tests/testthat/test-distribution-helpers.R

In gavinsimpson/gratia: Graceful 'ggplot'-Based Graphics and Other Functions for GAMs Fitted Using 'mgcv'

# Test rootograms() and related residuals functions

test_that("nb_theta works a nb() model", {
  expect_silent(theta <- nb_theta(m_nb))
})

test_that("nb_theta works a nb() model", {
  expect_silent(theta <- nb_theta(m_negbin))
  # this model sets theta == 25
  expect_equal(theta, 25)
})

test_that("nb_theta fails for a non NB model", {
  expect_error(theta <- nb_theta(m_gam),
    "Only negative binomial models are supported.",
    fixed = TRUE
  )
})

test_that("theta works a nb() model", {
  expect_silent(p <- theta(m_nb))
  expect_type(p, "double")
})

test_that("theta works a nb() model", {
  expect_silent(theta(m_nb))
})

test_that("theta fails for poisson() model", {
  expect_error(theta(b_pois),
    "No additional parameters available for this model",
    fixed = TRUE
  )
})

test_that("has_theta() returns TRUE for a nb() GAM", {
  expect_true(has_theta(m_nb))
})

test_that("has_theta() returns TRUE for a nb() GAM", {
  expect_true(has_theta(nb()))
})

test_that("has_theta() returns FALSE for a poisson() GAM", {
  expect_false(has_theta(b_pois))
})

test_that("has_theta() returns FALSE for a gaussian() GAM", {
  expect_false(has_theta(m_gam))
})