Warn user for invalid formula

[strengejacke]

The message when standardizing failed due to "problematic formulas" was not always clear. Now it should be clearer to users why model cannot be standardized.
Inspired by this SO post: https://stackoverflow.com/questions/79207876/variable-names-and-easystats-reports

data(mtcars)
m <- lm(mpg ~ hp, data = mtcars)
datawizard::standardise(m)
#> 
#> Call:
#> lm(formula = mpg ~ hp, data = data_std)
#> 
#> Coefficients:
#> (Intercept)           hp  
#>  -3.149e-17   -7.762e-01

colnames(mtcars)[1] <- "1_mpg"
m <- lm(`1_mpg` ~ hp, data = mtcars)
datawizard::standardise(m)
#> Warning: Looks like you are using invalid syntactically variables names, quoted
#>   in backticks: `1_mpg`. This may result in unexpected behaviour. Please
#>   rename your variables (e.g., `X1_mpg` instead of `1_mpg`) and fit the
#>   model again.
#> Model cannot be standardized.
#> 
#> Call:
#> lm(formula = `1_mpg` ~ hp, data = mtcars)
#> 
#> Coefficients:
#> (Intercept)           hp  
#>    30.09886     -0.06823

data(mtcars)
m <- lm(mtcars$mpg ~ mtcars$hp)
datawizard::standardise(m)
#> Warning: Using `$` in model formulas can produce unexpected results. Specify your
#>   model using the `data` argument instead.
#>   Try: mpg ~ hp, data = mtcars
#> Model cannot be standardized.
#> 
#> Call:
#> lm(formula = mtcars$mpg ~ mtcars$hp)
#> 
#> Coefficients:
#> (Intercept)    mtcars$hp  
#>    30.09886     -0.06823

m <- lm(mtcars[, 1] ~ hp, data = mtcars)
datawizard::standardise(m)
#> Warning: Using indexed data frames, such as `df[, 5]`, as model response can
#>   produce unexpected results. Specify your model using the literal name of
#>   the response variable instead.
#> Model cannot be standardized.
#> 
#> Call:
#> lm(formula = mtcars[, 1] ~ hp, data = mtcars)
#> 
#> Coefficients:
#> (Intercept)           hp  
#>    30.09886     -0.06823

^{Created on 2024-11-21 with reprex v2.1.1}

[etiennebacher]

Nice messages, very clear! I have a doubt on the behavior when formula is not ok, and there are some tests to fix but otherwise LGTM.

---

Edit: actually shouldn't it be "using syntactically invalid variables names" instead of "using invalid syntactically variables names"?

[strengejacke]

"using syntactically invalid variables names" instead of "using invalid syntactically variables names"?

yeah, but that's an issue in insight ;-) will fix.

[strengejacke]

Just saw, some tests are out of date, because these code-lines are now no longer reached for the exceptions. Will look at it.

[strengejacke]

data(mtcars)
m <- lm(mpg ~ hp, data = mtcars)
datawizard::standardise(m)
#> 
#> Call:
#> lm(formula = mpg ~ hp, data = data_std)
#> 
#> Coefficients:
#> (Intercept)           hp  
#>  -3.149e-17   -7.762e-01

colnames(mtcars)[1] <- "1_mpg"
m <- lm(`1_mpg` ~ hp, data = mtcars)
datawizard::standardise(m)
#> Error: Model cannot be standardized.
#>   Looks like you are using syntactically invalid variables names, quoted
#>   in backticks: `1_mpg`. This may result in unexpected behaviour. Please
#>   rename your variables (e.g., `X1_mpg` instead of `1_mpg`) and fit the
#>   model again.

data(mtcars)
m <- lm(mtcars$mpg ~ mtcars$hp)
datawizard::standardise(m)
#> Error: Model cannot be standardized.
#>   Using `$` in model formulas can produce unexpected results. Specify your
#>   model using the `data` argument instead.
#>   Try: mpg ~ hp, data = mtcars

m <- lm(mtcars[, 1] ~ hp, data = mtcars)
datawizard::standardise(m)
#> Error: Model cannot be standardized.
#>   Using indexed data frames, such as `df[, 5]`, as model response can
#>   produce unexpected results. Specify your model using the literal name of
#>   the response variable instead.

^{Created on 2024-11-21 with reprex v2.1.1}

[etiennebacher]

Thanks!

[strengejacke]

I fixed "variables names" in insight, should be "variable names" in the error message.