Another round of SBC checking validation

[wlandau]

The package has undergone many changes since I last ran the pipeline for https://openpharma.github.io/brms.mmrm/articles/sbc.html. In particular, the package now supports different correlation structures and different distributional parameter structures for sigma. We might need to cover these cases as new simulation scenarios. We might be able to make the computation easier with:

• non-subgroup
• no missing responses
• no nuisance covariates
• few treatment groups

We should probably have at least 4 time points.

I don't think we need to run informative prior archetypes through this because that is a data specification problem and the unit tests cover it well.

[wlandau]

Branch https://github.com/openpharma/brms.mmrm/tree/107 has an updated version of the pipeline in vignettes/sbc/. It seems to be working for a small number of simulations. Currently running it at scale. Should be done by around Monday if everything goes well.

[wlandau]

Fixed in #110