skip_if_not_installed("mgcv")
skip_if_not_installed("gamm4")
skip_if_not_installed("rstanarm")

set.seed(2) ## simulate some data...
void <- capture.output(
  dat <- mgcv::gamSim(1, n = 400, dist = "normal", scale = 2)
)

bt <- mgcv::gam(y ~ te(x0, x1, k = 7) + s(x2) + s(x3),
  data = dat,
  method = "REML"
)

test_that("find_smooth - gam", {
  expect_equal(find_smooth(bt), list(smooth_terms = c("te(x0, x1, k = 7)", "s(x2)", "s(x3)")))
  expect_equal(find_smooth(bt, flatten = TRUE), c("te(x0, x1, k = 7)", "s(x2)", "s(x3)"))
})

test_that("find_smooth - mgcv::gamm", {
  model <- mgcv::gamm(Petal.Length ~ Petal.Width + s(Sepal.Length), random = list(Species = ~1), data = iris)
  expect_equal(find_smooth(model, flatten = TRUE), "s(Sepal.Length)")
})

test_that("find_smooth - gamm4", {
  model <- gamm4::gamm4(Petal.Length ~ Petal.Width + s(Sepal.Length), random = ~ (1 | Species), data = iris)
  expect_equal(find_smooth(model, flatten = TRUE), "s(Sepal.Length)")
})

test_that("find_smooth - stan_gamm4", {
  skip_on_cran()
  model <- suppressWarnings(
    rstanarm::stan_gamm4(
      Petal.Length ~ Petal.Width + s(Sepal.Length),
      random = ~ (1 | Species),
      data = iris,
      iter = 100,
      chains = 1,
      refresh = 0
    )
  )
  expect_equal(find_smooth(model, flatten = TRUE), "s(Sepal.Length)")
})

# test_that("find_smooth - brms", {
#   model <- brms::brm(Petal.Length ~ Petal.Width + s(Sepal.Length) + (1|Species), data = iris, iter=100, chains=1, refresh=0)
#   expect_equal(find_smooth(model, flatten = TRUE), "s(Sepal.Length)")
# })