This should actually be relatively easy to do as the code is right there in Erik's previous Stan models. I'm thinking that the user would just enter a (scaled or unscaled) geographic distance matrix, with rows and column names equal to tip labels, as an argument to coev_fit(). Maybe geo_dist = NULL as a default. If a matrix is entered, the function scales it between 0 and 1 under the hood, ensures that it is in the same order as the phylogeny, and then includes it in the model with several Gaussian Processes, one for each variable.
This should actually be relatively easy to do as the code is right there in Erik's previous Stan models. I'm thinking that the user would just enter a (scaled or unscaled) geographic distance matrix, with rows and column names equal to tip labels, as an argument to
coev_fit(). Maybegeo_dist = NULLas a default. If a matrix is entered, the function scales it between 0 and 1 under the hood, ensures that it is in the same order as the phylogeny, and then includes it in the model with several Gaussian Processes, one for each variable.