R/factor_barchart_class.R

In computational-metabolomics/structtoolbox: Data processing & analysis tools for Metabolomics and other omics

#' @eval get_description('DatasetExperiment_factor_boxplot')
#' @examples
#' D = iris_DatasetExperiment()
#' C = DatasetExperiment_factor_boxplot(factor_names='Species',feature_to_plot='Petal.Width')
#' chart_plot(C,D)
#' @import struct
#' @import grid
#' @import gridExtra
#' @export DatasetExperiment_factor_boxplot
#' @include HSD_class.R
DatasetExperiment_factor_boxplot = function(feature_to_plot,factor_names,...) {
    out=struct::new_struct('DatasetExperiment_factor_boxplot',
        feature_to_plot=feature_to_plot,
        factor_names=factor_names,
        ...)
    return(out)
}


.DatasetExperiment_factor_boxplot<-setClass(
    "DatasetExperiment_factor_boxplot",
    contains='chart',
    slots=c(
        feature_to_plot='entity',
        factor_names='entity'
    ),
    prototype = list(name='Factor boxplot',
        description=paste0('Boxplot for a feature to visualise the ',
        'distribution of values within each group'),
        type="barchart",
        .params=c('factor_names','feature_to_plot'),

        feature_to_plot=entity(name='Feature to plot',
            value='V1',
            type=c('character','numeric','integer'),
            description='The name of the plotted feature.'
        ),
        factor_names=ents$factor_names
    )
)

#' @export
#' @template chart_plot
setMethod(f="chart_plot",
    signature=c("DatasetExperiment_factor_boxplot",'DatasetExperiment'),
    definition=function(obj,dobj)
    {
        X=dobj$data

        if (is.numeric(obj$feature_to_plot)) {
            varn=colnames(X)[obj$feature_to_plot]
        } else {
            varn=obj$feature_to_plot
        }

        f=obj$feature_to_plot
        X=dobj$data[,f]
        SM=dobj$sample_meta

        if (length(obj$factor_names)==1) {
            # single factor plot(s)

            # get color pallete using pmp
            clrs= createClassAndColors(class = SM[[obj$factor_names[1]]])
            SM[[obj$factor_names[1]]]=clrs$class

            # prep the data
            A=data.frame(x=SM[[obj$factor_names[1]]],y=X)

            g=ggplot(data=A,aes(x=x,y=y,color=x)) +
                geom_boxplot() +
                theme_Publication(base_size=10) +
                scale_colour_manual(values=clrs$manual_colors,name=obj$factor_names[1])+
                ylab('')+
                xlab(obj$factor_names[1])+
                theme(legend.position="none")
            return(g)
        }

        if (length(obj$factor_names)==2) {
            # dual factor plot(s)

            # get color pallete using pmp
            clrs= createClassAndColors(class = SM[[obj$factor_names[2]]])
            SM[[obj$factor_names[2]]]=clrs$class

            # prep the data
            A=data.frame(x=SM[[obj$factor_names[1]]],z=SM[[obj$factor_names[2]]],y=X)

            g=ggplot(data=A,aes(x=x,y=y,color=z)) +
                geom_boxplot() +
                theme_Publication(base_size=10) +
                geom_vline(xintercept=(1:(nlevels(A$x)-1))+0.5,linetype="dashed",color='grey') +
                scale_colour_manual(values=clrs$manual_colors,name=obj$factor_names[2])+
                ylab('')+
                xlab(obj$factor_names[1])
            return(g)
        }

        if (length(obj$factor_names)==3) {
            # dual factor plot(s)

            # get color pallete using pmp
            clrs= createClassAndColors(class = SM[[obj$factor_names[2]]])
            SM[[obj$factor_names[2]]]=clrs$class

            A=data.frame(x=SM[[obj$factor_names[1]]],z=SM[[obj$factor_names[2]]],y=X,a=SM[[obj$factor_names[3]]])

            lab=as.character(interaction(obj$factor_names[3],levels(SM[[obj$factor_names[3]]]),sep = ': '))
            names(lab)=levels(SM[[obj$factor_names[3]]])
            p=ggplot(data=A,aes(x=x,y=y,color=z)) +
                geom_boxplot() +
                theme_Publication(base_size=10) +
                geom_vline(xintercept=(1:(nlevels(A$x)-1))+0.5,linetype="dashed",color='grey') +
                scale_colour_manual(values=clrs$manual_colors,name=obj$factor_names[2]) +
                facet_grid(.~a,labeller=as_labeller(lab))+
                theme(panel.background = element_rect(fill = NA, color = "black"))+
                theme(strip.background =element_rect(fill='black'))+
                theme(strip.text = element_text(colour = 'white'))+
                ylab('')+
                xlab(obj$factor_names[1])

            # colour the facet labels
            ## https://github.com/tidyverse/ggplot2/issues/2096 ##
            g = ggplot_gtable(ggplot_build(p))
            stripr = which(grepl('strip-t', g$layout$name))
            fills = clrs$manual_colors
            k = 1
            for (i in stripr) {
                j <- which(grepl('rect', g$grobs[[i]]$grobs[[1]]$childrenOrder))
                g$grobs[[i]]$grobs[[1]]$children[[j]]$gp$fill <- fills[k]
                k = k+1
            }
            #grid.draw(g)
            ##

            return(invisible(g))
        }

    }
)