Modularity in likelihood and prior definition

Implement the possibility to define ONE distribution for the likelihood and ONE distribution for the prior. Make it possible to pass those distributions to the calculation of the loglikelihood and logprior respectively.