Fixes plotting clustree when best k is in CLUSTER_KMEANS_K parameter

Bug noticed by Lucie Pfeiferova. Error when tryint to create a tibble:  Column name `k1` must not be duplicated. Tibble wouldn't have unique column names because best would be a part of the k values. Fixed by taking only unique values.

R/visualization.R

@@ -577,6 +577,9 @@ dimred_plots_from_params_df <- function(sce_dimred, dimred_plots_params_df) {
 #' @concept sce_visualization
 #' @export
 plot_clustree <- function(cluster_list, params, prefix, title = deparse(substitute(cluster_list)), edge_arrow = FALSE, highlight_core = TRUE, ...) {
+  cluster_list <- unique(cluster_list)
+  params <- unique(params)
+
   assert_that_(length(cluster_list) == length(params))
 
   clustree_list <- cluster_list %>%