R/best_clust.R

In cnfoley/mrclust: Identifying Clustered Heterogeneity in Mendelian Randomization Analyses

best_clust <- function(dta) {
  res <- NULL
  unique.obs <- unique(dta$observation)
  for (i in unique.obs) {
    tmp <- dta$observation == i
    mx <- which.max(dta$probability[tmp])
    res <- rbind(res, dta[tmp, ][mx, ])
  }
  return(res)
}

# prevents R check throwing errors about unassigned global variables
utils::globalVariables(c("annotateText", "cluster_class", "hjustvar",
                         "probability", "runif", "vjustvar", "xpos", "ypos",
                         "num_c"))