tests/testthat/test_sampling-sample_quadrats().R

In mobsim: Spatial Simulation and Scale-Dependent Analysis of Biodiversity Changes

test_that("fails correctly", {
   sim_com1 <- sim_poisson_community(s_pool = 5L, n_sim = 50L)

   expect_error(sample_quadrats(data.frame(c(1:5), c(1:5))))
   expect_error(sample_quadrats(sim_com1, n_quadrats = 10.5))
   expect_error(sample_quadrats(sim_com1, quadrat_area = 0.01, method = "foo"))
})

test_that("classes are as expected", {
   comm_mat1 <- sample_quadrats(
      comm = sim_poisson_community(s_pool = 5L, n_sim = 50L),
      plot = FALSE, n_quadrats = 2L,
      method = "random", avoid_overlap = FALSE
   )

   expect_type(comm_mat1, "list")
   expect_s3_class(comm_mat1[[1]], "data.frame")
   expect_s3_class(comm_mat1[[2]], "data.frame")
})

test_that("dimensions are as expected", {
   S = 5L
   sim_com1 <- sim_poisson_community(
      s_pool = S, n_sim = 50L, fix_s_sim = TRUE)
   n_quadrats <- 10L
   comm_mat <- sample_quadrats(
      sim_com1, n_quadrats = n_quadrats,
      method = "random", plot = FALSE,
      avoid_overlap = TRUE
   )

   expect_vector(comm_mat, size = 2L)
   expect_equal(dim(comm_mat[[1]]), c(n_quadrats, S))
   expect_equal(dim(comm_mat[[2]]), c(n_quadrats, 2L))
})

test_that("default argument values are as expected", {
   comm <- sim_thomas_community(
      s_pool = 10L, n_sim = 100L,
      sad_type = "lnorm", sad_coef = list(cv_abund = 1), fix_s_sim = TRUE,
      sigma = 1, mother_points = 0L, xrange = c(0, 1), yrange = c(0, 1),
      seed = 42L)
   expect_snapshot_output(sample_quadrats(comm = comm, seed = 42L))
})

test_that("calls the correct subfunction if method is random, with overlap", {
   sim_com1 <- sim_poisson_community(s_pool = 5L, n_sim = 50L)
   quadrat_area <- 0.01
   expect_equal(
      as.matrix(
         suppressWarnings(
            sample_quadrats(
               comm = sim_com1, seed = 42L, plot = FALSE, n_quadrats = 10L,
               quadrat_area = quadrat_area,
               method = "random", avoid_overlap = FALSE)$xy_dat
         ),
         rownames.force = FALSE
      ),
      as.matrix(
         suppressWarnings(
            sampling_random_overlap(
               xmin = sim_com1$x_min_max[1L],
               xmax = sim_com1$x_min_max[2L] - sqrt(quadrat_area),
               ymin = sim_com1$y_min_max[1L],
               ymax = sim_com1$y_min_max[2L] - sqrt(quadrat_area),
               min_dist = sqrt(2*quadrat_area),
               n_quadrats = 10L, seed = 42L)
         )
      )
   )
})

test_that("calls the correct subfunction if method is random, without overlap, with spatstat", {
   testthat::skip_if_not(requireNamespace("spatstat.random", quietly = TRUE))
   sim_com1 <- sim_poisson_community(s_pool = 5L, n_sim = 50L)
   quadrat_area <- 0.01
   expect_equal(
      as.matrix(
         sample_quadrats(
            comm = sim_com1, seed = 42L, plot = FALSE, n_quadrats = 10L,
            quadrat_area = quadrat_area,
            method = "random", avoid_overlap = TRUE)$xy_dat,
         rownames.force = FALSE),
      as.matrix(
         sampling_random_spatstat(
            xmin = sim_com1$x_min_max[1L],
            xmax = sim_com1$x_min_max[2L] - sqrt(quadrat_area),
            ymin = sim_com1$y_min_max[1L],
            ymax = sim_com1$y_min_max[2L] - sqrt(quadrat_area),
            min_dist = sqrt(2*quadrat_area), n_quadrats = 10L, seed = 42L)
      )
   )
})

test_that("calls the correct subfunction if method is random, without overlap, without spatstat", {
   mockery::stub(where = sample_quadrats,
                 what = "requireNamespace", how = FALSE, depth = 1L)

