contributor_breadth_model task needs checkpoints

I am running a development instance on my laptop with a couple large repositories added for collection.

This means when i first start the instance I begin seeing log lines like `Processing cntrb XXXX of 31411`.

Sometimes the instance gets to ~25k events before i have to restart the instance, causing it to start over.

Is there a way we can implement this task to run in batches so that future runs can pick up approximately where a previous run left off?

This will require some deeper analysis to determine feasibility