R/blockSizeCluster.R

In MatMatt/MODIS: Acquisition and Processing of MODIS Products

blockSizeCluster <- function(x, var = 8) {
  # uses blockSize to get an idea of blockSize
  estim <- raster::blockSize(x)$nrows[1]
  at <- 1
  ind2 <- 0
  res <- list()
  while (at <= nrow(x)) {
    ind2 <- ind2 + 1
    atin <- at
    size <- sample(max(1, (estim-var)):(estim + var), 1) # min is 1 row
    at <- (at + size)
    if (at > nrow(x)) 
      size <- nrow(x)-(atin-1)

    res[[ind2]] <- list(row=atin,nrows=size)
  }
  res <- matrix(unlist(res),byrow=T,ncol=2)
  res <- list(row=res[,1],nrows=res[,2],n=nrow(res))
  return(res)
}