X-Git-Url: http://iramuteq.org/git?p=iramuteq;a=blobdiff_plain;f=Rscripts%2FRgraph.R;h=39fe61839b1585d5e683f161e36a5e529a804e5b;hp=a4dd0da5621184921750b40f9c888c0d57265693;hb=a6c0faa62f33e5e64ad7eab5b9d804be99c6759a;hpb=42a67a41b64a6e0cc3fd2a63a0749e9aa4b9374c diff --git a/Rscripts/Rgraph.R b/Rscripts/Rgraph.R index a4dd0da..39fe618 100644 --- a/Rscripts/Rgraph.R +++ b/Rscripts/Rgraph.R @@ -43,7 +43,7 @@ PlotDendroCut <- function(chd,filename,reso,clusternb) { # dev.off() #} -PlotAfc2dCoul<- function(afc,chisqrtable,filename, what='coord',col=FALSE, axetoplot=c(1,2), deb=0,fin=0, width=900, height=900, quality=100, reso=200, parcex=PARCEX, xlab = NULL, ylab = NULL, xmin=NULL, xmax=NULL, ymin=NULL, ymax=NULL) { +PlotAfc2dCoul<- function(afc,chisqrtable,filename, what='coord',col=FALSE, axetoplot=c(1,2), deb=0,fin=0, width=900, height=900, quality=100, reso=200, parcex=PARCEX, xlab = NULL, ylab = NULL, xmin=NULL, xmax=NULL, ymin=NULL, ymax=NULL, active = TRUE) { if (col) { if (what == 'coord') { rowcoord <- as.matrix(afc$colcoord) @@ -75,7 +75,7 @@ PlotAfc2dCoul<- function(afc,chisqrtable,filename, what='coord',col=FALSE, axeto if (!col) { classes <- as.matrix(apply(chitable,1,which.max)) cex.par <- norm.vec(apply(chitable,1,max), 0.8,3) - row.keep <- select.chi.classe(chitable, 60) + row.keep <- select.chi.classe(chitable, 80, active=active) rowcoord <- rowcoord[row.keep,] classes <- classes[row.keep] cex.par <- cex.par[row.keep] @@ -106,7 +106,12 @@ PlotAfc2dCoul<- function(afc,chisqrtable,filename, what='coord',col=FALSE, axeto table.in <- rowcoord[order(cex.par, decreasing = TRUE),] classes <- classes[order(cex.par, decreasing = TRUE)] cex.par <- cex.par[order(cex.par, decreasing = TRUE)] - table.in <- stopoverlap(table.in, cex.par=cex.par, xlim = c(xmin,xmax), ylim = c(ymin,ymax)) + table.out <- stopoverlap(table.in, cex.par=cex.par, xlim = c(xmin,xmax), ylim = c(ymin,ymax)) + table.in <- table.out$toplot + notplot <- table.out$notplot + if (! is.null(notplot)) { + write.csv2(notplot, file = paste(filename, '_notplotted.csv', sep = '')) + } classes <- classes[table.in[,4]] cex.par <- cex.par[table.in[,4]] make_afc_graph(table.in, classes, clnb, xlab, ylab, cex.txt = cex.par, xminmax=c(xmin,xmax), yminmax=c(ymin,ymax)) @@ -137,7 +142,6 @@ open_file_graph <- function (filename, width=800, height = 800, quality = 100, s svg(filename.to.svg(filename), width=width, height=height) } } else { - #library(RSvgDevice) if (svg) { svg(filename.to.svg(filename), width=width/74.97, height=height/74.97) } else { @@ -191,6 +195,7 @@ stopoverlap <- function(x, cex.par = NULL, xlim = NULL, ylim = NULL) { thetaStep <- .1 rStep <- .5 toplot <- NULL + notplot <- NULL # plot.new() plot(x[,1],x[,2], pch='', xlim = xlim, ylim = ylim) @@ -222,6 +227,7 @@ stopoverlap <- function(x, cex.par = NULL, xlim = NULL, ylim = NULL) { } else { if(r>sqrt(.5)){ print(paste(words[i], "could not be fit on page. It will not be plotted.")) + notplot <- rbind(notplot,c(words[i], x[i,1], x[i,2])) isOverlaped <- FALSE } theta <- theta+thetaStep @@ -232,10 +238,111 @@ stopoverlap <- function(x, cex.par = NULL, xlim = NULL, ylim = NULL) { } } row.names(toplot) <- words[toplot[,4]] - return(toplot) + return(list(toplot = toplot, notplot = notplot)) } ############################################################################### +getwordcloudcoord <- function(words,freq,scale=c(4,.5),min.freq=3,max.words=Inf,random.order=TRUE,random.color=FALSE, + rot.per=.1,colors="black",ordered.colors=FALSE,use.r.layout=FALSE,fixed.asp=TRUE,...) { + tails <- "g|j|p|q|y" + last <- 1 + + overlap <- function(x1, y1, sw1, sh1) { + if(!use.r.layout) + return(.overlap(x1,y1,sw1,sh1,boxes)) + s <- 0 + if (length(boxes) == 0) + return(FALSE) + for (i in c(last,1:length(boxes))) { + bnds <- boxes[[i]] + x2 <- bnds[1] + y2 <- bnds[2] + sw2 <- bnds[3] + sh2 <- bnds[4] + if (x1 < x2) + overlap <- x1 + sw1 > x2-s + else + overlap <- x2 + sw2 > x1-s + + if (y1 < y2) + overlap <- overlap && (y1 + sh1 > y2-s) + else + overlap <- overlap && (y2 + sh2 > y1-s) + if(overlap){ + last <<- i + return(TRUE) + } + } + FALSE + } + + ord <- rank(-freq, ties.method = "random") + words <- words[ord<=max.words] + freq <- freq[ord<=max.words] + + + ord <- order(freq,decreasing=TRUE) + words <- words[ord] + freq <- freq[ord] + words <- words[freq>=min.freq] + freq <- freq[freq>=min.freq] + if (ordered.colors) { + colors <- colors[ord][freq>=min.freq] + } + + thetaStep <- .1 + rStep <- .05 + plot.new() + + normedFreq <- freq/max(freq) + size <- (scale[1]-scale[2])*normedFreq + scale[2] + boxes <- list() + toplot <- NULL + + + for(i in 1:length(words)){ + rotWord <- runif(1)0 && y1-.5*ht>0 && + x1+.5*wid<1 && y1+.5*ht<1){ + toplot <- rbind(toplot, c(x1,y1,size[i], i)) + boxes[[length(boxes)+1]] <- c(x1-.5*wid,y1-.5*ht,wid,ht) + isOverlaped <- FALSE + }else{ + if(r>sqrt(.5)){ + warning(paste(words[i], + "could not be fit on page. It will not be plotted.")) + isOverlaped <- FALSE + } + theta <- theta+thetaStep + r <- r + rStep*thetaStep/(2*pi) + x1 <- .5+r*cos(theta) + y1 <- .5+r*sin(theta) + } + } + } + toplot <- cbind(toplot,norm.vec(freq[toplot[,4]], 1, 50)) + row.names(toplot) <- words[toplot[,4]] + toplot <- toplot[,-4] + return(toplot) +} + make_tree_tot <- function (chd) { library(ape) lf<-chd$list_fille @@ -309,8 +416,13 @@ select_point_chi <- function(tablechi, chi_limit) { row_keep } -select.chi.classe <- function(tablechi, nb) { +select.chi.classe <- function(tablechi, nb, active = TRUE) { rowkeep <- NULL + if (active & !is.null(debsup)) { + print(debsup) + print('###############################################################@') + tablechi <- tablechi[1:(debsup-1),] + } if (nb > nrow(tablechi)) { nb <- nrow(tablechi) } @@ -434,6 +546,29 @@ create_afc_table <- function(x) { res } +is.yellow <- function(my.color) { + if ((my.color[1] > 200) & (my.color[2] > 200) & (my.color[3] < 20)) { + return(TRUE) + } else { + return(FALSE) + } +} + +del.yellow <- function(colors) { + rgbs <- col2rgb(colors) + tochange <- apply(rgbs, 2, is.yellow) + tochange <- which(tochange) + if (length(tochange)) { + gr.col <- grey.colors(length(tochange), start = 0.5) + } + compt <- 1 + for (val in tochange) { + colors[val] <- gr.col[compt] + compt <- compt + 1 + } + colors +} + make_afc_graph <- function(toplot, classes, clnb, xlab, ylab, cex.txt = NULL, leg = FALSE, cmd = FALSE, black = FALSE, xminmax=NULL, yminmax=NULL) { rain <- rainbow(clnb) @@ -463,6 +598,8 @@ make_afc_graph <- function(toplot, classes, clnb, xlab, ylab, cex.txt = NULL, le if (is.null(cex.txt)) text(toplot[,1],toplot[,2],rownames(toplot),col=cl.color, offset=0) else + #require(wordcloud) + #textplot(toplot[,1],toplot[,2],rownames(toplot),col=cl.color, cex = cex.txt, xlim=xminmax, ylim = yminmax) text(toplot[,1],toplot[,2],rownames(toplot),col=cl.color, cex = cex.txt, offset=0) if (!cmd) { @@ -470,18 +607,131 @@ make_afc_graph <- function(toplot, classes, clnb, xlab, ylab, cex.txt = NULL, le } } +plot.dendro.prof <- function(tree, classes, chisqtable, nbbycl = 60, type.dendro = "phylogram", from.cmd = FALSE, bw = FALSE, lab = NULL) { + library(ape) + library(wordcloud) + classes<-classes[classes!=0] + classes<-as.factor(classes) + sum.cl<-as.matrix(summary(classes, maxsum=1000000)) + sum.cl<-(sum.cl/colSums(sum.cl)*100) + sum.cl<-round(sum.cl,2) + sum.cl<-cbind(sum.cl,as.matrix(100-sum.cl[,1])) + sum.cl <- sum.cl[,1] + tree.order<- as.numeric(tree$tip.label) + vec.mat<-NULL + row.keep <- select.chi.classe(chisqtable, nbbycl) + toplot <- chisqtable[row.keep,] + lclasses <- list() + for (classe in 1:length(sum.cl)) { + ntoplot <- toplot[,classe] + names(ntoplot) <- rownames(toplot) + ntoplot <- ntoplot[order(ntoplot, decreasing = TRUE)] + ntoplot <- round(ntoplot, 0) + ntoplot <- ntoplot[1:nbbycl] + #ntoplot <- ntoplot[order(ntoplot)] + #ntoplot <- ifelse(length(ntoplot) > nbbycl, ntoplot[1:nbbycl], ntoplot) + lclasses[[classe]] <- ntoplot + } + vec.mat <- matrix(1, nrow = 3, ncol = length(sum.cl)) + vec.mat[2,] <- 2 + vec.mat[3,] <- 3:(length(sum.cl)+2) + layout(matrix(vec.mat, nrow=3, ncol=length(sum.cl)),heights=c(2,1,6)) + if (! bw) { + col <- rainbow(length(sum.cl))[as.numeric(tree$tip.label)] + col <- del.yellow(col) + colcloud <- rainbow(length(sum.cl)) + colcloud <- del.yellow(colcloud) + } + label.ori<-tree[[2]] + if (!is.null(lab)) { + tree$tip.label <- lab + } else { + tree[[2]]<-paste('classe ',tree[[2]]) + } + par(mar=c(2,1,0,1)) + plot.phylo(tree,label.offset=0, tip.col=col, type=type.dendro, direction = 'downwards', srt=90, adj = 0.5, cex = 1.4, y.lim=c(-0.3,tree$Nnode)) + par(mar=c(0,0,0,0)) + d <- barplot(-sum.cl[tree.order], col=col, names.arg='', axes=FALSE, axisname=FALSE) + text(x=d, y=(-sum.cl[tree.order]+3), label=paste(round(sum.cl[tree.order],1),'%'), cex=1.4) + for (i in tree.order) { + par(mar=c(0,0,1,0),cex=0.7) + #wordcloud(names(lclasses[[i]]), lclasses[[i]], scale = c(1.5, 0.2), random.order=FALSE, colors = colcloud[i]) + yval <- 1.1 + plot(0,0,pch='', axes = FALSE) + vcex <- norm.vec(lclasses[[i]], 1.5, 2.5) + for (j in 1:length(lclasses[[i]])) { + yval <- yval-(strheight( names(lclasses[[i]])[j],cex=vcex[j])+0.02) + text(-0.9, yval, names(lclasses[[i]])[j], cex = vcex[j], col = colcloud[i], adj=0) + } + } + +} + +plot.dendro.cloud <- function(tree, classes, chisqtable, nbbycl = 60, type.dendro = "phylogram", from.cmd = FALSE, bw = FALSE, lab = NULL) { + library(wordcloud) + library(ape) + classes<-classes[classes!=0] + classes<-as.factor(classes) + sum.cl<-as.matrix(summary(classes, maxsum=1000000)) + sum.cl<-(sum.cl/colSums(sum.cl)*100) + sum.cl<-round(sum.cl,2) + sum.cl<-cbind(sum.cl,as.matrix(100-sum.cl[,1])) + sum.cl <- sum.cl[,1] + tree.order<- as.numeric(tree$tip.label) + vec.mat<-NULL + row.keep <- select.chi.classe(chisqtable, nbbycl) + toplot <- chisqtable[row.keep,] + lclasses <- list() + for (classe in 1:length(sum.cl)) { + ntoplot <- toplot[,classe] + ntoplot <- ntoplot[order(ntoplot, decreasing = TRUE)] + ntoplot <- round(ntoplot, 0) + ntoplot <- ntoplot[1:nbbycl] + ntoplot <- ntoplot[order(ntoplot)] + #ntoplot <- ifelse(length(ntoplot) > nbbycl, ntoplot[1:nbbycl], ntoplot) + lclasses[[classe]] <- ntoplot + } + for (i in 1:length(sum.cl)) vec.mat<-append(vec.mat,1) + v<-2 + for (i in 1:length(sum.cl)) { + vec.mat<-append(vec.mat,v) + v<-v+1 + } + layout(matrix(vec.mat,length(sum.cl),2),widths=c(1,2)) + if (! bw) { + col <- rainbow(length(sum.cl))[as.numeric(tree$tip.label)] + colcloud <- rainbow(length(sum.cl)) + } + par(mar=c(0,0,0,0)) + label.ori<-tree[[2]] + if (!is.null(lab)) { + tree$tip.label <- lab + } else { + tree[[2]]<-paste('classe ',tree[[2]]) + } + plot.phylo(tree,label.offset=0.1,tip.col=col, type=type.dendro) + for (i in rev(tree.order)) { + par(mar=c(0,0,1,0),cex=0.9) + wordcloud(names(lclasses[[i]]), lclasses[[i]], scale = c(4, 0.8), random.order=FALSE, colors = colcloud[i]) + } +} + plot.dendropr <- function(tree, classes, type.dendro="phylogram", histo=FALSE, from.cmd=FALSE, bw=FALSE, lab = NULL, tclasse=TRUE) { classes<-classes[classes!=0] classes<-as.factor(classes) - sum.cl<-as.matrix(summary(classes)) + sum.cl<-as.matrix(summary(classes, maxsum=1000000)) sum.cl<-(sum.cl/colSums(sum.cl)*100) sum.cl<-round(sum.cl,2) sum.cl<-cbind(sum.cl,as.matrix(100-sum.cl[,1])) tree.order<- as.numeric(tree$tip.label) + + if (! bw) { - col = rainbow(nrow(sum.cl))[as.numeric(tree$tip.label)] + col <- rainbow(nrow(sum.cl))[as.numeric(tree$tip.label)] + col <- del.yellow(col) col.bars <- col col.pie <- rainbow(nrow(sum.cl)) + col.pie <- del.yellow(col.pie) #col.vec<-rainbow(nrow(sum.cl))[as.numeric(tree[[2]])] } else { col = 'black' @@ -533,10 +783,24 @@ plot.dendropr <- function(tree, classes, type.dendro="phylogram", histo=FALSE, f } #tree <- tree.cut1$tree.cl #to.plot <- di -plot.dendro.lex <- function(tree, to.plot, bw=FALSE, lab=NULL, lay.width=c(3,3,2), cmd=FALSE) { +plot.dendro.lex <- function(tree, to.plot, bw=FALSE, lab=NULL, lay.width=c(3,3,2), colbar=NULL, classes=NULL, cmd=FALSE) { tree.order<- as.numeric(tree$tip.label) + if (!is.null(classes)) { + classes<-classes[classes!