mlr-org · mllg · Mar 1, 2022 · Mar 1, 2022 · Mar 1, 2022 · Mar 1, 2022
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -40,7 +40,8 @@ Suggests:
     mlr3proba,
     praznik,
     rpart,
-    testthat (>= 3.0.0)
+    testthat (>= 3.0.0),
+    withr
 Config/testthat/edition: 3
 Config/testthat/parallel: true
 Encoding: UTF-8

diff --git a/R/Filter.R b/R/Filter.R
@@ -6,13 +6,6 @@
 #' unimportant features get a small value. Note that filter scores may also be
 #' negative.
 #'
-#' @template field_id
-#' @template field_task_type
-#' @template field_task_properties
-#' @template field_param_set
-#' @template field_feature_types
-#' @template field_packages
-#'
 #' @details
 #' Some features support partial scoring of the feature set:
 #' If `nfeat` is not `NULL`, only the best `nfeat` features are guaranteed to
@@ -24,11 +17,38 @@
 #' @export
 Filter = R6Class("Filter",
   public = list(
+    #' @field id (`character(1)`)\cr
+    #'   Identifier of the object.
+    #'   Used in tables, plot and text output.
     id = NULL,
+
+    #' @field label (`character(1)`)\cr
+    #'   Label for this object.
+    #'   Can be used in tables, plot and text output instead of the ID.
+    label = NA_character_,
+
+    #' @field task_type (`character(1)`)\cr
+    #'   Task type, e.g. `"classif"` or `"regr"`.
+    #'   Can be set to `NA` to allow all task types.
+    #'
+    #'   For a complete list of possible task types (depending on the loaded packages),
+    #'   see [`mlr_reflections$task_types$type`][mlr_reflections].
     task_type = NULL,
+
+    #' @field task_properties (`character()`)\cr
+    #'   [mlr3::Task]task properties.
     task_properties = NULL,
+
+    #' @field param_set ([paradox::ParamSet])\cr
+    #'   Set of hyperparameters.
     param_set = NULL,
+
+    #' @field feature_types (`character()`)\cr
+    #'   Feature types of the filter.
     feature_types = NULL,
+
+    #' @field packages ([character()])\cr
+    #'   Packages which this filter is relying on.
     packages = NULL,
 
     #' @field man (`character(1)`)\cr
@@ -64,15 +84,18 @@ Filter = R6Class("Filter",
     #'   Set of required packages.
     #'   Note that these packages will be loaded via [requireNamespace()], and
     #'   are not attached.
+    #' @param label (`character(1)`)\cr
+    #'   Label for the new instance.
     #' @param man (`character(1)`)\cr
     #'   String in the format `[pkg]::[topic]` pointing to a manual page for
     #'   this object. The referenced help package can be opened via method
     #'   `$help()`.
     initialize = function(id, task_type, task_properties = character(),
       param_set = ps(), feature_types = character(),
-      packages = character(), man = NA_character_) {
+      packages = character(), label = NA_character_, man = NA_character_) {
 
       self$id = assert_string(id)
+      self$label = assert_string(label, na.ok = TRUE)
       if (!test_scalar_na(task_type)) {
         # we allow any task type here, otherwise we are not able to construct
         # the filter without loading additional packages like mlr3proba
@@ -103,15 +126,15 @@ Filter = R6Class("Filter",
     #' @description
     #' Printer for Filter class
     print = function() {
-      catf(format(self)) # nocov start
-      catf(str_indent("Task Types:", self$task_type))
-      catf(str_indent("Task Properties:", self$task_properties))
-      catf(str_indent("Packages:", self$packages))
-      catf(str_indent("Feature types:", self$feature_types))
+      catn(format(self), if (is.na(self$label)) "" else paste0(": ", self$label))
+      catn(str_indent("Task Types:", self$task_type))
+      catn(str_indent("Task Properties:", self$task_properties))
+      catn(str_indent("Packages:", self$packages))
+      catn(str_indent("Feature types:", self$feature_types))
       if (length(self$scores)) {
         print(as.data.table(self),
           nrows = 10L, topn = 5L, class = FALSE,
-          row.names = TRUE, print.keys = FALSE) # nocov end
+          row.names = TRUE, print.keys = FALSE)
       }
     },
 

