Merge branch 'main' into voting_methods

mlr-org · Nov 18, 2024 · 3bca203 · 3bca203
2 parents 0e2f93f + 39d0f6a
commit 3bca203
Show file tree

Hide file tree

Showing 12 changed files with 59 additions and 37 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,6 +1,6 @@
 Package: mlr3fselect
 Title: Feature Selection for 'mlr3'
-Version: 1.1.1.9000
+Version: 1.2.1.9000
 Authors@R: c(
     person("Marc", "Becker", , "[email protected]", role = c("aut", "cre"),
            comment = c(ORCID = "0000-0002-8115-0400")),
@@ -27,7 +27,7 @@ Depends:
     mlr3 (>= 0.21.1),
     R (>= 3.1.0)
 Imports:
-    bbotk (>= 1.0.0),
+    bbotk (>= 1.2.0),
     checkmate (>= 2.0.0),
     data.table,
     lgr,
@@ -43,8 +43,6 @@ Suggests:
     mlr3pipelines,
     rpart,
     testthat (>= 3.0.0)
-Remotes:
-    mlr-org/bbotk
 Config/testthat/edition: 3
 Config/testthat/parallel: true
 Encoding: UTF-8

diff --git a/NEWS.md b/NEWS.md
@@ -1,6 +1,13 @@
 # mlr3fselect (development version)
 
+# mlr3fselect 1.2.1
+
+* compatibility: mlr3 0.22.0
+
+# mlr3fselect 1.2.0
+
 * feat: Add internal tuning callback `mlr3fselect.internal_tuning`.
+* fix: Register mlr3fselect in the `mlr_reflections$loaded_packages` field.
 
 # mlr3fselect 1.1.1
 

diff --git a/R/AutoFSelector.R b/R/AutoFSelector.R
@@ -37,6 +37,7 @@
 #' @template param_check_values
 #' @template param_callbacks
 #' @template param_ties_method
+#' @template param_id
 #'
 #' @export
 #' @examples
@@ -123,7 +124,8 @@ AutoFSelector = R6Class("AutoFSelector",
       store_models = FALSE,
       check_values = FALSE,
       callbacks = NULL,
-      ties_method = "least_features"
+      ties_method = "least_features",
+      id = NULL
       ) {
       ia = list()
       self$fselector = assert_r6(fselector, "FSelector")$clone()
@@ -141,8 +143,10 @@ AutoFSelector = R6Class("AutoFSelector",
       ia$ties_method = assert_choice(ties_method, c("least_features", "random"))
       self$instance_args = ia
 
+      id = assert_string(id, null.ok = TRUE) %??% paste0(learner$id, ".fselector")
+
       super$initialize(
-        id = paste0(learner$id, ".fselector"),
+        id = id,
         task_type = learner$task_type,
         packages = c("mlr3fselect", learner$packages),
         feature_types = learner$feature_types,

diff --git a/R/auto_fselector.R b/R/auto_fselector.R
@@ -20,6 +20,7 @@
 #' @template param_check_values
 #' @template param_callbacks
 #' @template param_ties_method
+#' @template param_id
 #'
 #' @export
 #' @inherit AutoFSelector examples
@@ -36,7 +37,8 @@ auto_fselector = function(
   store_models = FALSE,
   check_values = FALSE,
   callbacks = NULL,
-  ties_method = "least_features"
+  ties_method = "least_features",
+  id = NULL
   ) {
   terminator = terminator %??% terminator_selection(term_evals, term_time)
 
@@ -51,5 +53,6 @@ auto_fselector = function(
     store_models = store_models,
     check_values = check_values,
     callbacks = callbacks,
-    ties_method = ties_method)
+    ties_method = ties_method,
+    id = id)
 }
diff --git a/R/mlr_callbacks.R b/R/mlr_callbacks.R
@@ -153,7 +153,7 @@ load_callback_one_se_rule = function() {
     label = "One Standard Error Rule Callback",
     man = "mlr3fselect::mlr3fselect.one_se_rule",
 
-    on_result = function(callback, context) {
+    on_optimization_end = function(callback, context) {
       archive = context$instance$archive
       data = as.data.table(archive)
       data[, "n_features" := map(get("features"), length)]
@@ -203,7 +203,7 @@ load_callback_internal_tuning = function() {
         callback$state$internal_search_space$aggr_internal_tuned_values(internal_tuned_values)
       })
 