=0] + classes<-as.factor(classes) + sum.cl<-as.matrix(summary(classes, maxsum=1000000)) + sum.cl<-(sum.cl/colSums(sum.cl)*100) + sum.cl<-round(sum.cl,2) + sum.cl<-cbind(sum.cl,as.matrix(100-sum.cl[,1])) + } par(mar=c(0,0,0,0)) - layout(matrix(c(1,2,3),1,byrow=TRUE), widths=lay.width,TRUE) + if (!is.null(classes)) { + matlay <- matrix(c(1,2,3,4),1,byrow=TRUE) + lay.width <- c(3,1,3,2) + } else { + matlay <- matrix(c(1,2,3),1,byrow=TRUE) + } + layout(matlay, widths=lay.width,TRUE) par(mar=c(3,0,2,0),cex=1) label.ori<-tree[[2]] if (!is.null(lab)) { @@ -546,15 +810,27 @@ plot.dendro.lex <- function(tree, to.plot, bw=FALSE, lab=NULL, lay.width=c(3,3,2 } to.plot <- matrix(to.plot[,tree.order], nrow=nrow(to.plot), dimnames=list(rownames(to.plot), colnames(to.plot))) if (!bw) { - col <- rainbow(ncol(to.plot)) - col.bars <- rainbow(nrow(to.plot)) + col <- rainbow(ncol(to.plot)) + col <- del.yellow(col) + if (is.null(colbar)) { + col.bars <- rainbow(nrow(to.plot)) + col.bars <- del.yellow(col.bars) + } else { + col.bars <- colbar + } } else { col <- 'black' col.bars <- grey.colors(nrow(to.plot),0,0.8) } col <- col[tree.order] plot.phylo(tree,label.offset=0.1,tip.col=col) - + if (!is.null(classes)) { + par(cex=0.7) + par(mar=c(3,0,2,1)) + to.plota <- sum.cl[tree.order,1] + d <- barplot(to.plota,horiz=TRUE, col=col, names.arg='', axes=FALSE, axisname=FALSE) + text(x=to.plota, y=d[,1], label=paste(round(to.plota,1),'%'), adj=1.2) + } par(mar=c(3,0,2,1)) d <- barplot(to.plot,horiz=TRUE, col=col.bars, beside=TRUE, names.arg='', space = c(0.1,0.6), axisname=FALSE) c <- colMeans(d) @@ -716,3 +992,135 @@ vire.nonascii <- function(rnames) { #d <- barplot(to.plot,horiz=TRUE, names.arg='', axes=FALSE, axisname=FALSE) #text(x=to.plot, y=d[,1], label=round(to.plot,1), adj=1.2) +make.afc.attributes <- function(rn, afc.table, contafc, clnb, column = FALSE, x=1, y=2) { + if (!column){ + nd <- clnb - 1 + afc.res <- afc.table$ligne + #tokeep <- which(row.names(afc.res) %in% rn) + afc.res <- afc.res[rn,] + debcor <- (nd*2) + 1 + cor <- afc.res[,debcor:(debcor+nd-1)][,c(x,y)] + debctr <- (nd*3) + 1 + ctr <- afc.res[,debctr:(debctr+nd-1)][,c(x,y)] + massdeb <- (nd*4) + 1 + mass <- afc.res[,massdeb] + chideb <- massdeb + 1 + chi <- afc.res[,chideb] + inertiadeb <- chideb + 1 + inertia <- afc.res[,inertiadeb] + frequence <- rowSums(contafc[rn,]) + } + res <- list(frequence=frequence, cor, ctr, mass = mass, chi=chi, inertia=inertia) + return(res) +} + + +afctogexf <- function(fileout, toplot, classes, clnb, sizes, nodes.attr=NULL) { + toplot <- toplot[,1:3] + toplot[,3] <- 0 + #toplot <- afc$rowcoord[1:100,1:3] + #toplot[,3] <- 0 + #rownames(toplot)<-afc$rownames[1:100] + cc <- rainbow(clnb)[classes] + cc <- t(sapply(cc, col2rgb, alpha=TRUE)) + #sizes <- apply(chistabletot[1:100,], 1, max) + + nodes <- data.frame(cbind(1:nrow(toplot), rownames(toplot))) + colnames(nodes) <- c('id', 