add butcher methods for xrf::xrf()

[simonpcouch]

Related to #234. Same story as #240 with NEWS entries.🐣

xrf just wraps glmnet and xgboost, but the glmnet object it carries around in the $glm slot is a bit modified (and given class glmnot) so we still need to do some subsetting in there beyond dispatching those packages' methods.

library(butcher)
library(xrf)

fit_xrf <- function() {
  boop <- runif(1e6)
  xrf(y ~ x, data.frame(y = rnorm(1e4), x = rnorm(1e4)), family = "gaussian")
}

xrf_fit <- fit_xrf()
xrf_res <- butcher(xrf_fit)

weigh(xrf_fit)
#> # A tibble: 43 × 2
#>    object                             size
#>    <chr>                             <dbl>
#>  1 glm.model.glmnet.fit.call       11.1   
#>  2 glm.formula                     10.1   
#>  3 rule_augmented_formula           8.36  
#>  4 base_formula                     8.17  
#>  5 glm.model.glmnet.fit.beta        0.490 
#>  6 xgb.raw                          0.117 
#>  7 rules.rule_id                    0.0518
#>  8 rules.split_id                   0.0494
#>  9 xgb.callbacks.cb.evaluation.log  0.0354
#> 10 rules.feature                    0.0155
#> # … with 33 more rows
weigh(xrf_res)
#> # A tibble: 43 × 2
#>    object                       size
#>    <chr>                       <dbl>
#>  1 glm.formula               1.93   
#>  2 glm.model.glmnet.fit.beta 0.490  
#>  3 rule_augmented_formula    0.146  
#>  4 xgb.raw                   0.117  
#>  5 rules.rule_id             0.0518 
#>  6 rules.split_id            0.0494 
#>  7 rules.feature             0.0155 
#>  8 rules.split               0.0155 
#>  9 rules.less_than           0.00775
#> 10 glm.model.glmnet.fit.a0   0.00669
#> # … with 33 more rows

^{Created on 2022-12-22 with reprex v2.0.2}

[juliasilge]

That is a pretty complicated structure inside! Thank you 🙏

[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.