library(lgpr)

# -------------------------------------------------------------------------

context("Inverse-gamma distribution")

test_that("inverse gamma density works and errors correctly", {
  a <- 2
  b <- 3
  diff <- abs(dinvgamma_stanlike(1, a, b) - 0.4480836)
  expect_lt(diff, 1e-6)
  expect_error(dinvgamma_stanlike(1, -1, 2))
  expect_error(dinvgamma_stanlike(1, 2, -1))
})

test_that("inverse gamma quantile works and errors correctly", {
  a <- 2
  b <- 3
  diff <- abs(qinvgamma_stanlike(0.9, a, b) - 5.641095)
  expect_lt(diff, 1e-6)
  expect_error(qinvgamma_stanlike(0.9, -1, 2))
  expect_error(qinvgamma_stanlike(0.9, 2, -1))
  expect_error(qinvgamma_stanlike(-0.1, 2, 2))
  expect_error(qinvgamma_stanlike(1.5, 2, 2))
})

test_that("plotting the inverse gamma distribution works", {
  a <- 2
  b <- 3
  p1 <- plot_invgamma(a, b)
  p2 <- plot_invgamma(a, b, return_quantiles = TRUE)
  c1 <- as.character(class(p1))
  c2 <- as.character(class(p2$plot))
  expect_equal(c1, c("gg", "ggplot"))
  expect_equal(c2, c("gg", "ggplot"))
  expect_error(plot_invgamma(a, b, IQR = 1.5))
})