The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

R-CMD-check CRAN

Introducing patrick

This package is an extension to testthat that enables parameterized unit testing in R.

Installing

The release version of patrick is available on CRAN. Install it in the usual manner:

install.packages("patrick")

The development version of patrick is currently only available on GitHub. Install it using devtools.

devtools::install_github("google/patrick")

To use patrick as a testing tool within your package, add it to your list of Suggests and Remotes within your package’s DESCRIPTION.

Suggests:
    patrick

Use

Many packages within R employ the following pattern when writing tests:

test_that("Data is a successfully converted: numeric", {
  input <- convert(numeric_data)
  expect_type(input, "double")
})

test_that("Data is a successfully converted: character", {
  input <- convert(character_data)
  expect_type(input, "character")
})

While explicit, recycling a test pattern like this is prone to user error and other issues, as it is a violation of the classic DNRY rule (do not repeat yourself). patrick eliminates this problem by creating test parameters.

with_parameters_test_that("Data is successfully converted:", {
    input <- convert(test_data)
    expect_type(input, type)
  },
  test_data = list(numeric_data, character_data),
  type = c("double", "character"),
  .test_name = type
)

Parameterized tests behave exactly the same as standard testthat tests. Per usual, you call all of your tests with devtools::test, and they’ll also run during package checks. Each executes independently and then your test report will produce a single report. A complete name for each test will be formed using the initial test description and the strings in the .test_name parameter.

Small sets of cases can be reasonably passed as parameters to with_parameters_test_that. This becomes less readable when the number of cases increases. To help mitigate this issue, patrick provides a case generator helper function.

with_parameters_test_that("Data is successfully converted:", {
    input <- convert(test_data)
    expect_type(input, type)
  },
  cases(
    double = list(test_data = numeric_data, type = "double"),
    character = list(test_data = character_data, type = "character")
  )
)

More complicated testing cases can be constructed using data frames. This is usually best handled within a helper function and in a helper-<test>.R file.

make_cases <- function() {
  tibble::tribble(
    ~ .test_name, ~ expr,      ~ numeric_value,
    "sin",       sin(pi / 4),     1 / sqrt(2),
    "cos",       cos(pi / 4),     1 / sqrt(2),
    "tan",       tan(pi / 4),               1
  )
}

with_parameters_test_that(
  "trigonometric functions match identities",
  {
    testthat::expect_equal(expr, numeric_value)
  },
  .cases = make_cases()
)

If you don’t provide test names when generating cases, patrick will generate them automatically from the test data.

Inspiration

This package is inspired by parameterized testing packages in other languages, notably the parameterized library in Python.

Contributing

Please read the CONTRIBUTING.md for details on how to contribute to this project.

Disclaimer

This is not an officially supported Google product.

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.