Add schema and code for cluster annealing

[lfrank]

Given a spikesorting across epochs, we need to find potential correspondences and save the bootstrapped distances between clusters, etc. in a table so that we can choose linking criteria and get the linked clusters.