tests/testthat/test-MRFcov.R

In nicholasjclark/MRFcov: Markov Random Fields with Additional Covariates

context("MRFcov")

#### Test basic error and warning messages using sample data ####
data("Bird.parasites")

test_that("family must be correctly specified", {
  expect_success(expect_error(MRFcov(data = Bird.parasites, n_nodes = 4, family = 'binmial'),
                              'Please select one of the three family options:
         "gaussian", "poisson", "binomial"'))
})

test_that("n_nodes must be a positive integer", {
expect_success(expect_error(MRFcov(data = Bird.parasites, n_nodes = -1,
                                   family = 'binomial'),
             'Please provide a positive integer for n_nodes'))
})


non.bin.dat <- Bird.parasites
non.bin.dat[1,1] <- 6

test_that("binary data should only have two possible values", {
  expect_success(expect_error(MRFcov(data = non.bin.dat, n_nodes = 4,
                                     family = 'binomial'),
                              'Non-binary variables detected'))
})

inf.dat <- Bird.parasites
inf.dat[1,1] <- Inf

test_that("infinite values are not allowed", {
  expect_success(expect_error(MRFcov(data = inf.dat, n_nodes = 4,
                                     family = 'binomial'),
                              'No infinite values permitted'))
})

test_that("no n_nodes argument should produce a warning", {
expect_success(expect_warning(MRFcov(data = Bird.parasites[, c(1:4)],
                                     family = 'binomial')))
})

# Run tests for poisson model
cov <- rnorm(500, 0.2)
cov2 <- rnorm(500, 4)
sp.2 <- rnorm(500, 1)
non.poiss.dat <- data.frame(sp.1 = ceiling(rnorm(500, 1) + cov2 * 1.5),
                        sp.2 = sp.2, sp.3 = ceiling((sp.2 * 2) + rnorm(500, 0.1)))

test_that("poisson variables must be integers", {
  expect_success(expect_error(MRFcov(data = non.poiss.dat, n_nodes = 3,
                                     family = 'poisson'),
                              'Non-integer variables detected'))
})

#### Run a model using the sample data ####
CRFmod <- MRFcov(data = Bird.parasites, n_nodes = 4,
                 family = 'binomial')

test_that("node coefficient graph must be symmetric", {
  expect_true(isSymmetric(CRFmod$graph, check.attributes = FALSE))
})

test_that("NA coefficients should not exist", {
  expect_false(any(is.na(CRFmod$graph)))
})

test_that("all indirect coefficient graphs must be symmetric", {
expect_true(unlist(lapply(seq_along(CRFmod$indirect_coefs), function(x){
  isSymmetric(CRFmod$indirect_coefs[[x]][[1]],check.attributes = FALSE)
})))
})

test_that("node coefficient outputs must have nrow() == n_nodes", {
  expect_equal(ncol(CRFmod$graph), 4)
  expect_equal(nrow(CRFmod$graph), 4)
})

test_that("node names must be carried over to model coefficient outputs", {
  expect_equal(rownames(CRFmod$graph), colnames(CRFmod$graph))
  expect_equal(rownames(CRFmod$direct_coefs), rownames(CRFmod$graph))
  expect_equal(rownames(CRFmod$graph), colnames(Bird.parasites)[1:4])
})

test_that("intercepts must be included in direct coefs", {
  expect_equal(CRFmod$intercepts, CRFmod$direct_coefs[,1])
})

test_that("coefficient names must be returned in param_names", {
  expect_equal(colnames(CRFmod$direct_coefs[, -1]), CRFmod$param_names)
})