fosr plot doesn't like it when fosr.obj$data is of class tbl_df

[jeff-goldsmith]

Usually not a problem, but it can come up if the covariate data are organized using dplyr and then the response matrix added using something like data$reponse = response_mat.

The first issues are in plot_shiny.fosr - the expression
fosr.obj$data[,names(attributes(terms(fosr.obj$terms))$dataClasses)[1]]
can be swapped with
fosr.obj$data[[ names(attributes(terms(fosr.obj$terms))$dataClasses)[1] ]]

However, there's another issue that comes up in making and baking the lasagna plot that I don't know how to fix.