`check_censor_model()` should error for models without mode `"censored regression"`

[hfrick]

.censoring_weights_graf.model_fit() checks the model fit via .check_censor_model() and since censoring weights only make sense for models with mode "censored regression", the check function should error for models with a different mode.

library(parsnip)
library(survival)

lung2 <- lung %>%
  dplyr::mutate(surv = Surv(time, status), .keep = "unused")

not_a_censor_model <- fit(linear_reg(), mpg ~ ., data = mtcars)

# this should error
parsnip:::.check_censor_model(not_a_censor_model)

# because it's used in and this now gives a different error
.censoring_weights_graf(not_a_censor_model, lung2)
#> Error:
#> ! The input should have a list column called `.pred`.
#> Backtrace:
#>     ▆
#>  1. ├─parsnip::.censoring_weights_graf(not_a_censor_model, lung2)
#>  2. └─parsnip:::.censoring_weights_graf.model_fit(...) at parsnip/R/survival-censoring-weights.R:188:2
#>  3.   └─parsnip:::.check_pred_col(predictions) at parsnip/R/survival-censoring-weights.R:223:2
#>  4.     └─rlang::abort(...) at parsnip/R/survival-censoring-weights.R:48:4

^{Created on 2023-05-17 with reprex v2.0.2}

[github-actions]

This issue 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.