stride <- function(seq,dis=NULL,tree=NULL)
{
    ncol <- length(seq)
    pnt <- 0

    if (!is.null(dis)) {
        res <- matrix(NA,nrow=attr(dis,'Size'),ncol=ncol)
        for (i in seq) {
            pnt <- pnt + 1
            res[,pnt] <- pam(dis,i)$clustering
        }
        out <- data.frame(res)
        names(out) <- as.character(seq)
        row.names(out) <- attr(dis,'Labels')
        out <- list(clustering=out,seq=seq)
        class(out) <- 'stride'
        out
    }
    else if (!is.null(tree)) {
        res <- matrix(NA,nrow=length(tree$order),ncol=ncol)
        for (i in seq) {
            pnt <- pnt + 1
            res[,pnt] <- cutree(tree,i)
        }
        out <- data.frame(res)
        names(out) <- as.character(seq)
        row.names(out) <- tree$labels
        out <- list(clustering=out,seq=seq)
        class(out) <- 'stride'
        out
    }
    else print('you must enter a dissimilarity or a tree')
}

plot.stride <- function (stride,dis,col2=4) 
{
    oldpar <- par(no.readonly=TRUE)
    par(mar=c(5,4,4,4)+0.1,bg='white',fg='black')
    plot(partana(stride,dis),xlab='Number of Clusters',
        ylab='Partana Ratio',type='b')
    par(new=TRUE)
    tmp <- silhouette(stride,dis)
    plot(tmp,axes=FALSE,xlab='',ylab='',col=col2,type='b')
    axis(4,at=pretty(range(tmp$sil_width)),col.axis=col2,col.tick=col2)
    mtext('Silhouette Width',4,2.5,col=col2)
    par(oldpar)
}

partana.stride <- function(stride,dis) 
{
    res <- rep(NA,ncol(stride$clustering))
    for (i in 1:ncol(stride$clustering)) {
        res[i] <- partana(stride$clustering[,i],dis)$ratio
    }
    clusters <- stride$seq
    ratio <- res
    out <- data.frame(clusters,ratio)
    out
}

    
silhouette.stride <- function(stride,dis)
{
    res <- rep(NA,ncol(stride$clustering))
    for (i in 1:ncol(stride$clustering)) {
        res[i] <-  mean(silhouette(stride$clustering[,i],dis)[,3])
    }
    clusters <- stride$seq
    sil_width <- res
    out <- data.frame(clusters,sil_width)
    out
}

indval.stride <- function(stride,taxa,numitr=1)
{
    res <- rep(NA,ncol(stride$clustering))
    for (i in 1:ncol(stride$clustering)) {
        res[i] <- mean(indval(taxa,stride$clustering[,i],numitr=numitr)$indcls)
    }
    clusters <- stride$seq 
    indval <- res
    out <- data.frame(clusters,indval)
    out
}

optpart.stride <- function(stride,dis)
{
    res <- rep(NA,ncol(stride$clustering))
    for (i in 1:ncol(stride$clustering)) {
        tmp <- optpart(dis,clustering=stride$clustering[,i]) 
        res[i] <- tmp$ratio[tmp$numitr]
    }
    clusters <- stride$seq
    ratio <- res
    out <- data.frame(clusters,ratio)
    out
}

diam.stride <- function(stride,dis,digits=3)
{
    res <- rep(NA,ncol(stride$clustering))
    dis <- as.matrix(dis)
    for (i in 1:ncol(stride$clustering)) {
        tmp <- disdiam(stride$clustering[,i],dis,digits=digits)
        res[i] <- tmp$mean
    }
    clusters <- stride$seq
    diameters <- res
    out <- data.frame(clusters,diameters)
    out
}