Fork me on GitHub

Overview

Teaching: 10 min
Exercises: 0 min
Questions
  • What tools are out there?
Objectives
  • Discover the tool that fits your needs.

Unit test frameworks

A test framework makes it easy to run tests across large amounts of code automatically. They provide more control than one single script which does some tests.


pytest

Very easy to set up: Anaconda, Miniconda, virtualenv, or pipenv.

Very easy to use: Prefix a function with “test_” and the test runner will execute it. No need to subclass anything.

def get_word_lengths(s):
    """
    Returns a list of integers representing
    the word lengths in string s.
    """
    return [len(w) for w in s.split()]

def test_get_word_lengths():
    text = "Three tomatoes are walking down the street"
    assert get_word_lengths(text) == [5, 8, 3, 7, 4, 3, 6]

testthat

  • Easily installed from CRAN with install.packages("testthat"), or from GitHub with devtools::install_github("r-lib/testthat")
  • Use in package development with usethis::use_testthat()
  • Add a new test file with usethis::use_test("test-name"), e.g.:

    # tests/testthat/test_example.R
    # file added by running `usethis::use_test("example")`
    
    context("Arithmetics")
    library("mypackage")
    
    test_that("square root function works", {
      expect_equal(my_sqrt(4), 2)
      expect_warning(my_sqrt(-4))
    })
    

    Tests consist of one or more expectations, and multiple tests can be grouped together in one test file. Test files are put in the directory tests/testthat/, and their file names are prefixed with test_.

  • Run all tests in package with devtools::test() (if you use RStudio, press Ctrl+Shift+T):

    > devtools::test()
    Loading mypackage
    Testing mypackage
    ✔ |  OK F W S | Context
    ✔ |   2       | Arithmetics
    
    ══ Results ═════════════════════════════════════════════════════════════════════
    OK:       2
    Failed:   0
    Warnings: 0
    Skipped:  0
    

More information in the Testing chapter of the book R Packages by Hadley Wickham.


Google Test

  • Widely used
  • Very rich in functionality
#include "gtest/gtest.h"
#include "example.h"

TEST(example, add)
{
    double res;
    res = add_numbers(1.0, 2.0);
    ASSERT_NEAR(res, 3.0, 1.0e-11);
}

pFUnit

  • Very rich in functionality
  • Requires modern Fortran compilers (uses F2003 standard)
@test
subroutine test_add_numbers()

   use hello
   use pfunit_mod

   implicit none

   real(8) :: res

   call add_numbers(1.0d0, 2.0d0, res)
   @assertEqual(res, 3.0d0)

end subroutine

Services to deploy testing and coverage

Key points

  • Python and C/C++ have better tooling for automated tests and you can use those also for Fortran projects (via iso_c_binding).