library(testthat)
library(AutoEDA)

#==========================================================
# missing_summary()
#==========================================================

test_that("missing_summary returns MissingSummary object", {

  ms <- missing_summary(iris)

  expect_s3_class(
    ms,
    "MissingSummary"
  )

})

test_that("missing_summary returns expected components", {

  ms <- missing_summary(iris)

  expect_named(
    ms,
    c(
      "TotalRows",
      "TotalColumns",
      "TotalCells",
      "TotalMissing",
      "OverallPercent",
      "VariablesWithMissing",
      "CompleteCases",
      "VariableSummary"
    )
  )

})

test_that("missing_summary works when no missing values exist", {

  ms <- missing_summary(iris)

  expect_equal(ms$TotalRows, 150)
  expect_equal(ms$TotalColumns, 5)
  expect_equal(ms$TotalMissing, 0)
  expect_equal(ms$VariablesWithMissing, 0)
  expect_equal(ms$CompleteCases, 150)

})

test_that("missing_summary detects missing values correctly", {

  dat <- iris

  dat$Sepal.Length[1:10] <- NA
  dat$Petal.Width[1:5] <- NA

  ms <- missing_summary(dat)

  expect_equal(
    ms$TotalMissing,
    15
  )

  expect_equal(
    ms$VariablesWithMissing,
    2
  )

})

test_that("missing_summary computes overall percentage correctly", {

  dat <- iris

  dat$Sepal.Length[1:15] <- NA

  ms <- missing_summary(dat)

  expect_equal(
    ms$OverallPercent,
    round(100 * 15 / (150 * 5), 2)
  )

})

test_that("missing_summary orders variables by missing percentage", {

  dat <- iris

  dat$Sepal.Length[1:20] <- NA
  dat$Petal.Width[1:5] <- NA

  ms <- missing_summary(dat)

  expect_equal(
    ms$VariableSummary$Variable[1],
    "Sepal.Length"
  )

})

test_that("missing_summary works with mixed variable types", {

  dat <- data.frame(

    x = c(1,2,NA,4),

    y = c("A","B","C",NA),

    z = factor(c("Low","High","Low","Medium"))

  )

  ms <- missing_summary(dat)

  expect_s3_class(
    ms,
    "MissingSummary"
  )

})

test_that("missing_summary stores variable classes", {

  dat <- data.frame(

    x = 1:5,

    y = letters[1:5],

    z = factor(letters[1:5])

  )

  ms <- missing_summary(dat)

  expect_equal(

    ms$VariableSummary$Type,

    c(
      "integer",
      "character",
      "factor"
    )

  )

})

#==========================================================
# Invalid input
#==========================================================

test_that("missing_summary rejects non-data.frame", {

  expect_error(

    missing_summary(
      1:10
    )

  )

})

test_that("missing_summary rejects empty data.frame", {

  expect_error(

    missing_summary(
      data.frame()
    )

  )

})

#==========================================================
# Return values
#==========================================================

test_that("VariableSummary is a data.frame", {

  ms <- missing_summary(iris)

  expect_s3_class(

    ms$VariableSummary,

    "data.frame"

  )

})

test_that("VariableSummary contains expected columns", {

  ms <- missing_summary(iris)

  expect_equal(

    names(ms$VariableSummary),

    c(
      "Variable",
      "Type",
      "Missing",
      "Percent",
      "Complete"
    )

  )

})