diff --git a/R/FilterAUC.R b/R/FilterAUC.R
@@ -35,6 +35,7 @@ FilterAUC = R6Class("FilterAUC",
         task_properties = "twoclass",
         feature_types = c("integer", "numeric"),
         packages = "mlr3measures",
+        label = "Area Under the ROC Curve Score",
         man = "mlr3filters::mlr_filters_auc"
       )
     }

diff --git a/R/FilterAnova.R b/R/FilterAnova.R
@@ -38,6 +38,7 @@ FilterAnova = R6Class("FilterAnova",
         packages = "stats",
         feature_types = c("integer", "numeric"),
         task_type = "classif",
+        label = "ANOVA F-Test",
         man = "mlr3filters::mlr_filters_anova"
       )
     }

diff --git a/R/FilterCMIM.R b/R/FilterCMIM.R
@@ -1,8 +1,8 @@
-#' @title Minimal Conditional Mutual Information Filter
+#' @title Minimal Conditional Mutual Information Maximization Filter
 #'
 #' @name mlr_filters_cmim
 #'
-#' @description Minimal conditional mutual information maximisation filter
+#' @description Minimal conditional mutual information maximization filter
 #' calling [praznik::CMIM()] from package \CRANpkg{praznik}.
 #'
 #' This filter supports partial scoring (see [Filter]).
@@ -41,6 +41,7 @@ FilterCMIM = R6Class("FilterCMIM",
         param_set = param_set,
         feature_types = c("integer", "numeric", "factor", "ordered"),
         packages = "praznik",
+        label = "Minimal Conditional Mutual Information Maximization",
         man = "mlr3filters::mlr_filters_cmim"
       )
     }

diff --git a/R/FilterCarScore.R b/R/FilterCarScore.R
@@ -43,6 +43,7 @@ FilterCarScore = R6Class("FilterCarScore",
         param_set = param_set,
         feature_types = "numeric",
         packages = "care",
+        label = "Correlation-Adjusted coRrelation Score",
         man = "mlr3filters::mlr_filters_carscore"
       )
     }

diff --git a/R/FilterCarSurvScore.R b/R/FilterCarSurvScore.R
@@ -35,6 +35,7 @@ FilterCarSurvScore = R6Class("FilterCarSurvScore",
         param_set = ps,
         feature_types = c("integer", "numeric"),
         task_type = "surv",
+        label = "Correlation-Adjusted coRrelation Survival Score",
         man = "mlr3filters::mlr_filters_carsurvscore"
       )
     }

diff --git a/R/FilterCorrelation.R b/R/FilterCorrelation.R
@@ -44,6 +44,7 @@ FilterCorrelation = R6Class("FilterCorrelation",
         param_set = param_set,
         feature_types = c("integer", "numeric"),
         packages = "stats",
+        label = "Correlation",
         man = "mlr3filters::mlr_filters_correlation"
       )
     }

diff --git a/R/FilterDISR.R b/R/FilterDISR.R
@@ -41,6 +41,7 @@ FilterDISR = R6Class("FilterDISR",
         param_set = param_set,
         feature_types = c("integer", "numeric", "factor", "ordered"),
         packages = "praznik",
+        label = "Double Input Symmetrical Relevance",
         man = "mlr3filters::mlr_filters_disr"
       )
     }

diff --git a/R/FilterFindCorrelation.R b/R/FilterFindCorrelation.R
@@ -47,6 +47,7 @@ FilterFindCorrelation = R6Class("FilterFindCorrelation",
         param_set = param_set,
         feature_types = c("integer", "numeric"),
         packages = "stats",
+        label = "Correlation-based Score",
         man = "mlr3filters::mlr_filters_find_correlation"
       )
     }

diff --git a/R/FilterImportance.R b/R/FilterImportance.R
@@ -38,6 +38,7 @@ FilterImportance = R6Class("FilterImportance",
         feature_types = learner$feature_types,
         packages = learner$packages,
         param_set = learner$param_set,
+        label = "Importance Score",
         man = "mlr3filters::mlr_filters_importance"
       )
     }

