Add custom log-normal race family

[strengejacke]

@DominiqueMakowski would it make sense / be possible to add that custom family implemented here (https://github.com/DominiqueMakowski/cogmod/blob/main/R/lnr_brms.R) to bayestestR?

That would make this family more widely available, and probably integrates more seamlessly into easystats?

[strengejacke]

(and maybe also the other families...)

[DominiqueMakowski]

I think it's too soon for that, the implementation is not yet tested or anything, so for now I'd best keep it separated before making it "official". Beside, I'm not sure bayestestR would be a good place to start piling up application-specific custom brms models, as there is no upper limit of what models we can include

[strengejacke]

Beside, I'm not sure bayestestR would be a good place to start piling up application-specific custom brms models, as there is no upper limit of what models we can include

True, but we already have custom contrasts, and I think this is somehow "similar", a kind of "extension" of modelling packages. And if this family is a Bayesian-specific or even brms-specific thin, why not? For user, it would mostly we something like family = bayestestR::lnr() or similar.

[DominiqueMakowski]

Makes sense, I'll think about it once I make sure it all works (it's the first time I'm doing this custom Stan thing, and I need to test it on my own analysis first to make sure it works :)

[mattansb]

True, but we already have custom contrasts

I think those have moved to datawizard?

---

I think if your family is useful to a general audience, you can make a PR on brms!

I also have some custom brms family / code - I've been thinking about that for a while...

[strengejacke]

I had this in mind
https://github.com/easystats/bayestestR/blob/main/R/contr.equalprior.R

I also have some custom brms family / code

Yes, why not?

[mattansb]

Oh I thought those were moved to datawizard to be together with contr.deviation.