context("Check the equivalence between manual and automatic scaling")

test_that("case numeric - scaling 'unitvar'", {
  set.seed(321)
  nsom1 <- trainSOM(iris[,1:4], type= "numeric", maxit= 10, scaling= "unitvar")
  iris.scale <- scale(iris[,1:4], center= TRUE, scale= TRUE)
  rownames(iris.scale) <- rownames(iris)
  set.seed(321)
  nsom2 <- trainSOM(iris.scale, type= "numeric", maxit= 10, scaling= "none")

  expect_equal(nsom1$clustering, nsom2$clustering)
})

test_that("case numeric - scaling 'center'", {
  set.seed(321)
  nsom1 <- trainSOM(iris[,1:4], type= "numeric", maxit= 10, scaling= "center")
  iris.scale <- scale(iris[,1:4], center= TRUE, scale= FALSE)
  rownames(iris.scale) <- rownames(iris)
  set.seed(321)
  nsom2 <- trainSOM(iris.scale, type= "numeric", maxit= 10, scaling= "none")
  expect_equal(nsom1$clustering, nsom2$clustering)
})

test_that("case relational - scaling 'cosine'", {
  data(lesmis)
  set.seed(321)
  rsom1 <- trainSOM(dissim.lesmis, type= "relational", maxit= 10, 
                    scaling= "cosine", nb.save=10)
  sim.matrix <- -.5* (diag(1, nrow(dissim.lesmis))-1/nrow(dissim.lesmis)) %*%
    dissim.lesmis %*% (diag(1, nrow(dissim.lesmis))-1/nrow(dissim.lesmis))
  sim.matrix <- round(sim.matrix, 10)
  scaled.ker <- sweep(sweep(sim.matrix,2,sqrt(diag(sim.matrix)),"/"),
                      1,sqrt(diag(sim.matrix)),"/")
  scaled.diss <- 2-2*scaled.ker
  colnames(scaled.diss) <- colnames(dissim.lesmis)
  rownames(scaled.diss) <- rownames(dissim.lesmis)
  lesmis.scale <- round(scaled.diss,10)
  set.seed(321)
  rsom2 <- trainSOM(lesmis.scale, type= "relational", maxit= 10, scaling= "none",
                    nb.save=10)
  expect_equal(rsom1$clustering, rsom2$clustering)
})