diff --git a/R/FilterInformationGain.R b/R/FilterInformationGain.R
@@ -48,6 +48,7 @@ FilterInformationGain = R6Class("FilterInformationGain",
         param_set = param_set,
         feature_types = c("integer", "numeric", "factor", "ordered"),
         packages = "FSelectorRcpp",
+        label = "Information Gain",
         man = "mlr3filters::mlr_filters_information_gain"
       )
     }

diff --git a/R/FilterJMI.R b/R/FilterJMI.R
@@ -42,6 +42,7 @@ FilterJMI = R6Class("FilterJMI",
         param_set = param_set,
         packages = "praznik",
         feature_types = c("integer", "numeric", "factor", "ordered"),
+        label = "Joint Mutual Information",
         man = "mlr3filters::mlr_filters_jmi"
       )
     }

diff --git a/R/FilterJMIM.R b/R/FilterJMIM.R
@@ -1,8 +1,8 @@
-#' @title Minimal Joint Mutual Information Maximisation Filter
+#' @title Minimal Joint Mutual Information Maximization Filter
 #'
 #' @name mlr_filters_jmim
 #'
-#' @description Minimal joint mutual information maximisation filter calling
+#' @description Minimal joint mutual information maximization filter calling
 #' [praznik::JMIM()] in package \CRANpkg{praznik}.
 #'
 #' This filter supports partial scoring (see [Filter]).
@@ -40,6 +40,7 @@ FilterJMIM = R6Class("FilterJMIM",
         param_set = param_set,
         packages = "praznik",
         feature_types = c("integer", "numeric", "factor", "ordered"),
+        label = "Minimal Joint Mutual Information Maximization",
         man = "mlr3filters::mlr_filters_jmim"
       )
     }

diff --git a/R/FilterKruskalTest.R b/R/FilterKruskalTest.R
@@ -42,6 +42,7 @@ FilterKruskalTest = R6Class("FilterKruskalTest",
         param_set = param_set,
         packages = "stats",
         feature_types = c("integer", "numeric"),
+        label = "Kruskal-Wallis Test",
         man = "mlr3filters::mlr_filters_kruskal_test"
       )
     }

diff --git a/R/FilterMIM.R b/R/FilterMIM.R
@@ -1,4 +1,4 @@
-#' @title Conditional Mutual Information Based Feature Selection Filter
+#' @title Mutual Information Maximization Filter
 #'
 #' @name mlr_filters_mim
 #'
@@ -41,6 +41,7 @@ FilterMIM = R6Class("FilterMIM",
         param_set = param_set,
         packages = "praznik",
         feature_types = c("integer", "numeric", "factor", "ordered"),
+        label = "Mutual Information Maximization",
         man = "mlr3filters::mlr_filters_mim"
       )
     }

diff --git a/R/FilterMRMR.R b/R/FilterMRMR.R
@@ -1,4 +1,4 @@
-#' @title Minimum redundancy maximal relevancy filter
+#' @title Minimum Redundancy Maximal Relevancy Filter
 #'
 #' @name mlr_filters_mrmr
 #'
@@ -41,6 +41,7 @@ FilterMRMR = R6Class("FilterMRMR",
         param_set = param_set,
         packages = "praznik",
         feature_types = c("integer", "numeric", "factor", "ordered"),
+        label = "Minimum Redundancy Maximal Relevancy",
         man = "mlr3filters::mlr_filters_mrmr"
       )
     }

diff --git a/R/FilterNJMIM.R b/R/FilterNJMIM.R
@@ -1,8 +1,8 @@
-#' @title Minimal Normalised Joint Mutual Information Maximisation Filter
+#' @title Minimal Normalised Joint Mutual Information Maximization Filter
 #'
 #' @name mlr_filters_njmim
 #'
-#' @description Minimal normalised joint mutual information maximisation filter
+#' @description Minimal normalised joint mutual information maximization filter
 #' calling [praznik::NJMIM()] from package \CRANpkg{praznik}.
 #'
 #' This filter supports partial scoring (see [Filter]).
@@ -40,6 +40,7 @@ FilterNJMIM = R6Class("FilterNJMIM",
         param_set = param_set,
         packages = "praznik",
         feature_types = c("integer", "numeric", "factor", "ordered"),
+        label = "Minimal Normalised Joint Mutual Information Maximization",
         man = "mlr3filters::mlr_filters_njmim"
       )
     }

