R/linkcomm_plotting.R

In alextkalinka/linkcomm: Tools for Generating, Visualizing, and Analysing Link Communities in Networks

#####################################################################################################
# Plotting-related functions for link communities generated by 'getLinkCommunities'.                #
#                                                                                                   #
# Author: Alex T. Kalinka (alex.t.kalinka@gmail.com)                                                #
#                                                                                                   #
# See: 												    #
#												    #
# Ahn et al. (2010). Link communities reveal multiscale complexity in networks. Nature 466:761-765. #
#												    #
# Kalinka & Tomancak (2011) linkcomm: an R package for  					    #
#   the generation, visualization, and analysis of link 					    #
#   communities in networks of arbitrary size and type. 					    #
#   Bioinformatics 27:2011-2012.								    #
#                                                                                                   #
#####################################################################################################


plot.linkcomm <- function(x, type = "", ...)
	# S3 method for "plot" generic function.
	# x is a "linkcomm" object.
	{
	switch(type,
		summary = plotLinkCommSumm(x, ...),
		members = plotLinkCommMembers(x, ...),
		dend = plotLinkCommDend(x, ...),
		graph = plotLinkCommGraph(x, ...),
		commsumm = plotLinkCommSummComm(x, ...)
		)
	}


plotLinkCommSumm <- function(x, col = TRUE, pal = brewer.pal(9,"Set1"), right = TRUE, droptrivial = TRUE, verbose = TRUE, ...)
	# x is a "linkcomm" object.
	{
	oldpar <- par(no.readonly = TRUE)
	# Set up and colour clusters in dendrogram.
	if(is.null(x$dendr)){
		dd <- as.dendrogram(x$hclust)
	}else{
		dd <- x$dendr
		}
	if(col && is.null(x$dendr)){
		cl <- unlist(x$clusters)
		crf <- colorRampPalette(pal,bias=1)
		cols <- crf(length(x$clusters))
		cols <- sample(cols,length(x$clusters),replace=FALSE)
		numnodes <- nrow(x$hclust$merge) + length(which(x$hclust$merge[,1]<0)) + length(which(x$hclust$merge[,2]<0))
		dd <- dendrapply(dd,.COL,height=x$pdmax,clusters=cl,cols=cols,labels=FALSE,numnodes=numnodes,droptrivial = droptrivial,verbose=verbose)
		if(verbose){cat("\n")}
		assign("i",0,environment(.COL))
		assign("memb",0,environment(.COL))
		assign("first",0,environment(.COL))
		assign("left",0,environment(.COL))
		}
	if(is.null(x$dendr) && right){ dd <- rev(dd)}
	grid.newpage()
	plot.new()
	# Set margin.
	margin<-unit(0.045,"npc")
	pushViewport(viewport(x=margin,y=margin,width=unit(1,"npc")-2*margin,height=unit(1,"npc")-2*margin,just=c("left","bottom")))
	pushViewport(viewport(layout=grid.layout(nrow=4,ncol=4,widths=unit(c(1,0.01,0.2,0.09),units=c("null",rep("native",3))),heights=unit(c(0.04,0.79,0.025,0.05),units=rep("npc",4)),respect=TRUE)))
	# Plot dendrogram using base plot function.
	pushViewport(viewport(layout.pos.row=1:3,layout.pos.col=1))
	#return(gridPLT())
	gpl <- c(0.009,0.71,0.13,0.902)
	par(oma=rep(0,4),mar=rep(0,4),ann=FALSE,omd=c(0,1,0,1),pty="m",mgp=rep(0,3),fig = gpl,xpd=NA,new=TRUE)
	plot(dd,axes=FALSE,leaflab="none")
	popViewport(1)
	# Title.
	pushViewport(viewport(layout.pos.row=1,layout.pos.col=1:3))
	title <- grid.text("Link Communities Dendrogram",x = unit(0.5,"npc"),y = unit(2,"npc"),draw=FALSE,name="title")
	title <- editGrob(title,gp = gpar(fontsize=14))
	grid.draw(title)
	popViewport(1)
	# Plot link partition densities.
	numzeros <- -1*log10(max(x$pdens[,2]))
	if(numzeros <= 1){ # Prevent partition density axis from being rounded to 0.
		rr <- 1
	}else{
		rr <- trunc(numzeros)+1
		}
	if(round(max(x$pdens[,2]), digits = rr) > max(x$pdens[,2])){
		xscale_add <- round(max(x$pdens[,2]), digits = rr) + 0.05*max(x$pdens[,2]) # Add 5% to part density x-axis.
		xaxs_max <- round(max(x$pdens[,2]), digits = rr)
	}else{
		xscale_add <- max(x$pdens[,2]) + 0.075*max(x$pdens[,2]) # 7.5% of max partition density added to x-axis.
		xaxs_max <- round(max(x$pdens[,2]), digits = (rr+1))
		}
	pushViewport(viewport(layout.pos.row=2, layout.pos.col=3, xscale=c(0,xscale_add),yscale=c(0,1)))
	ph <- x$pdens[,1]/max(x$pdens[,1])
	max <- x$pdmax/max(x$pdens[,1])

