change from depends to import

DESCRIPTION

@@ -12,6 +12,10 @@ URL: https://github.com/jonathan-columbiau/lionmap,
     https://jonathan-columbiau.github.io/lionmap/
 BugReports: https://github.com/jonathan-columbiau/lionmap/issues
 Depends: 
+Imports:
+    dplyr,
+    tidyr,
+    tidytree,
     caret,
     data.tree,
     e1071,
@@ -24,10 +28,6 @@ Depends:
     stats,
     stringr,
     testthat
-Imports:
-    dplyr,
-    tidyr,
-    tidytree
 Suggests: 
     kableExtra,
     ggplot2,

R/binomial_elastic_net.R

@@ -4,6 +4,8 @@
 #' @param celltype_labels Celltype labels in vector
 #'
 #' @return Binomial Elastic Net model trained on evenly split dataset (upsamples if classes aren't evenly split)
+#' @importFrom caret upSample
+#' @importFrom glmet cv.glmnet
 #' @keywords internal
 #'
 #' @examples
@@ -14,9 +16,9 @@
 #' dataset = dataset[,1:4]
 #' ex_model = binomial_elastic_net(dataset, labels)
 binomial_elastic_net <- function(reference_dataset, celltype_labels) {
-  upsampled_dataset <- caret::upSample(x = reference_dataset, y = celltype_labels, yname =  "celltype_labels", list = T)
+  upsampled_dataset <- upSample(x = reference_dataset, y = celltype_labels, yname =  "celltype_labels", list = T)
   celltype_labels <- upsampled_dataset[["y"]]
   reference_dataset <- upsampled_dataset[["x"]] %>% as.matrix()
-  pairwise_model <- glmnet::cv.glmnet(x = reference_dataset, y = celltype_labels, family = "binomial", alpha = .5)
+  pairwise_model <- cv.glmnet(x = reference_dataset, y = celltype_labels, family = "binomial", alpha = .5)
   pairwise_model
 }

R/binomial_lasso.R

@@ -4,6 +4,8 @@
 #' @param celltype_labels Celltype labels in vector
 #'
 #' @return Binomial lasso model trained on evenly split dataset (upsamples if classes aren't evenly split)
+#' @importFrom caret upSample
+#' @importFrom glmet cv.glmnet
 #' @keywords internal
 #'
 #' @examples
@@ -14,9 +16,9 @@
 #' dataset = dataset[,1:4]
 #' ex_model = binomial_lasso(dataset, labels)
 binomial_lasso <- function(reference_dataset, celltype_labels) {
-  upsampled_dataset <- caret::upSample(x = reference_dataset, y = celltype_labels, yname =  "celltype_labels", list = T)
+  upsampled_dataset <- upSample(x = reference_dataset, y = celltype_labels, yname =  "celltype_labels", list = T)
   celltype_labels <- upsampled_dataset[["y"]]
   reference_dataset <- upsampled_dataset[["x"]] %>% as.matrix()
-  pairwise_model <- glmnet::cv.glmnet(x = reference_dataset, y = celltype_labels, family = "binomial", alpha = 1)
+  pairwise_model <- cv.glmnet(x = reference_dataset, y = celltype_labels, family = "binomial", alpha = 1)
   pairwise_model
 }

R/binomial_ridge.R

@@ -4,6 +4,8 @@
 #' @param celltype_labels Celltype labels in vector
 #'
 #' @return Binomial ridge model trained on evenly split dataset (upsamples if classes aren't evenly split)
+#' @importFrom caret upSample
+#' @importFrom glmet cv.glmnet
 #' @keywords internal
 #'
 #' @examples
@@ -14,9 +16,9 @@
 #' dataset = dataset[,1:4]
 #' ex_model = binomial_ridge(dataset, labels)
 binomial_ridge <- function(reference_dataset, celltype_labels) {
-  upsampled_dataset <- caret::upSample(x = reference_dataset, y = celltype_labels, yname =  "celltype_labels", list = T)
+  upsampled_dataset <- upSample(x = reference_dataset, y = celltype_labels, yname =  "celltype_labels", list = T)
   celltype_labels <- upsampled_dataset[["y"]]
   reference_dataset <- upsampled_dataset[["x"]] %>% as.matrix()
-  pairwise_model <- glmnet::cv.glmnet(x = reference_dataset, y = celltype_labels, family = "binomial", alpha = 0)
+  pairwise_model <- cv.glmnet(x = reference_dataset, y = celltype_labels, family = "binomial", alpha = 0)
   pairwise_model
 }

