[mediasite] Improve extraction and code style, add support for DASH (closes #11185, closes #14343, refs #5428)

youtube_dl/extractor/generic.py

@@ -100,6 +100,7 @@ from .megaphone import MegaphoneIE
 from .vzaar import VzaarIE
 from .channel9 import Channel9IE
 from .vshare import VShareIE
+from .mediasite import MediasiteIE
 
 
 class GenericIE(InfoExtractor):
@@ -1925,6 +1926,18 @@ class GenericIE(InfoExtractor):
                 'title': 'vl14062007715967',
                 'ext': 'mp4',
             }
+        },
+        {
+            'url': 'http://www.heidelberg-laureate-forum.org/blog/video/lecture-friday-september-23-2016-sir-c-antony-r-hoare/',
+            'md5': 'aecd089f55b1cb5a59032cb049d3a356',
+            'info_dict': {
+                'id': '90227f51a80c4d8f86c345a7fa62bd9a1d',
+                'ext': 'mp4',
+                'title': 'Lecture: Friday, September 23, 2016 - Sir Tony Hoare',
+                'description': 'md5:5a51db84a62def7b7054df2ade403c6c',
+                'timestamp': 1474354800,
+                'upload_date': '20160920',
+            }
         }
         # {
         #     # TODO: find another test
@@ -2884,14 +2897,14 @@ class GenericIE(InfoExtractor):
                 vshare_urls, video_id, video_title, ie=VShareIE.ie_key())
 
         # Look for Mediasite embeds
-        mobj = re.search(r'''(?xi)
+        mediasite_urls = MediasiteIE._extract_urls(webpage)
-                <iframe[^>]+src="((?:https?://[a-z0-9\-\.:\[\]]+)?
+        if mediasite_urls:
-                    /Mediasite/Play/[0-9a-f]{32,34}(?:\?.*?)?)"
+            entries = [
-            ''', webpage)
+                self.url_result(smuggle_url(
-        if mobj is not None:
+                    compat_urlparse.urljoin(url, mediasite_url),
-            return self.url_result(smuggle_url(
+                    {'UrlReferrer': url}), ie=MediasiteIE.ie_key())
-                compat_urlparse.urljoin(url, unescapeHTML(mobj.group(1))),
+                for mediasite_url in mediasite_urls]
-                { 'UrlReferrer': url }), 'Livestream')
+            return self.playlist_result(entries, video_id, video_title)
 
         def merge_dicts(dict1, dict2):
             merged = {}

youtube_dl/extractor/mediasite.py

@@ -5,21 +5,22 @@ import re
 import json
 
 from .common import InfoExtractor
-from ..compat import compat_urlparse
+from ..compat import (
+    compat_str,
+    compat_urlparse,
+)
 from ..utils import (
     ExtractorError,
-    unsmuggle_url,
-    mimetype2ext,
     float_or_none,
+    mimetype2ext,
+    unescapeHTML,
+    unsmuggle_url,
+    urljoin,
 )
 
 
 class MediasiteIE(InfoExtractor):
-    _VALID_URL = r'''(?xi)
+    _VALID_URL = r'(?xi)https?://[^/]+/Mediasite/Play/(?P<id>[0-9a-f]{32,34})(?P<query>\?[^#]+|)'
-        https?://[a-z0-9\-\.:\[\]]+/Mediasite/Play/
-        (?P<id>[0-9a-f]{32,34})
-        (?P<QueryString>\?[^#]+|)
-    '''
     _TESTS = [
         {
             'url': 'https://hitsmediaweb.h-its.org/mediasite/Play/2db6c271681e4f199af3c60d1f82869b1d',
@@ -94,59 +95,88 @@ class MediasiteIE(InfoExtractor):
         5: 'video3',
     }
 
+    @staticmethod
+    def _extract_urls(webpage):
+        return [
+            unescapeHTML(mobj.group('url'))
+            for mobj in re.finditer(
+                r'(?xi)<iframe\b[^>]+\bsrc=(["\'])(?P<url>(?:(?:https?:)?//[^/]+)?/Mediasite/Play/[0-9a-f]{32,34}(?:\?.*?)?)\1',
+                webpage)]
+
     def _real_extract(self, url):
         url, data = unsmuggle_url(url, {})
         mobj = re.match(self._VALID_URL, url)
-        ResourceId = mobj.group('id')
+        resource_id = mobj.group('id')
-        QueryString = mobj.group('QueryString')
+        query = mobj.group('query')
 
-        webpage = self._download_webpage(url, ResourceId) # XXX: add UrlReferrer?
+        webpage, urlh = self._download_webpage_handle(url, resource_id)  # XXX: add UrlReferrer?
+        redirect_url = compat_str(urlh.geturl())
 
         # XXX: might have also extracted UrlReferrer and QueryString from the html
-        ServicePath = compat_urlparse.urljoin(url, self._html_search_regex(
+        service_path = compat_urlparse.urljoin(redirect_url, self._html_search_regex(
-            r'<div id="ServicePath">(.+?)</div>', webpage, ResourceId,
+            r'<div[^>]+\bid=["\']ServicePath[^>]+>(.+?)</div>', webpage, resource_id,
             default='/Mediasite/PlayerService/PlayerService.svc/json'))
 
