Use condense_control()

DESCRIPTION

@@ -27,7 +27,7 @@ Imports:
     GPfit,
     hardhat (>= 1.2.0),
     lifecycle (>= 1.0.0),
-    parsnip (>= 1.0.0),
+    parsnip (>= 1.0.1.9000),
     purrr (>= 0.3.2),
     recipes (>= 1.0.0),
     rlang (>= 1.0.2),
@@ -54,4 +54,6 @@ Encoding: UTF-8
 Language: en-US
 LazyData: true
 Roxygen: list(markdown = TRUE)
-RoxygenNote: 7.2.1.9000
+RoxygenNote: 7.2.0.9000
+Remotes:  
+    tidymodels/parsnip

NEWS.md

@@ -1,5 +1,7 @@
 # tune (development version)
 
+* `last_fit()`, `fit_resamples()`, `tune_grid()`, and `tune_bayes()` do not automatically error if  the wrong type of `control` object is passed. If the passed control object is not a superset of the one that is needed, the function will still error. As an example, passing `control_grid()` to `tune_bayes()` will fail but passing `control_bayes()` to `tune_grid()` will not. ([#449](https://github.com/tidymodels/tune/issues/449))
+
 # tune 1.0.0
 
 * `show_notes()` is a new function that can better help understand warnings and errors. 

R/control.R

@@ -1,8 +1,6 @@
 #' Control aspects of the grid search process
 #'
 #' @inheritParams control_bayes
-#' @param allow_par A logical to allow parallel processing (if a parallel
-#'   backend is registered).
 #'
 #' @details
 #'
@@ -33,6 +31,9 @@ control_grid <- function(verbose = FALSE, allow_par = TRUE,
                          pkgs = NULL, save_workflow = FALSE,
                          event_level = "first",
                          parallel_over = NULL) {
+  # Any added arguments should also be added in superset control functions
+  # in other packages
+
   # add options for  seeds per resample
 
   val_class_and_single(verbose, "logical", "control_grid()")
@@ -84,6 +85,9 @@ control_last_fit <- function(
     verbose = FALSE,
     event_level = "first"
 ) {
+  # Any added arguments should also be added in superset control functions
+  # in other packages
+
   extr <- function(x) x
   control <-
     control_resamples(
@@ -148,6 +152,8 @@ print.control_last_fit <- function(x, ...) {
 #'   `"everything"` describing how to use parallel processing. Alternatively,
 #'   `NULL` is allowed, which chooses between `"resamples"` and `"everything"`
 #'   automatically.
+#' @param allow_par A logical to allow parallel processing (if a parallel
+#'   backend is registered).
 #'
 #'   If `"resamples"`, then tuning will be performed in parallel over resamples
 #'   alone. Within each resample, the preprocessor (i.e. recipe or formula) is
@@ -201,7 +207,11 @@ control_bayes <-
            save_workflow = FALSE,
            save_gp_scoring = FALSE,
            event_level = "first",
-           parallel_over = NULL) {
+           parallel_over = NULL,
+           allow_par = TRUE) {
+    # Any added arguments should also be added in superset control functions
+    # in other packages
+
     # add options for seeds per resample
 
     val_class_and_single(verbose, "logical", "control_bayes()")
@@ -216,6 +226,8 @@ control_bayes <-
     val_class_or_null(pkgs, "character", "control_bayes()")
     val_class_and_single(event_level, "character", "control_bayes()")
     val_parallel_over(parallel_over, "control_bayes()")
+    val_class_and_single(allow_par, "logical", "control_bayes()")
+
 
     if (!is.infinite(uncertain) && uncertain > no_improve) {
       cli::cli_alert_warning(
@@ -226,6 +238,7 @@ control_bayes <-
     res <-
       list(
         verbose = verbose,
+        allow_par = allow_par,
         no_improve = no_improve,
         uncertain = uncertain,
         seed = seed,

R/last_fit.R

@@ -83,6 +83,8 @@ last_fit.model_spec <- function(object, preprocessor, split, ..., metrics = NULL
     ))
   }
 
+  control <- parsnip::condense_control(control, control_last_fit())
+
   empty_ellipses(...)
 
   wflow <- add_model(workflow(), object)
@@ -101,6 +103,9 @@ last_fit.model_spec <- function(object, preprocessor, split, ..., metrics = NULL
 #' @export
 last_fit.workflow <- function(object, split, ..., metrics = NULL, control = control_last_fit()) {
   empty_ellipses(...)
+
+  control <- parsnip::condense_control(control, control_last_fit())
+
   last_fit_workflow(object, split, metrics, control)
 }
 

R/resample.R

@@ -81,6 +81,8 @@ fit_resamples.model_spec <- function(object,
     ))
   }
 
+  control <- parsnip::condense_control(control, control_resamples())
+
   empty_ellipses(...)
 
   wflow <- add_model(workflow(), object)
@@ -109,6 +111,8 @@ fit_resamples.workflow <- function(object,
                                    control = control_resamples()) {
   empty_ellipses(...)
 
