1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-12-05 08:27:43 +00:00
youtube-dl/youtube_dl/extractor/internetvideoarchive.py

65 lines
2.4 KiB
Python
Raw Normal View History

2014-09-29 10:23:52 +00:00
from __future__ import unicode_literals
2019-10-28 18:45:42 +00:00
import json
import re
from .common import InfoExtractor
from ..compat import (
compat_parse_qs,
compat_urlparse,
)
class InternetVideoArchiveIE(InfoExtractor):
_VALID_URL = r'https?://video\.internetvideoarchive\.net/(?:player|flash/players)/.*?\?.*?publishedid.*?'
_TEST = {
'url': 'http://video.internetvideoarchive.net/player/6/configuration.ashx?customerid=69249&publishedid=194487&reporttag=vdbetatitle&playerid=641&autolist=0&domain=www.videodetective.com&maxrate=high&minrate=low&socialplayer=false',
2014-09-29 10:23:52 +00:00
'info_dict': {
'id': '194487',
2014-09-29 10:23:52 +00:00
'ext': 'mp4',
2019-10-28 18:45:42 +00:00
'title': 'Kick-Ass 2',
'description': 'md5:c189d5b7280400630a1d3dd17eaa8d8a',
},
'params': {
# m3u8 download
'skip_download': True,
},
}
@staticmethod
def _build_json_url(query):
return 'http://video.internetvideoarchive.net/player/6/configuration.ashx?' + query
def _real_extract(self, url):
2019-10-28 18:45:42 +00:00
query = compat_parse_qs(compat_urlparse.urlparse(url).query)
video_id = query['publishedid'][0]
data = self._download_json(
'https://video.internetvideoarchive.net/videojs7/videojs7.ivasettings.ashx',
video_id, data=json.dumps({
'customerid': query['customerid'][0],
'publishedid': video_id,
}).encode())
title = data['Title']
formats = self._extract_m3u8_formats(
data['VideoUrl'], video_id, 'mp4',
'm3u8_native', m3u8_id='hls', fatal=False)
file_url = formats[0]['url']
if '.ism/' in file_url:
replace_url = lambda x: re.sub(r'\.ism/[^?]+', '.ism/' + x, file_url)
formats.extend(self._extract_f4m_formats(
replace_url('.f4m'), video_id, f4m_id='hds', fatal=False))
formats.extend(self._extract_mpd_formats(
replace_url('.mpd'), video_id, mpd_id='dash', fatal=False))
formats.extend(self._extract_ism_formats(
replace_url('Manifest'), video_id, ism_id='mss', fatal=False))
self._sort_formats(formats)
return {
'id': video_id,
'title': title,
'formats': formats,
2019-10-28 18:45:42 +00:00
'thumbnail': data.get('PosterUrl'),
'description': data.get('Description'),
}