-        data.table::set(context$aggregated_performance, j = "internal_tuned_values", value = list(internal_tuned_values))
+      data.table::set(context$aggregated_performance, j = "internal_tuned_values", value = list(internal_tuned_values))
     },
 
     on_optimization_end = function(callback, context) {

diff --git a/man/AutoFSelector.Rd b/man/AutoFSelector.Rd
diff --git a/man/auto_fselector.Rd b/man/auto_fselector.Rd
diff --git a/tests/testthat/test_AutoFSelector.R b/tests/testthat/test_AutoFSelector.R
@@ -189,6 +189,7 @@ test_that("AutoFSelector get_base_learner method works", {
 
 test_that("AutoFSelector hash works #647 in mlr3", {
   afs_1 = AutoFSelector$new(
+    id = "afs_1",
     learner = lrn("classif.rpart"),
     resampling = rsmp("holdout"),
     measure = msr("classif.ce"),
@@ -197,6 +198,7 @@ test_that("AutoFSelector hash works #647 in mlr3", {
     store_benchmark_result = FALSE)
 
   afs_2 = AutoFSelector$new(
+    id = "afs_2",
     learner = lrn("classif.rpart"),
     resampling = rsmp("holdout"),
     measure = msr("classif.ce"),

diff --git a/tests/testthat/test_FSelectorShadowVariableSearch.R b/tests/testthat/test_FSelectorShadowVariableSearch.R
@@ -43,7 +43,7 @@ test_that("second selected feature is a shadow variable works", {
   expect_equal(instance$search_space, domain)
   expect_equal(instance$objective$domain, domain)
   # check that task is restored
-  expect_equal(instance$objective$task, task)
+  suppressWarnings(expect_equal(instance$objective$task, task))
 })
 
 test_that("search is terminated by terminator works", {

diff --git a/tests/testthat/test_extract_inner_fselect_archives.R b/tests/testthat/test_extract_inner_fselect_archives.R
@@ -17,8 +17,8 @@ test_that("extract_inner_fselect_archives function works with resample and repea
 })
 
 test_that("extract_inner_fselect_archives function works with benchmark and cv", {
-  at_1 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
-  at_2 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
+  at_1 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_1")
+  at_2 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_2")
   resampling_outer = rsmp("cv", folds = 2)
   grid = benchmark_grid(tsk("iris"), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = TRUE)
@@ -30,8 +30,8 @@ test_that("extract_inner_fselect_archives function works with benchmark and cv",
 })
 
 test_that("extract_inner_fselect_archives function works with benchmark and repeated cv", {
-  at_1 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
-  at_2 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
+  at_1 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_1")
+  at_2 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_2")
   resampling_outer = rsmp("repeated_cv", folds = 2, repeats = 3)
   grid = benchmark_grid(tsk("iris"), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = TRUE)
@@ -43,8 +43,8 @@ test_that("extract_inner_fselect_archives function works with benchmark and repe
 })
 
 test_that("extract_inner_fselect_archives function works with multiple tasks", {
-  at_1 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
-  at_2 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
+  at_1 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_1")
+  at_2 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_2")
   resampling_outer = rsmp("cv", folds = 2)
   grid = benchmark_grid(list(tsk("iris"), tsk("pima")), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = TRUE)
@@ -73,8 +73,8 @@ test_that("extract_inner_fselect_archives function works with no instance", {
 })
 
 test_that("extract_inner_fselect_archives function works with benchmark and no models", {
-  at_1 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
-  at_2 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
+  at_1 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_1")
+  at_2 = auto_fselector(fs("random_search", batch_size = 1), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_2")
   resampling_outer = rsmp("cv", folds = 2)
   grid = benchmark_grid(tsk("iris"), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = FALSE)
@@ -84,9 +84,9 @@ test_that("extract_inner_fselect_archives function works with benchmark and no m
 
 test_that("extract_inner_fselect_archives function works with mixed store instance", {
   at_1 = AutoFSelector$new(lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), trm("evals", n_evals = 4),
-  fselector = fs("random_search", batch_size = 1), store_fselect_instance = FALSE, store_benchmark_result = FALSE)
+  fselector = fs("random_search", batch_size = 1), store_fselect_instance = FALSE, store_benchmark_result = FALSE, id = "at_1")
   at_2 = AutoFSelector$new(lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), trm("evals", n_evals = 4),
-  fselector = fs("random_search", batch_size = 1))
+  fselector = fs("random_search", batch_size = 1), id = "at_2")
   resampling_outer = rsmp("cv", folds = 2)
   grid = benchmark_grid(tsk("iris"), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = TRUE)

