From 4d329c23a50f7d88b8e7c28543c2a285fe3f1649 Mon Sep 17 00:00:00 2001
From: benholmgren <44175897+benholmgren@users.noreply.github.com>
Date: Thu, 6 Dec 2018 11:04:49 -0700
Subject: [PATCH 1/3] Added documentation to knnDE function

Briefly explained parameters and return value with roxygen style comments, then added documentation to walk through code
---
 R/knnDE.R | 23 +++++++++++++++++++++--
 1 file changed, 21 insertions(+), 2 deletions(-)

diff --git a/R/knnDE.R b/R/knnDE.R
index 5f57519..1e64309 100644
--- a/R/knnDE.R
+++ b/R/knnDE.R
@@ -1,20 +1,39 @@
+#' @param X an n by d matrix of coordinates containing points used int the
+#' density estimation, where n is the number of points and d is the dimension
+#' @param Grid an m by d matrix of coordinates, where m is the number of points
+#' in the grid and d is the dimension
+#' @param k a number used as the smoothing parameter
+#' @return a vector of length m (number of points in Grid) containing the value
+#' of the knn Density Estimator for each point in the grid
+
 knnDE <-
 function(X, Grid, k){
-
+  
+  # ensure input X is a matrix of numbers
   if (!is.numeric(X) && !is.data.frame(X)) {
     stop("X should be a matrix of coordinates")
   }
+  # ensure input Grid is a matrix of numbers
   if (!is.numeric(Grid) && !is.data.frame(Grid)) {
     stop("Grid should be a matrix of coordinates")
   }
+  # ensure that k is a positive integer
   if (!is.numeric(k) || length(k) != 1 || k < 1) {
     stop("k should be a positive integer")
   }
 
+  # store the number of rows and columns in input matrix X
+  # (d being the dimension and n being number of points in X)
   d <- ncol(X)
   n <- nrow(X)
+  # find the Euclidean distances from k nearest neighbors using input matrix X, the query matrix
+  # Grid, a maximum number n of nearest neighbors to search, and the nearest neighbor searching
+  # algorithm ("kd_tree")
   r.k <- apply(FNN::knnx.dist(X, Grid, k = k, algorithm = "kd_tree"), 1, max)
+  # volume of the Euclidean d dimensional unit ball
   v.d <- pi^(d/2) /gamma(d/2+1)
+  # function in its final form, dividing smoothing parameter k by the number of points in X multiplied 
+  # by the volume of unit ball and the Euclidean distances from k nearest neighbors to the power of d dimensions
   out <- k / (n * v.d * r.k ^ d)  
   return(out)
-}
\ No newline at end of file
+}

From ff3a754fe81934074615436d0e3875724c90e5e9 Mon Sep 17 00:00:00 2001
From: benholmgren <44175897+benholmgren@users.noreply.github.com>
Date: Mon, 25 Feb 2019 10:35:58 -0700
Subject: [PATCH 2/3] Revised comments for knnDE

Streamlined comments
---
 R/knnDE.R | 15 +++++----------
 1 file changed, 5 insertions(+), 10 deletions(-)

diff --git a/R/knnDE.R b/R/knnDE.R
index 1e64309..b186dd1 100644
--- a/R/knnDE.R
+++ b/R/knnDE.R
@@ -9,31 +9,26 @@
 knnDE <-
 function(X, Grid, k){
   
-  # ensure input X is a matrix of numbers
   if (!is.numeric(X) && !is.data.frame(X)) {
     stop("X should be a matrix of coordinates")
   }
-  # ensure input Grid is a matrix of numbers
   if (!is.numeric(Grid) && !is.data.frame(Grid)) {
     stop("Grid should be a matrix of coordinates")
   }
-  # ensure that k is a positive integer
   if (!is.numeric(k) || length(k) != 1 || k < 1) {
     stop("k should be a positive integer")
   }
 
-  # store the number of rows and columns in input matrix X
-  # (d being the dimension and n being number of points in X)
+  # store d, the dimension of X, and n, the number of points in X
   d <- ncol(X)
   n <- nrow(X)
-  # find the Euclidean distances from k nearest neighbors using input matrix X, the query matrix
-  # Grid, a maximum number n of nearest neighbors to search, and the nearest neighbor searching
-  # algorithm ("kd_tree")
+  # find the Euclidean distances of k nearest neighbors from input matrix X and the query matrix
+  # Grid using the fast nearest neighbor searching algorithm "kd_tree"
   r.k <- apply(FNN::knnx.dist(X, Grid, k = k, algorithm = "kd_tree"), 1, max)
   # volume of the Euclidean d dimensional unit ball
   v.d <- pi^(d/2) /gamma(d/2+1)
-  # function in its final form, dividing smoothing parameter k by the number of points in X multiplied 
-  # by the volume of unit ball and the Euclidean distances from k nearest neighbors to the power of d dimensions
+  # final output incorporates smoothing parameter with volume of unit ball and euclidean k nearest neighbor
+  # distances to return output vector containing density estimators for each point in the grid
   out <- k / (n * v.d * r.k ^ d)  
   return(out)
 }

From 03c10c500ee0dd71fcfb38dece7048954f54919a Mon Sep 17 00:00:00 2001
From: benholmgren <44175897+benholmgren@users.noreply.github.com>
Date: Tue, 26 Feb 2019 12:45:40 -0700
Subject: [PATCH 3/3] Added title to knnDE roxygen

roxygen needs a title to make documentation reliably
---
 R/knnDE.R | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/R/knnDE.R b/R/knnDE.R
index b186dd1..d2e5c65 100644
--- a/R/knnDE.R
+++ b/R/knnDE.R
@@ -1,8 +1,13 @@
+#' @title knnDE
+#'
 #' @param X an n by d matrix of coordinates containing points used int the
 #' density estimation, where n is the number of points and d is the dimension
+#'
 #' @param Grid an m by d matrix of coordinates, where m is the number of points
 #' in the grid and d is the dimension
+#'
 #' @param k a number used as the smoothing parameter
+#'
 #' @return a vector of length m (number of points in Grid) containing the value
 #' of the knn Density Estimator for each point in the grid