From e59ded611b3c1c57e69a8f4d7286427c663929d8 Mon Sep 17 00:00:00 2001
From: Larisa Yanceva <larisa.yanceva@priorbank.by>
Date: Thu, 2 May 2019 20:11:10 +0300
Subject: [PATCH 1/9] Added documentation for windsorize()

---
 R/windsorize.R    | 11 ++++++++++-
 man/windsorize.Rd | 16 +++++++++++++++-
 2 files changed, 25 insertions(+), 2 deletions(-)

diff --git a/R/windsorize.R b/R/windsorize.R
index b4e15e6..c3fc6e3 100644
--- a/R/windsorize.R
+++ b/R/windsorize.R
@@ -1,6 +1,15 @@
 #' Windsorize
 #'
-#' Do some windsorization.
+#' set all outliers to a specified percentile of the data; 
+#' a 90% winsorization would see all data below the 5th percentile set 
+#' to the 5th percentile, 
+#' and data above the 95th percentile set to the 95th percentile.
+#' 
+#' @param x A vector.
+#' @param p A quantile.
+#' @return dataset with trimmed outliers with 10% percentile
+#' @examples
+#' windsorize(c(92, 19, 101, 58, 101, 91, 26, 78, 10, 13, −5, 101, 86, −5))
 #' @export
 windsorize <- function(x, p = .90) {
   q <- quantile(x, p)
diff --git a/man/windsorize.Rd b/man/windsorize.Rd
index 832c3cb..b060543 100644
--- a/man/windsorize.Rd
+++ b/man/windsorize.Rd
@@ -6,6 +6,20 @@
 \usage{
 windsorize(x, p = 0.9)
 }
+\arguments{
+\item{x}{A vector.}
+
+\item{p}{A quantile.}
+}
+\value{
+dataset with trimmed outliers with 10% percentile
+}
 \description{
-Do some windsorization.
+set all outliers to a specified percentile of the data; 
+a 90% winsorization would see all data below the 5th percentile set 
+to the 5th percentile, 
+and data above the 95th percentile set to the 95th percentile.
+}
+\examples{
+windsorize(c(92, 19, 101, 58, 101, 91, 26, 78, 10, 13, −5, 101, 86, −5))
 }

From 4c1393d5879a736994de2d656b4836c1ba52dd6c Mon Sep 17 00:00:00 2001
From: Larisa Yanceva <larisa.yanceva@priorbank.by>
Date: Thu, 2 May 2019 20:59:35 +0300
Subject: [PATCH 2/9] Added arguments' checks into windsorize() Fix usage
 example of windorize()

---
 R/windsorize.R    | 6 +++++-
 man/windsorize.Rd | 2 +-
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/R/windsorize.R b/R/windsorize.R
index c3fc6e3..fe0b4f7 100644
--- a/R/windsorize.R
+++ b/R/windsorize.R
@@ -9,9 +9,13 @@
 #' @param p A quantile.
 #' @return dataset with trimmed outliers with 10% percentile
 #' @examples
-#' windsorize(c(92, 19, 101, 58, 101, 91, 26, 78, 10, 13, −5, 101, 86, −5))
+#' windsorize(c(3,4,4,3,4,5,1))
 #' @export
 windsorize <- function(x, p = .90) {
+  if (length(x) == 0) stop("argument should not be a empty vector")
+  if (is.na(x)) {
+    stop("argument should not be a vector containing NA")
+  }
   q <- quantile(x, p)
   x[x >= q] <- q
   x
diff --git a/man/windsorize.Rd b/man/windsorize.Rd
index b060543..2a0ff5d 100644
--- a/man/windsorize.Rd
+++ b/man/windsorize.Rd
@@ -21,5 +21,5 @@ to the 5th percentile,
 and data above the 95th percentile set to the 95th percentile.
 }
 \examples{
-windsorize(c(92, 19, 101, 58, 101, 91, 26, 78, 10, 13, −5, 101, 86, −5))
+windsorize(c(3,4,4,3,4,5,1))
 }

From 6dcb8afd900eefde1f798f328c21f2206339659c Mon Sep 17 00:00:00 2001
From: Larisa Yanceva <larisa.yanceva@priorbank.by>
Date: Thu, 2 May 2019 21:09:49 +0300
Subject: [PATCH 3/9] Fixed is.na() for all(is.na())

---
 R/windsorize.R | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/R/windsorize.R b/R/windsorize.R
index fe0b4f7..0182bdd 100644
--- a/R/windsorize.R
+++ b/R/windsorize.R
@@ -13,8 +13,8 @@
 #' @export
 windsorize <- function(x, p = .90) {
   if (length(x) == 0) stop("argument should not be a empty vector")
-  if (is.na(x)) {
-    stop("argument should not be a vector containing NA")
+  if (all(is.na(x))) {
+    stop("argument should not be a vector containing only NA")
   }
   q <- quantile(x, p)
   x[x >= q] <- q

