Use successes rather than attempts as stop condition

[dhimmel]

Since only successful swaps advance the network randomization, it doesn't make a ton of sense to me to set an attempt threshold as the stop condition versus a success threshold. This has been bothering me since I initially implemented the attempt threshold in hetnetpy.

To prevent infinite running, we could enable a max_attempts parameter, perhaps that is a multiplier for based on the specified number of successes.