update

laderast · Jul 1, 2020 · 0d5b654 · 0d5b654
1 parent a9e2359
commit 0d5b654
Show file tree

Hide file tree

Showing 10 changed files with 68 additions and 81 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -31,7 +31,7 @@ Imports:
 Depends: 
     shiny
 Roxygen: list(markdown = TRUE)
-RoxygenNote: 6.1.1
+RoxygenNote: 7.0.2
 Suggests: 
     testthat
 URL: http://github.com/laderast/burro
diff --git a/NAMESPACE b/NAMESPACE
@@ -2,6 +2,7 @@
 
 export("%>%")
 export(build_shiny_app)
+export(check_data)
 export(explore_data)
 import(dplyr)
 import(ggplot2)

diff --git a/R/check_data.R b/R/check_data.R
@@ -1,26 +1,29 @@
-check_data <- function(dataset){
-  if(!is.null(covariates)) {
+#' Checks the Dataset and sets attributes on it for use in
+#'
+#' @param dataset
+#' @param covariates
+#' @param outcome_var
+#'
+#' @return
+#' @export
+#'
+#' @examples
+check_data <- function(dataset, covariates=NULL, outcome_var=NULL){
 
     covariates_in_data <- covariates %in% colnames(dataset)
+
     num_in_data <- which(covariates_in_data)
     not_in_data <- which(!covariates_in_data)
 
-    if(length(num_in_data) == 0) {
-      stop("Your covariates aren't in the dataset - make sure they correpond to column names in the data")
-    }
-
-    if(length(not_in_data) > 0) {
-      warning(
+    if(length(num_in_data) == 0 & length(covariates) > 0) {
         paste0("The following covariates weren't in the dataset:", paste(covariates[not_in_data]))
-      )
     }
 
-  }
-
-
   myDataFrame <- data.table::data.table(dataset)
 
-  if(!is.null(covariates)){
+  covariates <- covariates[covariates_in_data]
+
+  if(length(covariates) > 0){
     myDataFrame <- myDataFrame[,covariates,with=FALSE]
   }
 
@@ -34,16 +37,15 @@ check_data <- function(dataset){
   in_dataset <- length(which(outcome_var %in% colnames(dataset)))
 
   # need to check column names and outcome var names in data
-  if(in_dataset == 0){
-    stop("Your outcome variable is not the dataset - try using colnames(data) to select it")
+  if(in_dataset == 0 & length(outcome_var) > 0){
+    warning("Your outcome variable is not the dataset - using all categorical variables as outcome")
   }
+
   if(in_dataset < length(outcome_var)){
-    warning("Some of your outcomes weren't in the dataset")
+    warning("Some of your outcomes weren't in the dataset - using the ones we found")
   }
 
-  #myDataFrame <- burro:::sanitize_data_frame(myDataFrame, outcome_var)
   remove_categories <- outcome_var
-
   cat_no_outcome <- categoricalVars
 
   if(length(outcome_var) != length(categoricalVars)){
@@ -53,5 +55,11 @@ check_data <- function(dataset){
 
   numericVars <- sort(burro:::get_numeric_variables(myDataFrame))
 
+  attr(myDataFrame, "outcome_var") <-  outcome_var
+  attr(myDataFrame, "cat_no_outcome") <-  cat_no_outcome
+  attr(myDataFrame, "categoricalVars") <-  categoricalVars
+  attr(myDataFrame, "numericVars") <- numericVars
+
+  return(myDataFrame)
 
 }
diff --git a/R/helper.R b/R/helper.R
@@ -5,6 +5,11 @@ get_numeric_variables <- function(df){
   varClass <- sapply(df, class)
 
   numericVars <- names(varClass[varClass %in% c("numeric", "integer", "Date")])
+
+  if(length(numericVars) == 0){
+    numericVars <- NULL
+  }
+
   return(numericVars)
 }
 

diff --git a/R/xplor.R b/R/xplor.R
@@ -51,62 +51,12 @@ explore_data <- function(dataset, covariates=NULL,
   #Sys.setlocale("LC_CTYPE", "Chinese")
   dataset_name <- deparse(substitute(dataset))
 
+  myDataFrame <- check_data(dataset)
 
-  if(!is.null(covariates)) {
-
-    covariates_in_data <- covariates %in% colnames(dataset)
-    num_in_data <- which(covariates_in_data)
-    not_in_data <- which(!covariates_in_data)
-
-    if(length(num_in_data) == 0) {
-      stop("Your covariates aren't in the dataset - make sure they correpond to column names in the data")
-    }
-
-    if(length(not_in_data) > 0) {
-      warning(
-        paste0("The following covariates weren't in the dataset:", paste(covariates[not_in_data]))
-      )
-    }
-
-
-  }
-
-
-  myDataFrame <- data.table::data.table(dataset)
-
-  if(!is.null(covariates)){
-    myDataFrame <- myDataFrame[,covariates,with=FALSE]
-  }
-
-  categoricalVars <- sort(names(burro:::get_category_variables(myDataFrame)))
-  outcome_var <- outcome_var[outcome_var %in% categoricalVars]
-
-  #todo - just show two way dropdown if outcome_var = NULL
-  if(is.null(outcome_var)){
-    outcome_var <- categoricalVars
-  }
-
-  in_dataset <- length(which(outcome_var %in% colnames(dataset)))
-
-  # need to check column names and outcome var names in data
-  if(in_dataset == 0){
-    stop("Your outcome variable is not the dataset - try using colnames(data) to select it")
-  }
-  if(in_dataset < length(outcome_var)){
-    warning("Some of your outcomes weren't in the dataset")
-  }
-
-  #myDataFrame <- burro:::sanitize_data_frame(myDataFrame, outcome_var)
-  remove_categories <- outcome_var
-
-  cat_no_outcome <- categoricalVars
-
-  if(length(outcome_var) != length(categoricalVars)){
-    cat_no_outcome <-
-      setdiff(categoricalVars, remove_categories)
-  }
-
-  numericVars <- sort(burro:::get_numeric_variables(myDataFrame))
+  numericVars <- attr(myDataFrame, "numericVars")
+  categoricalVars <- attr(myDataFrame, "categoricalVars")
+  outcome_var <- attr(myDataFrame, "outcome_var")
+  cat_no_outcome <- attr(myDataFrame, "cat_no_outcome")
 
   ggplot2::theme_set(ggplot2::theme_classic(base_size = 15))
   #data_dictionary <- readr::read_csv("data/data_dictionary.csv") %>%

diff --git a/data/smoke_complete.xlsx b/data/smoke_complete.xlsx
diff --git a/man/check_data.Rd b/man/check_data.Rd
diff --git a/man/explore_data.Rd b/man/explore_data.Rd
diff --git a/tests/testthat/test-get_categorical_data.R b/tests/testthat/test-get_categorical_data.R
@@ -15,3 +15,10 @@ test_that("get category variables works", {
   expect_equal(names(cats1), "Species")
   expect_equal(NULL, cats4)
 })
+
+data1 <- check_data(mtcars)
+data2 <- check_data(diamonds)
+
+test_that("check_data works",{
+
+})
diff --git a/tests/testthat/test-get_continuous_data.R b/tests/testthat/test-get_continuous_data.R