R/output_graphs_single.R

In SLEMI: Statistical Learning Based Estimation of Mutual Information

#' Plotting output of capacity estimation. Auxiliary functions.
#' 
#' INPUT:
#' @param data is a data.frame object
#' @param signal is a character object that indicates columns of data that should be treated as channel's input
#' @param response is a character vector that indicates columns of data that should be treated as channel's output
#' @param side_variables is a character vector that indicates side variables' columns of data
#' @param cc_output a list that is the output of capacity_logreg_algorithm function
#' @param path character giving the directory, where graphs should be saved
#' @param height integer indicating the height of a single plot
#' @param width integer indicating the width of a single plot
#' @keywords internal
#' 
capacity_output_graph_boxplots<-function(data,signal,response,path,height=4,width=6){
  
  x=y=density=NULL

  data_colnames=colnames(data)
  response_length=length(response)
  signalNUM=paste(signal,"_RAW",sep="")
  
  if (any(data_colnames %in% (signalNUM))){
  signal=signalNUM
    }
  
  if (is.numeric(data[,signal])){
    data[,signal]<-factor(data[,signal])
  }
  
  dataPlot=reshape2::melt(data[,c(signal,response)],id.vars=c(signal))
  plot<-ggplot2::ggplot(data=dataPlot,ggplot2::aes_string(x=signal,y="value") ) + 
    ggplot2::geom_boxplot()+
    ggplot2::facet_grid(.~variable)+
    ggplot2::scale_y_continuous(name=paste("Output",sep="") )+
    ggplot2::scale_x_discrete(name=paste("Input",sep="") )+
    ggplot2::ggtitle("Boxplots")+
    aux_theme_publ(version=2)
  
  ggplot2::ggsave(plot,file=paste(path,'data_boxplots.pdf',sep=""),height=height,width=min(width*response_length,49.5) )
  
  plot
}


#' @rdname capacity_output_graph_boxplots
#' @keywords internal
capacity_output_graph_violinMean<-function(data,signal,response,path,height=4,width=6){
  
  x=y=density=NULL

  data_colnames=colnames(data)
  response_length=length(response)
  signalNUM=paste(signal,"_RAW",sep="")
  
  if (any(data_colnames %in% (signalNUM))){
    
    signal_classes=unique(data[[signalNUM]])
    maxSignal=max(data[[signalNUM]])
    minSignal=min(0,min(data[[signalNUM]]))
    rangeSignal=range(data[[signalNUM]])
    temp_coeff=(rangeSignal[2]-rangeSignal[1])/length(signal_classes)

    dataPlot=reshape2::melt(data[,c(signalNUM,response)],id.vars=c(signalNUM))
    plot<-ggplot2::ggplot(data=dataPlot,ggplot2::aes_string(x=signalNUM,y="value"))+ggplot2::geom_violin(ggplot2::aes_string(group=signalNUM),scale="width")+
      ggplot2::stat_summary(fun=mean,geom="line",size=1.15)+ggplot2::stat_summary(fun=mean,geom="point",size=1.5)+
      ggplot2::facet_grid(variable~.)+
      ggplot2::scale_y_continuous(paste("Output",sep="") )+
      ggplot2::scale_x_continuous(paste("Input",sep=""),limits=c(minSignal*0.9-1.5,maxSignal*1.1+1.5) )+
      ggplot2::ggtitle("Violin plots with means")+
      aux_theme_publ(version=2)
  } else {
    dataPlot=reshape2::melt(data[,c(signal,response)],id.vars=c(signal))
    plot<-ggplot2::ggplot(data=dataPlot,ggplot2::aes_string(x=signal,y="value"))+ggplot2::geom_violin(ggplot2::aes_string(group=signal),scale="width")+
      ggplot2::stat_summary(fun=mean,geom="point",size=1.5)+
      ggplot2::facet_grid(variable~.)+
      ggplot2::scale_y_continuous(paste("Output",sep="") )+
      ggplot2::scale_x_discrete(paste("Input",sep="") )+
      ggplot2::ggtitle("Violin plots with means")+
      aux_theme_publ(version=2)
  }
  
  ggplot2::ggsave(plot,file=paste(path,'data_MeanViolin.pdf',sep=""),height=min(49.5,height*response_length),width=width)
  
  plot
}



#' @rdname capacity_output_graph_boxplots
#' @keywords internal
capacity_output_graph_boxplotsSideVar<-function(data,signal,side_variables,path,height=4,width=6){
    
