# skip long running tests on CRAN
skip_on_cran()

#context("test partial dependence")

library(lavaan)

#
# test numeric predictors
#

# skip long running tests on CRAN
skip_on_cran()

# draw predictors randomly
set.seed(325)
N <- 200
pred1 <- factor(sample(c("red", "green", "blue"), N, replace = TRUE))
pred2 <- ordered(sample(c(0, 1, 2), N, replace = TRUE))
pred3 <- as.numeric(sample(1:20, size = N, replace = TRUE))

x <- rnorm(N) + ifelse(pred2 == "1", 10, 0)
x <- x + pred3 / 10
df <- data.frame(x, pred3)

model = "x ~~ var*x; x~ mu*0"
fitted_model <- lavaan(model, df)
tree = semtree(fitted_model, df, control = semtree.control(verbose = TRUE, report.level = 99))
plot(tree)


forest = semforest(fitted_model, df, control = semforest.control(num.trees =
                                                                   10))

pd = partialDependence(forest, reference.var = "pred3")
plot(pd, parameter="mu")
test_that("partial dependence works for numeric predictors", {
  expect_equal(class(forest), "semforest")
  expect_equal(class(pd), "partialDependence")
  expect_equal(dim(pd$samples), c(20, 3))
  expect_true(median(pd$samples$mu[pd$samples$pred3 < 1]) > median(pd$samples$mu[pd$samples$pred3 >=
                                                                                   1]))
})

plot(pd, "mu")

pd = partialDependence(forest, reference.var = "pred2")

#
# testing unordered factors
#
df <- data.frame(x, pred1)
forest = semforest(fitted_model, df, control = semforest.control(num.trees =
                                                                   10))
pd = partialDependence(forest, reference.var = "pred1")
plot(pd, "mu")
test_that("partial dependence works for numeric predictors", {
  expect_equal(class(forest), "semforest")
  expect_equal(class(pd), "partialDependence")
  expect_equal(dim(pd$samples), c(3, 3))
})

#
# testing ordered factors
#
df <- data.frame(x, pred2)
forest = semforest(fitted_model, df, control = semforest.control(num.trees =
                                                                   10))
pd = partialDependence(forest, reference.var = "pred2")
test_that("partial dependence works for numeric predictors", {
  expect_equal(class(forest), "semforest")
  expect_equal(class(pd), "partialDependence")
})

#
# expect error here
#
test_that("missing reference variable is detected", {
  expect_error(pd = partialDependence(forest, reference.var = "reallyweirdPredictor958Qz6"))
})