Adds prune argument to control handling of undirected (symmetric) matrices

.Rbuildignore

@@ -0,0 +1,6 @@
+Rphenograph_iris_cluster.png
+^appveyor\.yml$
+^\.travis\.yml$
+^codecov\.yml$
+^.*\.Rproj$
+^\.Rproj\.user$

.travis.yml

@@ -0,0 +1,11 @@
+# R for travis: see documentation at https://docs.travis-ci.com/user/languages/r
+
+language: R
+sudo: true
+cache: packages
+
+r_packages:
+  - covr
+
+after_success:
+  - Rscript -e 'covr::codecov()'

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: Rphenograph
 Type: Package
 Title: R implementation of the phenograph algorithm
-Version: 0.99.1
+Version: 0.99.1.9001
 Date: 2015-09-07
 Author: Hao Chen
 Maintainer: Hao Chen <[email protected]>
@@ -13,14 +13,15 @@ Description: A simple R implementation of the phenograph [PhenoGraph](http://
     sets, and then identifying communities using the well known ["Louvain method"]
     (https://sites.google.com/site/findcommunities/) in this graph.
 Depends:
-    R (>= 2.10.0), ggplot2, igraph
+    R (>= 2.10.0)
 Imports:
     igraph(>= 1.0.1),
     RANN(>= 2.5),
     Rcpp (>= 0.12.0),
     ggplot2
 LinkingTo: Rcpp
 Suggests:
-    testthat
+    testthat,
+    covr
 License: Artistic-2.0
-RoxygenNote: 5.0.1
+RoxygenNote: 6.1.1

NAMESPACE

@@ -4,8 +4,12 @@ export(Rphenograph)
 export(find_neighbors)
 import(ggplot2)
 importFrom(RANN,nn2)
+importFrom(Rcpp,sourceCpp)
+importFrom(igraph,as_adjacency_matrix)
 importFrom(igraph,cluster_louvain)
 importFrom(igraph,graph.data.frame)
+importFrom(igraph,graph_from_adjacency_matrix)
 importFrom(igraph,membership)
 importFrom(igraph,modularity)
 useDynLib(Rphenograph)
+useDynLib(Rphenograph, .registration = TRUE)

R/RcppExports.R

@@ -1,7 +1,7 @@
-# This file was generated by Rcpp::compileAttributes
+# Generated by using Rcpp::compileAttributes() -> do not edit by hand
 # Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
 
 jaccard_coeff <- function(idx) {
-    .Call('Rphenograph_jaccard_coeff', PACKAGE = 'Rphenograph', idx)
+    .Call('_Rphenograph_jaccard_coeff', PACKAGE = 'Rphenograph', idx)
 }
 

R/Rphenograph-package.R

@@ -0,0 +1,5 @@
+## usethis namespace: start
+#' @importFrom Rcpp sourceCpp
+#' @useDynLib Rphenograph, .registration = TRUE
+## usethis namespace: end
+NULL

