jadeTF:

plot_data <- function(counts, reads, pos, ylim=c(0, 1), cols=NULL,
                      range=NULL, sep.tab=NULL, bg.color="blue",
                      shapes = NULL,
                      xlab="Position", ylab="Methylation Proportion"){
  if(!is.matrix(counts)){
				K=1
				p <- length(counts)
				counts <- matrix(counts, nrow=p)
				reads <- matrix(reads, nrow=p)
  }else{
		p <- dim(counts)[1]
		K <- dim(counts)[2]
  }
	cat(p, K, "\n")
  if(K==1) sv <- TRUE
  else sv <- FALSE

	if(is.null(cols)){
	  cols=c("black", "red3", "blue2")
	  if(K >3) cols <- c(cols, 4:(K-3))
	}
	if(is.null(shapes)) shapes <- rep(20, K)
	if(is.null(range)){
		range <- range(pos)
		idx <- 1:p
	}else{
		idx <- which(pos <= max(range) & pos >= min(range))
		pos <- pos[idx]
		cat(dim(fits), "\n")
	}
	p <- length(pos)
  data.full <- (counts/reads)[idx,]

	dfdat <- data.frame(pos, data.full)
	names(dfdat) <- c("pos", paste("y", 1:K, sep=""))
	dfdat_long <- gather(dfdat, "sample", "prop", -pos)

	dfreads <- data.frame(pos, reads)
	names(dfreads) <- c("pos", paste0("y", 1:K))
	dfreads_long <- gather(dfreads, "sample", "reads", -pos)

  stopifnot(all(dfreads_long$pos == dfdat_long$pos))
  stopifnot(all(dfreads_long$sample == dfdat_long$sample))
  dfdat_long$reads <- dfreads_long$reads

  dfdat_long$sample <- factor(dfdat_long$sample, levels=paste0("y", 1:K))

	h <- ggplot(dfdat_long) + theme_bw()
	#Background colors
	if(!is.null(sep.tab)){
	  h <- h + gg_sep_rect(sep.tab, color = bg.color, alpha=0.2)
	}
	h <- h + geom_point(aes(x=pos, y=prop, col=sample, group=sample, shape=sample, size=reads)) +
	  scale_color_manual(values=cols) + scale_shape_manual(values=shapes) +
   scale_size_area() + ylim(ylim)+ labs(x=xlab, y=ylab) + theme(legend.position="none",
                                                                text=element_text(size=15))
	return(h)

}

#' Plot a JADE fit
#' @param fits A p by K matrix of fitted values (the \code{fits}) element of a JADE fit object.
#' @param pos Vector of positions.
#' @param y A p by K matrix of data. Optional. If provided, data points will be added to the plot.
#' Data can be plotted by itself using the \code{plot_data} function.
#' @param reads A p by K matrix of read counts or an indication of the relative amount of data for
#' each element of \code{y}. Line widths will be proportional to \code{reads}.
#' @param wsize Window size used to determine line widths.
#' @param sep.tab A table giving regions which are separated.
#' This can be obtained using \code{\link{get_separated_regions}}.
#' @return A \code{ggplot} object for the desred plot.
#' @export
plot_jade <- function(fits,  pos, reads=NULL, cols=NULL, ltys=NULL,
                       range=NULL, wsize=10, maxcov=Inf, take.log.cov=FALSE,
                       maxwidth=3, minwidth=0.5,  ylim=NULL, sep.tab=NULL,
                      ylab=NULL, xlab="Position", bg.color="blue"){

	if(class(fits)=="numeric"){
			p <- length(fits)
			K <- 1
			cat("Single variable.\n")
			if(!is.null(y)) y <- matrix(y, nrow=p)
			if(!is.null(reads)) reads <- matrix(reads, nrow=p)
			alpha <- 0.8
			sv<- TRUE
	}else{
		p <- dim(fits)[1]
		K <- dim(fits)[2]
		sv <- FALSE
		alpha <- 0.6
	}

