tests/testthat/test-acc_multivariate_outlier.R

In dataquieR: Data Quality in Epidemiological Research

test_that("acc_multivariate_outlier works with 3 args", {
  skip_on_cran() # slow and tested elsewhere
  skip_if_translated()
  meta_data <- prep_get_data_frame("meta_data")
  study_data <- prep_get_data_frame("study_data")
  expect_error(
    res1 <-
      acc_multivariate_outlier(
        study_data = study_data, meta_data = meta_data),
    regexp =
      "Argument variable_group is NULL",
    perl = TRUE
  )

  expect_error(
    res1 <-
      acc_multivariate_outlier(
        variable_group = "v00014",
        study_data = study_data, meta_data = meta_data),
    regexp =
      "Need at least two variables for multivariate outliers.",
    perl = TRUE
  )

  expect_message(
    res1 <-
      acc_multivariate_outlier(variable_group = c("v00014", "v00006"),
                         study_data = study_data, meta_data = meta_data),
    regexp =
      sprintf(
        "(%s|%s)",
        paste("As no ID-var has been specified the rownumbers will be used."),
        paste("Due to missing values in v00014, v00006 or dq_id N=602",
              "observations were excluded.")
      ),
    perl = TRUE,
    all = TRUE
  )

  expect_true(all(c("FlaggedStudyData",
                    "SummaryTable",
                    "SummaryPlot") %in% names(res1)))
  expect_lt(
    suppressWarnings(abs(sum(as.numeric(
      as.matrix(res1$FlaggedStudyData)),
      na.rm = TRUE) - 4492191)), 0.5
  )
  expect_equal(
    suppressWarnings(abs(sum(as.numeric(
      as.matrix(res1$SummaryTable)),
      na.rm = TRUE))), 182
  )
})

test_that("acc_multivariate_outlier works with label_col", {
  skip_on_cran() # slow and tested elsewhere
  meta_data <- prep_get_data_frame("meta_data")
  study_data <- prep_get_data_frame("study_data")
  expect_error(
    res1 <-
      acc_multivariate_outlier(
        label_col = LABEL,
        study_data = study_data, meta_data = meta_data),
    regexp =
      "Argument variable_group is NULL",
    perl = TRUE
  )

  expect_error(
    res1 <-
      acc_multivariate_outlier(
        label_col = LABEL,
        variable_group = "CRP_0",
        study_data = study_data, meta_data = meta_data),
    regexp =
      "Need at least two variables for multivariate outliers.",
    perl = TRUE
  )

  expect_message(
    res1 <-
      acc_multivariate_outlier(variable_group = c("CRP_0", "GLOBAL_HEALTH_VAS_0"),
                               label_col = LABEL,
                               study_data = study_data, meta_data = meta_data),
    regexp =
      sprintf(
        "(%s|%s)",
        paste("As no ID-var has been specified the rownumbers will be used."),
        paste("Due to missing values in CRP_0, GLOBAL_HEALTH_VAS_0 or",
              "dq_id N=602 observations were excluded.")
      ),
    perl = TRUE,
    all = TRUE
  )

  expect_true(all(c("FlaggedStudyData",
                    "SummaryTable",
                    "SummaryPlot") %in% names(res1)))
  expect_lt(
    suppressWarnings(abs(sum(as.numeric(
      as.matrix(res1$FlaggedStudyData)),
      na.rm = TRUE) - 4492191)), 0.5
  )
  expect_equal(
    suppressWarnings(abs(sum(as.numeric(
      as.matrix(res1$SummaryTable)),
      na.rm = TRUE))), 182
  )

  skip_on_cran()
  skip_if_not_installed("vdiffr")
  skip_if_not(capabilities()["long.double"])
  vdiffr::expect_doppelganger(
    "acc_mv_outlierCRP0GLOBHEAVA0",
                              res1$SummaryPlot)
})

test_that("criteria is robust", {

})