X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fgdcvault.py;h=3befd3e7b8f9ffeb48a31a284b5971b3a8a5cfab;hb=5c2266df4b9aeb7881ed8c026a038e2a25e43734;hp=74908da7708a79314619d5c43b89e3b9fc71ecad;hpb=546582ec3eab46712bead1ca7aeded09499d169a;p=youtube-dl diff --git a/youtube_dl/extractor/gdcvault.py b/youtube_dl/extractor/gdcvault.py index 74908da77..3befd3e7b 100644 --- a/youtube_dl/extractor/gdcvault.py +++ b/youtube_dl/extractor/gdcvault.py @@ -1,22 +1,26 @@ from __future__ import unicode_literals import re -import xml.etree.ElementTree from .common import InfoExtractor +from ..compat import compat_urllib_parse from ..utils import ( - compat_urllib_parse, - compat_urllib_request, + remove_end, + HEADRequest, + sanitized_Request, ) + class GDCVaultIE(InfoExtractor): - _VALID_URL = r'https?://(?:www\.)?gdcvault\.com/play/(?P\d+)/(?P(\w|-)+)' + _VALID_URL = r'https?://(?:www\.)?gdcvault\.com/play/(?P\d+)/(?P(\w|-)+)?' + _NETRC_MACHINE = 'gdcvault' _TESTS = [ { 'url': 'http://www.gdcvault.com/play/1019721/Doki-Doki-Universe-Sweet-Simple', 'md5': '7ce8388f544c88b7ac11c7ab1b593704', 'info_dict': { 'id': '1019721', + 'display_id': 'Doki-Doki-Universe-Sweet-Simple', 'ext': 'mp4', 'title': 'Doki-Doki Universe: Sweet, Simple and Genuine (GDC Next 10)' } @@ -25,13 +29,29 @@ class GDCVaultIE(InfoExtractor): 'url': 'http://www.gdcvault.com/play/1015683/Embracing-the-Dark-Art-of', 'info_dict': { 'id': '1015683', + 'display_id': 'Embracing-the-Dark-Art-of', 'ext': 'flv', 'title': 'Embracing the Dark Art of Mathematical Modeling in AI' }, - 'params' : { - 'skip_download': True, # Requires rtmpdump + 'params': { + 'skip_download': True, # Requires rtmpdump } }, + { + 'url': 'http://www.gdcvault.com/play/1015301/Thexder-Meets-Windows-95-or', + 'md5': 'a5eb77996ef82118afbbe8e48731b98e', + 'info_dict': { + 'id': '1015301', + 'display_id': 'Thexder-Meets-Windows-95-or', + 'ext': 'flv', + 'title': 'Thexder Meets Windows 95, or Writing Great Games in the Windows 95 Environment', + }, + 'skip': 'Requires login', + }, + { + 'url': 'http://gdcvault.com/play/1020791/', + 'only_matching': True, + } ] def _parse_mp4(self, xml_description): @@ -54,27 +74,41 @@ class GDCVaultIE(InfoExtractor): return video_formats def _parse_flv(self, xml_description): - video_formats = [] - akami_url = xml_description.find('./metadata/akamaiHost').text + formats = [] + akamai_url = xml_description.find('./metadata/akamaiHost').text + audios = xml_description.find('./metadata/audios') + if audios is not None: + for audio in audios: + formats.append({ + 'url': 'rtmp://%s/ondemand?ovpfv=1.1' % akamai_url, + 'play_path': remove_end(audio.get('url'), '.flv'), + 'ext': 'flv', + 'vcodec': 'none', + 'format_id': audio.get('code'), + }) slide_video_path = xml_description.find('./metadata/slideVideo').text - video_formats.append({ - 'url': 'rtmp://' + akami_url + '/' + slide_video_path, - 'format_note': 'slide deck video', - 'quality': -2, - 'preference': -2, - 'format_id': 'slides', - }) + formats.append({ + 'url': 'rtmp://%s/ondemand?ovpfv=1.1' % akamai_url, + 'play_path': remove_end(slide_video_path, '.flv'), + 'ext': 'flv', + 'format_note': 'slide deck video', + 'quality': -2, + 'preference': -2, + 'format_id': 'slides', + }) speaker_video_path = xml_description.find('./metadata/speakerVideo').text - video_formats.append({ - 'url': 'rtmp://' + akami_url + '/' + speaker_video_path, - 'format_note': 'speaker video', - 'quality': -1, - 'preference': -1, - 'format_id': 'speaker', - }) - return video_formats - - def _login(self, webpage_url, video_id): + formats.append({ + 'url': 'rtmp://%s/ondemand?ovpfv=1.1' % akamai_url, + 'play_path': remove_end(speaker_video_path, '.flv'), + 'ext': 'flv', + 'format_note': 'speaker video', + 'quality': -1, + 'preference': -1, + 'format_id': 'speaker', + }) + return formats + + def _login(self, webpage_url, display_id): (username, password) = self._get_login_info() if username is None or password is None: self.report_warning('It looks like ' + webpage_url + ' requires a login. Try specifying a username and password and try again.') @@ -89,11 +123,11 @@ class GDCVaultIE(InfoExtractor): 'password': password, } - request = compat_urllib_request.Request(login_url, compat_urllib_parse.urlencode(login_form)) + request = sanitized_Request(login_url, compat_urllib_parse.urlencode(login_form)) request.add_header('Content-Type', 'application/x-www-form-urlencoded') - login_content = self._download_webpage(request, video_id, 'Logging in') - start_page = self._download_webpage(webpage_url, video_id, 'Getting authenticated video page') - logout_content = self._download_webpage(logout_url, video_id, 'Logging out') + self._download_webpage(request, display_id, 'Logging in') + start_page = self._download_webpage(webpage_url, display_id, 'Getting authenticated video page') + self._download_webpage(logout_url, display_id, 'Logging out') return start_page @@ -101,27 +135,54 @@ class GDCVaultIE(InfoExtractor): mobj = re.match(self._VALID_URL, url) video_id = mobj.group('id') - webpage_url = 'http://www.gdcvault.com/play/' + video_id - start_page = self._download_webpage(webpage_url, video_id) + display_id = mobj.group('name') or video_id - xml_root = self._html_search_regex(r'', + start_page, 'xml root', default=None) if xml_root is None: # Probably need to authenticate - start_page = self._login(webpage_url, video_id) - if start_page is None: + login_res = self._login(webpage_url, display_id) + if login_res is None: self.report_warning('Could not login.') else: + start_page = login_res # Grab the url from the authenticated page - xml_root = self._html_search_regex(r'', + start_page, 'xml root') - xml_name = self._html_search_regex(r'', + start_page, 'xml filename', default=None) if xml_name is None: # Fallback to the older format xml_name = self._html_search_regex(r'