	grid.lines(x$pdens[,2],ph,gp=gpar(col='blue',lwd=2),default.units="native")
	xticks <- seq(0, xaxs_max, length.out=3)
	xa <- grid.xaxis(at = xticks,draw=FALSE,name="xa")
	xa <- editGrob(xa,gPath="ticks",y1 = unit(-0.02,"npc"))
	xa <- editGrob(xa,gPath="labels",gp = gpar(fontsize=10),y = unit(-0.04,"npc"))
	grid.draw(xa)
	xl <- grid.text("Partition Density",x=unit(0.5, "npc"), y = unit(-0.08, "npc"),draw=FALSE,name="xl")
	xl <- editGrob(xl,gp = gpar(fontsize=10))
	grid.draw(xl)
	popViewport(1)
	pushViewport(viewport(layout.pos.row=2,layout.pos.col=1:3))
	grid.lines(x=c(0,1),y=c(max,max),gp = gpar(col="red",lty=2,lwd=2))
	popViewport(1)
	pushViewport(viewport(layout.pos.row=2,layout.pos.col=4))
	yticks <- c(0,0.2,0.4,0.6,0.8,1)
	ya <- grid.yaxis(at = yticks,name="ya",draw=FALSE)
	ya <- editGrob(ya,gPath="ticks",x1 = unit(0.1,"npc"))
	ya <- editGrob(ya,gPath="labels",gp = gpar(fontsize=10),x = unit(0.2,"npc"))
	ya <- editGrob(ya,gPath="labels",just = c("left","centre"))
	if(max(x$pdens[,1]<1)){
		yu <- seq(0,round(max(x$pdens[,1]),2),length.out=6)
		roundS <- function(x){return(round(x,2))}
		yu <- sapply(yu,roundS)
		ya <- editGrob(ya,gPath="labels",label=as.character(yu))
		}
	grid.draw(ya)
	yl <- grid.text("Height",x=unit(0.8, "npc"), y = unit(0.5, "npc"),rot=90,draw=FALSE,name="yl")
	yl <- editGrob(yl,gp = gpar(fontsize=10))
	grid.draw(yl)
	popViewport(1)
	# Summary statistics.
	pushViewport(viewport(layout.pos.row=4,layout.pos.col=1))
	summ <- paste("# edges = ",x$numbers[1],",   ","# nodes = ",x$numbers[2],"\n# clusters = ",x$numbers[3],",   Largest cluster = ",x$clustsizes[1]," nodes\nHclust method: ",x$hclust$method)
	ne <- grid.text(summ,x=unit(0.5,"npc"),y=unit(0.1,"npc"),draw=FALSE,name="ne")
	ne <- editGrob(ne,gp = gpar(fontsize=11))
	grid.draw(ne)
	popViewport(1)
	# Return linkcomm object with dendrogram so we don't have to generate it again in the future.
	if(is.null(x$dendr)){
		x$dendr <- dd
		return(x)
		}
	popViewport(0)
	par(oldpar)
	}


