Problem
The quickstart vignette uses plan("multisession") which won't work for real datasets due to memory copying. New users don't have a clear template for how to actually run clustOpt in production.
Proposal
Add ready-to-adapt Rscript templates in inst/scripts/:
run_clustopt_multicore.R — local machine with plan("multicore")
run_clustopt_slurm.R — HPC with future.batchtools
These should show the real-world pattern: load data → sketch → free original object → SCTransform → PCA → run clust_opt → save results.
Users can copy and adapt these rather than translating the vignette.
Problem
The quickstart vignette uses
plan("multisession")which won't work for real datasets due to memory copying. New users don't have a clear template for how to actually run clustOpt in production.Proposal
Add ready-to-adapt Rscript templates in
inst/scripts/:run_clustopt_multicore.R— local machine withplan("multicore")run_clustopt_slurm.R— HPC withfuture.batchtoolsThese should show the real-world pattern: load data → sketch → free original object → SCTransform → PCA → run clust_opt → save results.
Users can copy and adapt these rather than translating the vignette.