wilcox_R: What the package does (short line)

rslopesm <-
function(x,y,fr=1,est=tmean,nmin=10,pts=x,plotit=FALSE,xlab="X",
ylab="Y",SEED=TRUE,nboot=40,xout=FALSE,RNA=TRUE,atr=.2,scat=TRUE,pyhat=TRUE,...){
#
#  For a regression line predicting Y given X
# Estimate slope at points in pts with bagging
# followed by a smooth.
#
# pyhat=T, returns estimated slopes corresponding to the sorted
# x values.
# fr controls amount of smoothing
# atr controls the amount of trimming.
#
# OUTPUT: by default, the estimated  slopes at
# X_1<=X_2<=...<=X_n
# That is, for the x values written in ascending order, the
# slope is estimated for each value. If the slope is not considered
# estimable, the estimate is set to NA.
#
# pts is used if the goal is to estimate the slope for some
# other collection of points.
#
# nmin controls how many points close to x are required when
# deciding that the slope is estimable.
# plotit=TRUE will plot the estimates.
#
# The plotted points are the estimates using rslope and
# the solid line gives the estimated values reported by this function
#
# Missing values are automatically removed.
#
if(SEED) set.seed(2)
temp<-cbind(x,y)
if(ncol(temp)!=2)stop("One predictor only is allowed")
temp<-elimna(temp) # Eliminate any rows with missing values
if(xout) {
                flag <- outfun(temp[, 1], plotit = FALSE)$keep
                temp <- temp[flag,  ]
x<-temp[,1]
y<-temp[,2]
}
flag<-order(x)
x<-x[flag]
y<-y[flag]
mat<-matrix(NA,nrow=nboot,ncol=length(pts))
vals<-NA
       for(it in 1:nboot) {
                idat <- sample(c(1:length(y)), replace = T)
                xx <- temp[idat, 1]
                yy <- temp[idat, 2]
#                mat[it,  ] <- runhat(xx, yy, pts = x, est = est, fr = fr, ...)
mat[it,]<-rslope(xx,yy,fr=fr,est=est,nmin=nmin,pts=x,plotit=FALSE)
        }
rmd<-apply(mat,2,mean,na.rm=RNA,tr=atr)
flag<-is.na(rmd)
rmdsm<-lplot(x,rmd,pyhat=TRUE,plotit=plotit)
output<-"Done"
if(pyhat){
temp<-rep(NA,length(x))
temp[!flag]<-rmdsm$yhat.values
output<-temp
}
output
}