tests/testthat/test-sparsebnData.R

In itsrainingdata/sparsebnUtils: Utilities for Learning Sparse Bayesian Networks

context("sparsebnData")

#
# THESE TESTS RELY ON A HELPER FILE
#

dat <- generate_fixed_data_frame()
dat_na <- generate_na_data_frame()
dat_disc <- generate_fixed_discrete_data_frame()

test_that("sparsebnData constructor fails if input not a list, data.frame, or matrix", {
    expect_error(sparsebnData(1L, type = "c"), "no applicable method for 'sparsebnData'")
    expect_error(sparsebnData(pi, type = "c"), "no applicable method for 'sparsebnData'")
    expect_error(sparsebnData(rep(1,5), type = "c"), "no applicable method for 'sparsebnData'")
})

test_that("sparsebnData constructor fails if type not specified or improperly specified", {
    expect_error(sparsebnData(x = dat), regexp = "type")
    expect_error(sparsebnData(x = dat, type = "contns"), regexp = "Invalid type")
    expect_error(sparsebnData(x = dat, type = "dicsrt"), regexp = "Invalid type")

    ### Check that partial matching works OK
    expect_error(sparsebnData(x = dat, type = "c"), NA)
    expect_error(sparsebnData(x = dat, type = "cont"), NA)
    expect_error(sparsebnData(x = dat, type = "d"), NA)
    expect_error(sparsebnData(x = dat, type = "disc"), NA)
})

test_that("sparsebnData constructor issues a warning if data has missing values, but does not fail", {
    expect_warning(sparsebnData(x = dat_na, type = "continuous"), regexp = "Data contains [0-9]+")
})

test_that("sparsebnData constructor correctly organizes discrete data", {
    sbd <- sparsebnData(dat_disc, type = "d")

    ### Check levels slot
    sbd_levels <- sbd$levels
    expect_equal(sbd_levels[[1]], c(0,1))
    expect_equal(sbd_levels[[2]], c(0,1))
    expect_equal(sbd_levels[[3]], c(0,1,2))

    ### Check ivn slot
    sbd_ivn <- sbd$ivn
    for(j in 1:ncol(sbd$data)){
        expect_true(is.null(sbd_ivn[[j]]))
    }

})

### Need to add more tests for 'ivn' slot (non-null, etc.)

test_that("print.sparsebnData functions properly", {
    expect_output(print(sparsebnData(x = dat, type = "continuous")), regexp = "5 total rows")
    expect_output(print(sparsebnData(x = dat, type = "continuous")), regexp = "Observational")
})