Add a serializer for content manifest files

kolibri/core/content/serializers.py

@@ -5,6 +5,7 @@
 from kolibri.core.content.models import ContentNode
 from kolibri.core.content.models import File
 from kolibri.core.content.models import Language
+from kolibri.core.content.utils.tree import get_channel_content_selection
 from kolibri.core.fields import create_timezonestamp
 
 
@@ -248,3 +249,36 @@ def get_updated_resource(self, instance):
 
     def get_is_leaf(self, instance):
         return instance.kind != content_kinds.TOPIC
+
+
+class ContentManifestChannelSerializer(serializers.ModelSerializer):
+    include_node_ids = serializers.SerializerMethodField()
+    exclude_node_ids = serializers.SerializerMethodField()
+
+    _content_selection = None
+
+    class Meta:
+        model = ChannelMetadata
+        fields = (
+            "id",
+            "version",
+            "include_node_ids",
+            "exclude_node_ids",
+        )
+
+    def get_include_node_ids(self, instance):
+        include_nodes, _ = self._get_content_selection(instance)
+        return [node.id for node in include_nodes]
+
+    def get_exclude_node_ids(self, instance):
+        _, exclude_nodes = self._get_content_selection(instance)
+        return [node.id for node in exclude_nodes]
+
+    def _get_content_selection(self, instance):
+        if not self._content_selection:
+            self._content_selection = get_channel_content_selection(instance)
+        return self._content_selection
+
+
+class ContentManifestSerializer(serializers.Serializer):
+    channels = serializers.ListField(child=ContentManifestChannelSerializer())

kolibri/core/content/utils/tree.py

@@ -18,3 +18,42 @@ def get_channel_node_depth(bridge, channel_id):
         return node_depth[0]
 
     return 0
+
+
+def get_channel_content_selection(channel_metadata):
+    include_nodes = list()
+    exclude_nodes = list()
+
+    available_nodes = ContentNode.objects.filter(
+        channel_id=channel_metadata.id, available=True
+    ).exclude(kind="topic")
+
+    available_nodes_queue = [channel_metadata.root]
+
+    while len(available_nodes_queue) > 0:
+        node = available_nodes_queue.pop(0)
+
+        # We could add nodes to exclude_nodes when less than half of the
+        # sibling nodes are missing. However, it is unclear if this would
+        # be useful.
+
+        if node.kind == "topic":
+            leaf_nodes = _get_leaf_nodes(node)
+            matching_leaf_nodes = set(leaf_nodes).intersection(available_nodes)
+            missing_leaf_nodes = set(leaf_nodes).difference(available_nodes)
+            if len(missing_leaf_nodes) == 0:
+                assert node not in include_nodes
+                include_nodes.append(node)
+            elif len(matching_leaf_nodes) > 0:
+                available_nodes_queue.extend(node.children.all())
+        elif node in available_nodes:
+            assert node not in include_nodes
+            include_nodes.append(node)
+
+    return include_nodes, exclude_nodes
+
+
+def _get_leaf_nodes(node):
+    return ContentNode.objects.filter(
+        lft__gte=node.lft, lft__lte=node.rght, channel_id=node.channel_id
+    ).exclude(kind="topic")