[vrt] Extract all videos on page

youtube_dl/extractor/vrt.py

@@ -1,6 +1,7 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import json
 import re
 
 from .common import InfoExtractor
@@ -54,34 +55,51 @@ class VRTIE(InfoExtractor):
     def _real_extract(self, url):
         site, display_id = re.match(self._VALID_URL, url).groups()
         webpage = self._download_webpage(url, display_id)
-        attrs = extract_attributes(self._search_regex(
-            r'(<[^>]+class="vrtvideo( [^"]*)?"[^>]*>)', webpage, 'vrt video'))
 
-        asset_id = attrs['data-video-id']
-        publication_id = attrs.get('data-publication-id')
-        if publication_id:
-            asset_id = publication_id + '$' + asset_id
-        client = attrs.get('data-client-code') or self._CLIENT_MAP[site]
-
-        title = strip_or_none(get_element_by_class(
+        page_title = strip_or_none(get_element_by_class(
             'vrt-title', webpage) or self._html_search_meta(
             ['og:title', 'twitter:title', 'name'], webpage))
-        description = self._html_search_meta(
+        page_description = self._html_search_meta(
             ['og:description', 'twitter:description', 'description'], webpage)
-        if description == '…':
-            description = None
-        timestamp = unified_timestamp(self._html_search_meta(
+        if page_description == '…':
+            page_description = None
+        page_timestamp = unified_timestamp(self._html_search_meta(
             'article:published_time', webpage))
 
-        return {
-            '_type': 'url_transparent',
-            'id': asset_id,
-            'display_id': display_id,
-            'title': title,
-            'description': description,
-            'thumbnail': attrs.get('data-posterimage'),
-            'timestamp': timestamp,
-            'duration': float_or_none(attrs.get('data-duration'), 1000),
-            'url': 'https://mediazone.vrt.be/api/v1/%s/assets/%s' % (client, asset_id),
-            'ie_key': 'Canvas',
-        }
+        def entries():
+            matches = re.findall(r'(<[^>]+class="vrtvideo(?: [^"]*)?"[^>]*>)', webpage)
+            if not matches:
+                self._downloader.report_warning('no vrt videos found on page')
+                return
+
+            for tag_string in matches:
+                attrs = extract_attributes(tag_string)
+                asset_id = attrs['data-video-id']
+                publication_id = attrs.get('data-publication-id')
+                if publication_id:
+                    asset_id = publication_id + '$' + asset_id
+                client = attrs.get('data-client-code') or self._CLIENT_MAP[site]
+
+                analytics_str = attrs.get('data-analytics')
+                if analytics_str:
+                    analytics = json.loads(analytics_str)
+                    timestamp = unified_timestamp(analytics.get("onTime")) or page_timestamp
+                    title = analytics.get("episode") or analytics.get("program") or page_title
+                else:
+                    timestamp = page_timestamp
+                    title = page_title
+
+                yield {
+                    '_type': 'url_transparent',
+                    'id': asset_id,
+                    'display_id': display_id,
+                    'title': title,
+                    'thumbnail': attrs.get('data-posterimage'),
+                    'timestamp': timestamp,
+                    'duration': float_or_none(attrs.get('data-duration'), 1000),
+                    'url': 'https://mediazone.vrt.be/api/v1/%s/assets/%s' % (client, asset_id),
+                    'ie_key': 'Canvas',
+                }
+
+        return self.playlist_result(
+            entries(), display_id, page_title, page_description)