description')),
2015-05-16 21:01:52 +00:00
'thumbnails': thumbnails,
'timestamp': parse_iso8601(asset.get('createTime')),
'duration': float_or_none(asset.get('accurateDuration') or asset.get('duration')),
'view_count': int_or_none(asset.get('views')),
'categories': asset.get('keywords', '').split(','),
2015-05-16 21:01:52 +00:00
'formats': formats,
}
class TV2ArticleIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?tv2\.no/(?:a|\d{4}/\d{2}/\d{2}(/[^/]+)+)/(?P<id>\d+)'
_TESTS = [{
'url': 'http://www.tv2.no/2015/05/16/nyheter/alesund/krim/pingvin/6930542',
'info_dict': {
'id': '6930542',
'title': 'Russen hetses etter pingvintyveri - innrømmer å ha åpnet luken på buret',
'description': 'De fire siktede nekter fortsatt for å ha stjålet pingvinbabyene, men innrømmer å ha åpnet luken til de små kyllingene.',
},
'playlist_count': 2,
}, {
'url': 'http://www.tv2.no/a/6930542',
'only_matching': True,
}]
def _real_extract(self, url):
playlist_id = self._match_id(url)
webpage = self._download_webpage(url, playlist_id)
# Old embed pattern (looks unused nowadays)
assets = re.findall(r'data-assetid=["\'](\d+)', webpage)
if not assets:
# New embed pattern
for v in re.findall(r'(?s)TV2ContentboxVideo\(({.+?})\)', webpage):
video = self._parse_json(
v, playlist_id, transform_source=js_to_json, fatal=False)
if not video:
continue
asset = video.get('assetId')
if asset:
assets.append(asset)
entries = [
self.url_result('http://www.tv2.no/v/%s' % asset_id, 'TV2')
for asset_id in assets]
title = remove_end(self._og_search_title(webpage), ' - TV2.no')
description = remove_end(self._og_search_description(webpage), ' - TV2.no')
return self.playlist_result(entries, playlist_id, title, description)
class KatsomoIE(TV2IE):
_VALID_URL = r'https?://(?:www\.)?(?:katsomo|mtv)\.fi/(?:#!/)?(?:[^/]+/[0-9a-z-]+-\d+/[0-9a-z-]+-|[^/]+/\d+/[^/]+/)(?P<id>\d+)'
_TEST = {
'url': 'https://www.mtv.fi/sarja/mtv-uutiset-live-33001002003/lahden-pelicans-teki-kovan-ratkaisun-ville-nieminen-pihalle-1181321',
'info_dict': {
'id': '1181321',
'ext': 'mp4',
'title': 'MTV Uutiset Live',
'description': 'Päätöksen teki Pelicansin hallitus.',
'timestamp': 1575116484,
'upload_date': '20191130',
'duration': 37.12,
'view_count': int,
'categories': list,
},
'params': {
# m3u8 download
'skip_download': True,
},
}
_API_DOMAIN = 'api.katsomo.fi'
_PROTOCOLS = ('HLS', 'MPD')
_GEO_COUNTRIES = ['FI']