feat: enforce ability to save and restore the current detector state

One will want to save / restore a given state point (e.g. for training, later potentially for inference)

Proposition: in TADLearner(Protocol), 
`def export_state(self, ...):`

`def load_state(self, ...):`