diff --git a/R/FilterPerformance.R b/R/FilterPerformance.R
@@ -51,6 +51,7 @@ FilterPerformance = R6Class("FilterPerformance",
         param_set = learner$param_set,
         feature_types = learner$feature_types,
         packages = packages,
+        label = "Predictive Performance",
         man = "mlr3filters::mlr_filters_performance"
       )
     }

diff --git a/R/FilterPermutation.R b/R/FilterPermutation.R
@@ -1,4 +1,4 @@
-#' @title Permutation Filter
+#' @title Permutation Score Filter
 #'
 #' @name mlr_filters_permutation
 #'
@@ -59,6 +59,7 @@ FilterPermutation = R6Class("FilterPermutation",
         feature_types = learner$feature_types,
         packages = packages,
         param_set = param_set,
+        label = "Permutation Score",
         man = "mlr3filters::mlr_filters_performance"
       )
     }

diff --git a/R/FilterRelief.R b/R/FilterRelief.R
@@ -1,4 +1,4 @@
-#' @title Information Gain Filter
+#' @title RELIEF Filter
 #'
 #' @name mlr_filters_relief
 #'
@@ -33,6 +33,7 @@ FilterRelief = R6Class("FilterRelief",
         param_set = param_set,
         feature_types = c("integer", "numeric", "factor", "ordered"),
         packages = "FSelectorRcpp",
+        label = "RELIEF",
         man = "mlr3filters::mlr_filters_relief"
       )
     }

diff --git a/R/FilterSelectedFeatures.R b/R/FilterSelectedFeatures.R
@@ -46,6 +46,7 @@ FilterSelectedFeatures = R6Class("FilterSelectedFeatures",
         feature_types = learner$feature_types,
         packages = learner$packages,
         param_set = learner$param_set,
+        label = "Embedded Feature Selection",
         man = "mlr3filters::mlr_filters_selected_features"
       )
     }

diff --git a/R/FilterVariance.R b/R/FilterVariance.R
@@ -39,6 +39,7 @@ FilterVariance = R6Class("FilterVariance",
         param_set = param_set,
         packages = "stats",
         feature_types = c("integer", "numeric"),
+        label = "Variance",
         man = "mlr3filters::mlr_filters_variance"
       )
     }

diff --git a/R/helper.R b/R/helper.R
@@ -8,3 +8,7 @@ call_praznik = function(self, task, fun, nfeat) {
 
   set_names(seq(from = 1, to = 0, length.out = length(selection)), names(selection))
 }
+
+catn = function(..., file = "") {
+  cat(paste0(..., collapse = "\n"), "\n", sep = "", file = file)
+}
diff --git a/R/mlr_filters.R b/R/mlr_filters.R
@@ -27,16 +27,16 @@ mlr_filters = DictionaryFilter = R6Class("DictionaryFilter",
 
 
 #' @export
-as.data.table.DictionaryFilter = function(x, ...) {
+as.data.table.DictionaryFilter = function(x, ..., objects = FALSE) {
+  assert_flag(objects)
+
   setkeyv(map_dtr(x$keys(), function(key) {
-    l = x$get(key)
-    list(
-      key = key,
-      task_type = list(l$task_type),
-      task_properties = list(l$task_properties),
-      param_set = list(l$param_set),
-      feature_types = list(l$feature_types),
-      packages = list(l$packages)
+    f = x$get(key)
+    insert_named(
+      list(key = key, label = f$label, task_type = list(f$task_type),
+        task_properties = list(f$task_properties), params = list(f$param_set$ids()),
+        feature_types = list(f$feature_types), packages = list(f$packages)),
+      if (objects) list(object = list(f))
     )
   }), "key")[]
 }
diff --git a/man-roxygen/field_feature_types.R b/man-roxygen/field_feature_types.R
diff --git a/man-roxygen/field_id.R b/man-roxygen/field_id.R
diff --git a/man-roxygen/field_packages.R b/man-roxygen/field_packages.R
diff --git a/man-roxygen/field_param_set.R b/man-roxygen/field_param_set.R
diff --git a/man-roxygen/field_task_properties.R b/man-roxygen/field_task_properties.R
diff --git a/man-roxygen/field_task_type.R b/man-roxygen/field_task_type.R