[tvp] Telewizja Polska: new extractor for tvp.pl, fixes #1719
Thanks-To: mplonski https://github.com/mplonski/linux/blob/master/tvp-dl.py
This commit is contained in:
parent
7f34001d57
commit
5137ebac0b
|
@ -132,6 +132,7 @@ from .tube8 import Tube8IE
|
|||
from .tudou import TudouIE
|
||||
from .tumblr import TumblrIE
|
||||
from .tutv import TutvIE
|
||||
from .tvp import TvpIE
|
||||
from .unistra import UnistraIE
|
||||
from .ustream import UstreamIE, UstreamChannelIE
|
||||
from .vbox7 import Vbox7IE
|
||||
|
|
|
@ -0,0 +1,60 @@
|
|||
# encoding: utf-8
|
||||
import re
|
||||
import json
|
||||
|
||||
from .common import InfoExtractor
|
||||
from ..utils import (
|
||||
determine_ext,
|
||||
ExtractorError,
|
||||
RegexNotFoundError,
|
||||
)
|
||||
|
||||
class TvpIE(InfoExtractor):
|
||||
IE_NAME = u'tvp.pl'
|
||||
_VALID_URL = r'https?://www\.tvp\.pl/.*?wideo/(?P<date>\d+)/(?P<id>\d+)'
|
||||
_INFO_URL = 'http://www.tvp.pl/pub/stat/videofileinfo?video_id=%s'
|
||||
|
||||
|
||||
_TEST = {
|
||||
u'url': u'http://www.tvp.pl/warszawa/magazyny/campusnews/wideo/31102013/12878238',
|
||||
u'file': u'31.10.2013-12878238.wmv',
|
||||
u'info_dict': {
|
||||
u'title': u'31.10.2013',
|
||||
u'description': u'31.10.2013',
|
||||
},
|
||||
}
|
||||
|
||||
def _real_extract(self, url):
|
||||
mobj = re.match(self._VALID_URL, url)
|
||||
video_id = mobj.group('id')
|
||||
webpage = self._download_webpage(url, video_id, "Downloading video webpage")
|
||||
json_params = self._download_webpage(self._INFO_URL % video_id, video_id, "Downloading video metadata")
|
||||
|
||||
try:
|
||||
params = json.loads(json_params)
|
||||
except:
|
||||
raise ExtractorError(u'Invalid JSON')
|
||||
|
||||
self.report_extraction(video_id)
|
||||
try:
|
||||
video_url = params['video_url']
|
||||
except KeyError:
|
||||
raise ExtractorError('Missing JSON parameter: ' + sys.exc_info()[1])
|
||||
|
||||
try:
|
||||
title = self._og_search_title(webpage)
|
||||
except RegexNotFoundError:
|
||||
title = video_id
|
||||
info = {
|
||||
'id': video_id,
|
||||
'title': title,
|
||||
'ext': 'wmv',
|
||||
'url': video_url,
|
||||
}
|
||||
try:
|
||||
info['description'] = self._og_search_description(webpage)
|
||||
info['thumbnail'] = self._og_search_thumbnail(webpage)
|
||||
except RegexNotFoundError:
|
||||
pass
|
||||
|
||||
return info
|
Loading…
Reference in New Issue