  x=y=density=NULL

  if (is.null(side_variables)) {
    plot=grid::textGrob(" ")  
  } else if (!all(sapply(data[,side_variables],function(x) is.numeric(x) ))) {
    plot=grid::textGrob(" ")  
  } else  {
    dataTemp=data[,colnames(data)%in% c(signal,side_variables)]
    dataMelt=reshape2::melt(dataTemp, id.vars=c(signal))
    
    plot<-ggplot2::ggplot(data=dataMelt,ggplot2::aes_string(x=signal,y="value") ) + 
      ggplot2::geom_boxplot()+ ggplot2::facet_grid(variable~.) +
      ggplot2::scale_y_continuous("Side variables value")+
      ggplot2::scale_x_discrete(paste("Input",signal,sep="") )+
      ggplot2::ggtitle("Boxplots of side variables")+
      aux_theme_publ(version=2)
    
    ggplot2::ggsave(plot,file=paste(path,'side_variables_boxplots.pdf',sep=""),height=height,width=width)
  }
  
  plot
}


#' @rdname capacity_output_graph_boxplots
#' @keywords internal
capacity_output_graph_capacity<-function(cc_output,path,height=4,width=6){
    
  x=y=density=NULL

  temp_name="Capacity"
  if (any(names(cc_output)=="mi")){
    temp_name="Mutual Information"
    cc_output$cc=cc_output$mi
  }

  if (is.null(cc_output$testing)){
    
    plot<-ggplot2::ggplot(data=data.frame(x=cc_output$cc,y=0),ggplot2::aes(x=x,y=y) ) + 
      ggplot2::geom_point(size=4,shape=15) + 
      ggplot2::geom_text(ggplot2::aes(label=round(x,digits=3),y=0.075),size=8)+
      ggplot2::geom_line(data=data.frame(x=seq(from=0.9*cc_output$cc,to=1.1*cc_output$cc,length=10),y=0),arrow = ggplot2::arrow())+
      ggplot2::scale_y_continuous("",limits=c(-0.05,0.1))+ggplot2::scale_x_continuous( paste0(temp_name," (bits)") )+
      ggplot2::ggtitle(temp_name)+
      aux_theme_publ(version=2)+
      ggplot2::theme(axis.ticks.y=ggplot2::element_blank(),
            axis.text.y=ggplot2::element_blank(),
            axis.title.y=ggplot2::element_blank(),
            panel.grid.major.y=ggplot2::element_blank(),
            axis.line.y=ggplot2::element_blank())
    
    ggplot2::ggsave(plot,file=paste(path,stringr::str_replace_all(stringr::str_to_lower(temp_name)," ","_"),'.pdf',sep=""),height=2,width=6)
    
  } else if (length(cc_output$testing)==2) {
    
    if (any(names(cc_output)=="mi")){
        cc_output$testing$bootstrap=lapply(cc_output$testing$bootstrap,function(x) {
            x$cc=x$mi
            x
        })
        cc_output$testing$traintest=lapply(cc_output$testing$traintest,function(x) {
            x$cc=x$mi
            x
        })
    }

    boot_results=sapply(cc_output$testing$bootstrap,function(x) x$cc)
    boot_sd=stats::sd(boot_results)
    boot_mean=mean(boot_results)
    
    plot1<-ggplot2::ggplot(data=data.frame(x=cc_output$cc,y=0),ggplot2::aes(x=x,y=y) ) + 
      ggplot2::geom_point(size=6,shape=15) + 
      ggplot2::geom_text(ggplot2::aes(label=round(x,digits=3),y=0.075),size=8)+
      ggplot2::geom_line(data=data.frame(x=seq(from=0.95*(cc_output$cc-boot_sd),to=1.05*(cc_output$cc+boot_sd),length=10),y=0),arrow = ggplot2::arrow())+
      ggplot2::geom_point(data=data.frame(x=c(boot_mean-boot_sd,boot_mean+boot_sd),y=c(0,0)),shape=124,colour="red",size=10)+ 
      ggplot2::geom_text(data=data.frame(x=c(0.985*(boot_mean-boot_sd),1.015*(boot_mean+boot_sd)),y=c(0.05,0.05)),ggplot2::aes(label=round(x,digits=3),y=0.025),size=6, colour="red")+
      ggplot2::geom_point(data=data.frame(x=c(boot_mean),y=c(0)),shape=15,colour="red",size=4)+
      ggplot2::scale_y_continuous("",limits=c(-0.05,0.1))+
      ggplot2::scale_x_continuous( paste0(temp_name," (bits)"),breaks=c(0.95*(cc_output$cc-boot_sd),1.05*(cc_output$cc+boot_sd)) )+
      ggplot2::ggtitle(temp_name)+
      aux_theme_publ(version=2)+
      ggplot2::theme(axis.ticks.y=ggplot2::element_blank(),
            axis.text.y=ggplot2::element_blank(),
            axis.title.y=ggplot2::element_blank(),
            panel.grid.major.y=ggplot2::element_blank(),
            axis.line.y=ggplot2::element_blank())
    
