new param no_collapse_above_absolute

R/PipeOpCollapseFactors.R

@@ -61,9 +61,10 @@ PipeOpCollapseFactors = R6Class("PipeOpCollapseFactors",
  initialize = function(id = "collapsefactors", param_vals = list()) {
  ps = ps(
  no_collapse_above_prevalence = p_dbl(0, 1, tags = c("train", "predict")),
+ no_collapse_above_absolute = p_int(1, special_vals = list(Inf), tags = c("train", "predict")),
  target_level_count = p_int(2, tags = c("train", "predict"))
  )
- ps$values = list(no_collapse_above_prevalence = 1, target_level_count = 2)
+ ps$values = list(no_collapse_above_prevalence = 1, no_collapse_above_count = Inf, target_level_count = 2)
  super$initialize(id, param_set = ps, param_vals = param_vals, feature_types = c("factor", "ordered"))
  }
  ),
@@ -74,6 +75,7 @@ PipeOpCollapseFactors = R6Class("PipeOpCollapseFactors",
  dt = task$data(cols = private$.select_cols(task))
 
  keep_fraction = self$param_set$values$no_collapse_above_prevalence
+ keep_absolute = self$param_set$values$no_collapse_above_absolute
  target_count = self$param_set$values$target_level_count
 
  collapse_map = sapply(dt, function(d) {
@@ -86,8 +88,10 @@ PipeOpCollapseFactors = R6Class("PipeOpCollapseFactors",
  dtable = table(d)
  fractions = sort(dtable, decreasing = TRUE) / sum(!is.na(d))
  keep_fraction = names(fractions)[fractions >= keep_fraction]
+ # TODO: test this
+ keep_absolute = names(fractions)[fractions >= keep_absolute]
  keep_count = names(fractions)[seq_len(target_count)] # at this point we know there are more levels than target_count
- keep = union(keep_fraction, keep_count)
+ keep = union(keep_fraction, setdiff(keep_count, keep_absolute))
  dont_keep = setdiff(levels(d), keep)
  if (is.ordered(d)) {
  cmap = stats::setNames(as.list(levels(d)), levels(d))