An improvement to the DR clustering algorithm

We show that mapping data from a bounded region onto a toroidal surface prior to running the DR (Dog-Rabbit) clustering algorithm greatly increases the rate of success of finding all the clusters in the region.