Sampler rework LinearRTONew and RegularizedLinearRTONew

[chaozg]

fixed #368

• [x] LinearRTO
  • [x] overwrite setter for target
  • [x] put precomputation-relevant code in _precompute
  • [x] put _precompute into target setter
• [x] RegularizedLinearRTO
  • [x] move stepsize-relevant code in _choose_stepsize
  • [x] check if prior has proximal in target setter
  • [x] Unit tests with both sample and warmup
  • [x] plain/usual targets with LinearRTO
  • [x] MultipleLikelihoodPosterior targets with LinearRTO
  • [x] regularized targets with RegularizedLinearRTO
  • [x] regularized MultipleLikelihoodPosterior targets with RegularizedLinearRTO