error on predict(type = "prob") with outcome level named "class"

[simonpcouch]

Closes #720.

One thing I'd appreciate an eye for from folks who have a better sense for model_spec objects than I do: do spec$lvl and spec$preproc$y_var always exist for relevant cases here? Do we need to unit test this more thoroughly with other model specs?

library(parsnip)
library(tibble)

x <- tibble(
  boop = factor(sample(c("class", "class_1"), 100, replace = TRUE)),
  bop = rnorm(100),
  beep = rnorm(100)
)

mod <- logistic_reg() %>%
  set_mode(mode = "classification") %>%
  fit(boop ~ bop + beep, data = x)

predict(mod, type = "prob", new_data = x)
#> Error:
#> ! The outcome variable `boop` has a level called 'class'. This level is reserved for parsnip's classification internals; please adjust the levels to use a different value.

^{Created on 2022-05-11 by the reprex package (v2.0.1)}

[juliasilge]

I believe that object$lvl is one of the required bits that gets returned in a model_fit. You can check that out more here.

[juliasilge]

Looks great! 🚀 Apologies for being so back-and-forth on the error message.

[simonpcouch]

No worries at all—thanks for the review!

[github-actions]

This pull request has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.