Update engine-specific args

[juliasilge]

This PR addresses tidymodels/tune#254.

Before this PR, update_dot_check() would error if anything (such as engine-specific parameters) showed up in the dots for update methods. This changes that, and then implements updating (from tune() ➡️ finalized parameters) for engine-specific parameters.

[juliasilge]

These are the results for random forest:

library(tidymodels)
#> ── Attaching packages ────────────────────────── tidymodels 0.1.1 ──
#> ✓ broom     0.7.0          ✓ recipes   0.1.13    
#> ✓ dials     0.0.8          ✓ rsample   0.0.7     
#> ✓ dplyr     1.0.0          ✓ tibble    3.0.3     
#> ✓ ggplot2   3.3.2          ✓ tidyr     1.1.0     
#> ✓ infer     0.5.3          ✓ tune      0.1.1.9000
#> ✓ modeldata 0.0.2          ✓ workflows 0.1.2     
#> ✓ parsnip   0.1.2.9000     ✓ yardstick 0.0.7     
#> ✓ purrr     0.3.4
#> ── Conflicts ───────────────────────────── tidymodels_conflicts() ──
#> x purrr::discard() masks scales::discard()
#> x dplyr::filter()  masks stats::filter()
#> x dplyr::lag()     masks stats::lag()
#> x recipes::step()  masks stats::step()

rf_mod1 <- 
  rand_forest(min_n = tune()) %>%
  set_mode("regression") %>%
  set_engine("ranger", 
             regularization.factor = tune(), 
             regularization.usedepth = tune())

rf_mod2 <- 
  rand_forest(min_n = tune()) %>%
  set_mode("regression") %>%
  set_engine("ranger")

set.seed(123)
car_boot <- bootstraps(mtcars, times = 5)

car_wf <- workflow() %>%
  add_formula(mpg ~ .) 

set.seed(123)
best1 <- car_wf %>%
  add_model(rf_mod1) %>%
  tune_grid(
    resamples = car_boot
  ) %>%
  select_best("rmse")
#> ! Bootstrap1: internal: A correlation computation is required, but `estimate` is const...
#> ! Bootstrap2: internal: A correlation computation is required, but `estimate` is const...
#> ! Bootstrap3: internal: A correlation computation is required, but `estimate` is const...
#> ! Bootstrap4: internal: A correlation computation is required, but `estimate` is const...
#> ! Bootstrap5: internal: A correlation computation is required, but `estimate` is const...

best2 <- car_wf %>%
  add_model(rf_mod2) %>%
  tune_grid(
    resamples = car_boot
  ) %>%
  select_best("rmse")
#> ! Bootstrap1: internal: A correlation computation is required, but `estimate` is const...
#> ! Bootstrap2: internal: A correlation computation is required, but `estimate` is const...
#> ! Bootstrap3: internal: A correlation computation is required, but `estimate` is const...
#> ! Bootstrap4: internal: A correlation computation is required, but `estimate` is const...
#> ! Bootstrap5: internal: A correlation computation is required, but `estimate` is const...

car_wf %>%
  add_model(rf_mod1) %>%
  finalize_workflow(best1)
#> ══ Workflow ════════════════════════════════════════════════════════
#> Preprocessor: Formula
#> Model: rand_forest()
#> 
#> ── Preprocessor ────────────────────────────────────────────────────
#> mpg ~ .
#> 
#> ── Model ───────────────────────────────────────────────────────────
#> Random Forest Model Specification (regression)
#> 
#> Main Arguments:
#>   min_n = 4
#> 
#> Engine-Specific Arguments:
#>   regularization.factor = 0.684913042467088
#>   regularization.usedepth = TRUE
#> 
#> Computational engine: ranger

car_wf %>%
  add_model(rf_mod2) %>%
  finalize_workflow(best2)
#> ══ Workflow ════════════════════════════════════════════════════════
#> Preprocessor: Formula
#> Model: rand_forest()
#> 
#> ── Preprocessor ────────────────────────────────────────────────────
#> mpg ~ .
#> 
#> ── Model ───────────────────────────────────────────────────────────
#> Random Forest Model Specification (regression)
#> 
#> Main Arguments:
#>   min_n = 5
#> 
#> Computational engine: ranger

^{Created on 2020-07-29 by the reprex package (v0.3.0.9001)}

[topepo]

You probably want to update from main; I just merged a PR that fixed some tests.

[juliasilge]

Here are some boost_tree() results:

library(tidymodels)
#> ── Attaching packages ──────────────────────────────── tidymodels 0.1.1 ──
#> ✓ broom     0.7.0          ✓ recipes   0.1.13    
#> ✓ dials     0.0.8          ✓ rsample   0.0.7     
#> ✓ dplyr     1.0.0          ✓ tibble    3.0.3     
#> ✓ ggplot2   3.3.2          ✓ tidyr     1.1.0     
#> ✓ infer     0.5.3          ✓ tune      0.1.1.9000
#> ✓ modeldata 0.0.2          ✓ workflows 0.1.2     
#> ✓ parsnip   0.1.2.9000     ✓ yardstick 0.0.7     
#> ✓ purrr     0.3.4
#> ── Conflicts ─────────────────────────────────── tidymodels_conflicts() ──
#> x purrr::discard() masks scales::discard()
#> x dplyr::filter()  masks stats::filter()
#> x dplyr::lag()     masks stats::lag()
#> x recipes::step()  masks stats::step()

data(penguins, package = "palmerpenguins")

set.seed(234)
folds <- vfold_cv(penguins, v = 5)

bands_grid <- tibble(bands = c(2, 50, 100))

c50_spec <- boost_tree() %>% 
  set_mode("classification") %>%
  set_engine("C5.0", rules = TRUE, bands = tune())

c50_wf <- workflow() %>%
  add_formula(species ~ .) %>%
  add_model(c50_spec)

best <- c50_wf %>%
  tune_grid(resamples = folds, grid = bands_grid) %>%
  select_best("roc_auc")

best
#> # A tibble: 1 x 2
#>   bands .config
#>   <dbl> <chr>  
#> 1     2 Model1

finalize_workflow(c50_wf, best)
#> ══ Workflow ══════════════════════════════════════════════════════════════
#> Preprocessor: Formula
#> Model: boost_tree()
#> 
#> ── Preprocessor ──────────────────────────────────────────────────────────
#> species ~ .
#> 
#> ── Model ─────────────────────────────────────────────────────────────────
#> Boosted Tree Model Specification (classification)
#> 
#> Engine-Specific Arguments:
#>   rules = TRUE
#>   bands = 2
#> 
#> Computational engine: C5.0

^{Created on 2020-07-30 by the reprex package (v0.3.0.9001)}

[topepo]

Looks great! I also tested it on tidymodels/tune#254

[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.