mlr3 upkeep

NEWS.md

@@ -1,5 +1,7 @@
 # mlr3torch dev
 
+* Don't use deprecated `data_formats` anymore
+
 # mlr3torch 0.1.1
 
 * fix(preprocessing): regarding the construction of some `PipeOps` such as `po("trafo_resize")`

R/DataBackendLazy.R

@@ -45,9 +45,6 @@
 #'  Desired data format, e.g. `"data.table"` or `"Matrix"`.
 #' @param na_rm (`logical(1)`)\cr
 #'   Whether to remove NAs or not.
-#' @param data_formats (`character()`)\cr
-#'   Set of supported data formats. E.g. `"data.table"`.
-#'   These must be a subset of the data formats of the lazily constructed backend.
 #' @param primary_key (`character(1)`)\cr
 #'   Name of the primary key column.
 #'
@@ -72,7 +69,6 @@
 #'   constructor = constructor,
 #'   rownames = 1:10,
 #'   col_info = column_info,
-#'   data_formats = "data.table",
 #'   primary_key = "row_id"
 #' )
 #'
@@ -94,7 +90,7 @@ DataBackendLazy = R6Class("DataBackendLazy",
   public = list(
     #' @description
     #' Creates a new instance of this [R6][R6::R6Class] class.
-    initialize = function(constructor, rownames, col_info, primary_key, data_formats) {
+    initialize = function(constructor, rownames, col_info, primary_key) {
       private$.rownames = assert_integerish(rownames, unique = TRUE)
       private$.col_info = assert_data_table(col_info, ncols = 3, min.rows = 1)
       assert_permutation(colnames(col_info), c("id", "type", "levels"))
@@ -103,7 +99,7 @@ DataBackendLazy = R6Class("DataBackendLazy",
       assert_choice(primary_key, col_info$id)
       private$.constructor = assert_function(constructor, args = "backend")
 
-      super$initialize(data = NULL, primary_key = primary_key, data_formats = data_formats)
+      super$initialize(data = NULL, primary_key = primary_key)
     },
 
     #' @description
@@ -114,8 +110,8 @@ DataBackendLazy = R6Class("DataBackendLazy",
     #' Duplicated row ids result in duplicated rows, duplicated column names lead to an exception.
     #'
     #' Accessing the data triggers the construction of the backend.
-    data = function(rows, cols, data_format = "data.table") {
-      self$backend$data(rows = rows, cols = cols, data_format = data_format)
+    data = function(rows, cols) {
+      self$backend$data(rows = rows, cols = cols)
     },
 
     #' @description
@@ -190,7 +186,6 @@ DataBackendLazy = R6Class("DataBackendLazy",
         f(test_permutation, backend$colnames, private$.colnames, "column names")
         f(test_equal_col_info, col_info(backend), private$.col_info, "column information")
         # need to reverse the order for correct error message
-        f(function(x, y) test_subset(y, x), backend$data_formats, self$data_formats, "data formats")
         private$.backend = backend
       }
       private$.backend

R/LearnerTorch.R

@@ -201,7 +201,6 @@ LearnerTorch = R6Class("LearnerTorch",
         param_set = self$param_set,
         predict_types = predict_types,
         properties = properties,
-        data_formats = "data.table",
         label = label,
         feature_types = feature_types,
         man = man

R/TaskClassif_mnist.R

@@ -75,8 +75,7 @@ load_task_mnist = function(id = "mnist") {
     constructor = cached_constructor,
     rownames = seq_len(70000),
     col_info = load_col_info("mnist"),
-    primary_key = "..row_id",
-    data_formats = "data.table"
+    primary_key = "..row_id"
   )
 
   task = TaskClassif$new(

R/TaskClassif_tiny_imagenet.R

@@ -97,8 +97,7 @@ load_task_tiny_imagenet = function(id = "tiny_imagenet") {
     constructor = cached_constructor,
     rownames = seq_len(120000),
     col_info = load_col_info("tiny_imagenet"),
-    primary_key = "..row_id",
-    data_formats = "data.table"
+    primary_key = "..row_id"
   )
 