From c2ec4977fd38488defdd6a7a7e1d37a5c45722cb Mon Sep 17 00:00:00 2001
From: Larisa Yanceva <larisa.yanceva@priorbank.by>
Date: Thu, 2 May 2019 21:21:29 +0300
Subject: [PATCH 4/9] Fixed windorize logic

---
 R/windsorize.R | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/R/windsorize.R b/R/windsorize.R
index 0182bdd..54c0c1b 100644
--- a/R/windsorize.R
+++ b/R/windsorize.R
@@ -16,8 +16,13 @@ windsorize <- function(x, p = .90) {
   if (all(is.na(x))) {
     stop("argument should not be a vector containing only NA")
   }
-  q <- quantile(x, p)
-  x[x >= q] <- q
+  if (p < 0 || p > 1) {
+    stop("p invalid percentale. Expected values from 0 to 1")
+  }
+  q_lower <- quantile(x, (1-p)/2)
+  q_upper <- quantile(x, 1 - (1-p)/2)
+  x[x <= q_lower] <- q_lower
+  x[x >= q_upper] <- q_upper
   x
 }
 

From b447bafb24eb3b010a8822287a958bc4e37427da Mon Sep 17 00:00:00 2001
From: Larisa Yanceva <larisa.yanceva@priorbank.by>
Date: Thu, 2 May 2019 22:11:10 +0300
Subject: [PATCH 5/9] Added additional checks for arguments

---
 R/windsorize.R | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/R/windsorize.R b/R/windsorize.R
index 54c0c1b..de18782 100644
--- a/R/windsorize.R
+++ b/R/windsorize.R
@@ -16,6 +16,8 @@ windsorize <- function(x, p = .90) {
   if (all(is.na(x))) {
     stop("argument should not be a vector containing only NA")
   }
+  if (!is.numeric(x)) stop("argument should be a numeric vector")
+  if (!is.numeric(p)) stop("argument should be a number from 0 to 1")
   if (p < 0 || p > 1) {
     stop("p invalid percentale. Expected values from 0 to 1")
   }

From 8260705f72a8e3b5d2e20ad344bd31822ace92d5 Mon Sep 17 00:00:00 2001
From: Larisa Yanceva <larisa.yanceva@priorbank.by>
Date: Thu, 2 May 2019 22:23:20 +0300
Subject: [PATCH 6/9] Add tests for windsorize()

---
 tests/testthat.R                |  4 ++++
 tests/testthat/test_windorize.R | 17 +++++++++++++++++
 2 files changed, 21 insertions(+)
 create mode 100644 tests/testthat.R
 create mode 100644 tests/testthat/test_windorize.R

diff --git a/tests/testthat.R b/tests/testthat.R
new file mode 100644
index 0000000..8adbfff
--- /dev/null
+++ b/tests/testthat.R
@@ -0,0 +1,4 @@
+library(testthat)
+library(datacleaner)
+
+test_check("datacleaner")
diff --git a/tests/testthat/test_windorize.R b/tests/testthat/test_windorize.R
new file mode 100644
index 0000000..5e0beb4
--- /dev/null
+++ b/tests/testthat/test_windorize.R
@@ -0,0 +1,17 @@
+context("Windsorize")
+library(datacleaner)
+test_that("windorizing is correct", {
+  expect_equal(windsorize(c(2,2,2,2,3),.9), c(2,2,2,2,2.8) )
+  expect_equal(windsorize(c(2,2,2,2,1),.9), c(2,2,2,2,1.2) )
+})
+
+test_that("unexpected parameters", {
+  expect_error(windsorize(c(NA,NA,NA), .9), "argument should not be a vector containing only NA")
+  expect_error(windsorize(c(), .9), "argument should not be a empty vector")
+  expect_error(windsorize(c(1,2,3,"4"), .9), "argument should be a numeric vector")
+  expect_error(windsorize(c(1,2,3,4), ".9"), "argument should be a number from 0 to 1")
+  expect_error(windsorize(c(1,2,3,4), -1), "p invalid percentale. Expected values from 0 to 1")
+  expect_error(windsorize(c(1,2,3,4), 1.2), "p invalid percentale. Expected values from 0 to 1")
+  expect_error(windsorize(c(1,2,3,NA), .9))
+  
+})

From e95213f6ad0c2ed646e8149a8428e65d8622afa7 Mon Sep 17 00:00:00 2001
From: Larisa Yanceva <larisa.yanceva@priorbank.by>
Date: Thu, 2 May 2019 23:22:24 +0300
Subject: [PATCH 7/9] Add tranform_log() together with tests