plotLinkCommMembers <- function(x, nodes = head(names(x$numclusters),10), pal = brewer.pal(11,"Spectral"), shape = "rect", total=TRUE, fontsize=11, nspace = 3.5, maxclusters = 20)
	# Plots community membership matrix using a community-specific colour scheme.
	# x is a "linkcomm" object.
	{
	# Construct community matrix.
	comms <- unique(x$nodeclusters[as.character(x$nodeclusters[,1])%in%nodes,2]) # Community (cluster) IDs.
	if(length(comms) > maxclusters){
		comms <- comms[1:maxclusters]
		}
	commatrix <- getCommunityMatrix(x,nodes=nodes)
	crf <- colorRampPalette(pal,bias=1)
	cols <- crf(length(comms))
	grid.newpage()
	# Set margin.
	if(total){
		C <- 2; R <- 3
		nodesums <- apply(commatrix,1,sum)
		commsums <- apply(commatrix,2,sum)
	}else{
		C <- 1; R <- 2
		}
	margin<-unit(0.1,"lines")
	pushViewport(viewport(x=1,y=1,width=unit(1,"npc")-2*margin,height=unit(1,"npc")-2*margin,just=c("right","top")))
	pushViewport(viewport(layout=grid.layout(nrow=length(nodes)+R,ncol=length(comms)+C,widths=unit(c(nspace,rep(1,length(comms)+C-1)),rep("null",length(comms)+C)),heights=unit(rep(1,length(nodes)+R),rep("null",length(nodes)+R)),respect=TRUE)))
	# Titles.
	pushViewport(viewport(layout.pos.row=1,layout.pos.col=2:length(comms)+1))
	ctitle <- grid.text("Community Membership",x = unit(0.5,"npc"),y = unit(0.5,"npc"),draw=FALSE,name="ctitle")
	ctitle <- editGrob(ctitle,gp = gpar(fontsize=14))
	grid.draw(ctitle)
	popViewport(1)
	# Draw membership coloured squares/circles/polygons.
	for(i in 1:(length(nodes)+R-2)){
		if(i != length(nodes)+1){
			pushViewport(viewport(layout.pos.row=i+2,layout.pos.col=1))
			nname <- grid.text(as.character(nodes[i]),x = unit(0.9,"npc"),y = unit(0.5,"npc"),draw=FALSE,name="nname")
			nname <- editGrob(nname,gp = gpar(fontsize=fontsize),just="right")
			grid.draw(nname)
			popViewport(1)
			}
		for(j in 1:(length(comms)+C-1)){
			if(total && j == length(comms)+1 && i != length(nodes)+1){
				pushViewport(viewport(layout.pos.row=i+2,layout.pos.col=j+1))
				ntot <- grid.text(nodesums[i],x = unit(0.5,"npc"),y = unit(0.5,"npc"),draw=FALSE,name="ntot")
				ntot <- editGrob(ntot,gp = gpar(fontsize=12))
				grid.draw(ntot)
				popViewport(1)
				if(i == 1){
					pushViewport(viewport(layout.pos.row=2,layout.pos.col=j+1))
					rt <- grid.text(expression(Sigma),x = unit(0.5,"npc"),y = unit(0.5,"npc"),draw=FALSE,name="rt")
					rt <- editGrob(rt,gp = gpar(fontsize=12))
					grid.draw(rt)
					popViewport(1)
					}
			}else{
			if(i == 1 && j != length(comms)+1){
				pushViewport(viewport(layout.pos.row=2,layout.pos.col=j+1))
				rtitle <- grid.text(comms[j],x = unit(0.5,"npc"),y = unit(0.5,"npc"),draw=FALSE,name="rtitle")
				rtitle <- editGrob(rtitle,gp = gpar(fontsize=12))
				grid.draw(rtitle)
				popViewport(1)
				}
			if(total && i == length(nodes)+1 && j != length(comms)+1){
				if(j==1){
					pushViewport(viewport(layout.pos.row=i+2,layout.pos.col=1))
					ct <- grid.text(expression(Sigma),x = unit(0.9,"npc"),y = unit(0.5,"npc"),draw=FALSE,name="ct")
					ct <- editGrob(ct,gp = gpar(fontsize=12))
					grid.draw(ct)
					popViewport(1)
					}
				pushViewport(viewport(layout.pos.row=i+2,layout.pos.col=j+1))
				ctot <- grid.text(commsums[j],x = unit(0.5,"npc"),y = unit(0.5,"npc"),draw=FALSE,name="ctot")
				ctot <- editGrob(ctot,gp = gpar(fontsize=12))
				grid.draw(ctot)
				popViewport(1)
			}else if(i != length(nodes)+1 && j != length(comms)+1){
				if(commatrix[i,j] == 1){
					fill <- cols[j]
				}else{
					fill <- "white"
					}
				if(shape=="rect"){
					pushViewport(viewport(layout.pos.row=i+2,layout.pos.col=j+1))
					grid.rect(gp=gpar(fill=fill,col="grey"),width = unit(0.9,"npc"), height = unit(0.9,"npc"),draw=TRUE)
					popViewport(1)
				}else if(shape=="circle"){
					pushViewport(viewport(layout.pos.row=i+2,layout.pos.col=j+1))
					grid.circle(x=0.5,y=0.5,r=0.45,gp=gpar(fill=fill,col="grey"),draw=TRUE)
					popViewport(1)
					}
				}
			}
			}
		}
	}