   task = TaskClassif$new(

tests/testthat/test_DataBackendLazy.R

@@ -21,8 +21,7 @@ test_that("DataBackendLazy works", {
     constructor = constructor,
     rownames = n:1,
     col_info = column_info,
-    primary_key = "row_id",
-    data_formats = "data.table"
+    primary_key = "row_id"
   )
 
   expect_r6(backend_lazy, c("DataBackend", "DataBackendLazy"))
@@ -97,8 +96,7 @@ test_that("DataBackendLazy works", {
         constructor = constructor,
         rownames = 1:10,
         col_info = col_info,
-        primary_key = "a",
-        data_formats = "data.table"
+        primary_key = "a"
       )$backend,
       regexp = regexp
     )
@@ -114,8 +112,7 @@ test_that("primary_key must be in col_info", {
     constructor = function(backend) NULL,
     col_info = data.table(id = "a", type = "integer", levels = list(NULL)),
     rownames = 1,
-    primary_key = "b",
-    data_formats = "data.table"
+    primary_key = "b"
   ), regexp = "Must be element of")
 })
 
@@ -130,8 +127,7 @@ test_that("primary_key must be the same for backends", {
     constructor = constructor,
     col_info = data.table(id = c("x", "y"), type = rep("integer", 2), levels = list(NULL, NULL)),
     rownames = 1:5,
-    primary_key = "y",
-    data_formats = "data.table"
+    primary_key = "y"
   )
   expect_error(backend_lazy$backend, "primary key")
 })
@@ -141,7 +137,6 @@ test_that("constructor must have argument backend", {
     constructor = function() NULL,
     col_info = data.table(id = c("x", "y"), type = rep("integer", 2), levels = list(NULL, NULL)),
     rownames = 1:5,
-    primary_key = "y",
-    data_formats = "data.table"
+    primary_key = "y"
   ), regexp = "formal arguments")
 })

tests/testthat/test_LearnerTorch.R

@@ -419,9 +419,10 @@ test_that("resample() works", {
 
 test_that("marshaling", {
   task = tsk("mtcars")$filter(1:5)
-  learner = lrn("regr.mlp", batch_size = 150, epochs = 1, device = "cpu", encapsulate = c(train = "callr"),
+  learner = lrn("regr.mlp", batch_size = 150, epochs = 1, device = "cpu",
     neurons = 20
   )
+  learner$encapsulate("callr", lrn("regr.featureless"))
   learner$train(task)
   expect_false(learner$marshaled)
   learner$marshal()$unmarshal()
@@ -431,9 +432,10 @@ test_that("marshaling", {
 test_that("callr encapsulation and marshaling", {
   skip_if_not_installed("callr")
   task = tsk("mtcars")$filter(1:5)
-  learner = lrn("regr.mlp", batch_size = 150, epochs = 1, device = "cpu", encapsulate = c(train = "callr"),
+  learner = lrn("regr.mlp", batch_size = 150, epochs = 1, device = "cpu",
     neurons = 20
   )
+  learner$encapsulate("callr", lrn("regr.featureless"))
   learner$train(task)
   expect_prediction(learner$predict(task))
 })
@@ -461,7 +463,7 @@ test_that("Input verification works during `$train()` (train-predict shapes work
   )
 
   # fallback learner cannot help in this case!
-  learner$fallback = lrn("classif.featureless")
+  learner$encapsulate("evaluate", fallback = lrn("classif.featureless"))
   rr_faulty = resample(task_invalid, learner, rsmp("holdout"))
   expect_true(nrow(rr_faulty$errors) == 1L)
   rr1 = resample(task, learner, rsmp("holdout"))

tests/testthat/test_PipeOpTorchModel.R

@@ -118,7 +118,7 @@ test_that("base_learner works", {
     po("torch_model_regr")
 
   glrn = as_learner(graph)
-  expect_equal(glrn$base_learner(return_po = TRUE)$id, "torch_model_regr")
+  expect_equal(glrn$base_learner(return_po = TRUE, recursive = 1)$id, "torch_model_regr")
 })
 
 test_that("internal_tuning", {