Merge pull request #485 from tidymodels/glmnet-fail-penalty

Add error for glmnet models if penalty is not exactly 1

Author: topepo

R/engines.R

@@ -82,7 +82,7 @@ load_libs <- function(x, quiet, attach = FALSE) {
 #' @examples
 #' # First, set general arguments using the standardized names
 #' mod <-
-#'   logistic_reg(mixture = 1/3) %>%
+#'   logistic_reg(penalty = 0.01, mixture = 1/3) %>%
 #'   # now say how you want to fit the model and another other options
 #'   set_engine("glmnet", nlambda = 10)
 #' translate(mod, engine = "glmnet")

R/linear_reg.R

@@ -112,6 +112,7 @@ translate.linear_reg <- function(x, engine = x$engine, ...) {
     # Since the `fit` information is gone for the penalty, we need to have an
     # evaluated value for the parameter.
     x$args$penalty <- rlang::eval_tidy(x$args$penalty)
+    check_glmnet_penalty(x)
   }
 
   x

R/logistic_reg.R

@@ -115,6 +115,7 @@ translate.logistic_reg <- function(x, engine = x$engine, ...) {
     # Since the `fit` information is gone for the penalty, we need to have an
     # evaluated value for the parameter.
     x$args$penalty <- rlang::eval_tidy(x$args$penalty)
+    check_glmnet_penalty(x)
   }
 
   if (engine == "LiblineaR") {

R/misc.R

@@ -323,4 +323,13 @@ stan_conf_int <- function(object, newdata) {
   rlang::eval_tidy(fn)
 }
 
-
+check_glmnet_penalty <- function(x) {
+  if (length(x$args$penalty) != 1) {
+    rlang::abort(c(
+      "For the glmnet engine, `penalty` must be a single number (or a value of `tune()`).",
+      glue::glue("There are {length(x$args$penalty)} values for `penalty`."),
+      "To try multiple values for total regularization, use the tune package.",
+      "To predict multiple penalties, use `multi_predict()`"
+    ))
+  }
+}

R/translate.R

@@ -38,7 +38,7 @@
 #' translate(lm_spec, engine = "spark")
 #'
 #' # with a placeholder for an unknown argument value:
-#' translate(linear_reg(mixture = varying()), engine = "glmnet")
+#' translate(linear_reg(penalty = varying(), mixture = varying()), engine = "glmnet")
 #'
 #' @export
 

man/contr_one_hot.Rd

@@ -10,16 +10,14 @@ Engines may have pre-set default arguments when executing the model fit call. Fo
 ```{r lm-reg}
 linear_reg() %>% 
   set_engine("lm") %>% 
-  set_mode("regression") %>% 
   translate()
 ```
 
 ## glmnet
 
 ```{r glmnet-csl}
-linear_reg() %>% 
+linear_reg(penalty = 0.1) %>% 
   set_engine("glmnet") %>% 
-  set_mode("regression") %>% 
   translate()
 ```
 
@@ -37,7 +35,6 @@ penalty results.
 ```{r stan-reg}
 linear_reg() %>% 
   set_engine("stan") %>% 
-  set_mode("regression") %>% 
   translate()
 ```
 
@@ -55,7 +52,6 @@ returned.
 ```{r spark-reg}
 linear_reg() %>% 
   set_engine("spark") %>% 
-  set_mode("regression") %>% 
   translate()
 ```
 
@@ -64,7 +60,6 @@ linear_reg() %>%
 ```{r keras-reg}
 linear_reg() %>% 
   set_engine("keras") %>% 
-  set_mode("regression") %>% 
   translate()
 ```
 

man/linear_reg.Rd

@@ -11,16 +11,14 @@ For this type of model, the template of the fit calls are below.
 ```{r glm-reg}
 logistic_reg() %>% 
   set_engine("glm") %>% 
-  set_mode("classification") %>% 
   translate()
 ```
 
 ## glmnet
 
 ```{r glmnet-csl}
-logistic_reg() %>% 
+logistic_reg(penalty = 0.1) %>% 
   set_engine("glmnet") %>% 
-  set_mode("classification") %>% 
   translate()
 ```
 
@@ -38,7 +36,6 @@ penalty results.
 ```{r liblinear-reg}
 logistic_reg() %>% 
   set_engine("LiblineaR") %>% 
-  set_mode("classification") %>% 
   translate()
 ```
 
@@ -54,7 +51,6 @@ regularized regression models do not, which will result in different parameter e
 ```{r stan-reg}
 logistic_reg() %>% 
   set_engine("stan") %>% 
-  set_mode("classification") %>% 
   translate()
 ```
 
@@ -72,7 +68,6 @@ returned.
 ```{r spark-reg}
 logistic_reg() %>% 
   set_engine("spark") %>% 
-  set_mode("classification") %>% 
   translate()
 ```
 
@@ -81,7 +76,6 @@ logistic_reg() %>%
 ```{r keras-reg}
 logistic_reg() %>% 
   set_engine("keras") %>% 
-  set_mode("classification") %>% 
   translate()
 ```
 

man/logistic_reg.Rd

@@ -9,9 +9,8 @@ For this type of model, the template of the fit calls are below.
 ## glmnet
 
 ```{r glmnet-cls}
-multinom_reg() %>% 
+multinom_reg(penalty = 0.1) %>% 
   set_engine("glmnet") %>% 
-  set_mode("classification") %>% 
   translate()
 ```
 
@@ -29,7 +28,6 @@ penalty results.
 ```{r nnet-cls}
 multinom_reg() %>% 
   set_engine("nnet") %>% 
-  set_mode("classification") %>% 
   translate()
 ```
 
@@ -38,7 +36,6 @@ multinom_reg() %>%
 ```{r spark-cls}
 multinom_reg() %>% 
   set_engine("spark") %>% 
-  set_mode("classification") %>% 
   translate()
 ```
 
@@ -47,7 +44,6 @@ multinom_reg() %>%
 ```{r keras-cls}
 multinom_reg() %>% 
   set_engine("keras") %>% 
-  set_mode("classification") %>% 
   translate()
 ```