plotLinkCommDend <- function(x, col=TRUE, pal = brewer.pal(9,"Set1"), height=x$pdmax, right = FALSE, labels=FALSE, plotcut=TRUE, droptrivial = TRUE, leaflab = "none", verbose = TRUE, ...)
	# x is a "linkcomm" object.
	{
	dd <- as.dendrogram(x$hclust)
	if(col){
		cl <- unlist(x$clusters)
		crf <- colorRampPalette(pal,bias=1)
		cols <- crf(length(x$clusters))
		cols <- sample(cols,length(x$clusters),replace=FALSE)
		numnodes <- nrow(x$hclust$merge) + length(which(x$hclust$merge[,1]<0)) + length(which(x$hclust$merge[,2]<0))
		dd <- dendrapply(dd, .COL, height=height, clusters=cl, cols=cols, labels=labels, numnodes = numnodes, droptrivial = droptrivial, verbose = verbose)
		cat("\n")
		assign("i",0,environment(.COL))
		assign("memb",0,environment(.COL))
		assign("first",0,environment(.COL))
		assign("left",0,environment(.COL))
		}
	if(right){
		dd <- rev(dd)
		}
	plot(dd,ylab="Height", leaflab = leaflab, ...)
	if(plotcut){
		abline(h=height,col='red',lty=2,lwd=2)
		}
	#ll <- sapply(x$clusters,length)
	#maxnodes <- length(unique(x$nodeclusters[x$nodeclusters[,2]%in%which(ll==max(ll)),1]))
	summ <- paste("# clusters = ",length(x$clusters),"\nLargest cluster = ",x$clustsizes[1]," nodes")
	mtext(summ, line = -28)
	}


.COL<-local({

memb <- 0
first <- 0
i <- 0
left <- 0

colorHclusters <<- function(x, height, clusters, cols, labels, numnodes, droptrivial, verbose)
	# Adds colours to edges that belong to clusters below "height" in the dendrogram.
	# Clusters gives leaf IDs for clusters that should be coloured.
	# x is a node in the tree.
	{
	left <<- left + 1
	if(verbose){
		out <- paste(c("   Colouring dendrogram... ",floor((left/numnodes)*100),"%"),collapse="")
		cat(out,"\r")
		flush.console()
		}

