set.seed(42)
n <- 100
x <- rnorm(n)
errors <- rgamma(n, shape = 2, scale = 1) - 2
y <- 2 + 1.5 * x + errors
test_data <- data.frame(x = x, y = y)

test_that("summary method runs without error", {
  fit <- lm_pmm2(y ~ x, data = test_data)

  expect_error(summary(fit), NA)
})

test_that("coef method extracts coefficients", {
  fit <- lm_pmm2(y ~ x, data = test_data)

  c <- coef(fit)

  expect_equal(c, fit@coefficients)
  expect_length(c, 2)
})

test_that("residuals method returns residual vector", {
  fit <- lm_pmm2(y ~ x, data = test_data)

  res <- residuals(fit)

  expect_true(is.numeric(res))
  expect_equal(length(res), n)
})

test_that("fitted method returns fitted values", {
  fit <- lm_pmm2(y ~ x, data = test_data)

  fitted_vals <- fitted(fit)

  expect_true(is.numeric(fitted_vals))
  expect_equal(length(fitted_vals), n)
})

test_that("predict method supports new data", {
  fit <- lm_pmm2(y ~ x, data = test_data)

  new_data <- data.frame(x = c(0, 1, -1))
  preds <- predict(fit, newdata = new_data)

  expect_true(is.numeric(preds))
  expect_equal(length(preds), 3)
})

test_that("plot method works for PMM2fit", {
  fit <- lm_pmm2(y ~ x, data = test_data)

  expect_error(plot(fit), NA)
})