Add tests for resuming optimization loop

[bpkroth]

trafficstars

Might just need to verify that we can reprime and finish an optimizer loop without that directory around. Not sure if we have a test for that yet ...

Originally posted by @bpkroth in https://github.com/microsoft/MLOS/pull/579#discussion_r1383873499

• [ ] In particular, we should make sure that optimizations that abort part way through and are picked up again (e.g., data reloaded) finish converging as expected.
• [ ] Additionally, we should have tests for disallowing reusing trials from an experiment that doesn't have compatible configs. This can happen for instance if you try to change the tunable params in an "incompatible" way for the same experiment.