test_that("empirical_pval() returns correct values", {
  tol <- 0.1
  expect_equal(
    empirical_pval(seq(1, 100), 95, alternative = "greater"),
    (5 + 1) / (100 + 1)
  )
  expect_lt(empirical_pval(rnorm(1e3), 2, alternative = "two.sided"), .05 + tol)
  expect_lt(empirical_pval(rnorm(1e3), -2, alternative = "less"), .025 + tol)
})

test_that("empirical_pval() identifies incorrect inputs", {
  matrix_input <- matrix(rnorm(1e3), ncol = 10)
  multiple_observations <- c(2, 3)
  expect_error(empirical_pval(matrix_input, 10))
  expect_error(empirical_pval(seq(1, 100), multiple_observations))
  expect_error(empirical_pval(95, seq(1, 100)))
})

yrep <- matrix(rnorm(1e3), ncol = 10)
y <- rep(2, 10)

test_that("post_pred_pval() returns a list containing a numeric value and a ggplot", {
  x <- post_pred_pval(yrep, y, plot = TRUE)
  expect_is(x, "list")
  expect_is(x[[1]], "numeric")
  expect_is(x[[2]], "ggplot")
})

test_that("post_pred_pval() identifies potentially incorrect inputs", {
  expect_warning(post_pred_pval(yrep, c(y, 2)))
})

test_that("post_pred_pval() identifies incorrect inputs", {
  expect_error(post_pred_pval(yrep, y, test_statistic = "mean"))
  expect_error(post_pred_pval(yrep, y, plot = "no"))
})