remove new_data argument from predict.censoring_model_reverse_km()

[EmilHvitfeldt]

To close #965

I removed new_data from the function signature since it is swalloed up by ... anyways. I also added a little warning, since using new_data for predict methods are quite a common action

library(censored)

mod_fit <-
  survival_reg() %>%
  fit(Surv(time, status) ~ age + sex, data = lung)

# can use different data (because reverse_km model uses ~ 1)
pred_newdata <- predict(mod_fit$censor_probs, time = (7:10) * 100,
                        new_data = mtcars)
#> Warning: `new_data` is ignored when predicting on `censoring_model_reverse_km()` models
#> as it doesn't affect the result.

^{Created on 2023-11-17 with reprex v2.0.2}

[hfrick]

predict.censoring_model_reverse_km() already went to CRAN so typically we'd deprecate instead of directly remove.

It's pretty obscure so I'd say a deprecate_stop() is okay instead of kicking things off with deprecate_warn().

If we're not using the dots, we should check that they are empty. But I don't mind if we make that another issue 🗒️

[EmilHvitfeldt]

I switched to deprecate_stop() now

library(censored)

mod_fit <-
  survival_reg() %>%
  fit(Surv(time, status) ~ age + sex, data = lung)

pred_times <- (7:10) * 100
pred_df <- predict(mod_fit$censor_probs, time = pred_times, new_data = lung)
#> Error:
#> ! The `new_data` argument of `predict.censoring_model_reverse_km()` was
#>   deprecated in parsnip 1.2.0 and is now defunct.

[hfrick]

💯 thank you! Could you add a NEWS bullet on this before merging?

[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.