Include the logprior evaluations to support use of priorsense package

[MilaniC]

Hi @dmphillippo. A request to modify the Stan code for multinma to include the logprior evaluations (lprior) and the loglikelihood specified so that the fitted model (via rstan interface) can be used with the priorsense R package (see List 1 in Kallioinen et al (2024) p57).

And so support assessing posterior sensitivity to perturbations of the network model prior and likelihood.

#Kallioinen N et al (2024) Detecting and diagnosing prior and likelihood sensitivity with power-scaling. Statistics and Comp
uting 34: 57. https://doi.org/10.1007/s11222-023-10366-5