classes -> class for #65

Author: topepo

R/boost_tree.R

@@ -391,7 +391,7 @@ xgb_by_tree <- function(tree, object, new_data, type, ...) {
     nms <- names(pred)
   } else {
     if (type == "class") {
-      pred <- boost_tree_xgboost_data$classes$post(pred, object)
+      pred <- boost_tree_xgboost_data$class$post(pred, object)
       pred <- tibble(.pred = factor(pred, levels = object$lvl))
     } else {
       pred <- boost_tree_xgboost_data$classprob$post(pred, object)

R/boost_tree_data.R

@@ -41,7 +41,7 @@ boost_tree_xgboost_data <-
           newdata = quote(new_data)
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = function(x, object) {
         if (is.vector(x)) {
@@ -97,7 +97,7 @@ boost_tree_C5.0_data <-
       func = c(pkg = "parsnip", fun = "C5.0_train"),
       defaults = list()
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = NULL,
       func = c(fun = "predict"),
@@ -152,7 +152,7 @@ boost_tree_spark_data <-
           dataset = quote(new_data)
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = format_spark_class,
       func = c(pkg = "sparklyr", fun = "ml_predict"),

R/logistic_reg_data.R

@@ -33,7 +33,7 @@ logistic_reg_glm_data <-
           family = expr(stats::binomial)
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = prob_to_class_2,
       func = c(fun = "predict"),
@@ -109,7 +109,7 @@ logistic_reg_glmnet_data <-
           family = "binomial"
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = organize_glmnet_class,
       func = c(fun = "predict"),
@@ -156,7 +156,7 @@ logistic_reg_stan_data <-
           family = expr(stats::binomial)
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = function(x, object) {
         x <- object$fit$family$linkinv(x)
@@ -268,7 +268,7 @@ logistic_reg_spark_data <-
           family = "binomial"
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = format_spark_class,
       func = c(pkg = "sparklyr", fun = "ml_predict"),

R/mars_data.R

@@ -34,7 +34,7 @@ mars_earth_data <-
           type = "response"
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = function(x, object) {
         x <- ifelse(x[,1] >= 0.5, object$lvl[2], object$lvl[1])

R/mlp_data.R

@@ -35,7 +35,7 @@ mlp_keras_data <-
           x = quote(as.matrix(new_data))
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = function(x, object) {
         object$lvl[x + 1]
@@ -92,7 +92,7 @@ mlp_nnet_data <-
           type = "raw"
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = NULL,
       func = c(fun = "predict"),

R/multinom_reg_data.R

@@ -28,7 +28,7 @@ multinom_reg_glmnet_data <-
           family = "multinomial"
         )
     ),
-    classes = list(
+    class = list(
       pre = check_glmnet_lambda,
       post = organize_multnet_class,
       func = c(fun = "predict"),
@@ -75,7 +75,7 @@ multinom_reg_spark_data <-
           family = "multinomial"
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = format_spark_class,
       func = c(pkg = "sparklyr", fun = "ml_predict"),

R/nearest_neighbor_data.R

@@ -42,7 +42,7 @@ nearest_neighbor_kknn_data <-
           type = "raw"
         )
     ),
-    classes = list(
+    class = list(
       pre = function(x, object) {
         if (!(object$fit$response %in% c("ordinal", "nominal"))) {
           stop("`kknn` model does not appear to use class predictions. Was ",

R/predict_class.R

@@ -13,23 +13,23 @@ predict_class.model_fit <- function (object, new_data, ...) {
     stop("`predict.model_fit` is for predicting factor outcomes.",
          call. = FALSE)
 
-  if (!any(names(object$spec$method) == "classes"))
+  if (!any(names(object$spec$method) == "class"))
     stop("No class prediction module defined for this model.", call. = FALSE)
 
   new_data <- prepare_data(object, new_data)
 
   # preprocess data
-  if (!is.null(object$spec$method$classes$pre))
-    new_data <- object$spec$method$classes$pre(new_data, object)
+  if (!is.null(object$spec$method$class$pre))
+    new_data <- object$spec$method$class$pre(new_data, object)
 
   # create prediction call
-  pred_call <- make_pred_call(object$spec$method$classes)
+  pred_call <- make_pred_call(object$spec$method$class)
 
   res <- eval_tidy(pred_call)
 
   # post-process the predictions
-  if(!is.null(object$spec$method$classes$post)) {
-    res <- object$spec$method$classes$post(res, object)
+  if(!is.null(object$spec$method$class$post)) {
+    res <- object$spec$method$class$post(res, object)
   }
 
   # coerce levels to those in `object`

R/rand_forest_data.R

@@ -123,7 +123,7 @@ rand_forest_ranger_data <-
           verbose = FALSE
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = ranger_class_pred,
       func = c(fun = "predict"),
@@ -200,7 +200,7 @@ rand_forest_randomForest_data <-
           newdata = quote(new_data)
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = NULL,
       func = c(fun = "predict"),
@@ -257,7 +257,7 @@ rand_forest_spark_data <-
           dataset = quote(new_data)
         )
     ),
-    classes = list(
+    class = list(
       pre = NULL,
       post = format_spark_class,
       func = c(pkg = "sparklyr", fun = "ml_predict"),