align null_model() with other model types

NEWS.md

@@ -1,5 +1,9 @@
 # parsnip (development version)
 
+* Aligned `null_model()` with other model types; the model type now has an
+  engine argument that defaults to `"parsnip"` and is checked with the same
+  machinery that checks other model types in the package (#1083).
+
 * New `extract_fit_time()` method has been added that returns the time it took to train the model (#853).
 
 # parsnip 1.2.1

R/nullmodel.R

@@ -130,12 +130,11 @@ predict.nullmodel <- function (object, new_data = NULL, type  = NULL, ...) {
 #' `null_model()` defines a simple, non-informative model. It doesn't have any
 #'  main arguments. This function can fit classification and regression models.
 #'
-#' @inheritParams set_new_model
-#' @details The model can be created using the `fit()` function using the
-#'  following _engines_:
-#' \itemize{
-#' \item \pkg{R}:  `"parsnip"`
-#' }
+#' @param mode A single character string for the type of model. The only
+#' possible values for this model are `"regression"` and `"classification"`.
+#' @param engine A single character string specifying what computational engine
+#' to use for fitting. Possible engines are listed below. The default for this
+#' model is `"parsnip"`.
 #'
 #' @includeRmd man/rmd/null-model.md details
 #'
@@ -144,28 +143,20 @@ predict.nullmodel <- function (object, new_data = NULL, type  = NULL, ...) {
 #' null_model(mode = "regression")
 #' @export
 null_model <-
-  function(mode = "classification") {
-    null_model_modes <- unique(get_model_env()$null_model$mode)
-    # Check for correct mode
-    if (!(mode %in% null_model_modes))
-      rlang::abort(
-        glue::glue(
-          "`mode` should be one of: ",
-          glue::glue_collapse(glue::glue("'{null_model_modes}'"), sep = ", ")
-          )
-        )
-
-    # Capture the arguments in quosures
+  function(mode = "classification", engine = "parsnip") {
     args <- list()
 
-    # Save some empty slots for future parts of the specification
-    out <- list(args = args, eng_args = NULL,
-                mode = mode, method = NULL, engine = NULL)
-
-    # set classes in the correct order
-    class(out) <- make_classes("null_model")
-    out
-  }
+    new_model_spec(
+      "null_model",
+      args = args,
+      eng_args = NULL,
+      mode = mode,
+      user_specified_mode = !missing(mode),
+      method = NULL,
+      engine = engine,
+      user_specified_engine = !missing(engine)
+    )
+}
 
 
 

tests/testthat/_snaps/model_basics.md

@@ -235,6 +235,8 @@
     Output
       Null Model Specification (classification)
 
+      Computational engine: parsnip 
+
 
 ---
 

tests/testthat/_snaps/nullmodel.md

@@ -5,6 +5,8 @@
     Output
       Null Model Specification (classification)
 
+      Computational engine: parsnip 
+
 
 ---
 

tests/testthat/test_nullmodel.R

@@ -15,46 +15,63 @@ test_that('bad input', {
 
 num_pred <- names(hpc)[1:3]
 hpc_bad_form <- as.formula(class ~ term)
-hpc_basic <- null_model(mode = "regression") %>% set_engine("parsnip")
 
 # ------------------------------------------------------------------------------
 
 test_that('nullmodel execution', {
 
-  expect_error(
+  expect_no_condition(
     res <- fit(
-      hpc_basic,
+      null_model(mode = "regression") %>% set_engine("parsnip"),
       compounds ~ log(input_fields) + class,
       data = hpc
-    ),
-    regexp = NA
+    )
   )
-  expect_error(
+  expect_no_condition(
+    res <- fit(
+      null_model(mode = "regression"),
+      compounds ~ log(input_fields) + class,
+      data = hpc
+    )
+  )
+  expect_no_condition(
     res <- fit_xy(
-      hpc_basic,
+      null_model(mode = "regression") %>% set_engine("parsnip"),
       x = hpc[, num_pred],
       y = hpc$num_pending
-    ),
-    regexp = NA
+    )
+  )
+  expect_no_condition(
+    res <- fit_xy(
+      null_model(mode = "regression"),
+      x = hpc[, num_pred],
+      y = hpc$num_pending
+    )
   )
 
   expect_error(
     res <- fit(
-      hpc_basic,
+      null_model(mode = "regression") %>% set_engine("parsnip"),
       hpc_bad_form,
       data = hpc
     )
   )
 
   ## multivariate y
 
-  expect_error(
+  expect_no_condition(
     res <- fit(
-      hpc_basic,
+      null_model(mode = "regression") %>% set_engine("parsnip"),
       cbind(compounds, input_fields) ~ .,
       data = hpc
-    ),
-    regexp = NA
+    )
+  )
+  expect_no_condition(
+    res <- fit(
+      null_model(mode = "regression"),
+      cbind(compounds, input_fields) ~ .,
+      data = hpc
+    )
   )
 
 })
@@ -67,7 +84,7 @@ test_that('nullmodel prediction', {
                     carb = rep(2.8125, 5))
 
   res_xy <- fit_xy(
-    hpc_basic,
+    null_model(mode = "regression") %>% set_engine("parsnip"),
     x = hpc[, num_pred],
     y = hpc$num_pending
   )
@@ -77,7 +94,7 @@ test_that('nullmodel prediction', {
                tolerance = .01)
 
   res_form <- fit(
-    hpc_basic,
+    null_model(mode = "regression") %>% set_engine("parsnip"),
     num_pending ~ log(compounds) + class,
     data = hpc
   )
@@ -87,7 +104,7 @@ test_that('nullmodel prediction', {
 
   # Multivariate y
   res <- fit(
-    hpc_basic,
+    null_model(mode = "regression") %>% set_engine("parsnip"),
     cbind(gear, carb) ~ .,
     data = mtcars
   )