Merge pull request #919 from tidymodels/check-outcome

More verbose check_outcome()

NEWS.md

@@ -20,6 +20,7 @@
 
 * `logistic_reg()` will now warn at `fit()` when the outcome has more than two levels (#545).
 
+* Functions now indicate what class the outcome was if the outcome is the wrong class (#887).
 
 # parsnip 1.0.4
 

R/misc.R

@@ -336,14 +336,22 @@ check_outcome <- function(y, spec) {
   if (spec$mode == "regression") {
     outcome_is_numeric <- if (is.atomic(y)) {is.numeric(y)} else {all(map_lgl(y, is.numeric))}
     if (!outcome_is_numeric) {
-      rlang::abort("For a regression model, the outcome should be numeric.")
+      cls <- class(y)[[1]]
+      abort(paste0(
+        "For a regression model, the outcome should be `numeric`, ",
+        "not a `", cls, "`."
+      ))
     }
   }
 
   if (spec$mode == "classification") {
     outcome_is_factor <- if (is.atomic(y)) {is.factor(y)} else {all(map_lgl(y, is.factor))}
     if (!outcome_is_factor) {
-      rlang::abort("For a classification model, the outcome should be a factor.")
+      cls <- class(y)[[1]]
+      abort(paste0(
+        "For a classification model, the outcome should be a `factor`, ",
+        "not a `", cls, "`."
+      ))
     }
 
     if (inherits(spec, "logistic_reg") && is.atomic(y) && length(levels(y)) > 2) {
@@ -361,7 +369,11 @@ check_outcome <- function(y, spec) {
   if (spec$mode == "censored regression") {
     outcome_is_surv <- inherits(y, "Surv")
     if (!outcome_is_surv) {
-      rlang::abort("For a censored regression model, the outcome should be a `Surv` object.")
+      cls <- class(y)[[1]]
+      abort(paste0(
+        "For a censored regression model, the outcome should be a `Surv` object, ",
+        "not a `", cls, "`."
+      ))
     }
   }
 

tests/testthat/_snaps/misc.md

@@ -132,3 +132,27 @@
       Error in `fn()`:
       ! Please use `new_data` instead of `newdata`.
 
+# check_outcome works as expected
+
+    Code
+      check_outcome(factor(1:2), reg_spec)
+    Condition
+      Error in `check_outcome()`:
+      ! For a regression model, the outcome should be `numeric`, not a `factor`.
+
+---
+
+    Code
+      check_outcome(1:2, class_spec)
+    Condition
+      Error in `check_outcome()`:
+      ! For a classification model, the outcome should be a `factor`, not a `integer`.
+
+---
+
+    Code
+      check_outcome(1:2, cens_spec)
+    Condition
+      Error in `check_outcome()`:
+      ! For a censored regression model, the outcome should be a `Surv` object, not a `integer`.
+

tests/testthat/test_misc.R

@@ -185,6 +185,51 @@ test_that('set_engine works as a generic', {
 test_that('check_for_newdata points out correct context', {
   fn <- function(...) {check_for_newdata(...); invisible()}
   expect_snapshot(error = TRUE,
-    fn(newdata = "boop!")
+                  fn(newdata = "boop!")
+  )
+})
+
+test_that('check_outcome works as expected', {
+  reg_spec <- linear_reg()
+
+  expect_no_error(
+    check_outcome(1:2, reg_spec)
+  )
+
+  expect_no_error(
+    check_outcome(mtcars, reg_spec)
+  )
+
+  expect_snapshot(
+    error = TRUE,
+    check_outcome(factor(1:2), reg_spec)
+  )
+
+  class_spec <- logistic_reg()
+
+  expect_no_error(
+    check_outcome(factor(1:2), class_spec)
+  )
+
+  expect_no_error(
+    check_outcome(lapply(mtcars, as.factor), class_spec)
+  )
+
+  expect_snapshot(
+    error = TRUE,
+    check_outcome(1:2, class_spec)
+  )
+
+  # Fake specification to avoid having to load {censored}
+  cens_spec <- logistic_reg()
+  cens_spec$mode <- "censored regression"
+
+  expect_no_error(
+    check_outcome(survival::Surv(1, 1), cens_spec)
+  )
+
+  expect_snapshot(
+    error = TRUE,
+    check_outcome(1:2, cens_spec)
   )
 })

tests/testthat/test_nearest_neighbor_kknn.R

@@ -24,7 +24,7 @@ test_that('kknn execution', {
       x = hpc[, num_pred],
       y = hpc$input_fields
     ),
-    regexp = "outcome should be a factor"
+    regexp = "outcome should be a `factor`"
   )
 
   # nominal