'label') + nodes[,1] <- as.character(nodes[,1]) + nodes[,2] <- as.character(nodes[,2]) + #nodes attributs + if (! is.null(nodes.attr)) { + nodesatt <- as.data.frame(nodes.attr) + } else { + nodesatt <- data.frame(cbind(toplot[,1],toplot[,2])) + } + #make axes + edges<-matrix(c(1,1),ncol=2) + xmin <- min(toplot[,1]) + xmax <- max(toplot[,1]) + ymin <- min(toplot[,2]) + ymax <- max(toplot[,2]) + nodes<-rbind(nodes, c(nrow(nodes)+1, 'F1')) + nodes<-rbind(nodes, c(nrow(nodes)+1, 'F1')) + nodes<-rbind(nodes, c(nrow(nodes)+1, 'F2')) + nodes<-rbind(nodes, c(nrow(nodes)+1, 'F2')) + nodesatt<-rbind(nodesatt, c(0,0)) + nodesatt<-rbind(nodesatt, c(0,0)) + nodesatt<-rbind(nodesatt, c(0,0)) + nodesatt<-rbind(nodesatt, c(0,0)) + toplot <- rbind(toplot, c(xmin, 0,0)) + toplot <- rbind(toplot, c(xmax,0,0)) + toplot <- rbind(toplot, c(0,ymin,0)) + toplot <- rbind(toplot, c(0,ymax,0)) + cc <- rbind(cc, c(255,255,255,1)) + cc <- rbind(cc, c(255,255,255,1)) + cc <- rbind(cc, c(255,255,255,1)) + cc <- rbind(cc, c(255,255,255,1)) + sizes <- c(sizes, c(0.5, 0.5, 0.5, 0.5)) + edges <- rbind(edges, c(nrow(nodes)-3, nrow(nodes)-2)) + edges <- rbind(edges, c(nrow(nodes)-1, nrow(nodes))) + write.gexf(nodes, edges, output=fileout, nodesAtt=nodesatt, nodesVizAtt=list(color=cc, position=toplot, size=sizes)) +} + +simi.to.gexf <- function(fileout, graph.simi, nodes.attr = NULL) { + lo <- graph.simi$layout + if (ncol(lo) == 3) { + lo[,3] <- 0 + } else { + lo <- cbind(lo, rep(0,nrow(lo))) + } + g <- graph.simi$graph + nodes <- data.frame(cbind(1:nrow(lo), V(g)$name)) + colnames(nodes) <- c('id', 'label') + print(nodes) + if (! is.null(nodes.attr)) { + nodesatt <- as.data.frame(nodes.attr) + } else { + nodesatt <- data.frame(cbind(lo[,1],lo[,2])) + } + edges <- as.data.frame(get.edges(g, c(1:ecount(g)))) + col <- rep('red', nrow(lo)) + col <- t(sapply(col, col2rgb, alpha=TRUE)) + write.gexf(nodes, edges, output=fileout, nodesAtt=nodesatt, nodesVizAtt=list(color=col,position=lo)) +} + + +graph.to.file <- function(grah.simi, nodesfile = NULL, edgesfile = NULL, community = FALSE, color = NULL, sweight = NULL) { + require(igraph) + g <- graph.simi$graph + V(g)$weight <- graph.simi$eff + V(g)$x <- graph.simi$layout[,1] + V(g)$y <- graph.simi$layout[,2] + if (ncol(graph.simi$layout) == 3) { + V(g)$z <- graph.simi$layout[,3] + } + if (community) { + member <- graph.simi$communities$membership + col <- rainbow(max(member)) + v.colors <- col[member] + v.colors <- col2rgb(v.colors) + V(g)$r <- v.colors[1,] + V(g)$g <- v.colors[2,] + V(g)$b <- v.colors[3,] + } + if (!is.null(color)) { + v.colors <- col2rgb(color) + V(g)$r <- v.colors[1,] + V(g)$g <- v.colors[2,] + V(g)$b <- v.colors[3,] + } + if (!is.null(sweight)) { + V(g)$sweight <- sweight + } + df <- get.data.frame(g, what='both') + if (!is.null(nodesfile)) { + write.table(df$vertices, nodesfile, sep='\t') + } + if (!is.null(edgesfile)) { + write.table(df$edges, edgesfile, sep='\t') + } + if (is.null(edgesfile) & is.null(nodesfile)) { + df + } +}