Better summary functions

We should have a summary function that mirrors the arviz summary object
https://python.arviz.org/en/latest/api/generated/arviz.summary.html

I'm not sure if this is directly applicable out of the box (since our objects are a bit higher dimensional than the examples).

Start by reading the ArviZ summary documentation at the linked arviz.summary entry point and compare its expected output with this project's higher-dimensional objects. No repository files or tests are named, so first locate the existing summary-related code and establish which dimensions and statistics must be supported. Done means a defined summary function with behavior appropriate for these objects and coverage for the chosen cases.

Written by the indexing model from the issue text.