1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-11-24 03:01:53 +00:00
youtube-dl/youtube_dl/extractor/ultimedia.py

104 lines
3.4 KiB
Python
Raw Normal View History

# coding: utf-8
from __future__ import unicode_literals
2015-03-18 19:06:39 +00:00
import re
from .common import InfoExtractor
from ..utils import (
ExtractorError,
qualities,
unified_strdate,
clean_html,
)
class UltimediaIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?ultimedia\.com/default/index/video[^/]+/id/(?P<id>[\d+a-z]+)'
_TESTS = [{
# news
'url': 'https://www.ultimedia.com/default/index/videogeneric/id/s8uk0r',
'md5': '276a0e49de58c7e85d32b057837952a2',
'info_dict': {
'id': 's8uk0r',
'ext': 'mp4',
'title': 'Loi sur la fin de vie: le texte prévoit un renforcement des directives anticipées',
'description': 'md5:3e5c8fd65791487333dda5db8aed32af',
'thumbnail': 're:^https?://.*\.jpg',
'upload_date': '20150317',
},
}, {
# music
'url': 'https://www.ultimedia.com/default/index/videomusic/id/xvpfp8',
'md5': '2ea3513813cf230605c7e2ffe7eca61c',
'info_dict': {
'id': 'xvpfp8',
'ext': 'mp4',
'title': "Two - C'est la vie (Clip)",
'description': 'Two',
'thumbnail': 're:^https?://.*\.jpg',
'upload_date': '20150224',
},
}]
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
deliver_url = self._search_regex(
r'<iframe[^>]+src="(https?://(?:www\.)?ultimedia\.com/deliver/[^"]+)"',
webpage, 'deliver URL')
deliver_page = self._download_webpage(
deliver_url, video_id, 'Downloading iframe page')
if '>This video is currently not available' in deliver_page:
raise ExtractorError(
'Video %s is currently not available' % video_id, expected=True)
player = self._parse_json(
self._search_regex(
r"jwplayer\('player(?:_temp)?'\)\.setup\(({.+?})\)\.on", deliver_page, 'player'),
video_id)
quality = qualities(['flash', 'html5'])
2015-03-18 19:06:39 +00:00
formats = []
for mode in player['modes']:
video_url = mode.get('config', {}).get('file')
if not video_url:
continue
if re.match(r'https?://www\.youtube\.com/.+?', video_url):
return self.url_result(video_url, 'Youtube')
formats.append({
'url': video_url,
'format_id': mode.get('type'),
'quality': quality(mode.get('type')),
})
self._sort_formats(formats)
thumbnail = player.get('image')
title = clean_html((
self._html_search_regex(
r'(?s)<div\s+id="catArticle">.+?</div>(.+?)</h1>',
2015-03-28 22:35:55 +00:00
webpage, 'title', default=None) or
self._search_regex(
r"var\s+nameVideo\s*=\s*'([^']+)'",
deliver_page, 'title')))
description = clean_html(self._html_search_regex(
r'(?s)<span>Description</span>(.+?)</p>', webpage,
'description', fatal=False))
upload_date = unified_strdate(self._search_regex(
r'Ajouté le\s*<span>([^<]+)', webpage,
'upload date', fatal=False))
return {
'id': video_id,
'title': title,
'description': description,
'thumbnail': thumbnail,
'upload_date': upload_date,
'formats': formats,
}