	if(round(attributes(x)$height,digits=5) > height){
		return(x)
	}else{
		if(is.leaf(x)){
			if(is.na(match(as.numeric(attributes(x)$label),clusters))){
				if(!labels){
					attributes(x)$label <- NULL
					}
				return(x)
				}
			}
		a <- attributes(x)
		if(memb == 0){
			memb <<- attributes(x)$members
			if(droptrivial == TRUE && memb == 2){
				memb <<- 0
				first <<- 1
			}else{
				i <<- i+1
				first <<- 1 # Because we don't colour the edge leading to the first node in a cluster.
				}
			}
		if(first == 0){
			attr(x,"edgePar") <- c(a$edgePar,list(col = cols[i], lwd = 2))
			}
		if(is.leaf(x)){
			if(!labels){
				attributes(x)$label <- NULL
				}
			memb <<- memb-1
			}
		first <<- 0
		}
	
	return(x)
	}

})


plotLinkCommGraph <- function(x, clusterids = 1:length(x$clusters), nodes = NULL, layout = layout.fruchterman.reingold, pal = brewer.pal(7,"Set2"), random = TRUE, node.pies = TRUE, pie.local = TRUE, vertex.radius = 0.03, scale.vertices = 0.05, edge.color = NULL, vshape = "none", vsize = 15, ewidth = 3, margin = 0, vlabel.cex = 0.8, vlabel.color = "black", vlabel.family = "Helvetica", vertex.color = "palegoldenrod", vlabel = TRUE, col.nonclusters = "black", jitter = 0.2, circle = TRUE, printids = TRUE, cid.cex = 1, shownodesin = 0, showall = FALSE, verbose = TRUE, ...)
	# x is a "linkcomm" object.
	{
	if(length(nodes) > 0){
		clusterids <- which.communities(x, nodes = nodes)
		}
	clusters <- x$clusters[clusterids]
	miss <- setdiff(x$hclust$order,unlist(clusters))
	crf <- colorRampPalette(pal,bias=1)
	cols <- crf(length(clusters))
	if(random){
		cols <- sample(cols,length(clusters),replace=FALSE)
		}
	if(showall){
		# Add single edge "clusters".
		single <- setdiff(1:x$numbers[1],unlist(clusters))
		ll <- length(clusters)
		for(i in 1:length(single)){
			clusters[[(i+ll)]] <- single[i]
			}
		cols <- append(cols, rep(col.nonclusters, length(single)))
		}
	drawcircle <- FALSE
	if(inherits(layout,"character")){
		if(layout == "spencer.circle"){
			if(length(clusters) > length(x$clusters[1:x$numbers[3]])){
				clusterids <- 1:x$numbers[3]
				}
			ord <- orderCommunities(x, clusterids = clusterids, verbose = FALSE)
			clusters <- ord$ordered
			clusterids <- ord$clusids
			layout <- layout.spencer.circle(x, clusterids = clusterids, jitter = jitter, verbose = verbose)$nodes
			drawcircle <- TRUE
			}
		}
	names(cols) <- clusterids
	if(length(unlist(clusters)) < nrow(x$edgelist) || length(miss) == 0){
		# Convert old clus ids into new ones.
		edges <- x$edgelist[unlist(clusters),]
		ig <- graph.edgelist(edges, directed=x$directed)
		clen <- sapply(clusters,length)
		j<-1
		# Colour edges according to community membership.
		for(i in 1:length(clusters)){
			newcids <- j:sum(clen[1:i])
			E(ig)[newcids]$color <- cols[i]
			j <- tail(newcids,1)+1
			}
	}else{
		ig <- x$igraph
		for(i in 1:length(clusters)){
			E(ig)[clusters[[i]]]$color <- cols[i]
			}
		}
	
	if(shownodesin == 0){
		vnames <- V(ig)$name
	}else{ # Show nodes that belong to more than x number of communities.
		vnames <- V(ig)$name
		inds <- NULL
		for(i in 1:length(vnames)){
			if(x$numclusters[which(names(x$numclusters)==vnames[i])] < shownodesin){
				inds <- append(inds,i)
				}
			}
		vnames[inds] <- ""
		}
	if(vlabel==FALSE){
		vnames = NA
		}

