related to "proof of concept for posterior_pit" PR in brms

[florence-bockting]

Description

This PR is related to the brms PR proof of concept for posterior_pit#1857 (I can't push directly to the PR as I am not a maintainer).

The original PR in brms aims at improving the posterior_predict function to provide the possibility to return different types of values. Currently, only random draws ("random") is supported. Future support shall include "probability", "density", and "quantile".

TODOs

In this PR I provide first changes to the above described PR including

• creating a wrapper function that supports the cases "probability" and "random" (.predict_continuous_helper)
• include the wrapper in the functions posterior_predict_gaussian and posterior_predict_student as test cases
• add a tests for posterior_predict_gaussian to test the functionality of the implemented cases
• add tests for posterior_predict_student
• support truncated distributions