diff --git a/tests/testthat/test_extract_inner_fselect_result.R b/tests/testthat/test_extract_inner_fselect_result.R
@@ -17,8 +17,8 @@ test_that("extract_inner_fselect_results function works with resample and repeat
 })
 
 test_that("extract_inner_fselect_results function works with benchmark and cv", {
-  at_1 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
-  at_2 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
+  at_1 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_1")
+  at_2 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_2")
   resampling_outer = rsmp("cv", folds = 2)
   grid = benchmark_grid(tsk("iris"), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = TRUE)
@@ -30,8 +30,8 @@ test_that("extract_inner_fselect_results function works with benchmark and cv",
 })
 
 test_that("extract_inner_fselect_results function works with benchmark and repeated cv", {
-  at_1 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
-  at_2 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
+  at_1 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_1")
+  at_2 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_2")
   resampling_outer = rsmp("repeated_cv", folds = 2, repeats = 3)
   grid = benchmark_grid(tsk("iris"), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = TRUE)
@@ -43,8 +43,8 @@ test_that("extract_inner_fselect_results function works with benchmark and repea
 })
 
 test_that("extract_inner_fselect_results function works with multiple tasks", {
-  at_1 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
-  at_2 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
+  at_1 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_1")
+  at_2 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_2")
   resampling_outer = rsmp("cv", folds = 2)
   grid = benchmark_grid(list(tsk("iris"), tsk("pima")), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = TRUE)
@@ -73,8 +73,8 @@ test_that("extract_inner_fselect_results function works no instance", {
 })
 
 test_that("extract_inner_fselect_results function works with benchmark and no models", {
-  at_1 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
-  at_2 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
+  at_1 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_1")
+  at_2 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_2")
   resampling_outer = rsmp("cv", folds = 2)
   grid = benchmark_grid(tsk("iris"), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = FALSE)
@@ -84,9 +84,9 @@ test_that("extract_inner_fselect_results function works with benchmark and no mo
 
 test_that("extract_inner_fselect_results function works with mixed store instance", {
   at_1 = AutoFSelector$new(lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), trm("evals", n_evals = 4),
-  fselector = fs("random_search"), store_fselect_instance = FALSE, store_benchmark_result = FALSE)
+  fselector = fs("random_search"), store_fselect_instance = FALSE, store_benchmark_result = FALSE, id = "at_1")
   at_2 = AutoFSelector$new(lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), trm("evals", n_evals = 4),
-  fselector = fs("random_search"))
+  fselector = fs("random_search"), id = "at_2")
   resampling_outer = rsmp("cv", folds = 2)
   grid = benchmark_grid(tsk("iris"), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = TRUE)
@@ -118,8 +118,8 @@ test_that("extract_inner_fselect_results function works with resample and return
 })
 
 test_that("extract_inner_fselect_results function works with benchmark and return of instance", {
-  at_1 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
-  at_2 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4)
+  at_1 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_1")
+  at_2 = auto_fselector(fs("random_search"), lrn("classif.rpart"), rsmp("holdout"), msr("classif.ce"), term_evals = 4, id = "at_2")
   resampling_outer = rsmp("cv", folds = 2)
   grid = benchmark_grid(tsk("iris"), list(at_1, at_2), resampling_outer)
   bmr = benchmark(grid, store_models = TRUE)

diff --git a/tests/testthat/test_fsi.R b/tests/testthat/test_fsi.R
@@ -38,7 +38,7 @@ test_that("fsi and FSelectInstanceBatchSingleCrit are equal", {
     instance_1 = FSelectInstanceBatchSingleCrit$new(task, learner, resampling, measures, terminator, store_benchmark_result, store_models, check_values, callbacks)
     instance_2 = fsi(task, learner, resampling, measures, terminator, store_benchmark_result, store_models, check_values, callbacks)
 
-    expect_equal(instance_1, instance_2)
+    suppressWarnings(expect_equal(instance_1, instance_2))
 })
 
 test_that("fsi and FSelectInstanceBatchMultiCrit are equal", {
@@ -61,5 +61,5 @@ test_that("fsi and FSelectInstanceBatchMultiCrit are equal", {
     instance_1 = FSelectInstanceBatchMultiCrit$new(task, learner, resampling, measures, terminator, store_benchmark_result, store_models, check_values, callbacks)
     instance_2 = fsi(task, learner, resampling, measures, terminator, store_benchmark_result, store_models, check_values, callbacks)
 
-    expect_equal(instance_1, instance_2)
+    suppressWarnings(expect_equal(instance_1, instance_2))
 })