	dev.hold(); on.exit(dev.flush())
	oldpar <- par(no.readonly = TRUE)
	par(mar = c(4,4,2,2))

	if(!node.pies){
		plot(ig, layout=layout, vertex.shape=vshape, edge.width=ewidth, vertex.label=vnames, vertex.label.family=vlabel.family, vertex.label.color=vlabel.color, vertex.size=vsize, vertex.color=vertex.color, margin=margin, vertex.label.cex = vlabel.cex, ...)
	}else{
		nodes <- V(ig)$name
		# Get node community membership by edges.
		if(pie.local){
			edge.memb <- numberEdgesIn(x, clusterids = clusterids, nodes = nodes)
		}else{
			edge.memb <- numberEdgesIn(x, nodes = nodes)
			}

		cat("   Getting node layout...")
		if(inherits(layout,"function")){
			lay <- layout(ig)
		}else{
			lay <- layout
			}
		lay <- layout.norm(lay, xmin=-1, xmax=1, ymin=-1, ymax=1)
		rownames(lay) <- V(ig)$name
		cat("\n")
		node.pies <- .nodePie(edge.memb=edge.memb, layout=lay, nodes=nodes, edges=100, radius=vertex.radius, scale=scale.vertices)
		cat("\n")
		# Plot graph.
		if(is.null(edge.color)){
			plot(ig, layout=lay, vertex.shape="none", vertex.label=NA, vertex.label.dist=1, edge.width=ewidth, vertex.label.color=vlabel.color, ...)
		}else{
			plot(ig, layout=lay, vertex.shape="none", vertex.label=NA, vertex.label.dist=1, edge.width=ewidth, vertex.label.color=vlabel.color, edge.color=edge.color, ...)
			}
		labels <- list()
		# Plot node pies and node names.
		for(i in 1:length(node.pies)){
			yp <- NULL
			for(j in 1:length(node.pies[[i]])){
				seg.col <- cols[which(names(cols)==names(edge.memb[[i]])[j])]
				polygon(node.pies[[i]][[j]][,1], node.pies[[i]][[j]][,2], col = seg.col)
				yp <- append(yp, node.pies[[i]][[j]][,2])
				}
			lx <- lay[which(rownames(lay)==names(node.pies[i])),1] + 0.1
			ly <- max(yp) + 0.02 # Highest point of node pie.
			labels[[i]] <- c(lx, ly)
			}
		# Plot node names after nodes so they overlay them.
		for(i in 1:length(labels)){
			text(labels[[i]][1], labels[[i]][2], labels = vnames[which(nodes==names(node.pies[i]))], cex = vlabel.cex, col = vlabel.color)
			}
		}

	if(circle && drawcircle){
		# Add circle for Spencer layout.
		cx<-NULL; for(i in 1:100){cx[i]<-1.25*cos(i*(2*pi)/100)}
		cy<-NULL; for(i in 1:100){cy[i]<-1.25*sin(i*(2*pi)/100)}
		polygon(cx-0.08,cy-0.08, border="grey",lwd=2)
		# Add community anchor points and cluster IDs.
		for(i in 1:length(clusters)){
			px <- 1.1*cos(i*(2*pi)/length(clusters))
			py <- 1.1*sin(i*(2*pi)/length(clusters))
			points(px-0.08,py-0.08, pch = 20, col = cols[i])
			if(printids){
				tx <- 1.3*cos(i*(2*pi)/length(clusters))
				ty <- 1.3*sin(i*(2*pi)/length(clusters))
				text(tx-0.08,ty-0.08, labels = clusterids[i], col = cols[i], cex = cid.cex, font=2)
				}
			}
		}
	par(oldpar)
	