+  control <- parsnip::condense_control(control, control_resamples())
+
   res <-
     resample_workflow(
       workflow = object,

R/tune_bayes.R

@@ -160,6 +160,8 @@ tune_bayes.model_spec <- function(object,
     ))
   }
 
+  control <- parsnip::condense_control(control, control_bayes())
+
   wflow <- add_model(workflow(), object)
 
   if (is_recipe(preprocessor)) {
@@ -190,6 +192,8 @@ tune_bayes.workflow <-
            initial = 5,
            control = control_bayes()) {
 
+    control <- parsnip::condense_control(control, control_bayes())
+
     res <-
       tune_bayes_workflow(
         object,

R/tune_grid.R

@@ -257,6 +257,8 @@ tune_grid.model_spec <- function(object, preprocessor, resamples, ...,
     ))
   }
 
+  control <- parsnip::condense_control(control, control_grid())
+
   empty_ellipses(...)
 
   wflow <- add_model(workflow(), object)
@@ -283,6 +285,8 @@ tune_grid.workflow <- function(object, resamples, ..., param_info = NULL,
                                grid = 10, metrics = NULL, control = control_grid()) {
   empty_ellipses(...)
 
+  control <- parsnip::condense_control(control, control_grid())
+
   # Disallow `NULL` grids in `tune_grid()`, as this is the special signal
   # used when no tuning is required
   if (is.null(grid)) {

tests/testthat/_snaps/bayes.md

@@ -127,46 +127,25 @@
       ! validation: internal: A correlation computation is required, but `estimate` is constant and ha...
       ! For the rsq estimates, 1 missing value was found and removed before fitting
         the Gaussian process model.
-      ! The Gaussian process model is being fit using 1 features but only has 2
-        data points to do so. This may cause errors or a poor model fit.
-      ! Gaussian process model: X should be in range (0, 1)
-      ! For the rsq estimates, 1 missing value was found and removed before fitting
-        the Gaussian process model.
-      ! Gaussian process model: X should be in range (0, 1)
       ! validation: internal: A correlation computation is required, but `estimate` is constant and ha...
       ! For the rsq estimates, 2 missing values were found and removed before
         fitting the Gaussian process model.
-      ! Gaussian process model: X should be in range (0, 1)
       ! validation: internal: A correlation computation is required, but `estimate` is constant and ha...
       ! For the rsq estimates, 3 missing values were found and removed before
         fitting the Gaussian process model.
-      ! Gaussian process model: X should be in range (0, 1)
       ! validation: internal: A correlation computation is required, but `estimate` is constant and ha...
       ! For the rsq estimates, 4 missing values were found and removed before
         fitting the Gaussian process model.
-      ! Gaussian process model: X should be in range (0, 1)
       ! validation: internal: A correlation computation is required, but `estimate` is constant and ha...
       ! For the rsq estimates, 5 missing values were found and removed before
         fitting the Gaussian process model.
-      ! Gaussian process model: X should be in range (0, 1)
       ! validation: internal: A correlation computation is required, but `estimate` is constant and ha...
       ! For the rsq estimates, 6 missing values were found and removed before
         fitting the Gaussian process model.
-      ! Gaussian process model: X should be in range (0, 1)
       ! validation: internal: A correlation computation is required, but `estimate` is constant and ha...
       ! For the rsq estimates, 7 missing values were found and removed before
         fitting the Gaussian process model.
-      ! Gaussian process model: X should be in range (0, 1)
-      ! validation: internal: A correlation computation is required, but `estimate` is constant and ha...
-      ! For the rsq estimates, 8 missing values were found and removed before
-        fitting the Gaussian process model.
-      ! Gaussian process model: X should be in range (0, 1)
-      ! validation: internal: A correlation computation is required, but `estimate` is constant and ha...
-      ! For the rsq estimates, 9 missing values were found and removed before
-        fitting the Gaussian process model.
-      ! Gaussian process model: X should be in range (0, 1)
       ! validation: internal: A correlation computation is required, but `estimate` is constant and ha...
-      ! No improvement for 10 iterations; returning current results.
 
 ---
 

tests/testthat/test-condense_control.R

@@ -0,0 +1,52 @@
+test_that("control functions respect hierarchy with condense_control", {
+
+  ctrl <- parsnip::condense_control(
+    control_grid(),
+    control_resamples()
+  )
+
+  expect_equal(
+    ctrl,
+    control_resamples()
+  )
+
+  ctrl <- parsnip::condense_control(
+    control_last_fit(),
+    control_grid()
+  )
+
+  expect_equal(
+    ctrl,
+    control_grid(extract = control_last_fit()$extract, save_pred = TRUE)
+  )
+
+  ctrl <- parsnip::condense_control(
+    control_last_fit(),
+    control_resamples()
+  )
+
+  expect_equal(
+    ctrl,
+    control_resamples(extract = control_last_fit()$extract, save_pred = TRUE)
+  )
+
+  ctrl <- parsnip::condense_control(
+    control_bayes(),
+    control_grid()
+  )
+
+  expect_equal(
+    ctrl,
+    control_grid()
+  )
+
+  ctrl <- parsnip::condense_control(
+    control_bayes(),
+    control_resamples()
+  )
+
+  expect_equal(
+    ctrl,
+    control_resamples()
+  )
+})