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

[eitb] fix info extraction

Browse source
This commit is contained in:
remitamine 2015-10-31 04:02:49 +01:00 committed by Sergey M․
parent 30eecc6a04
commit 80dcee5cd5
1 changed files with 44 additions and 21 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

65
youtube_dl/extractor/eitb.py
View file

@ -1,39 +1,62 @@
# encoding: utf-8 # encoding: utf-8
from __future__ import unicode_literals from __future__ import unicode_literals
import re
from .common import InfoExtractor from .common import InfoExtractor
from .brightcove import BrightcoveIE from ..compat import compat_urllib_request
from ..utils import ExtractorError from ..utils import (
int_or_none,
unified_strdate,
)
class EitbIE(InfoExtractor): class EitbIE(InfoExtractor):
IE_NAME = 'eitb.tv' IE_NAME = 'eitb.tv'
_VALID_URL = r'https?://www\.eitb\.tv/(eu/bideoa|es/video)/[^/]+/(?P<playlist_id>\d+)/(?P<chapter_id>\d+)' _VALID_URL = r'https?://www\.eitb\.tv/(eu/bideoa|es/video)/[^/]+/\d+/(?P<id>\d+)'
_TEST = { _TEST = {
'add_ie': ['Brightcove'], 'url': 'http://www.eitb.tv/es/video/60-minutos-60-minutos-2013-2014/4104995148001/4090227752001/lasa-y-zabala-30-anos/',
'url': 'http://www.eitb.tv/es/video/60-minutos-60-minutos-2013-2014/2677100210001/2743577154001/lasa-y-zabala-30-anos/',
'md5': 'edf4436247185adee3ea18ce64c47998', 'md5': 'edf4436247185adee3ea18ce64c47998',
'info_dict': { 'info_dict': {
'id': '2743577154001', 'id': '4090227752001',
'ext': 'mp4', 'ext': 'mp4',
'title': '60 minutos (Lasa y Zabala, 30 años)', 'title': '60 minutos (Lasa y Zabala, 30 años)',
# All videos from eitb has this description in the brightcove info 'description': '',
'description': '.', 'duration': 3996760,
'uploader': 'Euskal Telebista', 'upload_date': '20131014',
}, },
} }
def _real_extract(self, url): def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url) video_id = self._match_id(url)
chapter_id = mobj.group('chapter_id') video_data = self._download_json('http://mam.eitb.eus/mam/REST/ServiceMultiweb/Video/MULTIWEBTV/%s/' % video_id, video_id)['web_media'][0]
webpage = self._download_webpage(url, chapter_id)
bc_url = BrightcoveIE._extract_brightcove_url(webpage) formats = []
if bc_url is None: for rendition in video_data['RENDITIONS']:
raise ExtractorError('Could not extract the Brightcove url') formats.append({
# The BrightcoveExperience object doesn't contain the video id, we set 'url': rendition['PMD_URL'],
# it manually 'width': int_or_none(rendition.get('FRAME_WIDTH')),
bc_url += '&%40videoPlayer={0}'.format(chapter_id) 'height': int_or_none(rendition.get('FRAME_HEIGHT')),
return self.url_result(bc_url, BrightcoveIE.ie_key()) 'tbr': int_or_none(rendition.get('ENCODING_RATE')),
})
# TODO: parse f4m manifest
request = compat_urllib_request.Request(
'http://mam.eitb.eus/mam/REST/ServiceMultiweb/DomainRestrictedSecurity/TokenAuth/',
headers={'Referer': url})
token_data = self._download_json(request, video_id, fatal=False)
if token_data:
m3u8_formats = self._extract_m3u8_formats('%s?hdnts=%s' % (video_data['HLS_SURL'], token_data['token']), video_id, m3u8_id='hls', fatal=False)
if m3u8_formats:
formats.extend(m3u8_formats)
self._sort_formats(formats)
return {
'id': video_id,
'title': video_data['NAME_ES'],
'description': video_data.get('SHORT_DESC_ES'),
'thumbnail': video_data.get('STILL_URL'),
'duration': int_or_none(video_data.get('LENGTH')),
'upload_date': unified_strdate(video_data.get('BROADCST_DATE')),
'formats': formats,
}