Open WardBrian opened 3 months ago
This was requested at StanCon by @wds15
@paul-buerkner mentioned that some additional changes in BRMS would be necessary if we wanted to be able to take advantage of all the cool downstream stuff BRMS enables for its models, so this issue can also serve as a place to discuss if anything would be needed on our end
For basic uses, it looks like this is already possible, e.g.:
install.packages('brms')
library(brms)
make_stancode(rating ~ treat + period + carry + (1|subject),
data = inhaler, family = "cumulative")
data <- make_standata(rating ~ treat + period + carry + (1|subject),
data = inhaler, family = "cumulative")
in the data generation window will set the data and print the Stan code, which you can copy over yourself. A bit clunky, and missing all of brms' niceties around PPC, etc
The primary use case of this would be to support brms. This supports a
make_stancode
andmake_standata
functions which are paired together.Related to #139, with the same questions about whether or not it belongs in the same scripts.
There is not really an equivalent need/desire for this in Python