1 from __future__ import unicode_literals
4 import xml.etree.ElementTree
6 from .common import InfoExtractor
12 class GDCVaultIE(InfoExtractor):
13 _VALID_URL = r'https?://(?:www\.)?gdcvault\.com/play/(?P<id>\d+)/(?P<name>(\w|-)+)'
16 'url': 'http://www.gdcvault.com/play/1019721/Doki-Doki-Universe-Sweet-Simple',
17 'md5': '7ce8388f544c88b7ac11c7ab1b593704',
21 'title': 'Doki-Doki Universe: Sweet, Simple and Genuine (GDC Next 10)'
25 'url': 'http://www.gdcvault.com/play/1015683/Embracing-the-Dark-Art-of',
29 'title': 'Embracing the Dark Art of Mathematical Modeling in AI'
32 'skip_download': True, # Requires rtmpdump
37 def _parse_mp4(self, xml_description):
39 mp4_video = xml_description.find('./metadata/mp4video')
43 mobj = re.match(r'(?P<root>https?://.*?/).*', mp4_video.text)
44 video_root = mobj.group('root')
45 formats = xml_description.findall('./metadata/MBRVideos/MBRVideo')
46 for format in formats:
47 mobj = re.match(r'mp4\:(?P<path>.*)', format.find('streamName').text)
48 url = video_root + mobj.group('path')
49 vbr = format.find('bitrate').text
50 video_formats.append({
56 def _parse_flv(self, xml_description):
58 akami_url = xml_description.find('./metadata/akamaiHost').text
59 slide_video_path = xml_description.find('./metadata/slideVideo').text
60 video_formats.append({
61 'url': 'rtmp://' + akami_url + '/' + slide_video_path,
62 'format_note': 'slide deck video',
65 'format_id': 'slides',
67 speaker_video_path = xml_description.find('./metadata/speakerVideo').text
68 video_formats.append({
69 'url': 'rtmp://' + akami_url + '/' + speaker_video_path,
70 'format_note': 'speaker video',
73 'format_id': 'speaker',
77 def _login(self, webpage_url, video_id):
78 (username, password) = self._get_login_info()
79 if username is None or password is None:
80 self.report_warning('It looks like ' + webpage_url + ' requires a login. Try specifying a username and password and try again.')
83 mobj = re.match(r'(?P<root_url>https?://.*?/).*', webpage_url)
84 login_url = mobj.group('root_url') + 'api/login.php'
85 logout_url = mobj.group('root_url') + 'logout'
92 request = compat_urllib_request.Request(login_url, compat_urllib_parse.urlencode(login_form))
93 request.add_header('Content-Type', 'application/x-www-form-urlencoded')
94 login_content = self._download_webpage(request, video_id, 'Logging in')
95 start_page = self._download_webpage(webpage_url, video_id, 'Getting authenticated video page')
96 logout_content = self._download_webpage(logout_url, video_id, 'Logging out')
100 def _real_extract(self, url):
101 mobj = re.match(self._VALID_URL, url)
103 video_id = mobj.group('id')
104 webpage_url = 'http://www.gdcvault.com/play/' + video_id
105 start_page = self._download_webpage(webpage_url, video_id)
107 xml_root = self._html_search_regex(r'<iframe src="(?P<xml_root>.*?)player.html.*?".*?</iframe>', start_page, 'xml root', None, False)
110 # Probably need to authenticate
111 start_page = self._login(webpage_url, video_id)
112 if start_page is None:
113 self.report_warning('Could not login.')
115 # Grab the url from the authenticated page
116 xml_root = self._html_search_regex(r'<iframe src="(?P<xml_root>.*?)player.html.*?".*?</iframe>', start_page, 'xml root')
118 xml_name = self._html_search_regex(r'<iframe src=".*?\?xml=(?P<xml_file>.+?\.xml).*?".*?</iframe>', start_page, 'xml filename', None, False)
120 # Fallback to the older format
121 xml_name = self._html_search_regex(r'<iframe src=".*?\?xmlURL=xml/(?P<xml_file>.+?\.xml).*?".*?</iframe>', start_page, 'xml filename')
123 xml_decription_url = xml_root + 'xml/' + xml_name
124 xml_description = self._download_xml(xml_decription_url, video_id)
126 video_title = xml_description.find('./metadata/title').text
127 video_formats = self._parse_mp4(xml_description)
128 if video_formats is None:
129 video_formats = self._parse_flv(xml_description)
133 'title': video_title,
134 'formats': video_formats,