1
0
mirror of https://codeberg.org/polarisfm/youtube-dl synced 2024-11-09 03:34:32 +01:00
youtube-dl/youtube_dl/extractor/dbtv.py
2014-12-21 16:57:07 +01:00

76 lines
2.6 KiB
Python

# coding: utf-8
from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..compat import compat_str
from ..utils import (
float_or_none,
int_or_none,
clean_html,
)
class DBTVIE(InfoExtractor):
_VALID_URL = r'http://dbtv\.no/(?P<id>[0-9]+)#(?P<display_id>.+)'
_TEST = {
'url': 'http://dbtv.no/3649835190001#Skulle_teste_ut_fornøyelsespark,_men_kollegaen_var_bare_opptatt_av_bikinikroppen',
'md5': 'b89953ed25dacb6edb3ef6c6f430f8bc',
'info_dict': {
'id': '33100',
'display_id': 'Skulle_teste_ut_fornøyelsespark,_men_kollegaen_var_bare_opptatt_av_bikinikroppen',
'ext': 'mp4',
'title': 'Skulle teste ut fornøyelsespark, men kollegaen var bare opptatt av bikinikroppen',
'description': 'md5:1504a54606c4dde3e4e61fc97aa857e0',
'thumbnail': 're:https?://.*\.jpg$',
'timestamp': 1404039863.438,
'upload_date': '20140629',
'duration': 69.544,
'view_count': int,
'categories': list,
}
}
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('id')
display_id = mobj.group('display_id')
data = self._download_json(
'http://api.dbtv.no/discovery/%s' % video_id, display_id)
video = data['playlist'][0]
formats = [{
'url': f['URL'],
'vcodec': f.get('container'),
'width': int_or_none(f.get('width')),
'height': int_or_none(f.get('height')),
'vbr': float_or_none(f.get('rate'), 1000),
'filesize': int_or_none(f.get('size')),
} for f in video['renditions'] if 'URL' in f]
if not formats:
for url_key, format_id in [('URL', 'mp4'), ('HLSURL', 'hls')]:
if url_key in video:
formats.append({
'url': video[url_key],
'format_id': format_id,
})
self._sort_formats(formats)
return {
'id': compat_str(video['id']),
'display_id': display_id,
'title': video['title'],
'description': clean_html(video['desc']),
'thumbnail': video.get('splash') or video.get('thumb'),
'timestamp': float_or_none(video.get('publishedAt'), 1000),
'duration': float_or_none(video.get('length'), 1000),
'view_count': int_or_none(video.get('views')),
'categories': video.get('tags'),
'formats': formats,
}