From 2741cdc5c86dd2e1f47e45d5bf9db8d5c9db1c91 Mon Sep 17 00:00:00 2001
From: Daniel <mail@danielluedecke.de>
Date: Thu, 21 Nov 2024 19:55:19 +0100
Subject: [PATCH] Make `standardize()` error messages clearer (#562)

* Warn user for invalif formula

* add tests

* fix tests
---
 DESCRIPTION                              |  2 +-
 R/standardize.models.R                   |  8 ++++++
 tests/testthat/test-standardize_models.R | 32 ++++++++++++++++++++----
 3 files changed, 36 insertions(+), 6 deletions(-)

diff --git a/DESCRIPTION b/DESCRIPTION
index ba821b0ba..2325c062d 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,7 +1,7 @@
 Type: Package
 Package: datawizard
 Title: Easy Data Wrangling and Statistical Transformations
-Version: 0.13.0.12
+Version: 0.13.0.13
 Authors@R: c(
     person("Indrajeet", "Patil", , "patilindrajeet.science@gmail.com", role = "aut",
            comment = c(ORCID = "0000-0003-1995-6531")),
diff --git a/R/standardize.models.R b/R/standardize.models.R
index a92ffe243..cf6062c78 100644
--- a/R/standardize.models.R
+++ b/R/standardize.models.R
@@ -78,6 +78,14 @@ standardize.default <- function(x,
     return(x)
   }
 
+  # check model formula. Some notations don't work when standardizing data
+  insight::formula_ok(
+    x,
+    action = "error",
+    prefix_msg = "Model cannot be standardized.",
+    verbose = verbose
+  )
+
   data_std <- NULL # needed to avoid note
   .standardize_models(x,
     robust = robust, two_sd = two_sd,
diff --git a/tests/testthat/test-standardize_models.R b/tests/testthat/test-standardize_models.R
index 706a4e6e7..d61caf450 100644
--- a/tests/testthat/test-standardize_models.R
+++ b/tests/testthat/test-standardize_models.R
@@ -31,6 +31,29 @@ test_that("standardize | errors", {
 })
 
 
+test_that("standardize | problematic formulas", {
+  data(mtcars)
+  m <- lm(mpg ~ hp, data = mtcars)
+  expect_equal(
+    coef(standardise(m)),
+    c(`(Intercept)` = -3.14935717633686e-17, hp = -0.776168371826586),
+    tolerance = 1e-4
+  )
+
+  colnames(mtcars)[1] <- "1_mpg"
+  m <- lm(`1_mpg` ~ hp, data = mtcars)
+  expect_error(standardise(m), regex = "Looks like")
+
+  # works interactive only
+  # data(mtcars)
+  # m <- lm(mtcars$mpg ~ mtcars$hp)
+  # expect_error(standardise(m), regex = "model formulas")
+
+  m <- lm(mtcars[, 1] ~ hp, data = mtcars)
+  expect_error(standardise(m), regex = "indexed data")
+})
+
+
 # Transformations ---------------------------------------------------------
 test_that("transformations", {
   skip_if_not_installed("effectsize")
@@ -206,15 +229,14 @@ test_that("standardize non-Gaussian response", {
 # variables evaluated in the environment $$$ ------------------------------
 test_that("variables evaluated in the environment", {
   m <- lm(mtcars$mpg ~ mtcars$cyl + am, data = mtcars)
-  w <- capture_warnings(standardize(m))
-  expect_true(any(grepl("mtcars$mpg", w, fixed = TRUE)))
+  w <- capture_error(standardize(m))
+  expect_true(any(grepl("Using `$`", w, fixed = TRUE)))
 
   ## Note:
   # No idea why this is suddenly not giving a warning on older R versions.
   m <- lm(mtcars$mpg ~ mtcars$cyl + mtcars$am, data = mtcars)
-  warns <- capture_warnings(standardize(m))
-  expect_true(any(grepl("mtcars$mpg", warns, fixed = TRUE)))
-  expect_true(any(grepl("No variables", warns, fixed = TRUE)))
+  w <- capture_error(standardize(m))
+  expect_true(any(grepl("Using `$`", w, fixed = TRUE)))
 })