---
 R/transform_log.R                  | 19 +++++++++++++++++++
 man/transform_log.Rd               | 24 ++++++++++++++++++++++++
 tests/testthat/test_tranform_log.R | 11 +++++++++++
 3 files changed, 54 insertions(+)
 create mode 100644 R/transform_log.R
 create mode 100644 man/transform_log.Rd
 create mode 100644 tests/testthat/test_tranform_log.R

diff --git a/R/transform_log.R b/R/transform_log.R
new file mode 100644
index 0000000..004f883
--- /dev/null
+++ b/R/transform_log.R
@@ -0,0 +1,19 @@
+#' transform_log 
+#' Transform numerical values into their log values
+#' @param x A vector
+#' @return logarithm of x
+#' @export 
+#' @examples
+#' transform_log(1)
+#' transform_log(c(1, 2, 3, 4, 5))
+#' 
+transform_log <- function(x) { 
+  if (!is.numeric(x)) {
+    warning("transform_log: Expecting numeric argument")
+  }
+  x_badval <- is.na(suppressWarnings(as.numeric(x)))
+  x[x_badval] <- 1
+  y <- log(as.numeric(x))
+  y[x_badval] <- NA
+  y
+} 
\ No newline at end of file
diff --git a/man/transform_log.Rd b/man/transform_log.Rd
new file mode 100644
index 0000000..c5b287c
--- /dev/null
+++ b/man/transform_log.Rd
@@ -0,0 +1,24 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/transform_log.R
+\name{transform_log}
+\alias{transform_log}
+\title{transform_log 
+Transform numerical values into their log values}
+\usage{
+transform_log(x)
+}
+\arguments{
+\item{x}{A vector}
+}
+\value{
+logarithm of x
+}
+\description{
+transform_log 
+Transform numerical values into their log values
+}
+\examples{
+transform_log(1)
+transform_log(c(1, 2, 3, 4, 5))
+
+}
diff --git a/tests/testthat/test_tranform_log.R b/tests/testthat/test_tranform_log.R
new file mode 100644
index 0000000..70bae1c
--- /dev/null
+++ b/tests/testthat/test_tranform_log.R
@@ -0,0 +1,11 @@
+context("tranform_log")
+library(datacleaner)
+test_that("tranform_log is correct", {
+  expect_equal(as.character(transform_log(c(2, 2, 2, 2, 3))), as.character(c(0.693147180559945, 0.693147180559945, 0.693147180559945, 0.693147180559945, 1.09861228866811)))
+  expect_equal(as.character(transform_log(c(2, 3, NA))), as.character(c(0.693147180559945, 1.09861228866811, NA)))
+})
+
+test_that("unexpected parameters", {
+  expect_warning(t <- transform_log(c(2, 3, "NA")), "transform_log: Expecting numeric argument")
+  expect_equal(as.character(t), as.character(c(0.693147180559945, 1.09861228866811, NA)))
+})
\ No newline at end of file

From 049c52df57ec9b60763c265edbb5854a682a21e0 Mon Sep 17 00:00:00 2001
From: Larisa Yanceva <larisa.yanceva@priorbank.by>
Date: Fri, 3 May 2019 21:49:52 +0300
Subject: [PATCH 8/9] Add parament for meanimpute()

---
 R/meanimpute.R    | 1 +
 man/meanimpute.Rd | 3 +++
 2 files changed, 4 insertions(+)

diff --git a/R/meanimpute.R b/R/meanimpute.R
index cc7cf5e..f7f79e5 100644
--- a/R/meanimpute.R
+++ b/R/meanimpute.R
@@ -1,4 +1,5 @@
 #' Meanimputation
+#' @param x A vector
 #' @export
 meanimpute <- function(x) {
   x[is.na(x)] <- mean(x, na.rm = TRUE)
diff --git a/man/meanimpute.Rd b/man/meanimpute.Rd
index 8139e8f..77f2fbc 100644
--- a/man/meanimpute.Rd
+++ b/man/meanimpute.Rd
@@ -6,6 +6,9 @@
 \usage{
 meanimpute(x)
 }
+\arguments{
+\item{x}{A vector}
+}
 \description{
 Meanimputation
 }

From 5c2ce83b438d1e269ee6d83489f093fd5da35cf1 Mon Sep 17 00:00:00 2001
From: Larisa Yanceva <larisa.yanceva@priorbank.by>
Date: Fri, 3 May 2019 22:08:38 +0300
Subject: [PATCH 9/9] Import quantile() for windsorize()

---
 NAMESPACE | 1 +
 1 file changed, 1 insertion(+)

diff --git a/NAMESPACE b/NAMESPACE
index d75f824..c1a7cea 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -1 +1,2 @@
 exportPattern("^[[:alpha:]]+")
+importFrom("stats", "quantile")