For the curation task where some expert annotators assign labels to audio clips, we use a tree for representing the taxonomy. This tree holds information about the categories in the AudioSet Ontology. Some of this information is used for enabling or forbidding categories to be used in the curation task (Omitted curation task property).
Currently, this tree has to be computed offline, which is not practicle when we update the properties of the categories (taxonomy nodes) from the Django administration page for instance.
It would be nice to pre-compute this tree automatically, so that any change in the categories properties are automatically reflected in the tree.
I am not sure what is the best way, but we could for instance compute it async and store it in redis. It might be possible also to send a signal when some properties of the categories are updated to trigger the update of the tree.
For the curation task where some expert annotators assign labels to audio clips, we use a tree for representing the taxonomy. This tree holds information about the categories in the AudioSet Ontology. Some of this information is used for enabling or forbidding categories to be used in the curation task (Omitted curation task property).
Currently, this tree has to be computed offline, which is not practicle when we update the properties of the categories (taxonomy nodes) from the Django administration page for instance. It would be nice to pre-compute this tree automatically, so that any change in the categories properties are automatically reflected in the tree.
I am not sure what is the best way, but we could for instance compute it async and store it in redis. It might be possible also to send a signal when some properties of the categories are updated to trigger the update of the tree.