kumi/yt-dlp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

[la7] extract more info from a kaltura custom server

Browse source
This commit is contained in:
Remita Amine 2016-07-04 17:59:03 +01:00
parent 81953d1ae5
commit dafafe7cf1
1 changed files with 10 additions and 19 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

29
youtube_dl/extractor/la7.py
View file

@ -3,8 +3,8 @@ from __future__ import unicode_literals
from .common import InfoExtractor from .common import InfoExtractor
from ..utils import ( from ..utils import (
determine_ext,
js_to_json, js_to_json,
smuggle_url,
) )
@ -18,13 +18,16 @@ class LA7IE(InfoExtractor):
_TESTS = [{ _TESTS = [{
# 'src' is a plain URL # 'src' is a plain URL
'url': 'http://www.la7.it/crozza/video/inccool8-02-10-2015-163722', 'url': 'http://www.la7.it/crozza/video/inccool8-02-10-2015-163722',
'md5': '6054674766e7988d3e02f2148ff92180', 'md5': '8b613ffc0c4bf9b9e377169fc19c214c',
'info_dict': { 'info_dict': {
'id': 'inccool8-02-10-2015-163722', 'id': 'inccool8-02-10-2015-163722',
'ext': 'mp4', 'ext': 'mp4',
'title': 'Inc.Cool8', 'title': 'Inc.Cool8',
'description': 'Benvenuti nell\'incredibile mondo della INC. COOL. 8. dove “INC.” sta per “Incorporated” “COOL” sta per “fashion” ed Eight sta per il gesto atletico', 'description': 'Benvenuti nell\'incredibile mondo della INC. COOL. 8. dove “INC.” sta per “Incorporated” “COOL” sta per “fashion” ed Eight sta per il gesto atletico',
'thumbnail': 're:^https?://.*', 'thumbnail': 're:^https?://.*',
'uploader_id': 'kdla7pillole@iltrovatore.it',
'timestamp': 1443814869,
'upload_date': '20151002',
}, },
}, { }, {
# 'src' is a dictionary # 'src' is a dictionary
@ -49,26 +52,14 @@ class LA7IE(InfoExtractor):
self._search_regex(r'videoLa7\(({[^;]+})\);', webpage, 'player data'), self._search_regex(r'videoLa7\(({[^;]+})\);', webpage, 'player data'),
video_id, transform_source=js_to_json) video_id, transform_source=js_to_json)
source = player_data['src']
source_urls = source.values() if isinstance(source, dict) else [source]
formats = []
for source_url in source_urls:
ext = determine_ext(source_url)
if ext == 'm3u8':
formats.extend(self._extract_m3u8_formats(
source_url, video_id, ext='mp4',
entry_protocol='m3u8_native', m3u8_id='hls'))
else:
formats.append({
'url': source_url,
})
self._sort_formats(formats)
return { return {
'_type': 'url_transparent',
'url': smuggle_url('kaltura:103:%s' % player_data['vid'], {
'service_url': 'http://kdam.iltrovatore.it',
}),
'id': video_id, 'id': video_id,
'title': player_data['title'], 'title': player_data['title'],
'description': self._og_search_description(webpage, default=None), 'description': self._og_search_description(webpage, default=None),
'thumbnail': player_data.get('poster'), 'thumbnail': player_data.get('poster'),
'formats': formats, 'ie_key': 'Kaltura',
} }