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

[arte] Add support for multiple formats

Browse source
This commit is contained in:
Philipp Hagemeister 2014-03-24 21:36:26 +01:00
parent 515bbe4b5b
commit 878d11ec29
1 changed files with 17 additions and 21 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

38
youtube_dl/extractor/arte.py
View file

@ -21,7 +21,7 @@ from ..utils import (
class ArteTvIE(InfoExtractor): class ArteTvIE(InfoExtractor):
_VALID_URL = r'(?:http://)?videos\.arte\.tv/(?P<lang>fr|de)/.*-(?P<id>.*?)\.html' _VALID_URL = r'http://videos\.arte\.tv/(?P<lang>fr|de)/.*-(?P<id>.*?)\.html'
IE_NAME = 'arte.tv' IE_NAME = 'arte.tv'
def _real_extract(self, url): def _real_extract(self, url):
@ -31,28 +31,24 @@ class ArteTvIE(InfoExtractor):
ref_xml_url, video_id, note='Downloading metadata') ref_xml_url, video_id, note='Downloading metadata')
config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang) config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
config_xml_url = config_node.attrib['ref'] config_xml_url = config_node.attrib['ref']
config_xml = self._download_webpage( config = self._download_xml(
config_xml_url, video_id, note='Downloading configuration') config_xml_url, video_id, note='Downloading configuration')
video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml)) formats = [{
def _key(m): 'forma_id': q.attrib['quality'],
quality = m.group('quality') 'url': q.text,
if quality == 'hd': 'quality': 2 if q.attrib['quality'] == 'hd' else 1,
return 2 } for q in config.findall('.//quality')]
else: self._sort_formats(formats)
return 1
# We pick the best quality title = config.find('.//name').text
video_urls = sorted(video_urls, key=_key) thumbnail = config.find('.//firstThumbnailUrl').text
video_url = list(video_urls)[-1].group('url') return {
'id': video_id,
title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title') 'title': title,
thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>', 'thumbnail': thumbnail,
config_xml, 'thumbnail') 'url': video_url,
return {'id': video_id, 'ext': 'flv',
'title': title,
'thumbnail': thumbnail,
'url': video_url,
'ext': 'flv',
} }