tests/testthat/test-collapse_cart.R

In embed: Extra Recipes for Encoding Predictors

test_that("collapsing factors", {
  skip_if_not_installed("modeldata")
  data(ames, package = "modeldata")
  ames$Sale_Price <- log10(ames$Sale_Price)

  expect_error(
    {
      rec_1 <-
        recipe(Sale_Price ~ ., data = ames) %>%
        step_collapse_cart(
          Neighborhood, Central_Air,
          outcome = vars(Sale_Price)
        ) %>%
        prep()
    },
    regex = NA
  )

  expect_true(length(rec_1$steps[[1]]$results) == 1)
  expect_equal(names(rec_1$steps[[1]]$results), "Neighborhood")

  expect_true(
    length(unique(rec_1$steps[[1]]$results$Neighborhood$.group)) <
      length(levels(ames$Neighborhood))
  )

  expect_equal(
    ames %>% select(-Neighborhood, -Sale_Price),
    bake(rec_1, new_data = NULL) %>% select(-Neighborhood, -Sale_Price)
  )

  expect_false(
    isTRUE(
      all.equal(bake(rec_1, new_data = NULL)$Neighborhood, ames$Neighborhood)
    )
  )

  expect_error(
    {
      rec_2 <-
        recipe(Sale_Price ~ ., data = ames) %>%
        step_collapse_cart(Neighborhood, Central_Air,
          outcome = vars(Sale_Price),
          min_n = 100, cost_complexity = 0.1
        ) %>%
        prep()
    },
    regex = NA
  )

  expect_true(
    length(levels(rec_2$steps[[1]]$results$Neighborhood$.group)) <
      length(levels(rec_1$steps[[1]]$results$Neighborhood$.group))
  )
})

test_that("failed collapsing", {
  skip_if_not_installed("modeldata")
  data(ames, package = "modeldata")

  # model fails
  ames$Sale_Price2 <- Inf
  expect_error(
    {
      rec_3 <-
        recipe(Sale_Price2 ~ ., data = ames) %>%
        step_collapse_cart(
          Neighborhood, Central_Air,
          outcome = vars(Sale_Price2)
        ) %>%
        prep()
    },
    regex = NA
  )

  expect_true(length(rec_3$steps[[1]]$results) == 0)

  # too many splits
  expect_error(
    {
      rec_4 <-
        recipe(Sale_Price ~ ., data = ames) %>%
        step_collapse_cart(
          Neighborhood,
          outcome = vars(Sale_Price),
          cost_complexity = 0, min_n = 1
        ) %>%
        prep()
    },
    regex = NA
  )

  expect_true(length(rec_4$steps[[1]]$results) == 0)

  # too many splits
  expect_error(
    {
      rec_5 <-
        recipe(Sale_Price ~ ., data = ames) %>%
        step_collapse_cart(Central_Air, outcome = vars(Sale_Price)) %>%
        prep()
    },
    regex = NA
  )

  expect_true(length(rec_5$steps[[1]]$results) == 0)
})

# Infrastructure ---------------------------------------------------------------

test_that("bake method errors when needed non-standard role columns are missing", {
  skip_if_not_installed("modeldata")
  data(ames, package = "modeldata")
  
  rec <- recipe(Sale_Price ~ ., data = ames) %>%
    step_collapse_cart(MS_SubClass, outcome = vars(Sale_Price)) %>%
    update_role(MS_SubClass, new_role = "potato") %>%
    update_role_requirements(role = "potato", bake = FALSE)
  
  rec_trained <- prep(rec, training = ames, verbose = FALSE)
  
  expect_error(
    bake(rec_trained, new_data = ames[, -1]),
    class = "new_data_missing_column"
  )
})

test_that("empty printing", {
  rec <- recipe(mpg ~ ., mtcars)
  rec <- step_collapse_cart(rec)
  
  expect_snapshot(rec)
  
  rec <- prep(rec, mtcars)
  
  expect_snapshot(rec)
})

test_that("empty selection prep/bake is a no-op", {
  rec1 <- recipe(mpg ~ ., mtcars)
  rec2 <- step_collapse_cart(rec1)
  
  rec1 <- prep(rec1, mtcars)
  rec2 <- prep(rec2, mtcars)
  
  baked1 <- bake(rec1, mtcars)
  baked2 <- bake(rec2, mtcars)
  
  expect_identical(baked1, baked2)
})

test_that("empty selection tidy method works", {
  rec <- recipe(mpg ~ ., mtcars)
  rec <- step_collapse_cart(rec)
  
  expect <- tibble(terms = character(), value = double(), id = character())
  
  expect_identical(tidy(rec, number = 1), expect)
  
  rec <- prep(rec, mtcars)
  
  expect_identical(tidy(rec, number = 1), expect)
})

test_that("printing", {
  skip_if_not_installed("modeldata")
  data(ames, package = "modeldata")
  
  rec <- recipe(Sale_Price ~ ., data = ames) %>%
    step_collapse_cart(Neighborhood, Central_Air, outcome = vars(Sale_Price))
  
  expect_snapshot(print(rec))
  expect_snapshot(prep(rec))
})