test_that("ai_dashboard computes weighted overall correctly with complete scores", {
  d <- ai_dashboard(reliability = 0.8, fairness = 0.8, robustness = 0.8,
                     calibration = 0.8, hallucination = 0.8)
  expect_equal(round(d$overall, 6), 0.8)
})

test_that("ai_dashboard renormalizes weights when a score is missing, with a warning", {
  expect_warning(
    d <- ai_dashboard(reliability = 0.9, fairness = NA, robustness = 0.9,
                       calibration = 0.9, hallucination = 0.9),
    "Renormalizing"
  )
  # with one NA out of five equal-weighted scores, remaining 4 should
  # renormalize and overall should equal 0.9, not be deflated
  expect_equal(round(d$overall, 6), 0.9)
})

## ---- print method and grade boundaries -----------------------------

test_that("print.aiEvalR_dashboard renders bars and a grade", {
  d <- ai_dashboard(reliability = 0.94, fairness = 0.82, robustness = 0.70,
                     calibration = 0.88, hallucination = 0.81)
  expect_output(print(d), "AI Evaluation Dashboard")
  expect_output(print(d), "Overall")
})

test_that("ai_dashboard assigns letter grades by overall score", {
  hi <- ai_dashboard(0.95, 0.95, 0.95, 0.95, 0.95)
  lo <- ai_dashboard(0.50, 0.50, 0.50, 0.50, 0.50)
  expect_equal(hi$grade, "A")
  expect_equal(lo$grade, "F")
})

test_that("ai_dashboard honors custom weights that sum to 1", {
  w <- c(Reliability = 0.6, Fairness = 0.1, Robustness = 0.1,
         Calibration = 0.1, Hallucination = 0.1)
  d <- ai_dashboard(1.0, 0, 0, 0, 0, weights = w)
  expect_equal(round(d$overall, 6), 0.6)
})

test_that("ai_dashboard rejects weights that do not sum to 1", {
  w <- c(Reliability = 0.6, Fairness = 0.1, Robustness = 0.1,
         Calibration = 0.1, Hallucination = 0.5)
  expect_error(ai_dashboard(1, 1, 1, 1, 1, weights = w))
})

test_that("ai_dashboard rejects scores outside [0, 1]", {
  expect_error(ai_dashboard(1.5, 0.5, 0.5, 0.5, 0.5))
})