R/SamplerJointIndep.R

In paradox: Define and Work with Parameter Spaces for Complex Algorithms

#' @title SamplerJointIndep Class
#'
#' @description
#' Create joint, independent sampler out of multiple other samplers.
#'
#' @family Sampler
#' @include Sampler.R
#' @export
SamplerJointIndep = R6Class("SamplerJointIndep", inherit = Sampler,
  public = list(
    #' @field samplers (`list()`)\cr
    #' List of [Sampler] objects.
    samplers = NULL,

    #' @description
    #' Creates a new instance of this [R6][R6::R6Class] class.
    #'
    #' @param samplers (`list()`)\cr
    #'   List of [Sampler] objects.
    initialize = function(samplers) {
      assert_list(samplers, types = "Sampler")
      self$samplers = samplers
      pss = map(samplers, "param_set")
      # FIXME: maybe we should use a paramset collection here?
      pss[[1L]] = pss[[1]]$clone() # we need to clone, add will clone later, too, otherwise we change the 1set in place
      self$param_set = Reduce(function(ps1, ps2) ps1$add(ps2), pss)
      # must_bounded and untyped should be check by the sapler, or if the sampler still works, then ok
      assert_param_set(self$param_set, no_deps = TRUE)
    }
  ),

  private = list(
    # FIXME: would be nice if we could call .sample here instead of sample, for less type
    # conversion and peed, but .sample is private. make it public? also not great...
    .sample = function(n) map_dtc(self$samplers, function(s) s$sample(n)$data),
    .print = function() catf("Independent comps: %i", length(self$samplers))
  )
)