Check prediction `type` for `multi_predict()`

[hfrick]

This should error.

library(poissonreg)
#> Loading required package: parsnip

poisson_reg(penalty = 0.01) %>%
  set_engine("glmnet") %>%
  fit(mpg ~ ., data = mtcars) %>%
  multi_predict(mtcars, type = "class")
#> # A tibble: 32 × 1
#>    .pred           
#>    <list>          
#>  1 <tibble [1 × 2]>
#>  2 <tibble [1 × 2]>
#>  3 <tibble [1 × 2]>
#>  4 <tibble [1 × 2]>
#>  5 <tibble [1 × 2]>
#>  6 <tibble [1 × 2]>
#>  7 <tibble [1 × 2]>
#>  8 <tibble [1 × 2]>
#>  9 <tibble [1 × 2]>
#> 10 <tibble [1 × 2]>
#> # ℹ 22 more rows

^{Created on 2026-04-19 with reprex v2.1.1}

In parsnip's multi_predict() for glment models, we call parsnip:::check_pred_type() which we might want to export for this.

> linear_reg(penalty = 0.01) %>%
+       set_engine("glmnet") %>%
+       fit(mpg ~ ., data = mtcars) %>%
+       multi_predict(mtcars, type = "class")
Error in `multi_predict()`:
! For class predictions, the object should be a classification model.
Hide Traceback
    ▆
 1. ├─... %>% multi_predict(mtcars, type = "class") at tests/testthat/test-glmnet-linear.R:322:5
 2. ├─parsnip::multi_predict(., mtcars, type = "class")
 3. └─parsnip:::multi_predict._elnet(., mtcars, type = "class")
 4.   └─parsnip:::check_pred_type(object, type)
 5.     └─cli::cli_abort(...)
 6.       └─rlang::abort(...)