var.subset argument for plot.gllvm()

[gerverska]

Hi folks,

While I like the n.plot option in plot.gllvm(), it seems that it might be worth having a var.subset option similar to plot.manyglm() that allows the column index of a response variable to be specified. I've noticed that by default, when n.plot = 2 and the residuals of model are plotted, it seems the 2 response variables chosen are random (and it seems to be random no matter what the value of n.plot is. I know I can always extract the residuals and plot them myself (that's what I'm currently doing), but I figured I'd suggest anyways.

# First run ####
model |> plot(which = 1, n.plot = 2)

# Second run ####
model |> plot(which = 1, n.plot = 2)