tests/testthat/test-gender_fixer.R

In JayAchar/tbcleanr: tbcleanr: tools for cleaning TB datade

context("test-gender_fixer")
library(tbcleanr)

# Load data
        k6_raw <- system.file("testdata", "gender_k6.rds", package = "tbcleanr") %>% 
                        readRDS()

# generate correct output        
        k6_output <- gender_fixer(k6_raw)

test_that("k6 correct", {
  expect_true(all(names(k6_output) %in% c("registrationnb", "gender")))
  expect_equal(nrow(k6_raw), nrow(k6_output))
  expect_true(length(unique(k6_raw$registrationnb)) == 
                  length(unique(k6_output$registrationnb)))
  expect_true(is.factor(k6_output$gender))
  expect_true(is.character(k6_output$registrationnb))
  expect_equal(levels(k6_output$gender), c("Male", "Female"))
  expect_true(all(k6_output$gender[k6_raw$gender == 1] == "Male"))
  expect_true(all(k6_output$gender[k6_raw$gender == 2] == "Female"))
})

# generate raw data with NA
        k6_NA <- rbind(k6_raw, 
                       data.frame(registrationnb = as.character(11), 
                                  gender = NA_integer_, 
                                  stringsAsFactors = F))
# generate NA output
        k6_na_output <- gender_fixer(k6_NA)

test_that("k6 NA correct", {
        expect_equal(nrow(k6_NA), nrow(k6_na_output))
        expect_equal(levels(k6_na_output$gender), c("Male", "Female"))
        expect_true(is.na(k6_na_output$gender[which(is.na(k6_NA$gender))]))
})


# generate raw data with additional gender level
        k6_abn_level <- rbind(k6_raw, 
                       data.frame(registrationnb = as.character(11), 
                                  gender = as.numeric(3), 
                                  stringsAsFactors = F))
# test abnormal gender levels
test_that("k6 abnormal gender", {
    expect_message(gender_fixer(k6_abn_level), 
                     "Unrecognised gender levels detected - coerced to NA")
    expect_equal(gender_fixer(k6_abn_level) %>% 
                     is.na(.) %>%
                     sum(), 1)
})


## Epiinfo
# Load data
epi_raw <- system.file("testdata", "gender_epi.rds", package = "tbcleanr") %>% 
        readRDS()

# generate correct output        
epi_output <- gender_fixer(epi_raw)

test_that("epiinfo correct", {
        expect_true(all(names(epi_output) %in% c("APID", "SEX")))
        expect_equal(nrow(epi_raw), nrow(epi_output))
        expect_true(length(unique(epi_raw$APID)) == length(unique(epi_output$APID)))
        expect_true(is.factor(epi_output$SEX))
        expect_true(is.character(epi_output$APID))
        expect_equal(levels(epi_output$SEX), c("Male", "Female"))
        expect_true(all(epi_output$SEX[epi_raw$SEX == "M"] == "Male"))
        expect_true(all(epi_output$SEX[epi_raw$SEX == "F"] == "Female"))
})


# generate raw data with NA
epi_NA <- rbind(epi_raw, 
               data.frame(APID = as.character(11), 
                          SEX = NA_character_, 
                          stringsAsFactors = F))
# generate NA output
epi_na_output <- gender_fixer(epi_NA)

test_that("epi NA correct", {
        expect_equal(nrow(epi_NA), nrow(epi_na_output))
        expect_equal(levels(epi_output$SEX), c("Male", "Female"))
        expect_true(is.na(epi_na_output$SEX[which(is.na(epi_NA$SEX))]))
})

# generate raw data with additional gender level
epi_abn_level <- rbind(epi_raw, 
                      data.frame(APID = as.character(11), 
                                 SEX = as.character("N"), 
                                 stringsAsFactors = F))
# test abnormal gender levels
test_that("epi abnormal gender", {
        expect_message(gender_fixer(epi_abn_level), 
                     "Unrecognised gender levels detected - coerced to NA")
        expect_error(gender_fixer(c("X")))
})