mirror of
https://github.com/yt-dlp/yt-dlp.git
synced 2024-11-27 03:03:01 +00:00
Merge branch 'mcepl-tvnoe'
This commit is contained in:
commit
abf2c79f95
3 changed files with 56 additions and 0 deletions
|
@ -1,3 +1,9 @@
|
|||
version <unreleased>
|
||||
|
||||
Extractors
|
||||
+ [tvnoe] New extractor (#10524)
|
||||
|
||||
|
||||
version 2016.09.04.1
|
||||
|
||||
Core
|
||||
|
|
|
@ -916,6 +916,7 @@
|
|||
)
|
||||
from .tvigle import TvigleIE
|
||||
from .tvland import TVLandIE
|
||||
from .tvnoe import TVNoeIE
|
||||
from .tvp import (
|
||||
TVPEmbedIE,
|
||||
TVPIE,
|
||||
|
|
49
youtube_dl/extractor/tvnoe.py
Normal file
49
youtube_dl/extractor/tvnoe.py
Normal file
|
@ -0,0 +1,49 @@
|
|||
# coding: utf-8
|
||||
from __future__ import unicode_literals
|
||||
|
||||
from .jwplatform import JWPlatformBaseIE
|
||||
from ..utils import (
|
||||
clean_html,
|
||||
get_element_by_class,
|
||||
js_to_json,
|
||||
)
|
||||
|
||||
|
||||
class TVNoeIE(JWPlatformBaseIE):
|
||||
_VALID_URL = r'https?://(www\.)?tvnoe\.cz/video/(?P<id>[0-9]+)'
|
||||
_TEST = {
|
||||
'url': 'http://www.tvnoe.cz/video/10362',
|
||||
'md5': 'aee983f279aab96ec45ab6e2abb3c2ca',
|
||||
'info_dict': {
|
||||
'id': '10362',
|
||||
'ext': 'mp4',
|
||||
'series': 'Noční univerzita',
|
||||
'title': 'prof. Tomáš Halík, Th.D. - Návrat náboženství a střet civilizací',
|
||||
'description': 'md5:f337bae384e1a531a52c55ebc50fff41',
|
||||
}
|
||||
}
|
||||
|
||||
def _real_extract(self, url):
|
||||
video_id = self._match_id(url)
|
||||
webpage = self._download_webpage(url, video_id)
|
||||
|
||||
iframe_url = self._search_regex(
|
||||
r'<iframe[^>]+src="([^"]+)"', webpage, 'iframe URL')
|
||||
|
||||
ifs_page = self._download_webpage(iframe_url, video_id)
|
||||
jwplayer_data = self._parse_json(
|
||||
self._find_jwplayer_data(ifs_page),
|
||||
video_id, transform_source=js_to_json)
|
||||
info_dict = self._parse_jwplayer_data(
|
||||
jwplayer_data, video_id, require_title=False, base_url=iframe_url)
|
||||
|
||||
info_dict.update({
|
||||
'id': video_id,
|
||||
'title': clean_html(get_element_by_class(
|
||||
'field-name-field-podnazev', webpage)),
|
||||
'description': clean_html(get_element_by_class(
|
||||
'field-name-body', webpage)),
|
||||
'series': clean_html(get_element_by_class('title', webpage))
|
||||
})
|
||||
|
||||
return info_dict
|
Loading…
Reference in a new issue