    plot2<-ggplot2::ggplot( data=data.frame(x=sapply(cc_output$testing$bootstrap,function(x) x$cc)),ggplot2::aes_string(x="x",y="..density..") ) + 
      ggplot2::geom_histogram(bins=100 )+
      ggplot2::geom_point(data=data.frame(x=cc_output$cc,y=0),ggplot2::aes(x=x,y=y),size=4,shape=15) +
      ggplot2::ggtitle("Bootstrap")+ggplot2::scale_x_continuous(paste0(temp_name," (bits)"))+
      aux_theme_publ(version=2)
    
    plot4<-ggplot2::ggplot( data=data.frame(x=sapply(cc_output$testing$traintest,function(x) x$cc)),ggplot2::aes_string(x="x",y="..density..") ) + 
      ggplot2::geom_histogram(bins=100 )+
      ggplot2::geom_point(data=data.frame(x=cc_output$cc,y=0),ggplot2::aes(x=x,y=y),size=4,shape=15) +
      ggplot2::ggtitle("TrainTest")+ggplot2::scale_x_continuous(paste0(temp_name," (bits)"))+
      aux_theme_publ(version=2)
    
    plot5=grid::textGrob(paste("PV-left:",round(cc_output$testing_pv$bootstrap[1],digits=3),sep=" " ))
     plot6=grid::textGrob(paste("PV-right:",round(cc_output$testing_pv$bootstrap[2],digits=3),sep=" " )) 

    plot9=grid::textGrob(paste("PV-left:",round(cc_output$testing_pv$traintest[1],digits=3),sep=" " ))
    plot10=grid::textGrob(paste("PV-right:",round(cc_output$testing_pv$traintest[2],digits=3),sep=" " )) 

    
    plot=gridExtra::grid.arrange(plot1,plot2,plot4,plot5,plot6,plot9,plot10,
                      layout_matrix=rbind(c(1,1),c(1,1),c(2,4),c(2,5),c(3,6),c(3,7)))
    
    
    ggplot2::ggsave(plot,file=paste(path,stringr::str_replace_all(stringr::str_to_lower(temp_name)," ","_"),'.pdf',sep=""),height=1.5*height,width=width)
    
  } else {

    if (any(names(cc_output)=="mi")){
        cc_output$testing$bootstrap=lapply(cc_output$testing$bootstrap,function(x) {
            x$cc=x$mi
            x
        })
        cc_output$testing$traintest=lapply(cc_output$testing$traintest,function(x) {
            x$cc=x$mi
            x
        })
       cc_output$testing$resamplingMorph=lapply(cc_output$testing$resamplingMorph,function(x) {
            x$cc=x$mi
            x
        })
        cc_output$testing$bootResampMorph=lapply(cc_output$testing$bootResampMorph,function(x) {
            x$cc=x$mi
            x
        })
    }

    boot_results=sapply(cc_output$testing$bootstrap,function(x) x$cc)
    boot_sd=stats::sd(boot_results)
    boot_mean=mean(boot_results)
    
    plot1<-ggplot2::ggplot(data=data.frame(x=cc_output$cc,y=0),ggplot2::aes(x=x,y=y) ) + 
      ggplot2::geom_point(size=6,shape=15) + 
      ggplot2::geom_text(ggplot2::aes(label=round(x,digits=3),y=0.075),size=8)+
      ggplot2::geom_line(data=data.frame(x=seq(from=0.95*(cc_output$cc-boot_sd),to=1.05*(cc_output$cc+boot_sd),length=10),y=0),arrow = ggplot2::arrow())+
      ggplot2::geom_point(data=data.frame(x=c(boot_mean-boot_sd,boot_mean+boot_sd),y=c(0,0)),shape=124,colour="red",size=10)+ 
      ggplot2::geom_text(data=data.frame(x=c(0.985*(boot_mean-boot_sd),1.015*(boot_mean+boot_sd)),y=c(0.05,0.05)),ggplot2::aes(label=round(x,digits=3),y=0.025),size=6, colour="red")+
      ggplot2::geom_point(data=data.frame(x=c(boot_mean),y=c(0)),shape=15,colour="red",size=4)+
      ggplot2::scale_y_continuous("",limits=c(-0.05,0.1))+
      ggplot2::scale_x_continuous(  paste0(temp_name," (bits)"),breaks=c(0.95*(cc_output$cc-boot_sd),1.05*(cc_output$cc+boot_sd)) )+
      ggplot2::ggtitle(temp_name)+
      aux_theme_publ(version=2)+
      ggplot2::theme(axis.ticks.y=ggplot2::element_blank(),
                     axis.text.y=ggplot2::element_blank(),
                     axis.title.y=ggplot2::element_blank(),
                     panel.grid.major.y=ggplot2::element_blank(),
                     axis.line.y=ggplot2::element_blank())
    