-        PlayerOptions = self._download_json(
+        player_options = self._download_json(
-            '%s/GetPlayerOptions' % (ServicePath), ResourceId,
+            '%s/GetPlayerOptions' % service_path, resource_id,
             headers={
                 'Content-type': 'application/json; charset=utf-8',
                 'X-Requested-With': 'XMLHttpRequest',
             },
             data=json.dumps({
                 'getPlayerOptionsRequest': {
-                    'ResourceId': ResourceId,
+                    'ResourceId': resource_id,
-                    'QueryString': QueryString,
+                    'QueryString': query,
                     'UrlReferrer': data.get('UrlReferrer', ''),
                     'UseScreenReader': False,
                 }
-            }).encode('utf-8'))
+            }).encode('utf-8'))['d']
-        Presentation = PlayerOptions['d']['Presentation']
+
-        if Presentation is None:
+        presentation = player_options['Presentation']
-            raise ExtractorError('Mediasite says: %s' %
+        title = presentation['Title']
-                (PlayerOptions['d']['PlayerPresentationStatusMessage'],),
+
+        if presentation is None:
+            raise ExtractorError(
+                'Mediasite says: %s' % player_options['PlayerPresentationStatusMessage'],
                 expected=True)
 
         thumbnails = []
         formats = []
-        for snum, Stream in enumerate(Presentation['Streams']):
+        for snum, Stream in enumerate(presentation['Streams']):
-            stream_type = self._STREAM_TYPES.get(
+            stream_type = Stream.get('StreamType')
-                Stream['StreamType'], 'type%u' % Stream['StreamType'])
+            if stream_type is None:
-
-            stream_formats = []
-            for unum, VideoUrl in enumerate(Stream['VideoUrls']):
-                url = VideoUrl['Location']
-                # XXX: if Stream.get('CanChangeScheme', False), switch scheme to HTTP/HTTPS
-
-                if VideoUrl['MediaType'] == 'SS':
-                    stream_formats.extend(self._extract_ism_formats(
-                        url, ResourceId, ism_id='%s-%u.%u' % (stream_type, snum, unum)))
                 continue
 
+            video_urls = Stream.get('VideoUrls')
+            if not isinstance(video_urls, list):
+                video_urls = []
+
+            stream_id = self._STREAM_TYPES.get(
+                stream_type, 'type%u' % stream_type)
+
+            stream_formats = []
+            for unum, VideoUrl in enumerate(video_urls):
+                video_url = VideoUrl.get('Location')
+                if not video_url or not isinstance(video_url, compat_str):
+                    continue
+                # XXX: if Stream.get('CanChangeScheme', False), switch scheme to HTTP/HTTPS
+
+                media_type = VideoUrl.get('MediaType')
+                if media_type == 'SS':
+                    stream_formats.extend(self._extract_ism_formats(
+                        video_url, resource_id,
+                        ism_id='%s-%u.%u' % (stream_id, snum, unum),
+                        fatal=False))
+                elif media_type == 'Dash':
+                    stream_formats.extend(self._extract_mpd_formats(
+                        video_url, resource_id,
+                        mpd_id='%s-%u.%u' % (stream_id, snum, unum),
+                        fatal=False))
+                else:
                     stream_formats.append({
-                    'format_id': '%s-%u.%u' % (stream_type, snum, unum),
+                        'format_id': '%s-%u.%u' % (stream_id, snum, unum),
-                    'url': url,
+                        'url': video_url,
-                    'ext': mimetype2ext(VideoUrl['MimeType']),
+                        'ext': mimetype2ext(VideoUrl.get('MimeType')),
                     })
 
             # TODO: if Stream['HasSlideContent']:
@@ -155,16 +185,16 @@ class MediasiteIE(InfoExtractor):
             # this will require writing a custom downloader...
 
             # disprefer 'secondary' streams
-            if Stream['StreamType'] != 0:
+            if stream_type != 0:
                 for fmt in stream_formats:
                     fmt['preference'] = -1
 
-            ThumbnailUrl = Stream.get('ThumbnailUrl')
+            thumbnail_url = Stream.get('ThumbnailUrl')
-            if ThumbnailUrl:
+            if thumbnail_url:
                 thumbnails.append({
-                    'id': '%s-%u' % (stream_type, snum),
+                    'id': '%s-%u' % (stream_id, snum),
-                    'url': compat_urlparse.urljoin(url, ThumbnailUrl),
+                    'url': urljoin(redirect_url, thumbnail_url),
-                    'preference': -1 if Stream['StreamType'] != 0 else 0,
+                    'preference': -1 if stream_type != 0 else 0,
                 })
             formats.extend(stream_formats)
 
@@ -174,11 +204,11 @@ class MediasiteIE(InfoExtractor):
         # XXX: Presentation['Transcript']
 
         return {
-            'id': ResourceId,
+            'id': resource_id,
-            'title': Presentation['Title'],
+            'title': title,
-            'description': Presentation.get('Description'),
+            'description': presentation.get('Description'),
-            'duration': float_or_none(Presentation.get('Duration'), 1000),
+            'duration': float_or_none(presentation.get('Duration'), 1000),
-            'timestamp': float_or_none(Presentation.get('UnixTime'), 1000),
+            'timestamp': float_or_none(presentation.get('UnixTime'), 1000),
             'formats': formats,
             'thumbnails': thumbnails,
         }