   sim_com1 <- sim_poisson_community(s_pool = 5L, n_sim = 50L)
   quadrat_area <- 0.01
   expect_equal(
      as.matrix(
         sample_quadrats(
            comm = sim_com1, seed = 42L, plot = FALSE, n_quadrats = 10L, quadrat_area = quadrat_area,
            method = "random", avoid_overlap = TRUE)$xy_dat,
         rownames.force = FALSE),
      as.matrix(
         sampling_random_bruteforce(
            xmin = sim_com1$x_min_max[1L],
            xmax = sim_com1$x_min_max[2L] - sqrt(quadrat_area),
            ymin = sim_com1$y_min_max[1L],
            ymax = sim_com1$y_min_max[2L] - sqrt(quadrat_area),
            min_dist = sqrt(2*quadrat_area), n_quadrats = 10L, seed = 42L)
      )
   )
})

test_that("calls the correct subfunction if method is transects", {
   sim_com1 <- sim_poisson_community(s_pool = 5L, n_sim = 50L)
   quadrat_area <- 0.01
   expect_equal(
      as.matrix(
         sample_quadrats(
            comm = sim_com1, plot = FALSE, n_quadrats = 5L,
            quadrat_area = quadrat_area,
            method = "transect")$xy_dat,
         rownames.force = FALSE),
      as.matrix(
         sampling_transects(xmin = sim_com1$x_min_max[1L], xmax = sim_com1$x_min_max[2L] - sqrt(quadrat_area),
                            ymin = sim_com1$y_min_max[1L], ymax = sim_com1$y_min_max[2L] - sqrt(quadrat_area),
                            x0 = 0, y0 = 0, delta_x = 0.1, delta_y = 0.1,
                            quadrat_size = sqrt(quadrat_area), n_quadrats = 5L)
      )
   )
})
test_that("calls the correct subfunction if method is grid", {
   sim_com1 <- sim_poisson_community(s_pool = 5L, n_sim = 50L)
   quadrat_area <- 0.01
   expect_equal(
      as.matrix(
         sample_quadrats(sim_com1, plot = FALSE, n_quadrats = 5L, quadrat_area = quadrat_area,
                         method = "grid")$xy_dat,
         rownames.force = FALSE),
      as.matrix(
         sampling_grids(
            xmin = sim_com1$x_min_max[1L],
            xmax = sim_com1$x_min_max[2L] - sqrt(quadrat_area),
            ymin = sim_com1$y_min_max[1L],
            ymax = sim_com1$y_min_max[2L] - sqrt(quadrat_area),
            x0 = 0, y0 = 0, delta_x = 0.1, delta_y = 0.1,
            quadrat_size = sqrt(quadrat_area), n_quadrats = 5L),
         rownames.force = FALSE)
   )
})

test_that("calls the correct subfunction if edge case where n_quadrats == 1", {
   sim_com1 <- sim_poisson_community(s_pool = 5L, n_sim = 50L)
   quadrat_area <- 0.01
   expect_equal(
      as.numeric(
         sample_quadrats(
            comm = sim_com1, seed = 42L, plot = FALSE, n_quadrats = 1L,
            quadrat_area = quadrat_area,
            method = "random", avoid_overlap = TRUE)$xy_dat
      ),
      as.numeric(
         sampling_one_quadrat(
            xmin = sim_com1$x_min_max[1L],
            xmax = sim_com1$x_min_max[2L] - sqrt(quadrat_area),
            ymin = sim_com1$y_min_max[1L],
            ymax = sim_com1$y_min_max[2L] - sqrt(quadrat_area),
            seed = 42L)
      )
   )
})

test_that("plotting works", {
   sim_com1 <- sim_poisson_community(s_pool = 5L, n_sim = 50L)
   expect_snapshot_output(
      sample_quadrats(
         comm = sim_com1, n_quadrats = 15L, quadrat_area = 0.001,
         seed = 42L, plot = TRUE, method = "random", avoid_overlap = TRUE
      )
   )
   expect_snapshot_output(
      sample_quadrats(
         comm = sim_com1, n_quadrats = 25L, quadrat_area = 0.001,
         seed = 42L, plot = TRUE, method = "grid"
      )
   )
})