    plot2<-ggplot2::ggplot( data=data.frame(x=sapply(cc_output$testing$bootstrap,function(x) x$cc)),ggplot2::aes_string(x="x",y="..density..") ) + 
      ggplot2::geom_histogram(bins=100 )+
      ggplot2::geom_point(data=data.frame(x=cc_output$cc,y=0),ggplot2::aes(x=x,y=y),size=4,shape=15) +
      ggplot2::ggtitle("Bootstrap")+ggplot2::scale_x_continuous(  paste0(temp_name," (bits)") )+
      aux_theme_publ(version=2)
    
    plot3<-ggplot2::ggplot( data=data.frame(x=sapply(cc_output$testing$resamplingMorph,function(x) x$cc)),ggplot2::aes_string(x="x",y="..density..") ) + 
      ggplot2::geom_histogram(bins=100 )+
      ggplot2::geom_point(data=data.frame(x=cc_output$cc,y=0),ggplot2::aes(x=x,y=y),size=4,shape=15) +
      ggplot2::ggtitle("SideVar Resampling")+ggplot2::scale_x_continuous(  paste0(temp_name," (bits)") )+
      aux_theme_publ(version=2)
    
    plot4<-ggplot2::ggplot( data=data.frame(x=sapply(cc_output$testing$traintest,function(x) x$cc)),ggplot2::aes_string(x="x",y="..density..") ) + 
      ggplot2::geom_histogram(bins=100 )+
      ggplot2::geom_point(data=data.frame(x=cc_output$cc,y=0),ggplot2::aes(x=x,y=y),size=4,shape=15) +
      ggplot2::ggtitle("TrainTest")+ggplot2::scale_x_continuous(  paste0(temp_name," (bits)") )+
      aux_theme_publ(version=2)
    
    plot3b<-ggplot2::ggplot( data=data.frame(x=sapply(cc_output$testing$bootResampMorph,function(x) x$cc)),ggplot2::aes_string(x="x",y="..density..") ) + 
      ggplot2::geom_histogram(bins=100 )+
      ggplot2::geom_point(data=data.frame(x=cc_output$cc,y=0),ggplot2::aes(x=x,y=y),size=4,shape=15) +
      ggplot2::ggtitle("Boot SideVar Resampling")+ggplot2::scale_x_continuous(  paste0(temp_name," (bits)") )+
      aux_theme_publ(version=2)
    
    plot5=grid::textGrob(paste("PV-left:",round(cc_output$testing_pv$bootstrap[1],digits=3),sep=" " ))
    plot6=grid::textGrob(paste("PV-right:",round(cc_output$testing_pv$bootstrap[2],digits=3),sep=" " )) 
    plot7=grid::textGrob(paste("PV-left:",round(cc_output$testing_pv$resamplingMorph[1],digits=3),sep=" " ))
    plot8=grid::textGrob(paste("PV-right:",round(cc_output$testing_pv$resamplingMorph[2],digits=3),sep=" " )) 
    plot9=grid::textGrob(paste("PV-left:",round(cc_output$testing_pv$traintest[1],digits=3),sep=" " ))
    plot10=grid::textGrob(paste("PV-right:",round(cc_output$testing_pv$traintest[2],digits=3),sep=" " )) 
    plot11=grid::textGrob(paste("PV-left:",round(cc_output$testing_pv$bootResampMorph[1],digits=3),sep=" " ))
    plot12=grid::textGrob(paste("PV-right:",round(cc_output$testing_pv$bootResampMorph[2],digits=3),sep=" " )) 
    
    plot=gridExtra::grid.arrange(plot1,plot2,plot3,plot4,plot3b,plot5,plot6,plot7,plot8,plot9,plot10,plot11,plot12,
                                 layout_matrix=rbind(c(1,1),c(1,1),c(2,6),c(2,7),c(3,8),c(3,9),c(4,10),c(4,11),c(5,12),c(5,13)))
    
    
    ggplot2::ggsave(plot,file=paste(path,stringr::str_replace_all(stringr::str_to_lower(temp_name)," ","_"),'.pdf',sep=""),height=2*height,width=width)
  }
  
  plot
}