tests/testthat/test-prepare_df_plot.R

In seabbs/getTBinR: Access and Summarise World Health Organization Tuberculosis Data

context("prepare_df_plot")

## No country
df_all <- prepare_df_plot(download_data = TRUE, save = TRUE)

## Supply data
tb_data <- get_tb_burden()
df_supp_all <- prepare_df_plot(tb_data)

## Single country exact match
df_exact_country <- prepare_df_plot(countries = "United Kingdom of Great Britain and Northern Ireland")
df_exact_country$df$country <- factor("United Kingdom")

## Single country fuzzy matching
df_fuzzy_country <- prepare_df_plot(countries = "United Kingdom")

## Compare to region
df_region <- prepare_df_plot(
  countries = "United Kingdom",
  compare_to_region = TRUE
)

## Add country facet
df_country_facet <- prepare_df_plot(facet = "g_whoregion")


eur_countries <- c(
  "Albania", "Andorra",
  "Armenia", "Austria",
  "Azerbaijan", "Belarus",
  "Belgium", "Bosnia and Herzegovina",
  "Bulgaria", "Croatia"
)

test_that("TB burden data is correctly processed ready for plotting", {
  skip_on_cran()
  expect_equal(3, length(df_all))
  expect_null(df_all$facet)
  expect_equal(
    "Estimated incidence (all forms) per 100 000 population",
    df_all$metric_label
  )
  expect_equal("tbl_df", class(df_all$df)[1])
  expect_true(1 <= nrow(df_all$df))
  expect_true(1 <= ncol(df_all$df))
  expect_equal(0, sum(is.na(df_all$df$country)))
  expect_true(!any(duplicated(colnames(df_all$df))))
})

test_that("TB burden data is correctly proccessed when using local data", {
  expect_equal(df_all, df_supp_all)
})


test_that("TB burden data is correctly proccessed when using a single country", {
  expect_true(1 <= length(df_exact_country))
  expect_null(df_exact_country$facet)
  expect_equal(
    "Estimated incidence (all forms) per 100 000 population",
    df_exact_country$metric_label
  )
  expect_equal("tbl_df", class(df_exact_country$df)[1])
  expect_true(1 <= nrow(df_exact_country$df))
  expect_true(1 <= ncol(df_exact_country$df))
  expect_equal(
    "United Kingdom",
    as.character(unique(df_exact_country$df$country))
  )
})

test_that("Fuzzy country matching correctly selects a country", {
  expect_equal(df_exact_country, df_fuzzy_country)
})


test_that("TB burden data is correctly proccessed when comparing to the region", {
  expect_true(1 <= length(df_region))
  expect_equal("g_whoregion", df_region$facet)
  expect_equal(
    "Estimated incidence (all forms) per 100 000 population",
    df_region$metric_label
  )
  expect_equal("tbl_df", class(df_region$df)[1])
  expect_true(1 <= nrow(df_region$df))
  expect_true(1 <= ncol(df_region$df))
  expect_equal(
    eur_countries,
    as.character(unique(df_region$df$country))[1:10]
  )
})

test_that("Add a facet works correctly", {
  expect_equal("g_whoregion", df_country_facet$facet)
})

test_that("Adding a transform label to the metric label works correctly", {
  df_lab <- prepare_df_plot(metric_label = "test", trans = "suffix")

  exp_lab <- "test (suffix)"

  expect_equal(exp_lab, df_lab$metric_label)
})

## Tests using dummy data
test_df <- tibble::tibble(
  country = "test", year = 2000:2001,
  e_inc_100k = c(10, 12), e_inc_100k_lo = c(8, 10),
  e_inc_100k_hi = c(20, 22)
)

result_df <- test_df
result_df$country <- factor(result_df$country)
result_df$`Estimated incidence (all forms) per 100 000 population` <- result_df$e_inc_100k
result_df$Year <- result_df$year

test_that("prepare_df_plot correctly formats the input data.", {
  skip_on_cran()
  expect_equal(result_df, prepare_df_plot(
    metric = "e_inc_100k",
    conf = c("_lo", "_hi"),
    df = test_df,
    annual_change = FALSE
  )$df)
})



test_that("annual_change correctly transforms metric and confidence intervals", {
  result_df <- result_df[-1, ]
  result_df <- result_df[, (1:(ncol(result_df) - 2))]
  result_df$e_inc_100k <- 0.2
  result_df$e_inc_100k_lo <- 0.25
  result_df$e_inc_100k_hi <- 0.1
  result_df$`Percentage annual change: Estimated incidence (all forms) per 100 000 population` <- 0.2
  result_df$Year <- result_df$year

  df_annual_change <- prepare_df_plot(
    metric = "e_inc_100k", conf = c("_lo", "_hi"),
    df = test_df, annual_change = TRUE
  )
  skip_on_cran()
  expect_equal(result_df, prepare_df_plot(
    metric = "e_inc_100k",
    conf = c("_lo", "_hi"),
    df = test_df,
    annual_change = TRUE
  )$df)
})

test_that("years are correctly filtered for.", {
  result_df <- dplyr::filter(result_df, year == 2000)

  skip_on_cran()
  expect_equal(result_df, prepare_df_plot(
    metric = "e_inc_100k",
    conf = c("_lo", "_hi"),
    years = 2000,
    df = test_df,
    annual_change = FALSE
  )$df)
})

test_that("Using a metric that does not occur in the dataset and 
          no metric label leads to the metric being used as the default label.", {
  skip_on_cran()
  test_df <- tb_data
  test_df$test_metric <- test_df$e_inc_100k

  expect_equal("test_metric", prepare_df_plot(
    metric = "test_metric",
    years = 2000,
    df = test_df,
    annual_change = FALSE
  )$metric_label)
})