Updating for linear_reg() engine specific args

Author: juliasilge

R/linear_reg.R

@@ -131,7 +131,8 @@ update.linear_reg <-
            parameters = NULL,
            penalty = NULL, mixture = NULL,
            fresh = FALSE, ...) {
-    update_dot_check(...)
+
+    eng_args <- update_engine_parameters(object$eng_args, ...)
 
     if (!is.null(parameters)) {
       parameters <- check_final_param(parameters)
@@ -145,12 +146,15 @@ update.linear_reg <-
 
     if (fresh) {
       object$args <- args
+      object$eng_args <- eng_args
     } else {
       null_args <- map_lgl(args, null_value)
       if (any(null_args))
         args <- args[!null_args]
       if (length(args) > 0)
         object$args[names(args)] <- args
+      if (length(eng_args) > 0)
+        object$eng_args[names(eng_args)] <- eng_args
     }
 
     new_model_spec(

tests/testthat/test_linear_reg.R

@@ -166,11 +166,12 @@ test_that('updating', {
   expr1     <- linear_reg() %>% set_engine("lm", model = FALSE)
   expr1_exp <- linear_reg(mixture = 0) %>% set_engine("lm", model = FALSE)
 
-  expr2     <- linear_reg(mixture = varying()) %>% set_engine("glmnet")
-  expr2_exp <- linear_reg(mixture = varying()) %>% set_engine("glmnet", nlambda = 10)
+  expr2     <- linear_reg() %>% set_engine("glmnet", nlambda = varying())
+  expr2_exp <- linear_reg() %>% set_engine("glmnet", nlambda = 10)
 
-  expr3     <- linear_reg(mixture = 0, penalty = varying()) %>% set_engine("glmnet")
-  expr3_exp <- linear_reg(mixture = 1) %>% set_engine("glmnet")
+  expr3     <- linear_reg(mixture = 0, penalty = varying()) %>% set_engine("glmnet", nlambda = varying())
+  expr3_exp <- linear_reg(mixture = 0, penalty = varying()) %>% set_engine("glmnet", nlambda = 10)
+  expr3_fre <- linear_reg(mixture = 1) %>% set_engine("glmnet", nlambda = 10)
 
   expr4     <- linear_reg(mixture = 0) %>% set_engine("glmnet", nlambda = 10)
   expr4_exp <- linear_reg(mixture = 0) %>% set_engine("glmnet", nlambda = 10, pmax = 2)
@@ -179,7 +180,9 @@ test_that('updating', {
   expr5_exp <- linear_reg(mixture = 1) %>% set_engine("glmnet", nlambda = 10, pmax = 2)
 
   expect_equal(update(expr1, mixture = 0), expr1_exp)
-  expect_equal(update(expr3, mixture = 1, fresh = TRUE), expr3_exp)
+  expect_equal(update(expr2, nlambda = 10), expr2_exp)
+  expect_equal(update(expr3, mixture = 1, fresh = TRUE, nlambda = 10), expr3_fre)
+  expect_equal(update(expr3, nlambda = 10), expr3_exp)
 
   param_tibb <- tibble::tibble(mixture = 1/3, penalty = 1)
   param_list <- as.list(param_tibb)