small tweaks to reduce verbosity.

R/lightgbm.R

@@ -272,7 +272,7 @@ train_lightgbm <- function(x, y, max_depth = 17, num_iterations = 10, learning_r
     data = prepare_df_lgbm(x),
     label = y,
     categorical_feature = categorical_columns(x),
-    feature_pre_filter = FALSE
+    params = list(feature_pre_filter = FALSE)
   )
 
   main_args <- list(
@@ -345,7 +345,8 @@ predict_lightgbm_classification_raw <- function(object, new_data, ...) {
 #' @export
 predict_lightgbm_regression_numeric <- function(object, new_data, ...) {
   # train_colnames <- object$fit$.__enclos_env__$private$train_set$get_colnames()
-  p <- stats::predict(object$fit, prepare_df_lgbm(new_data), reshape = TRUE, predict_disable_shape_check=TRUE, ...)
+  p <- stats::predict(object$fit, prepare_df_lgbm(new_data), reshape = TRUE,
+                      params = list(predict_disable_shape_check=TRUE), ...)
   p
 }
 

tests/testthat/helper-model.R

@@ -5,7 +5,7 @@ mtcars_class_binary$vs <- as.factor(mtcars$vs)
 
 expect_all_modes_works <- function(model, engine) {
   if(engine == "lightgbm") {
-    model <- parsnip::set_engine(model, engine, verbosity = -1L)
+    model <- parsnip::set_engine(model, engine, verbose = -1L)
   } else {
     model <- parsnip::set_engine(model, engine)
   }

tests/testthat/test-lightgbm.R

@@ -1,20 +1,3 @@
-test_that("simple test", {
-
-  library(lightgbm)
-  data(agaricus.train, package='lightgbm')
-  train <- agaricus.train
-  dtrain <- lgb.Dataset(train$data, label = train$label)
-  model <- lgb.cv(
-    params = list(
-      objective = "regression"
-      , metric = "l2"
-    )
-    , data = dtrain
-  )
-
-  expect_s3_class(model, "lgb.CVBooster")
-})
-
 test_that("lightgbm", {
 
   model <- parsnip::boost_tree(mtry = 1, trees = 50, tree_depth = 15, min_n = 1)