tests/testthat/test-latentvec.R

In bbuchsbaum/neuroim2: Data Structures and IO for Neuroimaging Data with an Emphasis on Functional Magnetic Resonance Imaging (FMRI)

library(purrr)
gmask <- system.file("extdata", "global_mask.nii", package="neuroim2")


gen_dat <- function(d1 = 12,
                    d2 = 12,
                    d3 = 12,
                    d4 = 4,
                    rand = FALSE) {
  if (rand) {
    dat <- array(rnorm(d1 * d2 * d3 * d4), c(d1, d2, d3, d4))
  } else {
    dat <- array(0, c(d1, d2, d3, d4))
  }
  spc <- NeuroSpace(c(d1, d2, d3, d4))
  DenseNeuroVec(dat, spc)
}

gen_latent_vec <- function(d1,d2,d3,d4) {
  bv <- gen_dat(12,12,12,4, rand=TRUE)
  mat <- bv@.Data
  dim(mat) <- c(12*12*12,4)
  mat <- t(mat)
  mask <- bv[[1]]
  mask[] <- 1
  mask <- as.logical(mask)
  pres <- prcomp(mat)
  svec <- LatentNeuroVec(pres$x, pres$rotation, space=space(bv), mask=mask, offset=colMeans(mat))

}

context("latentneurovec")

test_that("can construct a LatentNeuroVec", {
  bv <- gen_dat(12,12,12,4, rand=TRUE)
  mat <- bv@.Data
  dim(mat) <- c(12*12*12,4)
  mat <- t(mat)
  mask <- bv[[1]]
  mask[] <- 1
  mask <- as.logical(mask)
  pres <- prcomp(mat)
  svec <- LatentNeuroVec(pres$x, pres$rotation, space=space(bv), mask=mask, offset=colMeans(mat))

  expect_true(!is.null(svec))
  expect_equal(svec[1,1,1,1], bv[1,1,1,1])
  expect_equal(bv[12,12,12,4], svec[12,12,12,4])
  expect_equal(dim(svec), c(12,12,12,4))
  expect_equal(series(svec, 1:2), series(bv,1:2))

  ## TODO should series drop dimensions?
  expect_equal(as.vector(series(svec, 3)), series(bv,3))

})

# test_that("can write a LatentNeuroVec to h5", {
#   bv <- gen_dat(12,12,12,4, rand=TRUE)
#   mat <- bv@.Data
#   dim(mat) <- c(12*12*12,4)
#   mat <- t(mat)
#   mask <- bv[[1]]
#   mask[] <- 1
#   mask <- as.logical(mask)
#   pres <- prcomp(mat)
#   svec <- LatentNeuroVec(pres$x, pres$rotation, space=space(bv), mask=mask, offset=colMeans(mat))
#   tmp <- paste0(tempfile())
#   write_vec(svec, tmp)
# })

test_that("can extract a single volume from a LatentNeuroVec", {
  bv1 <- gen_latent_vec()
  bv2 <- gen_latent_vec()

  bv3 <- concat(bv1, bv2)

  expect_equal(as.vector(series(bv1,1)), series(bv3,1)[1:4])
  expect_equal(as.vector(series(bv2,1)), series(bv3,1)[5:8])
})