context("car")

skip_on_cran()

skip_if_not_installed("modeltests")
library(modeltests)

skip_if_not_installed("survival")
library(survival)

skip_if_not_installed("car")

test_that("tidy.durbinWatsonTest", {
  check_arguments(tidy.durbinWatsonTest)

  dw <- car::durbinWatsonTest(lm(mpg ~ wt, data = mtcars))
  td <- tidy(dw)
  gl <- glance(dw)

  check_tidy_output(td)
  check_glance_outputs(gl)

  check_dims(td, 1, 5)
})

test_that("tidy.leveneTest", {
  skip_if_not_installed("carData")

  mod1 <- with(carData::Moore, leveneTest(conformity, fcategory))
  mod2 <- with(carData::Moore, leveneTest(conformity, interaction(fcategory, partner.status)))
  mod3 <- leveneTest(conformity ~ fcategory * partner.status, data = Moore)
  mod4 <- leveneTest(lm(conformity ~ fcategory * partner.status, data = Moore))
  mod5 <- leveneTest(conformity ~ fcategory * partner.status, data = Moore, center = mean)
  mod6 <- leveneTest(conformity ~ fcategory * partner.status, data = Moore, center = mean, trim = 0.1)

  # This is a tidy method, but the model object is very simple and the output
  # is a 1-row tibble with `df` and `df.residual` columns. `tidy.htest` and
  # `glance.htest` also return the same things.
  check_glance_outputs(
    tidy(mod1),
    tidy(mod2),
    tidy(mod3),
    tidy(mod4),
    tidy(mod5),
    tidy(mod6)
  )
})

test_that("tidy car::Anova glm", {
  fit <- glm(am ~ mpg, mtcars, family = "binomial")
  fit2 <- glm(am ~ mpg + wt, mtcars, family = "binomial")

  car_anova <- car::Anova(fit, test.statistic = "LR")
  car_anova2 <- car::Anova(fit2, test.statistic = "LR")

  td <- tidy(car_anova)
  td2 <- tidy(car_anova2)

  check_tidy_output(td)
  check_tidy_output(td2)

  check_dims(td, expected_rows = 1, expected_cols = 4)
  check_dims(td2, expected_rows = 2, expected_cols = 4)
})

test_that("tidy car::Anova coxph", {
  fit <- coxph(Surv(time, status) ~ differ, data = colon)
  fit2 <- coxph(Surv(time) ~ differ, data = colon)
  fit3 <- coxph(Surv(time) ~ differ + extent, data = colon)

  car_anova_coxph <- car::Anova(fit)
  car_anova_coxph2 <- car::Anova(fit2)
  car_anova_coxph3 <- car::Anova(fit3)

  td <- tidy(car_anova_coxph)
  td2 <- tidy(car_anova_coxph2)
  td3 <- tidy(car_anova_coxph3)

  check_tidy_output(td)
  check_tidy_output(td2)
  check_tidy_output(td3)

  check_dims(td, expected_rows = 2, expected_cols = 5)
  check_dims(td2, expected_rows = 2, expected_cols = 5)
  check_dims(td3, expected_rows = 2, expected_cols = 4)
})

test_that("tidy car::linearHypothesis with long formulas (#1171)", {
  reg_long <- 
    lm(
      Fertility ~ Agriculture + Examination + Education + Catholic + Infant.Mortality, 
      data = swiss
    )
  
  test_long <- car::linearHypothesis(reg_long, hypothesis.matrix = c(0,1, 0, 0, 0, -1))
  
  td <- broom::tidy(test_long)
  check_dims(td, expected_rows = 1, expected_cols = 10)
})