R-yan/l1ving_youtube-dl
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

[Heise] Add support for embedded youtube videos

Browse Source
This commit is contained in:
Kay B 2017-09-03 14:37:07 +02:00
parent 5113b69124
commit c636e0fd25
1 changed files with 13 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
youtube_dl/extractor/heise.py
View File

@ -9,6 +9,8 @@ from ..utils import (
xpath_text, xpath_text,
) )
import re
class HeiseIE(InfoExtractor): class HeiseIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?heise\.de/(?:[^/]+/)+[^/]+-(?P<id>[0-9]+)\.html' _VALID_URL = r'https?://(?:www\.)?heise\.de/(?:[^/]+/)+[^/]+-(?P<id>[0-9]+)\.html'
@ -40,6 +42,17 @@ class HeiseIE(InfoExtractor):
video_id = self._match_id(url) video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id) webpage = self._download_webpage(url, video_id)
title = self._html_search_meta('fulltitle', webpage, default=None)
if not title or title == "c't":
title = self._search_regex(
r'<div[^>]+class="videoplayerjw"[^>]+data-title="([^"]+)"',
webpage, 'title')
yt_videos = re.findall(
r'<iframe[^>]+class="yt_video"[^>]+src="//([^"]+)', webpage)
if yt_videos:
return self.playlist_from_matches(yt_videos, title, 'Youtube')
container_id = self._search_regex( container_id = self._search_regex(
r'<div class="videoplayerjw"[^>]+data-container="([0-9]+)"', r'<div class="videoplayerjw"[^>]+data-container="([0-9]+)"',
webpage, 'container ID') webpage, 'container ID')
@ -47,12 +60,6 @@ class HeiseIE(InfoExtractor):
r'<div class="videoplayerjw"[^>]+data-sequenz="([0-9]+)"', r'<div class="videoplayerjw"[^>]+data-sequenz="([0-9]+)"',
webpage, 'sequenz ID') webpage, 'sequenz ID')
title = self._html_search_meta('fulltitle', webpage, default=None)
if not title or title == "c't":
title = self._search_regex(
r'<div[^>]+class="videoplayerjw"[^>]+data-title="([^"]+)"',
webpage, 'title')
doc = self._download_xml( doc = self._download_xml(
'http://www.heise.de/videout/feed', video_id, query={ 'http://www.heise.de/videout/feed', video_id, query={
'container': container_id, 'container': container_id,