R/phenograph.R

@@ -9,6 +9,9 @@
 #' 
 #' @param data matrix; input data matrix
 #' @param k integer; number of nearest neighbours (default:30)
+#' @param directed logical; whether to use a symmetric (default) or asymmetric ("directed") graph (default:FALSE)
+#' @param prune logical; determines whether symmetric matrix is generated by iGraph (FALSE) or element-wise multiplication (TRUE) with the transpose (defaults:FALSE)
+#' @param clust_fun function; community detection algorithm. Defaults to cluster_louvian. Other options: cluster_walktrap, cluster_spinglass, cluster_leading_eigen, cluster_edge_betweenness, cluster_fast_greedy, cluster_label_prop  
 #'
 #' @return a list contains an igraph graph object for \code{graph_from_data_frame} and a communities object, the operations of this class contains:
 #' \item{print}{returns the communities object itself, invisibly.}
@@ -28,25 +31,33 @@
 #'  
 #' @references Jacob H. Levine and et.al. Data-Driven Phenotypic Dissection of AML Reveals Progenitor-like Cells that Correlate with Prognosis. Cell, 2015.                
 #' @examples
+#' library(igraph)
 #' iris_unique <- unique(iris) # Remove duplicates
 #' data <- as.matrix(iris_unique[,1:4])
 #' Rphenograph_out <- Rphenograph(data, k = 45)
 #' modularity(Rphenograph_out[[2]])
 #' membership(Rphenograph_out[[2]])
 #' iris_unique$phenograph_cluster <- factor(membership(Rphenograph_out[[2]]))
-#' ggplot(iris_unique, aes(x=Sepal.Length, y=Sepal.Width, col=Species, shape=phenograph_cluster)) + geom_point(size = 3)+theme_bw()
+#' library(ggplot2)
+#' ggplot(iris_unique, aes(x=Sepal.Length, y=Sepal.Width, col=Species, shape=phenograph_cluster)) +
+#'     geom_point(size = 3) + theme_bw()
 #' 
-#' @importFrom igraph graph.data.frame cluster_louvain modularity membership
+#' @importFrom igraph graph.data.frame cluster_louvain modularity membership as_adjacency_matrix graph_from_adjacency_matrix
 #' @import ggplot2
 #' @useDynLib Rphenograph
 #' 
 #' @export
-Rphenograph <- function(data, k=30){
+Rphenograph <- function(data, k=30, directed=FALSE, prune=FALSE, clust_fun=NULL){
+    if(prune){
+      print("Setting directed to FALSE because prune is TRUE")
+      directed <- FALSE
+    }
+
     if(is.data.frame(data))
         data <- as.matrix(data)
 
     if(!is.matrix(data))
-        stop("Wrong input data, should be a data frame of matrix!")
+        stop("Wrong input data, should be a data frame or matrix!")
 
     if(k<1){
         stop("k must be a positive integer!")
@@ -61,20 +72,32 @@ Rphenograph <- function(data, k=30){
     cat("  Finding nearest neighbors...")
     t1 <- system.time(neighborMatrix <- find_neighbors(data, k=k+1)[,-1])
     cat("DONE ~",t1[3],"s\n", " Compute jaccard coefficient between nearest-neighbor sets...")
-    t2 <- system.time(links <- jaccard_coeff(neighborMatrix))
-
+    t2 <- system.time(links <- Rphenograph:::jaccard_coeff(neighborMatrix))
+    
     cat("DONE ~",t2[3],"s\n", " Build undirected graph from the weighted links...")
     links <- links[links[,1]>0, ]
     relations <- as.data.frame(links)
-    colnames(relations)<- c("from","to","weight")
-    t3 <- system.time(g <- graph.data.frame(relations, directed=FALSE))
-
+    colnames(relations) <- c("from","to","weight")
+    t3 <- system.time({
+      if(prune){
+        cat("Removing non-symmetric links...")
+        g <- graph.data.frame(relations, directed=TRUE)
+        mat <- as_adjacency_matrix(g, sparse = FALSE)
+        mat <- mat * t(mat)
+        g <- graph_from_adjacency_matrix(mat, mode = "undirected")
+      } else {
+        g <- graph.data.frame(relations, directed=directed)
+      }
+    })
+
     # Other community detection algorithms: 
     #    cluster_walktrap, cluster_spinglass, 
     #    cluster_leading_eigen, cluster_edge_betweenness, 
     #    cluster_fast_greedy, cluster_label_prop  
+    if(is.null(clust_fun)) clust_fun <- cluster_louvain
+
     cat("DONE ~",t3[3],"s\n", " Run louvain clustering on the graph ...")
-    t4 <- system.time(community <- cluster_louvain(g))
+    t4 <- system.time(community <- clust_fun(g))
     cat("DONE ~",t4[3],"s\n")
 
     message("Run Rphenograph DONE, totally takes ", sum(c(t1[3],t2[3],t3[3],t4[3])), "s.")
@@ -84,7 +107,6 @@ Rphenograph <- function(data, k=30){
     return(list(g, community))
 }
 
