cleanup

config.jsonc

@@ -287,6 +287,16 @@ examples (config.SITENAME.jsonc) that are part of this pan.do/ra distribution.
             "find": true,
             "sort": true
         },
+        {
+            "id": "topic",
+            "title": "Topic",
+            "type": ["string"],
+            "autocomplete": true,
+            "columnWidth": 120,
+            "filter": true,
+            "find": true,
+            "sort": true
+        },
         {
             "id": "color",
             "title": "Color",

data.py

@@ -28,6 +28,7 @@ def update_item(d, add_only=False):
         'description': 'summary',
         'id': 'archive',
         'mp4': 'stream',
+        'subject': 'topic',
     }.get(k,k):d[k] for k in d}
     for k in ('type', 'flv', 'video'):
         if k in info:
@@ -68,12 +69,16 @@ def update_item(d, add_only=False):
         info['color'] = [{
             'color': 'Color',
             'c': 'Color',
+            'b': 'B&W',
+            'w': 'B&W',
             'sepia': 'Sepia',
             'tinted': 'Tinted',
+            'tinted B&W': 'B&W (tinted)',
             'black and white': 'B&W',
+            'black & white': 'B&W',
             'bw': 'B&W',
             'b&w': 'B&W',
-        }.get(c.lower(), c) for c in info['color'].replace('b/w', 'b&w').split('/')]
+        }.get(c.lower(), c) for c in info['color'].replace('b/w', 'b&w').replace('B/W', 'B&W').split('/')]
 
     i, created = get_item(info['archive'])
     if add_only and not created:
@@ -91,11 +96,10 @@ def update_items(add_only=False):
     for d in data:
         update_item(d, add_only)
 
-def update_item_stream(i):
-    if update_stream(i):
+def update_item_stream(i, force=False):
+    if update_stream(i) or force:
         for s in i.streams():
             s.make_timeline()
-        i.level = 0
         i.save()
         i.update_timeline()
         i.make_poster(True)

get_data.py

@@ -24,7 +24,8 @@ for e in internetarchive.search.Search('collection:prelinger'):
             'title', 'description', 'year',
             'publisher', 'addeddate', 'sound',
             'creator', 'color', 'credits', 'publisher',
-            'sponsor', 'uploader', 'licenseurl',
+            'sponsor', 'uploader', 'licenseurl', 'subject',
+            'language'
             #needed?
             'date',
         ):