0
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-01 23:12:40 +00:00
yt-dlp/youtube_dl/extractor/radiobremen.py
2015-01-05 18:14:09 +01:00

57 lines
2.3 KiB
Python

# -*- coding: utf-8 -*-
from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..utils import parse_duration
class RadioBremenIE(InfoExtractor):
_VALID_URL = r'http?://(?:www\.)?radiobremen\.de/mediathek/(?:index\.html)?\?id=(?P<id>[0-9]+)'
IE_NAME = 'radiobremen'
_TEST = {
'url': 'http://www.radiobremen.de/mediathek/index.html?id=114720',
'info_dict': {
'id': '114720',
'ext': 'mp4',
'duration': 1685,
'width': 512,
'title': 'buten un binnen vom 22. Dezember',
'thumbnail': 're:https?://.*\.jpg$',
'description': 'Unter anderem mit diesen Themen: 45 Flüchtlinge sind in Worpswede angekommen +++ Freies Internet für alle: Bremer arbeiten an einem flächendeckenden W-Lan-Netzwerk +++ Aktivisten kämpfen für das Unibad +++ So war das Wetter 2014 +++',
},
}
def _real_extract(self, url):
video_id = self._match_id(url)
meta_url = "http://www.radiobremen.de/apps/php/mediathek/metadaten.php?id=%s" % video_id
meta_doc = self._download_webpage(meta_url, video_id, 'Downloading metadata')
title = self._html_search_regex("<h1.*>(?P<title>.+)</h1>", meta_doc, "title")
description = self._html_search_regex("<p>(?P<description>.*)</p>", meta_doc, "description")
duration = parse_duration(
self._html_search_regex("L&auml;nge:</td>\s+<td>(?P<duration>[0-9]+:[0-9]+)</td>", meta_doc, "duration"))
page_doc = self._download_webpage(url, video_id, 'Downloading video information')
pattern = "ardformatplayerclassic\(\'playerbereich\',\'(?P<width>[0-9]+)\',\'.*\',\'(?P<video_id>[0-9]+)\',\'(?P<secret>[0-9]+)\',\'(?P<thumbnail>.+)\',\'\'\)"
mobj = re.search(pattern, page_doc)
video_url = (
"http://dl-ondemand.radiobremen.de/mediabase/%s/%s_%s_%s.mp4" %
(video_id, video_id, mobj.group("secret"), mobj.group('width')))
formats = [{
'url': video_url,
'ext': 'mp4',
'width': int(mobj.group("width")),
}]
return {
'id': video_id,
'title': title,
'description': description,
'duration': duration,
'formats': formats,
'thumbnail': mobj.group('thumbnail'),
}