-
 #' K Nearest Neighbour Search
 #'
 #' Uses a kd-tree to find the p number of near neighbours for each point in an input/output dataset.
@@ -94,8 +116,7 @@ Rphenograph <- function(data, k=30){
 #' to within a specified error bound. For more information on the ANN library please 
 #' visit http://www.cs.umd.edu/~mount/ANN/.
 #' 
-#' @param data matrix; input data matrix
-#' @param k integer; number of nearest neighbours
+#' @describeIn Rphenograph Uses a kd-tree to find the p number of near neighbours for each point in an input/output dataset.
 #' 
 #' @return a n-by-k matrix of neighbor indices
 #' 
@@ -107,6 +128,7 @@ Rphenograph <- function(data, k=30){
 #' @importFrom RANN nn2
 #' @export
 find_neighbors <- function(data, k){
-    nearest <- nn2(data, data, k, searchtype = "standard")
-    return(nearest[[1]])
-}
+  nearest <- nn2(data, data, k, searchtype = "standard")
+  return(nearest[[1]])
+}
+

README.md

@@ -2,6 +2,10 @@ R implementation of the PhenoGraph algorithm
 ===============
 
 ### Rphenograph
+[![AppVeyor Build Status](https://ci.appveyor.com/api/projects/status/github/TomKellyGenetics/Rphenograph?branch=master&svg=true)](https://ci.appveyor.com/project/TomKellyGenetics/Rphenograph)
+[![Travis-CI Build Status](https://travis-ci.org/TomKellyGenetics/Rphenograph.svg?branch=master)](https://travis-ci.org/TomKellyGenetics/Rphenograph)
+[![Project Status: Active – The project has reached a stable, usable state and is being actively developed.](http://www.repostatus.org/badges/latest/active.svg)](http://www.repostatus.org/#active)
+[![Coverage Status](https://img.shields.io/codecov/c/github/JinmiaoChenLab/Rphenograph/master.svg)](https://codecov.io/github/JinmiaoChenLab/Rphenograph?branch=master)
 
 A simple R implementation of the [PhenoGraph](http://www.cell.com/cell/abstract/S0092-8674(15)00637-6) [1] algorithm, which is a clustering method designed for high-dimensional single-cell data analysis. It works by creating a graph ("network") representing phenotypic similarities between cells by calclating the Jaccard coefficient between nearest-neighbor sets, and then identifying communities using the well known [Louvain method](https://sites.google.com/site/findcommunities/) in this graph. 
 

Rphenograph.Rproj

@@ -0,0 +1,17 @@
+Version: 1.0
+
+RestoreWorkspace: Default
+SaveWorkspace: Default
+AlwaysSaveHistory: Default
+
+EnableCodeIndexing: Yes
+UseSpacesForTab: Yes
+NumSpacesForTab: 2
+Encoding: UTF-8
+
+RnwWeave: Sweave
+LaTeX: pdfLaTeX
+
+BuildType: Package
+PackageUseDevtools: Yes
+PackageInstallArgs: --no-multiarch --with-keep.source

appveyor.yml

@@ -0,0 +1,45 @@
+# DO NOT CHANGE the "init" and "install" sections below
+
+# Download script file from GitHub
+init:
+  ps: |
+        $ErrorActionPreference = "Stop"
+        Invoke-WebRequest http://raw.github.com/krlmlr/r-appveyor/master/scripts/appveyor-tool.ps1 -OutFile "..\appveyor-tool.ps1"
+        Import-Module '..\appveyor-tool.ps1'
+
+install:
+  ps: Bootstrap
+
+cache:
+  - C:\RLibrary
+
+# Adapt as necessary starting from here
+
+build_script:
+  - travis-tool.sh install_deps
+
+test_script:
+  - travis-tool.sh run_tests
+
+on_failure:
+  - 7z a failure.zip *.Rcheck\*
+  - appveyor PushArtifact failure.zip
+
+artifacts:
+  - path: '*.Rcheck\**\*.log'
+    name: Logs
+
+  - path: '*.Rcheck\**\*.out'
+    name: Logs
+
+  - path: '*.Rcheck\**\*.fail'
+    name: Logs
+
+  - path: '*.Rcheck\**\*.Rout'
+    name: Logs
+
+  - path: '\*_*.tar.gz'
+    name: Bits
+
+  - path: '\*_*.zip'
+    name: Bits

codecov.yml

@@ -0,0 +1 @@
+comment: false

src/.gitignore

@@ -0,0 +1,3 @@
+*.o
+*.so
+*.dll