I think you'll need to
- Put INLA in Suggests as opposed to Imports, this way remotes (and CRAN) won't try to install it.
- Add a step in the GitHub Actions workflow that'll install INLA. See Non-standard Remotes package INLA in R package for suggestions on how to do so (e.g. the first solution is
install.packages("INLA",repos=c(getOption("repos"),INLA="https://inla.r-inla-download.org/R/stable"), dep=TRUE)). - Add the
Additional_repositoriesfield like https://github.com/cran/INLABMA/blob/8bf2a402746946f0ab6512fbe49dc8a6122d23b5/DESCRIPTION#L15 which is not for installation but just for R CMD check to "check availability". - (getting off topic) For your users, adding some sort of check that the package is installed e.g. rlang::check_installed() with some explanation (in that error and in the docs) on how to install it.