Prove Call/Result Structure

[traeger]

Rewrite our current Problem/ProblemVariant prover call from:

• ProveProcess deriving Process
• ProveScheduler deriving ThreadedProcessExecuter

to

• a new Prover class-
  • defining how a prover (e.g. Leo) is called
  • proving a method prove(problem, timeout, **prover_call_kwargs) to get a related prover call with optional additional parameter to start call the underlaying prover
• a new ProverCall class:
  • managing its own szsStatus, result parsing, stdout, stderr, timer
  • derivable for each supported TheormProver and TestSetup which is implemented:
  • using concurrent.Process, s.t. it supports termination and concurrent execution using run(), timer?
• the ProverCall can be used directly in the main-thread using a wait() pattern
• or used in a multitheaded environment s.t. a ProveScheduler
• ProblemVariant:
  • ProverCall is introduced as member
  • process, szsStatus, and such is removed

This should provide a good theorem-prover-call abstraction which may be used in any serial, concurrent or parallel environment.

Reminder:

• the package structure needs be reworked
  • concurrent -> utils.concurrent
  • new ProverCall and its derivations -> prover