num_comp not updated with option_add() for discrim_flexible workflows #157
Description
The problem
The usual method of finalizing tunable parameters in a workflow set is not working with num_comp for discrim_flexible workflows
Reproducible example
library(tidymodels)
library(tidyverse)
library(discrim)
#>
#> Attaching package: 'discrim'
#> The following object is masked from 'package:dials':
#>
#> smoothness
library(AppliedPredictiveModeling)
library(future)
set.seed(321)
data <- quadBoundaryFunc(2000) %>% select(A = X1, B = X2, class)
data_splits1 <- initial_split(data, prop = .85, strata = class)
train_datas1 <- training(data_splits1)
test_datas1 <- testing(data_splits1)
foldss1 <- vfold_cv(train_datas1, v = 5, repeats = 3, strata = class)
biv_rec <-
recipe(class ~ ., data = train_datas1) %>%
step_normalize(all_predictors())
discrim_flexible_spec <-
discrim_flexible(num_terms = tune::tune(),
prod_degree = tune::tune(),
prune_method = tune::tune()) %>%
set_engine('earth') %>%
set_mode('classification')
normalizeds1 <-
workflow_set(
preproc = list(norm = biv_rec),
models = list(FD = discrim_flexible_spec)
)
normalizeds1 %>% extract_workflow("norm_FD") %>% extract_parameter_set_dials()
#> Collection of 3 parameters for tuning
#>
#> identifier type object
#> num_terms num_terms nparam[?]
#> prod_degree prod_degree nparam[+]
#> prune_method prune_method dparam[+]
#>
#> Model parameters needing finalization:
#> # Model Terms ('num_terms')
#>
#> See `?dials::finalize` or `?dials::update.parameters` for more information.
pars <- normalizeds1 %>% extract_workflow("norm_FD") %>% extract_parameter_set_dials() %>% finalize(x = train_datas1 %>% select(-class))
pars
#> Collection of 3 parameters for tuning
#>
#> identifier type object
#> num_terms num_terms nparam[+]
#> prod_degree prod_degree nparam[+]
#> prune_method prune_method dparam[+]
normalizeds1 <- normalizeds1 %>%
option_add(param_info = pars, id = "norm_FD")
normalizeds1 %>% extract_workflow("norm_FD") %>% extract_parameter_set_dials()
#> Collection of 3 parameters for tuning
#>
#> identifier type object
#> num_terms num_terms nparam[?]
#> prod_degree prod_degree nparam[+]
#> prune_method prune_method dparam[+]
#>
#> Model parameters needing finalization:
#> # Model Terms ('num_terms')
#>
#> See `?dials::finalize` or `?dials::update.parameters` for more information.
# 'num_comp' is not getting updated
bayes_ctrl <-
control_bayes(
save_pred = TRUE,
parallel_over = "everything",
save_workflow = TRUE,
verbose = T
)
plan(multisession)
tune_bayes(normalizeds1 %>% extract_workflow("norm_FD"), seed = 1503, resamples = foldss1, metrics = metric_set(roc_auc, brier_class, kap, accuracy), iter = 25, verbose = T, initial = 11, control = bayes_ctrl)
#> Error in `dials::grid_latin_hypercube()`:
#> ✖ This argument contains unknowns: `num_terms`.
#> ℹ See the `dials::finalize()` function.
# Error: This argument contains unknowns: `num_terms`.
plan(sequential)Created on 2024-06-20 with reprex v2.1.0
Session info
sessionInfo()
#> R version 4.4.0 (2024-04-24)
#> Platform: aarch64-apple-darwin20
#> Running under: macOS Sonoma 14.5
#>
#> Matrix products: default
#> BLAS: /Library/Frameworks/R.framework/Versions/4.4-arm64/Resources/lib/libRblas.0.dylib
#> LAPACK: /Library/Frameworks/R.framework/Versions/4.4-arm64/Resources/lib/libRlapack.dylib; LAPACK version 3.12.0
#>
#> locale:
#> [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
#>
#> time zone: UTC
#> tzcode source: internal
#>
#> attached base packages:
#> [1] stats graphics grDevices utils datasets methods base
#>
#> other attached packages:
#> [1] future_1.33.2 AppliedPredictiveModeling_1.1-7
#> [3] discrim_1.0.1 lubridate_1.9.3
#> [5] forcats_1.0.0 stringr_1.5.1
#> [7] readr_2.1.5 tidyverse_2.0.0
#> [9] yardstick_1.3.1 workflowsets_1.1.0
#> [11] workflows_1.1.4 tune_1.2.1
#> [13] tidyr_1.3.1 tibble_3.2.1
#> [15] rsample_1.2.1 recipes_1.0.10
#> [17] purrr_1.0.2 parsnip_1.2.1
#> [19] modeldata_1.3.0 infer_1.0.7
#> [21] ggplot2_3.5.1 dplyr_1.1.4
#> [23] dials_1.2.1 scales_1.3.0
#> [25] broom_1.0.6 tidymodels_1.2.0
#>
#> loaded via a namespace (and not attached):
#> [1] rlang_1.1.4 magrittr_2.0.3 furrr_0.3.1
#> [4] rpart.plot_3.1.2 compiler_4.4.0 vctrs_0.6.5
#> [7] reshape2_1.4.4 lhs_1.1.6 pkgconfig_2.0.3
#> [10] fastmap_1.2.0 backports_1.5.0 utf8_1.2.4
#> [13] rmarkdown_2.27 prodlim_2023.08.28 tzdb_0.4.0
#> [16] xfun_0.44 reprex_2.1.0 styler_1.10.3
#> [19] parallel_4.4.0 cluster_2.1.6 R6_2.5.1
#> [22] CORElearn_1.57.3 stringi_1.8.4 parallelly_1.37.1
#> [25] rpart_4.1.23 Rcpp_1.0.12 iterators_1.0.14
#> [28] knitr_1.47 future.apply_1.11.2 R.utils_2.12.3
#> [31] Matrix_1.7-0 splines_4.4.0 nnet_7.3-19
#> [34] R.cache_0.16.0 timechange_0.3.0 tidyselect_1.2.1
#> [37] rstudioapi_0.16.0 yaml_2.3.8 timeDate_4032.109
#> [40] codetools_0.2-20 listenv_0.9.1 lattice_0.22-6
#> [43] plyr_1.8.9 withr_3.0.0 evaluate_0.23
#> [46] survival_3.7-0 pillar_1.9.0 foreach_1.5.2
#> [49] ellipse_0.5.0 generics_0.1.3 hms_1.1.3
#> [52] munsell_0.5.1 plotmo_3.6.3 globals_0.16.3
#> [55] class_7.3-22 glue_1.7.0 mda_0.5-4
#> [58] tools_4.4.0 data.table_1.15.4 gower_1.0.1
#> [61] fs_1.6.4 grid_4.4.0 plotrix_3.8-4
#> [64] ipred_0.9-14 colorspace_2.1-0 earth_5.3.3
#> [67] Formula_1.2-5 cli_3.6.2 DiceDesign_1.10
#> [70] fansi_1.0.6 lava_1.8.0 gtable_0.3.5
#> [73] R.methodsS3_1.8.2 GPfit_1.0-8 digest_0.6.35
#> [76] htmltools_0.5.8.1 R.oo_1.26.0 lifecycle_1.0.4
#> [79] hardhat_1.4.0 MASS_7.3-60.2