# Shared test helpers for neuroim2
# Automatically loaded by testthat before tests run.

#' Create a NeuroSpace with sensible defaults
#' @note NeuroSpace uses a 3D affine internally. For >3D spaces, only the
#'   first 3 elements of spacing/origin define the spatial transform.
make_space <- function(dim = c(10L, 10L, 10L), spacing = c(1, 1, 1),
                       origin = c(0, 0, 0), trans = NULL) {
  dim <- as.integer(dim)
  if (!is.null(trans)) {
    NeuroSpace(dim, trans = trans)
  } else {
    NeuroSpace(dim, spacing = spacing, origin = origin)
  }
}

#' Create a DenseNeuroVol with random data
make_vol <- function(dim = c(10L, 10L, 10L), data = NULL, label = "", ...) {
  sp <- make_space(dim = dim, ...)
  if (is.null(data)) data <- rnorm(prod(dim))
  DenseNeuroVol(data, sp, label = label)
}

#' Create a DenseNeuroVec (4D) with random data
make_vec <- function(dim = c(10L, 10L, 10L), ntime = 20L, data = NULL, ...) {
  sdim <- as.integer(dim)
  ntime <- as.integer(ntime)
  sp4 <- make_space(dim = c(sdim, ntime), ...)
  if (is.null(data)) data <- array(rnorm(prod(sdim) * ntime), c(sdim, ntime))
  DenseNeuroVec(data, sp4)
}

#' Create a LogicalNeuroVol mask
make_mask <- function(dim = c(10L, 10L, 10L), frac = 0.3, ...) {
  sp <- make_space(dim = dim, ...)
  m <- array(runif(prod(dim)) < frac, as.integer(dim))
  LogicalNeuroVol(m, sp)
}

#' Create a SparseNeuroVol with random nonzero entries
make_sparse_vol <- function(dim = c(10L, 10L, 10L), frac = 0.3, ...) {
  sp <- make_space(dim = dim, ...)
  n <- prod(dim)
  idx <- sort(sample.int(n, floor(n * frac)))
  sv <- Matrix::sparseVector(x = rnorm(length(idx)), i = idx, length = n)
  SparseNeuroVol(sv, sp, indices = idx)
}

#' Create a SparseNeuroVec with random data
make_sparse_vec <- function(dim = c(10L, 10L, 10L), ntime = 20L, frac = 0.3, ...) {
  sdim <- as.integer(dim)
  ntime <- as.integer(ntime)
  sp4 <- make_space(dim = c(sdim, ntime), ...)
  mask <- array(runif(prod(sdim)) < frac, sdim)
  nk <- sum(mask)
  dat <- matrix(rnorm(ntime * nk), nrow = ntime, ncol = nk)
  SparseNeuroVec(dat, sp4, mask = mask)
}

#' Write a temporary NIfTI file from a volume, return the path
make_temp_nifti <- function(vol, ext = ".nii.gz") {
  tmp <- tempfile(fileext = ext)
  write_vol(vol, tmp)
  tmp
}