Leverage cluster diameter in seed finding

[stephenswat]

We know from the work on the GBTS seed finding that the cluster diameter can be used to help find seeds, as the width of a cluster correlates with the pseudorapidity of the corresponding particle. Particles with low $|\eta|$ pass through barrel modules more perpendicularly, creating a smaller cluster while particles with high $|\eta|$ produce larger clusters. Thus, the cluster diameter projected on $z$ correlates with $|\eta|$, and the cluster diameter projected on $r$ correlates inversely with $|\eta|$.

#983 implements cluster diameter finding, but we need to figure out how we can use this in the seeding.