library(componentSkeleton)
library(biclust)

# test <- matrix(rbinom(400, 50, 0.4), 20, 20)

execute <- function( cf ) {
  # Input and parameters
  matr <- CSV.read(get.input(cf, 'in'))
  delta <- get.parameter(cf, 'delta', 'float')
  alpha <- get.parameter(cf, 'alpha', 'float')
  number <- get.parameter(cf, 'number', 'float')
  plot <- get.parameter(cf, 'plot', 'boolean')
  idColumn <- get.parameter(cf, 'idColumn', 'string')

  # Check input data
  if (idColumn != ""){
    idx <- which(colnames(matr) == idColumn)
    tab <- as.matrix(matr[,-idx])
    row.names(tab) <- matr[,idx]
  } else {
    tab <- as.matrix(matr)
  }

  bic <- biclust(tab, method=BCCC(), delta=delta,  alpha=alpha, number=number)
  clustcols <- apply( t(bic@NumberxCol), 2, as.numeric)
  colnames(clustcols) <- paste0('B',1:ncol(clustcols))
  rownames(clustcols) <- colnames(tab)

  clustrows <- apply( bic@RowxNumber, 2, as.numeric)
  colnames(clustrows) <- paste0('B',1:ncol(clustrows))
  rownames(clustrows) <- row.names(tab)

  CSV.write( get.output( cf, 'rowClusts' ), clustrows )
  CSV.write( get.output( cf, 'colClusts' ), clustcols )

  # Default values
  #diagnosticTest(BCresult, data, number = 1:BCresult@Number, verbose = TRUE, statistics = c("F", "Tukey"), sampling = TRUE, samplingtypes = NULL, nSim = 1000, alpha = 0.05, save_F = FALSE)
  stat <- diagnosticTest(BCresult=bic, data=tab, statistics="F",samplingtypes="Permutation")
  getResult <- function(x) {
    data.frame(test=paste0(x$table$Method, x$table$Type), pvalue=x$table$pVal)
  }
  pvals <- lapply(stat, FUN=getResult)
  statistics <- do.call(rbind,pvals)
  statistics$bicluster <- unlist(lapply(row.names(test3), function(x) substr(x,1, nchar(x)-2))

  CSV.write( get.output( cf, 'statistics' ), statistics )
}
main( execute )

## Possible methods within biclust. For now choose BCCC()
  # ‘BCCC’: delta, alpha, number
  # ‘BCXmotifs’: ns, nd, sd, alpha, number
  # ‘BCPlaid’: cluster, fit.model, rs, background, background.layer, background.df, shuffle, iter.startup, iter.layer, back.fit, row.release, col.release, max.layers
  # ‘BCSpectral’: normalization, numberOfEigenvalues, minr, minc, withinVar,
  # ‘BCBimax’: minr, minc, number, maxc
  # ‘BCQuest’: ns, nd, sd, alpha, number, d, quant, vari