skip_on_cran()
require(testthat)
context("HSLS data reads in correctly")
require(EdSurvey)
options(width = 500)
options(useFancyQuotes = FALSE)


source("REF-8-HSLS.R") # has REF output in it

if (!dir.exists(edsurveyHome)) {
  dir.create(edsurveyHome)
}

# able to toggle 'forceReread' for recaching the data if necessary
if (!exists("forceCacheUpdate")) {
  forceCacheUpdate <- FALSE
}

test_that("HSLS data reads in correctly", {
  expect_silent(downloadHSLS(years = 2009, root = edsurveyHome, verbose = FALSE))
  expect_silent(hsls <<- readHSLS(file.path(edsurveyHome, "HSLS", "2009"), filename = "hsls_16_student_v1_0.sav", verbose = FALSE, forceReread = forceCacheUpdate))

  expect_is(hsls, "edsurvey.data.frame")
  expect_equal(dim(hsls), c(23503, 8510))

  expect_silent(hslsSchl <<- readHSLS(file.path(edsurveyHome, "HSLS", "2009"), filename = "hsls_09_school_v1_0.sav", wgtFilename = NULL, verbose = FALSE, forceReread = forceCacheUpdate))
  expect_is(hslsSchl, "edsurvey.data.frame")
  expect_equal(dim(hslsSchl), c(944, 688))
})

context("HSLS getData")
test_that("HSLS getData", {
  withr::with_options(list(digits = 4), dat1 <- getData(hsls, c("stu_id", "x1sex", "x1race")))
  expect_equal(dim(dat1), c(22496, 3))
  withr::with_options(list(digits = 7), co <- capture.output(summary(dat1)))
  expect_equal(co[-4], dat1Summary.Ref[-4]) # medians may agree and round differently, ignore them

  # test HSLS School variables
  withr::with_options(list(digits = 4), dat2 <- getData(hslsSchl, c("sch_id", "x1region", "x1locale"), dropOmittedLevels = FALSE))
  expect_equal(dim(dat2), c(944, 3))
  withr::with_options(list(digits = 7), co <- capture.output(summary(dat2)))
  expect_equal(co, dat2Summary.Ref)

  # test with some continuous variables that have dropOmittedLevels
  withr::with_options(list(digits = 4), dat3 <- getData(hsls, c("stu_id", "x1ses", "x2ses")))
  expect_equal(dim(dat3), c(18948, 3))
  withr::with_options(list(digits = 7), co <- capture.output(summary(dat3)))
  expect_equal(co[-4], dat3Summary.Ref[-4])

  # ensure that values having nearly all omitted levels and '0=Zero' have the '0=Zero' label dropped and return numeric accurately
  dat4 <- hsls$x3tcredeng
  expect_equal(class(dat4), "numeric")

  # levelsSDF test for labels having '=' in them
  tLvl <- levelsSDF("x1famincome", hsls)
  refLbl <- c(
    "MISSING", "UNIT NON-RESPONSE", "ITEM LEGITIMATE SKIP/NA",
    "FAMILY INCOME LESS THAN OR EQUAL TO $15,000", "FAMILY INCOME > $15,000 AND <= $35,000",
    "FAMILY INCOME > $35,000 AND <= $55,000", "FAMILY INCOME > $55,000 AND <= $75,000",
    "FAMILY INCOME > $75,000 AND <= $95,000", "FAMILY INCOME > $95,000 AND <= $115,000",
    "FAMILY INCOME > $115,000 AND <= $135,000", "FAMILY INCOME > $135,000 AND <= $155,000",
    "FAMILY INCOME > $155,000 AND <=$175,000", "FAMILY INCOME > $175,000 AND <= $195,000",
    "FAMILY INCOME > $195,000 AND <= $215,000", "FAMILY INCOME > $215,000 AND <= $235,000",
    "FAMILY INCOME > $235,000"
  )
  expect_equal(tLvl$x1famincome$labels, refLbl)
})

context("HSLS rename.sdf")
test_that("HSLS rename.sdf", {
  hslsRename <- rename.sdf(hsls,
    oldnames = "x1txmth",
    newnames = "mathThetaScore"
  )
  expect_equal("mathThetaScore" %in% colnames(hslsRename), TRUE)
})

context("HSLS weights")
test_that("HSLS weights", {
  hslsWgtNames <- names(hsls$weights)
  expect_equal(all(hslsWgtNames %in% wgtNames.Ref), TRUE)
})