w/youtube-dl
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

[firstpost] Fix title extraction and add description

Browse source
This commit is contained in:
Sergey M․ 2014-06-22 01:20:40 +07:00
parent 18061bbab0
commit d2824416aa
1 changed files with 6 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
youtube_dl/extractor/firstpost.py
View file

@ -15,6 +15,7 @@ class FirstpostIE(InfoExtractor):
'id': '1025403', 'id': '1025403',
'ext': 'mp4', 'ext': 'mp4',
'title': 'India to launch indigenous aircraft carrier INS Vikrant today', 'title': 'India to launch indigenous aircraft carrier INS Vikrant today',
'description': 'md5:feef3041cb09724e0bdc02843348f5f4',
} }
} }
@ -22,13 +23,16 @@ class FirstpostIE(InfoExtractor):
mobj = re.match(self._VALID_URL, url) mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('id') video_id = mobj.group('id')
page = self._download_webpage(url, video_id)
title = self._html_search_meta('twitter:title', page, 'title')
description = self._html_search_meta('twitter:description', page, 'title')
data = self._download_xml( data = self._download_xml(
'http://www.firstpost.com/getvideoxml-%s.xml' % video_id, video_id, 'http://www.firstpost.com/getvideoxml-%s.xml' % video_id, video_id,
'Downloading video XML') 'Downloading video XML')
item = data.find('./playlist/item') item = data.find('./playlist/item')
thumbnail = item.find('./image').text thumbnail = item.find('./image').text
title = item.find('./title').text
formats = [ formats = [
{ {
@ -42,6 +46,7 @@ class FirstpostIE(InfoExtractor):
return { return {
'id': video_id, 'id': video_id,
'title': title, 'title': title,
'description': description,
'thumbnail': thumbnail, 'thumbnail': thumbnail,
'formats': formats, 'formats': formats,
} }