release: 1.4.0

.Rbuildignore

@@ -27,3 +27,4 @@
 ^.cursor$
 ^local_attic$
 ^tests/testthat/_object_snapshots$
+^.claude$

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: mlr3
 Title: Machine Learning in R - Next Generation
-Version: 1.3.0.9000
+Version: 1.4.0
 Authors@R:
   c(
     person("Michel", "Lang", , "michellang@gmail.com", role = "aut",

NEWS.md

@@ -1,6 +1,13 @@
-# mlr3 (development version)
+# mlr3 1.4.0
 
 * feat: Add `$native_model` active binding to `Learner` to access the model from the upstream package.
+* feat: Learner store the condition of warnings and errors in the `log` field.
+* feat: No supervised tasks with missing target values are allowed anymore.
+* refactor: Validate the input of fields.
+* fix: Assert list input in `assert_learners`, `assert_tasks`, `assert_measures` and `assert_resamplings`.
+* fix: `convert_task` converts the internal validation task now.
+* fix: disable printing of class ratios for large tasks.
+* fix: Make quantiles resettable.
 
 # mlr3 1.3.0
 

R/as_task_classif.R

@@ -47,7 +47,8 @@ as_task_classif.data.frame = function(x, target, id = deparse1(substitute(x)), p
   }
 
   if (anyMissing(x[[target]])) {
-    error_input("Target column '%s' must not contain missing values", target)
+    messagef("Target column '%s' contains missing values. This will be an error in the future.", target)
+    # error_input("Target column '%s' must not contain missing values", target)
   }
 
   y = x[[target]]

R/as_task_regr.R

@@ -45,7 +45,8 @@ as_task_regr.data.frame = function(x, target, id = deparse1(substitute(x)), labe
   }
 
   if (anyMissing(x[[target]])) {
-    error_input("Target column '%s' must not contain missing values", target)
+    messagef("Target column '%s' contains missing values. This will be an error in the future.", target)
+    # error_input("Target column '%s' must not contain missing values", target)
   }
 
   TaskRegr$new(id = id, backend = x, target = target, label = label)

R/mlr_reflections.R

@@ -3,7 +3,7 @@
 #' @format [environment].
 #' @description
 #' Environment which stores various information to allow objects to examine and introspect their structure and properties
-#' (c.f. [Reflections](https://www.wikiwand.com/en/Reflection_(computer_programming))).
+#' (c.f. [Reflections](https://en.wikipedia.org/wiki/Reflective_programming)).
 #'
 #' This environment be modified by third-party packages, e.g. by adding information about new task types
 #' or by extending the set of allowed feature types.

tests/testthat/_object_snapshots/snapshot.R

@@ -1,7 +1,7 @@
-task_classif = tsk("pima") # mlr3 1.2.0
+task_classif = tsk("pima") # mlr3 1.4.0
 saveRDS(task_classif, "tests/testthat/_object_snapshots/task_classif.rds")
 
-task_regr = tsk("mtcars") # mlr3 1.2.0
+task_regr = tsk("mtcars") # mlr3 1.4.0
 saveRDS(task_regr, "tests/testthat/_object_snapshots/task_regr.rds")
 
 learner_classif = lrn("classif.rpart") # mlr3 1.0.0
@@ -16,7 +16,7 @@ resampling = rsmp("cv", folds = 3) # mlr3 1.0.0
 resampling$instantiate(task_classif)
 saveRDS(resampling, "tests/testthat/_object_snapshots/resampling.rds")
 
-rr = resample(task_classif, learner_classif, resampling) # mlr3 1.2.0
+rr = resample(task_classif, learner_classif, resampling) # mlr3 1.4.0
 saveRDS(rr, "tests/testthat/_object_snapshots/rr.rds")
 
 design = benchmark_grid(
@@ -25,7 +25,7 @@ design = benchmark_grid(
   resamplings = list(resampling)
 )
 
-bmr = benchmark(design) # mlr3 1.2.0
+bmr = benchmark(design) # mlr3 1.4.0
 saveRDS(bmr, "tests/testthat/_object_snapshots/bmr.rds")
 
 measure = msr("classif.ce") # mlr3 1.0.0

tests/testthat/test_TaskClassif.R

@@ -44,7 +44,7 @@ test_that("Target is character/factor", {
 
 test_that("TaskClassif rejects NAs in target", {
   dt = data.table::data.table(x = 1:10, y = factor(c(NA, letters[1:9])))
-  expect_error(as_task_classif(dt, target = "y"), "contain missing")
+  expect_message(as_task_classif(dt, target = "y"), "contains missing")
 })
 
 test_that("0 feature task", {

tests/testthat/test_TaskRegr.R

@@ -17,7 +17,7 @@ test_that("Target is numeric", {
 
 test_that("TaskRegr rejects NAs in target", {
   dt = data.table::data.table(x = 1:10, y = c(NA_real_, 2:10))
-  expect_error(as_task_regr(dt, target = "y"), "contain missing")
+  expect_message(as_task_regr(dt, target = "y"), "contains missing")
 })
 
 test_that("TaskRegr: 0 feature task", {

tests/testthat/test_backward_compatibility.R

@@ -17,7 +17,7 @@ test_that("learner classif backward compatibility", {
   expect_learner(learner_classif)
 
   task = tsk("pima")
-  pred = learner_classif$predict(task)
+  suppressWarnings({pred = learner_classif$predict(task)})
   expect_prediction_classif(pred, task)
 
   learner_classif$train(task)
@@ -29,7 +29,7 @@ test_that("learner regr backward compatibility", {
   expect_learner(learner_regr)
 
   task = tsk("mtcars")
-  pred = learner_regr$predict(task)
+  suppressWarnings({pred = learner_regr$predict(task)})
   expect_prediction_regr(pred)
 
   learner_regr$train(task)