R/knn.R

@@ -4,6 +4,8 @@
 #' @param celltype_labels Celltype labels in vector
 #'
 #' @return knn model trained on evenly split dataset (upsamples if classes aren't evenly split)
+#' @importFrom caret upSample
+#' @importFrom caret knn3
 #' @keywords internal
 #'
 #' @examples
@@ -15,7 +17,7 @@
 #' ex_model = knn(dataset, labels)
 knn <- function(reference_dataset, celltype_labels) {
   #upsample minority class to make class frequencies equal
-  reference_dataset <- caret::upSample(x = reference_dataset, y = celltype_labels, yname = "celltype_labels")
-  pairwise_model <- caret::knn3(celltype_labels ~ ., data = reference_dataset, k = 5)
+  reference_dataset <- upSample(x = reference_dataset, y = celltype_labels, yname = "celltype_labels")
+  pairwise_model <- knn3(celltype_labels ~ ., data = reference_dataset, k = 5)
   pairwise_model
 }

R/linear_da.R

@@ -4,6 +4,8 @@
 #' @param celltype_labels Celltype labels in vector
 #'
 #' @return linear_da model trained on evenly split dataset (upsamples if classes aren't evenly split)
+#' @importFrom caret upSample
+#' @importFrom MASS lda
 #' @keywords internal
 #'
 #' @examples
@@ -15,7 +17,7 @@
 #' ex_model = linear_da(dataset, labels)
 linear_da <- function(reference_dataset, celltype_labels) {
   #upsample minority class to make class frequencies equal
-  reference_dataset <- caret::upSample(x = reference_dataset, y = celltype_labels, yname = "celltype_labels")
-  pairwise_model <- MASS::lda(celltype_labels ~ ., data = reference_dataset)
+  reference_dataset <- upSample(x = reference_dataset, y = celltype_labels, yname = "celltype_labels")
+  pairwise_model <- lda(celltype_labels ~ ., data = reference_dataset)
   pairwise_model
 }

R/linear_svm.R

@@ -4,6 +4,8 @@
 #' @param celltype_labels Celltype labels in vector
 #'
 #' @return linear_svm model trained on evenly split dataset (upsamples if classes aren't evenly split)
+#' @importFrom caret upSample
+#' @importFrom e1071 svm
 #' @keywords internal
 #'
 #' @examples
@@ -14,7 +16,7 @@
 #' dataset = dataset[,1:4]
 #' ex_model = linear_svm(dataset, labels)
 linear_svm <- function(reference_dataset, celltype_labels) {
-  reference_dataset <- caret::upSample(x = reference_dataset, y = celltype_labels, yname =  "celltype_labels")
-  pairwise_model <- e1071::svm(celltype_labels ~ ., data = reference_dataset, kernel = "linear", scale = F)
+  reference_dataset <- upSample(x = reference_dataset, y = celltype_labels, yname =  "celltype_labels")
+  pairwise_model <- svm(celltype_labels ~ ., data = reference_dataset, kernel = "linear", scale = F)
   pairwise_model
 }

