Trackpy gets stuck when subnetwork_size too large

[freemansw1]

From a discussion between me and @wreckdump in #320, we should document that sometimes trackpy hangs on tracking when subnetwork_size is too large. I've had trouble reliably replicating this, though. We should probably add a bit to the API docs saying to try that when/if it hangs.

Originally posted by @wreckdump in #320 (comment)

[JuliaKukulies]

Good point, yes we should do that.

[w-k-jones]

It might also be helpful to look into catching more exceptions in trackpy (such as the inability to create a velocity field because the search distance is too small to connect any features) and raising them with more meaningful info for use with tobac