Use condense_control()

[EmilHvitfeldt]

This PR reflects the changes made in tidymodels/parsnip#808.

This PR should close #449, the error isn't perfect, but it is a lot better

library(tidymodels)

rec_no_tune_1 <-
  recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_predictors())

set.seed(363)
mt_folds <- vfold_cv(mtcars, v = 5)

glmn_mod <- decision_tree(tree_depth = tune(), min_n = 2) %>% 
  set_mode("regression") %>%
  set_engine("rpart")

workflow() %>%
  add_recipe(rec_no_tune_1) %>%
  add_model(glmn_mod) %>%
  tune_bayes(mt_folds, control = control_grid())
#> Error in `parsnip::condense_control()` at tune/R/tune_bayes.R:195:4:
#> ! Object of class `control_grid` cannot be corresed to object of class `control_bayes`.
#> • The following arguments are missing:
#> • 'no_improve', 'uncertain', 'seed', 'time_limit', and 'save_gp_scoring'

^{Created on 2022-09-14 by the reprex package (v2.0.1)}

This means that we get more flexibility in cases where we need to pass a control object from function to function.

library(tidymodels)

rec_no_tune_1 <-
  recipe(mpg ~ ., data = mtcars) %>%
  step_normalize(all_predictors())

mt_folds <- vfold_cv(mtcars, v = 5)

tree_mod_no_tune <- decision_tree(min_n = 3) %>% 
  set_mode("regression") %>%
  set_engine("rpart")

tree_wf_no_tune <- workflow() %>%
  add_recipe(rec_no_tune_1) %>%
  add_model(tree_mod_no_tune)

# fit_resamples
tree_wf_no_tune %>%
  fit_resamples(mt_folds, control = control_bayes())
#> # Resampling results
#> # 5-fold cross-validation 
#> # A tibble: 5 × 4
#>   splits         id    .metrics         .notes          
#>   <list>         <chr> <list>           <list>          
#> 1 <split [25/7]> Fold1 <tibble [2 × 4]> <tibble [0 × 3]>
#> 2 <split [25/7]> Fold2 <tibble [2 × 4]> <tibble [0 × 3]>
#> 3 <split [26/6]> Fold3 <tibble [2 × 4]> <tibble [0 × 3]>
#> 4 <split [26/6]> Fold4 <tibble [2 × 4]> <tibble [0 × 3]>
#> 5 <split [26/6]> Fold5 <tibble [2 × 4]> <tibble [0 × 3]>

^{Created on 2022-09-14 by the reprex package (v2.0.1)}

[EmilHvitfeldt]

As far as I can tell, the CI errors are fixed in #541

[EmilHvitfeldt]

Changes has been merged :)

[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.