R/naive_bayes.R

@@ -4,6 +4,8 @@
 #' @param celltype_labels Celltype labels in vector
 #'
 #' @return naive_bayes model trained on evenly split dataset (upsamples if classes aren't evenly split)
+#' @importFrom caret upSample
+#' @importFrom e1071 naiveBayes
 #' @keywords internal
 #'
 #' @examples
@@ -15,7 +17,7 @@
 #' ex_model = naive_bayes(dataset, labels)
 naive_bayes <- function(reference_dataset, celltype_labels) {
   #upsample minority class to make class frequencies equal
-  reference_dataset <- caret::upSample(x = reference_dataset, y = celltype_labels, yname = "celltype_labels")
-  pairwise_model <- e1071::naiveBayes(celltype_labels ~ ., data = reference_dataset)
+  reference_dataset <- upSample(x = reference_dataset, y = celltype_labels, yname = "celltype_labels")
+  pairwise_model <- naiveBayes(celltype_labels ~ ., data = reference_dataset)
   pairwise_model
 }

R/polynomial_svm.R

@@ -4,6 +4,8 @@
 #' @param celltype_labels Celltype labels in vector
 #'
 #' @return polynomial_svm model trained on evenly split dataset (upsamples if classes aren't evenly split)
+#' @importFrom caret upSample
+#' @importFrom e1071 svm
 #' @keywords internal
 #'
 #' @examples
@@ -14,7 +16,7 @@
 #' dataset = dataset[,1:4]
 #' ex_model = polynomial_svm(dataset, labels)
 polynomial_svm <- function(reference_dataset, celltype_labels) {
-  reference_dataset <- caret::upSample(x = reference_dataset, y = celltype_labels, yname =  "celltype_labels")
-  pairwise_model <- e1071::svm(celltype_labels ~ ., data = reference_dataset, kernel = "polynomial", scale = F)
+  reference_dataset <- upSample(x = reference_dataset, y = celltype_labels, yname =  "celltype_labels")
+  pairwise_model <- svm(celltype_labels ~ ., data = reference_dataset, kernel = "polynomial", scale = F)
   pairwise_model
 }

R/quadratic_da.R

@@ -4,6 +4,8 @@
 #' @param celltype_labels Celltype labels in vector
 #'
 #' @return quadratic_da model trained on evenly split dataset (upsamples if classes aren't evenly split)
+#' @importFrom caret upSample
+#' @importFrom MASS qda
 #' @keywords internal
 #'
 #' @examples
@@ -15,7 +17,7 @@
 #' ex_model = quadratic_da(dataset, labels)
 quadratic_da <- function(reference_dataset, celltype_labels) {
   #upsample minority class to make class frequencies equal
-  reference_dataset <- caret::upSample(x = reference_dataset, y = celltype_labels, yname = "celltype_labels")
-  pairwise_model <- MASS::qda(celltype_labels ~ ., data = reference_dataset)
+  reference_dataset <- upSample(x = reference_dataset, y = celltype_labels, yname = "celltype_labels")
+  pairwise_model <- qda(celltype_labels ~ ., data = reference_dataset)
   pairwise_model
 }

R/random_forest.R

@@ -4,6 +4,8 @@
 #' @param celltype_labels Celltype labels in vector
 #'
 #' @return rf model trained on evenly split dataset (upsamples if classes aren't evenly split)
+#' @importFrom caret upSample
+#' @importFrom ranger ranger
 #' @keywords internal
 #'
 #' @examples
@@ -15,7 +17,7 @@
 #' ex_model = random_forest(dataset, labels)
 random_forest <- function(reference_dataset, celltype_labels) {
   #upsample minority class to make class frequencies equal
-  reference_dataset <- caret::upSample(x = reference_dataset, y = celltype_labels, yname = "celltype_labels")
-  pairwise_model <- ranger::ranger(celltype_labels ~ ., data = reference_dataset, num.trees = 500, classification = T, replace = T)
+  reference_dataset <- upSample(x = reference_dataset, y = celltype_labels, yname = "celltype_labels")
+  pairwise_model <- ranger(celltype_labels ~ ., data = reference_dataset, num.trees = 500, classification = T, replace = T)
   pairwise_model
 }