### Read Output of Mathematica
library(exametrika)

# Mathematica 参照データをシートごとに読み込み
sheets <- c(
  "Item", "JSS", "JCRR",
  "CCRR", "Item_Lift", "Mutual_Information",
  "Phi_Coefficient", "Tetrachoric_Correlation"
)

data_list <- lapply(sheets, function(sheet) {
  df <- read.csv(
    test_path(
      "fixtures", "mathematica_reference",
      paste0("Chapter03CTT_", sheet, ".csv")
    ),
    check.names = FALSE
  )
  # select numeric columns only
  df <- df[, sapply(df, is.numeric), drop = FALSE]
  as.matrix(df)
})

names(data_list) <- c(
  "Ch03Items", "Ch03JSS", "Ch03JCRR", "Ch03CCRR",
  "Ch03IL", "Ch03MI", "Ch03Phi", "Ch03Tet"
)

## read raw data (J20S400 package data is pre-formatted as nominal)
dat <- read.csv(
  test_path("fixtures", "sample_data", "J20S400.csv"),
  check.names = FALSE
)

# Test Section ------------------------------------------------------------

test_that("Correct Response Rate", {
  result <- crr(dat, na = -99) |> as.vector()
  expect <- data_list$Ch03Items[, 2] |> as.vector()
  expect_equal(result, expect)
})


test_that("Item Odds", {
  result <- ItemOdds(dat, na = -99) |> as.vector()
  expect <- data_list$Ch03Items[, 3] |> as.vector()
  expect_equal(result, expect)
})

test_that("Item Threshold", {
  result <- ItemThreshold(dat, na = -99) |> as.vector()
  expect <- data_list$Ch03Items[, 4] |> as.vector()
  expect_equal(result, expect)
})

test_that("Item Entropy", {
  result <- ItemEntropy(dat, na = -99) |> as.vector()
  expect <- data_list$Ch03Items[, 5] |> as.vector()
  expect_equal(result, expect)
})

# Between Items Section ---------------------------------------------------

test_that("Joint sample size", {
  result <- JointSampleSize(dat, na = -99) |>
    unclass() |>
    unname()
  expect <- data_list$Ch03JSS |> unname()
  expect_equal(result, expect)
})

test_that("Joint Correct Response Rate", {
  tmp <- dataFormat(dat, na = -99)
  result <- JCRR(tmp) |>
    unclass() |>
    unname()
  expect <- data_list$Ch03JCRR |> unname()
  expect_equal(result, expect)
})


test_that("Conditional Correct Response Rate", {
  tmp <- dataFormat(dat, na = -99)
  result <- CCRR(tmp) |>
    unclass() |>
    unname()
  expect <- data_list$Ch03CCRR |> unname()
  expect_equal(result, expect)
})


test_that("Item Lift", {
  result <- ItemLift(dat, na = -99) |>
    unclass() |>
    unname()
  expect <- data_list$Ch03IL |> unname()
  expect_equal(result, expect)
})

test_that("Mutual Information", {
  result <- MutualInformation(dat, na = -99) |>
    unclass() |>
    unname()
  expect <- data_list$Ch03MI |> unname()
  expect_equal(result, expect)
})

test_that("Phi Coefficient", {
  result <- PhiCoefficient(dat, na = -99) |>
    unclass() |>
    unname()
  expect <- data_list$Ch03Phi |> unname()
  expect_equal(result, expect)
})

test_that("Tetrachoric Correlation Matrix", {
  result <- TetrachoricCorrelationMatrix(dat, na = -99) |>
    unclass() |>
    unname()
  expect <- data_list$Ch03Tet |> unname()
  expect_equal(expected = expect, object = result, tolerance = 1e-4)
})

test_that("Item Total Correlation", {
  result <- ItemTotalCorr(dat, na = -99)
  expect <- data_list$Ch03Items[, 6]
  expect_equal(expected = expect, object = as.vector(result))
})


test_that("Item Total Biserial Correlation", {
  tmp <- dataFormat(dat, na = -99)
  result <- ITBiserial(tmp) |> as.vector()
  expect <- data_list$Ch03Items[, 7] |> as.vector()
  expect_equal(result, expect, tolerance = 1e-4)
})