[folketinget] Add extractor (Fixes #4262)

master
Philipp Hagemeister 9 years ago
parent 71069d2157
commit 6127693ed9

@ -115,6 +115,7 @@ from .fktv import (
FKTVPosteckeIE,
)
from .flickr import FlickrIE
from .folketinget import FolketingetIE
from .fourtube import FourTubeIE
from .franceculture import FranceCultureIE
from .franceinter import FranceInterIE

@ -0,0 +1,75 @@
# -*- coding: utf-8 -*-
from __future__ import unicode_literals
from .common import InfoExtractor
from ..compat import compat_parse_qs
from ..utils import (
int_or_none,
parse_duration,
parse_iso8601,
xpath_text,
)
class FolketingetIE(InfoExtractor):
IE_DESC = 'Folketinget (ft.dk; Danish parliament)'
_VALID_URL = r'https?://(?:www\.)?ft\.dk/webtv/video/[^?#]*?\.(?P<id>[0-9]+)\.aspx'
_TEST = {
'url': 'http://www.ft.dk/webtv/video/20141/eru/td.1165642.aspx?as=1#player',
'info_dict': {
'id': '1165642',
'ext': 'mp4',
'title': 'Åbent samråd i Erhvervsudvalget',
'description': 'Åbent samråd med erhvervs- og vækstministeren om regeringens politik på teleområdet',
'view_count': int,
'width': 768,
'height': 432,
'tbr': 928000,
'timestamp': 1416493800,
'upload_date': '20141120',
'duration': 3960,
},
'params': {
'skip_download': 'rtmpdump required',
}
}
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
title = self._og_search_title(webpage)
description = self._html_search_regex(
r'(?s)<div class="video-item-agenda"[^>]*>(.*?)<',
webpage, 'description', fatal=False)
player_params = compat_parse_qs(self._search_regex(
r'<embed src="http://ft\.arkena\.tv/flash/ftplayer\.swf\?([^"]+)"',
webpage, 'player params'))
xml_url = player_params['xml'][0]
doc = self._download_xml(xml_url, video_id)
timestamp = parse_iso8601(xpath_text(doc, './/date'))
duration = parse_duration(xpath_text(doc, './/duration'))
width = int_or_none(xpath_text(doc, './/width'))
height = int_or_none(xpath_text(doc, './/height'))
view_count = int_or_none(xpath_text(doc, './/views'))
formats = [{
'format_id': n.attrib['bitrate'],
'url': xpath_text(n, './url', fatal=True),
'tbr': int_or_none(n.attrib['bitrate']),
} for n in doc.findall('.//streams/stream')]
self._sort_formats(formats)
return {
'id': video_id,
'title': title,
'formats': formats,
'description': description,
'timestamp': timestamp,
'width': width,
'height': height,
'duration': duration,
'view_count': view_count,
}
Loading…
Cancel
Save