tests/testthat/testpreproc.R

In julianhatwell/arulesimp: Association Rules Based Method for Imputation

find_char <- function(pat, dt) {
  any(!(is.na(sapply(dt, match, x = pat))))
}

test_that("char_to_na returns correctly converted data"
          , { expect_false(
                find_char(""
                  , char_to_na(
                    data.frame(
                      x = c("foo", "")
                      , b = c("", "bar")))))
              expect_false(
                find_char("_"
                  , char_to_na(char_string = "_",
                    data.frame(
                      x = c("foo", "_")
                      , b = c("", "bar")))))
          })


test_that("all_factor returns only factors"
          , { expect_true(
            all(sapply(
              all_factor(
                data.frame(
                  x = c("foo", "bar")
                  , y = c(10, 20)))
              , class) == "factor")
          )
          })

test_that("all_factor with ord = TRUE returns all ordered factors"
          , { expect_true(
            all(sapply(
              all_factor(
                data.frame(
                  x = c(1, 2)
                  , y = c(10, 20))
                , ord = TRUE)
              , class) == c("ordered", "factor"))
          )
          }) # only works on numbers

test_that("all_factor with ord = c(TRUE, FALSE) returns correct ordered factors"
          , { expect_identical(
            sapply(
              all_factor(
                data.frame(
                  x = c(1, 2)
                  , y = c(10, 20))
                , ord = c(TRUE, FALSE))
              , class)
            , list(x = c("ordered", "factor")
                , y = "factor")
              )
          }) # only works on numbers


test_that("missing_values finds missing values"
          , { expect_equal(2
            , sum(
              missing_values(
              data.frame(
                x = c("foo", NA)
                , y = c(NA, "bar")))))
              expect_equal(3
              , sum(
                missing_values(
                data.frame(
                  x = c("foo", NA)
                  , y = c(NA, NA)))))
          })

data("responses")
resp <- char_to_na(responses)
resp <- all_factor(resp)
mv_sorted <- missing_values(resp)
resp <- as(resp, "transactions")

rules <- arules::apriori(resp
                         , parameter = list(
                           support = 0.5))
test_that("make_cars finds rules for each target var"
          , { expect_identical(names(make_cars(rules, names(mv_sorted)))
                             , names(mv_sorted))
          })