]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/gdcvault.py
Prepate to upload
[youtubedl] / youtube_dl / extractor / gdcvault.py
index 89d5994eef02623222b73fa04e0287a264b01333..f71d9092e5371d5d2c143058ded24cb5f4ca1958 100644 (file)
@@ -4,18 +4,22 @@ import re
 
 from .common import InfoExtractor
 from ..utils import (
-    compat_urllib_parse,
-    compat_urllib_request,
+    HEADRequest,
+    sanitized_Request,
+    urlencode_postdata,
 )
 
+
 class GDCVaultIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?gdcvault\.com/play/(?P<id>\d+)/(?P<name>(\w|-)+)'
+    _VALID_URL = r'https?://(?:www\.)?gdcvault\.com/play/(?P<id>\d+)/(?P<name>(\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)'
             }
@@ -24,6 +28,7 @@ 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'
             },
@@ -31,49 +36,61 @@ class GDCVaultIE(InfoExtractor):
                 '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,
+        },
+        {
+            # Hard-coded hostname
+            'url': 'http://gdcvault.com/play/1023460/Tenacious-Design-and-The-Interface',
+            'md5': 'a8efb6c31ed06ca8739294960b2dbabd',
+            'info_dict': {
+                'id': '1023460',
+                'ext': 'mp4',
+                'display_id': 'Tenacious-Design-and-The-Interface',
+                'title': 'Tenacious Design and The Interface of \'Destiny\'',
+            },
+        },
+        {
+            # Multiple audios
+            'url': 'http://www.gdcvault.com/play/1014631/Classic-Game-Postmortem-PAC',
+            'info_dict': {
+                'id': '1014631',
+                'ext': 'flv',
+                'title': 'How to Create a Good Game - From My Experience of Designing Pac-Man',
+            },
+            'params': {
+                'skip_download': True,  # Requires rtmpdump
+                'format': 'jp',  # The japanese audio
+            }
+        },
+        {
+            # gdc-player.html
+            'url': 'http://www.gdcvault.com/play/1435/An-American-engine-in-Tokyo',
+            'info_dict': {
+                'id': '1435',
+                'display_id': 'An-American-engine-in-Tokyo',
+                'ext': 'flv',
+                'title': 'An American Engine in Tokyo:/nThe collaboration of Epic Games and Square Enix/nFor THE LAST REMINANT',
+            },
+            'params': {
+                'skip_download': True,  # Requires rtmpdump
+            },
+        },
     ]
 
-    def _parse_mp4(self, xml_description):
-        video_formats = []
-        mp4_video = xml_description.find('./metadata/mp4video')
-        if mp4_video is None:
-            return None
-
-        mobj = re.match(r'(?P<root>https?://.*?/).*', mp4_video.text)
-        video_root = mobj.group('root')
-        formats = xml_description.findall('./metadata/MBRVideos/MBRVideo')
-        for format in formats:
-            mobj = re.match(r'mp4\:(?P<path>.*)', format.find('streamName').text)
-            url = video_root + mobj.group('path')
-            vbr = format.find('bitrate').text
-            video_formats.append({
-                'url': url,
-                'vbr': int(vbr),
-            })
-        return video_formats
-
-    def _parse_flv(self, xml_description):
-        video_formats = []
-        akami_url = xml_description.find('./metadata/akamaiHost').text
-        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',
-        })
-        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):
+    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.')
@@ -88,11 +105,11 @@ class GDCVaultIE(InfoExtractor):
             'password': password,
         }
 
-        request = compat_urllib_request.Request(login_url, compat_urllib_parse.urlencode(login_form))
+        request = sanitized_Request(login_url, urlencode_postdata(login_form))
         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
-        self._download_webpage(request, video_id, 'Logging in')
-        start_page = self._download_webpage(webpage_url, video_id, 'Getting authenticated video page')
-        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
 
@@ -100,35 +117,58 @@ class GDCVaultIE(InfoExtractor):
         mobj = re.match(self._VALID_URL, url)
 
         video_id = mobj.group('id')
+        display_id = mobj.group('name') or video_id
+
         webpage_url = 'http://www.gdcvault.com/play/' + video_id
-        start_page = self._download_webpage(webpage_url, video_id)
+        start_page = self._download_webpage(webpage_url, display_id)
+
+        direct_url = self._search_regex(
+            r's1\.addVariable\("file",\s*encodeURIComponent\("(/[^"]+)"\)\);',
+            start_page, 'url', default=None)
+        if direct_url:
+            title = self._html_search_regex(
+                r'<td><strong>Session Name</strong></td>\s*<td>(.*?)</td>',
+                start_page, 'title')
+            video_url = 'http://www.gdcvault.com' + direct_url
+            # resolve the url so that we can detect the correct extension
+            head = self._request_webpage(HEADRequest(video_url), video_id)
+            video_url = head.geturl()
+
+            return {
+                'id': video_id,
+                'display_id': display_id,
+                'url': video_url,
+                'title': title,
+            }
 
-        xml_root = self._html_search_regex(r'<iframe src="(?P<xml_root>.*?)player.html.*?".*?</iframe>', start_page, 'xml root', None, False)
+        PLAYER_REGEX = r'<iframe src="(?P<xml_root>.+?)/(?:gdc-)?player.*?\.html.*?".*?</iframe>'
 
+        xml_root = self._html_search_regex(
+            PLAYER_REGEX, 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'<iframe src="(?P<xml_root>.*?)player.html.*?".*?</iframe>', start_page, 'xml root')
+                xml_root = self._html_search_regex(
+                    PLAYER_REGEX, start_page, 'xml root')
 
-        xml_name = self._html_search_regex(r'<iframe src=".*?\?xml=(?P<xml_file>.+?\.xml).*?".*?</iframe>', start_page, 'xml filename', None, False)
+        xml_name = self._html_search_regex(
+            r'<iframe src=".*?\?xml=(.+?\.xml).*?".*?</iframe>',
+            start_page, 'xml filename', default=None)
         if xml_name is None:
             # Fallback to the older format
-            xml_name = self._html_search_regex(r'<iframe src=".*?\?xmlURL=xml/(?P<xml_file>.+?\.xml).*?".*?</iframe>', start_page, 'xml filename')
-
-        xml_decription_url = xml_root + 'xml/' + xml_name
-        xml_description = self._download_xml(xml_decription_url, video_id)
-
-        video_title = xml_description.find('./metadata/title').text
-        video_formats = self._parse_mp4(xml_description)
-        if video_formats is None:
-            video_formats = self._parse_flv(xml_description)
+            xml_name = self._html_search_regex(
+                r'<iframe src=".*?\?xmlURL=xml/(?P<xml_file>.+?\.xml).*?".*?</iframe>',
+                start_page, 'xml filename')
 
         return {
+            '_type': 'url_transparent',
             'id': video_id,
-            'title': video_title,
-            'formats': video_formats,
+            'display_id': display_id,
+            'url': '%s/xml/%s' % (xml_root, xml_name),
+            'ie_key': 'DigitallySpeaking',
         }