add extract_fit_time() for tuning results

[simonpcouch]

Parking these changes here with no intention to merge. This PR would introduce built-in support for extract_fit_time() without requiring the user to pass extract_fit_time() to control_*(), but it's likely not worth the added complexity.

library(tidymodels)

bt_res <-
  fit_resamples(
    workflow(
      recipe(tip ~ ., taxi) %>% step_dummy(all_nominal_predictors()),
      boost_tree("classification")
    ),
    vfold_cv(taxi)
  )

# Note: no changes to the contents of the tibble
bt_res
#> # Resampling results
#> # 10-fold cross-validation 
#> # A tibble: 10 × 4
#>    splits              id     .metrics         .notes          
#>    <list>              <chr>  <list>           <list>          
#>  1 <split [9000/1000]> Fold01 <tibble [3 × 4]> <tibble [0 × 4]>
#>  2 <split [9000/1000]> Fold02 <tibble [3 × 4]> <tibble [0 × 4]>
#>  3 <split [9000/1000]> Fold03 <tibble [3 × 4]> <tibble [0 × 4]>
#>  4 <split [9000/1000]> Fold04 <tibble [3 × 4]> <tibble [0 × 4]>
#>  5 <split [9000/1000]> Fold05 <tibble [3 × 4]> <tibble [0 × 4]>
#>  6 <split [9000/1000]> Fold06 <tibble [3 × 4]> <tibble [0 × 4]>
#>  7 <split [9000/1000]> Fold07 <tibble [3 × 4]> <tibble [0 × 4]>
#>  8 <split [9000/1000]> Fold08 <tibble [3 × 4]> <tibble [0 × 4]>
#>  9 <split [9000/1000]> Fold09 <tibble [3 × 4]> <tibble [0 × 4]>
#> 10 <split [9000/1000]> Fold10 <tibble [3 × 4]> <tibble [0 × 4]>

extract_fit_time(bt_res)
#> # A tibble: 30 × 5
#>    id     stage      process_id          time .config             
#>    <chr>  <chr>      <chr>              <dbl> <chr>               
#>  1 Fold01 preprocess prep.dummy_osY1q 0.00700 Preprocessor1_Model1
#>  2 Fold01 preprocess bake.dummy_osY1q 0.0190  Preprocessor1_Model1
#>  3 Fold01 model      boost_tree       0.089   Preprocessor1_Model1
#>  4 Fold02 preprocess prep.dummy_osY1q 0.00100 Preprocessor1_Model1
#>  5 Fold02 preprocess bake.dummy_osY1q 0.0170  Preprocessor1_Model1
#>  6 Fold02 model      boost_tree       0.087   Preprocessor1_Model1
#>  7 Fold03 preprocess prep.dummy_osY1q 0.00200 Preprocessor1_Model1
#>  8 Fold03 preprocess bake.dummy_osY1q 0.0170  Preprocessor1_Model1
#>  9 Fold03 model      boost_tree       0.0880  Preprocessor1_Model1
#> 10 Fold04 preprocess prep.dummy_osY1q 0.00200 Preprocessor1_Model1
#> # ℹ 20 more rows

extract_fit_time(bt_res, summarize = TRUE)
#> # A tibble: 20 × 4
#>    id     stage        time .config             
#>    <chr>  <chr>       <dbl> <chr>               
#>  1 Fold01 model      0.089  Preprocessor1_Model1
#>  2 Fold01 preprocess 0.0260 Preprocessor1_Model1
#>  3 Fold02 model      0.087  Preprocessor1_Model1
#>  4 Fold02 preprocess 0.0180 Preprocessor1_Model1
#>  5 Fold03 model      0.0880 Preprocessor1_Model1
#>  6 Fold03 preprocess 0.0190 Preprocessor1_Model1
#>  7 Fold04 model      0.0880 Preprocessor1_Model1
#>  8 Fold04 preprocess 0.0180 Preprocessor1_Model1
#>  9 Fold05 model      0.0900 Preprocessor1_Model1
#> 10 Fold05 preprocess 0.0190 Preprocessor1_Model1
#> # ℹ 10 more rows

^{Created on 2024-03-27 with reprex v2.1.0}

[no ci]

[simonpcouch]

Related to tidymodels/workflows#191 and friends.

[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.