Merge remote-tracking branch 'David-Development/rtl2.py'

pull/8/head
Philipp Hagemeister 10 years ago
commit d4f64cabf4

@ -104,6 +104,7 @@ class RtmpFD(FileDownloader):
live = info_dict.get('rtmp_live', False)
conn = info_dict.get('rtmp_conn', None)
protocol = info_dict.get('rtmp_protocol', None)
no_resume = info_dict.get('no_resume', False)
self.report_destination(filename)
tmpfilename = self.temp_name(filename)
@ -141,7 +142,10 @@ class RtmpFD(FileDownloader):
basic_args += ['--conn', conn]
if protocol is not None:
basic_args += ['--protocol', protocol]
args = basic_args + [[], ['--resume', '--skip', '1']][not live and self.params.get('continuedl', False)]
if not no_resume:
basic_args += ['--resume']
args = basic_args + [[], ['--skip', '1']][not live and self.params.get('continuedl', False)]
if sys.platform == 'win32' and sys.version_info < (3, 0):
# Windows subprocess module does not actually support Unicode

@ -349,6 +349,7 @@ from .rtbf import RTBFIE
from .rte import RteIE
from .rtlnl import RtlXlIE
from .rtlnow import RTLnowIE
from .rtl2 import RTL2IE
from .rtp import RTPIE
from .rts import RTSIE
from .rtve import RTVEALaCartaIE, RTVELiveIE

@ -0,0 +1,100 @@
# encoding: utf-8
from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..utils import (
ExtractorError,
clean_html,
unified_strdate,
int_or_none,
)
class RTL2IE(InfoExtractor):
"""Information Extractor for RTL2"""
_VALID_URL = r'http?://(?P<url>(?P<domain>(www\.)?rtl2\.de)/.*/(?P<video_id>.*))/'
_TESTS = [{
'url': 'http://www.rtl2.de/sendung/grip-das-motormagazin/folge/folge-203-0',
'info_dict': {
'id': 'folge-203-0',
'ext': 'f4v',
'title': 'GRIP sucht den Sommerkönig',
'description' : 'Matthias, Det und Helge treten gegeneinander an.'
},
'params': {
# rtmp download
'skip_download': True,
},
},
{
'url': 'http://www.rtl2.de/sendung/koeln-50667/video/5512-anna/21040-anna-erwischt-alex/',
'info_dict': {
'id': '21040-anna-erwischt-alex',
'ext': 'mp4',
'title': 'Anna erwischt Alex!',
'description' : 'Anna ist Alex\' Tochter bei Köln 50667.'
},
'params': {
# rtmp download
'skip_download': True,
},
},
]
def _real_extract(self, url):
#Some rtl2 urls have no slash at the end, so append it.
if not url.endswith("/"):
url += '/'
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('video_id')
webpage = self._download_webpage(url, video_id)
vico_id = self._html_search_regex(r'vico_id\s*:\s*([0-9]+)', webpage, 'vico_id not found');
vivi_id = self._html_search_regex(r'vivi_id\s*:\s*([0-9]+)', webpage, 'vivi_id not found');
info_url = 'http://www.rtl2.de/video/php/get_video.php?vico_id=' + vico_id + '&vivi_id=' + vivi_id
webpage = self._download_webpage(info_url, '')
video_info = self._download_json(info_url, video_id)
download_url = video_info["video"]["streamurl"]
title = video_info["video"]["titel"]
description = video_info["video"]["beschreibung"]
thumbnail = video_info["video"]["image"]
download_url = download_url.replace("\\", "")
stream_url = 'mp4:' + self._html_search_regex(r'ondemand/(.*)', download_url, '%s')
#Debug output
#print('URL: ' + url)
#print('DL URL: ' + download_url)
#print('Stream URL: ' + stream_url)
#print('Title: ' + title)
#print('Description: '+ description)
#print('Video ID: ' + video_id)
formats = [{
'url' : download_url,
#'app': 'ondemand?_fcs_vhost=cp108781.edgefcs.net',
'play_path': stream_url,
'player_url': 'http://www.rtl2.de/flashplayer/vipo_player.swf',
'page_url': url,
'flash_version' : "LNX 11,2,202,429",
'rtmp_conn' : ["S:connect", "O:1", "NS:pageUrl:" + url, "NB:fpad:0", "NN:videoFunction:1", "O:0"],
'no_resume' : True,
}]
return {
'id': video_id,
'title': title,
'thumbnail' : thumbnail,
'description' : description,
'formats': formats,
}
Loading…
Cancel
Save