###############
# TEST PRIORS #
###############
prior_model = list(
  y ~ 1 + x,
  1 + (1|id) ~ rel(1) + rel(x),
  rel(1) ~ 0
)

bad_prior = list(
  list(
    cp_1 = "dirichlet(1)",  # Has to be all-dirichlet
    cp_2 = "dnorm(3, 10)"
  ),
  list(
    cp_1 = "dirichlet(1)",
    cp_2 = "dirichlet(0)"  # alpha has to be > 0
  )
)

for (prior in bad_prior) {
  test_name = paste0("Bad priors: ", paste0(prior, collapse=", "))
  testthat::test_that(test_name, {
    testthat::expect_error(test_runs(prior_model, sample = FALSE, prior = prior))
  })
}


good_prior_essential = list(
  list(  # Fixed values and non-default change point
    int_2 = "int_1",
    cp_1 = "dnorm(3, 10)",
    x_2 = "-0.5"
  ),

  list(
    cp_1 = "dirichlet(1)",  # Dirichlet prior on change points
    cp_2 = "dirichlet(1)"
  )
)
good_prior_extensive = list(
  list(  # Changepoint outside of the observed range is allowed
    cp_1 = "dunif(-100, -90)",
    cp_2 = "dnorm(100, 20) T(100, 110)"
  ),

  list(
    cp_1 = "dirichlet(3)",  # Dirichlet prior on change points
    cp_2 = "dirichlet(2)"
  )
)

for (prior in good_prior_essential) {
  test_name = paste0("Good priors (essential): ", paste0(prior, collapse=", "))
  testthat::test_that(test_name, {
    test_runs(prior_model, prior = prior)
  })
}

if (is.null(getOption("test_mcp_allmodels")) == FALSE) {
  for (prior in good_prior_extensive) {
    test_name = paste0("Good priors (extensive): ", paste0(prior, collapse=", "))
    testthat::test_that(test_name, {
      test_runs(prior_model, prior = prior)
    })
  }
}