require(componentSkeleton)
require(ggplot2)
require(gridExtra)

CHROMOSOMES <- c(1:22, "X", "Y")

execute <- function(cf) {
    
    ## Read input and parameters
    cna <- CSV.read(get.input(cf, 'cna'))
    rearrangements <- CSV.read(get.input(cf, 'rearrangements'))
    binwidth <- as.numeric(get.parameter(cf, 'binwidth', type="int"))
	
	## Clean chromosomal rearrangements data
    names(rearrangements)  <- c("type","chr","start","end","id")
	rearrangements         <- subset(rearrangements, chr %in% CHROMOSOMES)
	rearrangements         <- droplevels(rearrangements)
	rearrangements$chr     <- factor(rearrangements$chr, CHROMOSOMES)
	## Set location of inversion as start site
	inversion              <- subset(rearrangements, type == "inversions")
	inversion$type         <- "inversion"
	inversion$location     <- inversion$start
	## Add inversion end points as separate locations
	inversion.end          <- transform(inversion, location = end, id = paste(id,"_end",sep=""))
	inversion              <- rbind(inversion, inversion.end)
	## Use center of translocations as location
	translocation          <- subset(rearrangements, type %in% c("translocationEnd","translocationStart"))
	translocation$location <- (translocation$start + translocation$end)/2
	## Remove distinction of start and end of translocations
	translocation          <- transform(translocation, type = "translocation")

	## Clean copy number data
	names(cna)   <- c("id","chr","start","ratio","medianRatio","cn","type")
	cna          <- subset(cna, chr %in% CHROMOSOMES)
	cna          <- droplevels(cna)
	cna$chr      <- factor(cna$chr, CHROMOSOMES)
	cna$location <- cna$start
	cna$end      <- cna$start
	cna$type     <- tolower(cna$type)

	## Combine cleaned data
	cols             <- c("id","chr","location","type")
	breakpoints      <- rbind(inversion[,cols], translocation[,cols], cna[,cols])
	types            <- sort(unique(breakpoints$type))
	breakpoints$type <- factor(breakpoints$type, types)
	## Add a reversed order type for nicer plot
	breakpoints$revtype <- factor(breakpoints$type, rev(types))
	## Convert location to Mb
	breakpoints$location <- as.double(breakpoints$location)*1e-6


	#### Plotting
	theme_set(theme_bw())
	## Create output folder for the component.
    output.dir <- get.output(cf, 'report')
    dir.create(output.dir, recursive=TRUE)
    instance.name <- get.metadata(cf, 'instanceName')
    tex <- character()

	## Plot bar diagrams of break point counts per chromosome
	filename <- sprintf('barplot-%s.pdf', instance.name)
    file     <- file.path(output.dir, filename)
	p <- ggplot(breakpoints)
	p <- p + geom_bar(aes(x=chr, fill=type)) + facet_wrap(~ type, ncol=2)
	p <- p + scale_fill_discrete(drop=FALSE, guide=FALSE)
	invisible(ggsave(file, p, width=7, height=4, units="in"))
	fig.caption <- "Counts of breakpoints in each chromosome grouped by breakpoint type."
	tex <- c(tex, latex.figure(file, caption=fig.caption, fig.label=sprintf("fig:%s-barplot",instance.name)))

	for(crom in CHROMOSOMES) {
	    data <- subset(breakpoints, chr == crom)
	    ## Duplicate data for faceting
	    #data  <- transform(data, plottype="breakpoints")
	    #data2 <- transform(data, plottype="histogram")

	    ## Facet method, aligned x axes, but ugly y axis scale in histogram
		#p <- ggplot(rbind(data,data2), aes(x=location)) + facet_wrap(~plottype, ncol=1, scales="free_y")
		#p <- p + geom_jitter(data=data, aes(y=revtype,color=type), alpha=0.3)
		#p <- p + geom_bar(data=data2, aes(y=..count..), binwidth=BINWIDTH*1e-6)
		#p <- p + scale_y_continuous(breaks=pretty)
		#p <- p + scale_colour_discrete(drop=FALSE, guide=FALSE)
		#p <- p + labs(title=sprintf("chr %s",crom), x="location (Mb)", y="")

		## Grid arrange method, needs less memory, nicer y scale, but unaligned x axes
		p <- ggplot(data, aes(x=location))
		p1 <- p  + geom_jitter(aes(y=revtype,color=type), alpha=0.4, size=2)
		p1 <- p1 + scale_colour_discrete(drop=FALSE, guide=FALSE)
		p1 <- p1 + labs(title=sprintf("chr %s",crom), x=NULL, y=NULL)
		p2 <- p + geom_bar(aes(y=..count..), binwidth=binwidth*1e-6)
		p2 <- p2 + labs(title=NULL, x="location (Mb)")

		filename <- sprintf("breakpointLocations-chr%s-%s.png", crom, instance.name)
    	file     <- file.path(output.dir, filename)
		#invisible(ggsave(file, p, width=7, height=4, units="in"))
		png(file, width=10, height=8, units="in", res=300)
		#align_plots(p1, p2, width=unit(7,'in'), height=unit(3,'in'))
		grid.arrange(p1, p2)
		invisible(dev.off())
		fig.caption <- sprintf("Breakpoint locations in chromosome %s.", crom)
		tex <- c(tex, latex.figure(file, caption=fig.caption, fig.label=sprintf("fig:%s-breakpoints-chr%s",instance.name,crom)))
	}
	latex.write.main(cf, 'report', tex)
	return(0)
}

main(execute)