	if(is.null(cols)){
	  cols=c("black", "red3", "blue2")
	  if(K >3) cols <- c(cols, 4:(K-3))
	}


	if(is.null(range)){
		range <- range(pos)
		idx <- 1:p
	}else{
		idx <- which(pos <= max(range) & pos >= min(range))
		fits <- fits[idx,, drop=FALSE]
		pos <- pos[idx]
		cat(dim(fits), "\n")
	}
	p <- length(pos)


  if(!is.null(reads)) reads.full <- reads[idx,, drop=FALSE]
	  else reads.full <- matrix(1, nrow=p, ncol=K)

	if(is.null(ylim)) ylim=range(fits)

	window.coverage <- matrix(NA, nrow=p, ncol=K)
	for(i in 1:p){
		t <- pos[i]
		if(sum(pos >= t-(wsize/2) & pos <= t+(wsize/2)) ==1){
			 C <- reads.full[i, ]
		}else{
			C <- apply(reads.full[ pos >= t-(wsize/2) & pos <= t+(wsize/2),,drop=FALSE] , MARGIN=2, FUN=sum)
		}
		window.coverage[i, ] <- C
	}
	window.coverage[window.coverage > maxcov] <- maxcov
	window.coverage <- window.coverage + 2
	if(take.log.cov){
		window.coverage=log10(window.coverage)
		window.coverage[window.coverage== -Inf] <- 0
	}

	dffit <- data.frame(pos, fits)
	names(dffit) <- c("pos", paste0("y", 1:K))
	dffit_long <- gather(dffit, "sample", "y", -pos)

	dfcov <- data.frame(pos, window.coverage)
	names(dfcov) <- c("pos", paste0("y", 1:K))
	dfcov_long <- gather(dfcov, "sample", "cov", -pos)

	stopifnot(all(dfcov_long$pos==dffit_long$pos))
	stopifnot(all(dfcov_long$sample==dffit_long$sample))
	dffit_long$cov <- dfcov_long$cov


	h <- ggplot(dffit_long) + theme_bw()

	#Background colors
	if(!is.null(sep.tab)){
		h <- h + gg_sep_rect(sep.tab, color = bg.color, alpha=0.2)
	}

  if(is.null(ltys)){
	  h <- h +  geom_line(aes(x=pos, y=y, size=cov, col=sample, group=sample), alpha=alpha) +
	      scale_color_manual(values=cols) + scale_size(range=c(minwidth, maxwidth))
  }else{
    cat("Warning: If using dotted lines, line width must be constant\n")
    h <- h +  geom_line(aes(x=pos, y=y, linetype=sample, col=sample, group=sample),
                        alpha=alpha, lwd=1.5) +
      scale_color_manual(values=cols) + scale_linetype_manual(values=ltys)
  }

	R <- h + scale_x_continuous(minor_breaks=pos) + ylim(ylim)
	if(is.null(xlab) & is.null(ylab)){
	  R <- R + theme(legend.position="none",
	                 axis.title.x=element_blank(),
	                 axis.title.y=element_blank(),
	                 text=element_text(size=15))
	}else if(is.null(xlab)){
	  R <- R + theme(legend.position="none",
	                 axis.title.x=element_blank(),
	                 text=element_text(size=15)) + labs(y=ylab)
	}else if(is.null(ylab)){
	  R <- R + theme(legend.position="none",
	                 axis.title.y=element_blank(),
	                 text=element_text(size=15)) + labs(x=xlab)
	}else{
	  R <- R + theme(legend.position="none",
	                 text=element_text(size=15)) + labs(x=xlab, y=ylab)
	}
	return(R)
}


gg_sep_rect <- function(sep.tab, color="blue", alpha=0.2){
  m.lower <- sep.tab$PlotStart
  m.upper <- sep.tab$PlotStop
  M <- data.frame(m.lower, m.upper)
  names(M) <- c("m.lower", "m.upper")

 return(geom_rect(data=M, aes(xmin=m.lower,xmax=m.upper,ymin=-Inf,ymax=Inf),
                     fill=color, alpha=alpha))
}