Ontological typing uses an unperformant algorithm

[liamhuber]

Fundamentally, we need to check the actual realized workflow graphs against the tree of available ontological graphs. Right now each port does this in the most naive way:

• Generate all possible workflows for the ontology, and check my graph.
• Repeat for each node in my tree.

This is only updated when new connections are made, but still. With graphs the size of the examples the delay is not particularly cumbersome, but the scaling is not good and a more intelligent comparison of the two trees will be needed in the future.