	}



layout.spencer.circle <- function(x, clusterids = 1:x$numbers[3], verbose = TRUE, jitter = 0.2)
	# Returns x-y node coordinates for Rob Spencer's circular layout of link communities together with x-y coordinates for the community anchors.
	# x is a "linkcomm" object.
	{
	clusters <- x$clusters[clusterids]
	edges <- x$edgelist[unlist(clusters),]
	ig <- graph.edgelist(edges, directed=FALSE)
	# Put communities in dendrogram order.
	clusters <- orderCommunities(x, clusterids = clusterids, verbose = verbose)$ordered
	# Set up community anchor points in Cartesian coordinates around unit circle (communities evenly spaced).
	xy_anchors <- matrix(0,length(clusters),2)
	for(i in 1:length(clusters)){
		xy_anchors[i,] <- c(cos(i*(2*pi)/length(clusters)), sin(i*(2*pi)/length(clusters)))
		}
	# Calculate community membership percentages per node.
	nodes <- c(x$edgelist[unlist(clusters),1],x$edgelist[unlist(clusters),2])
	node_names <- unique(nodes)
	xy_nodes <- matrix(0,length(node_names),2)
	for(i in 1:length(node_names)){
		if(verbose){
			mes <- paste(c("   Calculating node co-ordinates for Spencer circle...",floor(i/(length(node_names))*100),"%"),collapse="")
			cat(mes,'\r')
			flush.console()
			}
		freqs <- NULL
		total <- length(which(nodes==node_names[i]))
		for(j in 1:length(clusters)){
			freqs <- length(which(c(x$edgelist[clusters[[j]],1],x$edgelist[clusters[[j]],2])==node_names[i]))/total
			# Update x-y coordinates for this node.
			xy_nodes[i,1] <- sum(xy_nodes[i,1], freqs*xy_anchors[j,1])
			xy_nodes[i,2] <- sum(xy_nodes[i,2], freqs*xy_anchors[j,2])
			}
		# Add random jitter if this node has identical x-y coordinates to an earlier node.
		if(duplicated(xy_nodes)[i]){
			xy_nodes[i,1] <- sum(xy_nodes[i,1], runif(1, min = -jitter, max = jitter))
			xy_nodes[i,2] <- sum(xy_nodes[i,2], runif(1, min = -jitter, max = jitter))
			}
		}

	rownames(xy_nodes) <- node_names
	
	xy_nodes <- xy_nodes[match(V(ig)$name,rownames(xy_nodes)),]
	xy_nodes <- xy_nodes[!is.na(xy_nodes[,1]),]

	if(verbose){cat("\n")}
	
	xy <- list()
	xy$nodes <- xy_nodes
	xy$anchors <- xy_anchors

	return(xy)

	}


plotLinkCommSummComm <- function(x, clusterids = 1:x$numbers[3], summary = "conn", pie = FALSE, col = TRUE, pal = brewer.pal(11,"Spectral"), random = FALSE, verbose = TRUE, ...) 
	# Plots pie or bar chart summarising sizes of communities in terms of nodes, link density, community connectedness, or community modularity.
	# x is a "linkcomm" object.
	{
	if(col){
		crf <- colorRampPalette(pal,bias=1)
		cols <- crf(length(clusterids))
		if(random){
			cols <- sample(cols,length(clusterids),replace=FALSE)
			}
	}else{
		cols <- "lightblue"
		}
	# Extract number of nodes per community.
	nums <- NULL
	if(summary == "nodes"){
		for(i in 1:length(clusterids)){
			nums[i] <- length(unique(c(x$edgelist[x$clusters[[clusterids[i]]],1],x$edgelist[x$clusters[[clusterids[i]]],2])))
			}
		main <- "Node density per community"
	}else if(summary == "ld"){
		nums <- LinkDensities(x, clusterids = clusterids)
		main <- "Link density per community"
	}else{
		nums <- getCommunityConnectedness(x, clusterids = clusterids, conn = summary, verbose = verbose)
		if(summary == "conn"){
			main <- "Community Connectedness"
		}else{
			main <- "Community Modularity"
			}
		}
	names(nums) <- clusterids

	if(pie){
		pie(nums, col = cols, main = main, ...)
	}else{
		barplot(nums, xlab = "Community", ylab = main, col = cols)
		abline(h=0)
		}

	}