Track total SCF calls across restarts

[ltimmerman3]

Is your feature request related to a problem? Please describe.
The socket is being restarted regularly when using the NVTBerendsen thermostat for MD. The SocketSCFCount and elecgs_Count both get reset when the socket is restarted. This makes tracking simulation progress impossible in SPARC C and leads to misleading .out and .static files.

Describe the solution you'd like
An SCF count on the python side that can be passed to SPARC if a restart must be performed.

Additional context
This leads to other problems in the context of MLFF. If a user was training a FF from scratch, it will be restarted from scratch no matter where they were in the simulation.