Testing is an important part of writing reusable code and having confidence when deployment production code. How should we write tests?

In many analytics applications, code is generally in the form of scripts, and there is rarely a need to completely “package” an application; especially when you are still prototyping code or in a “discovery” or “experiment” phase. In this post, I will look at how we can create simple templates in Python and R in order to be able to write and develop tests with confidence when we have simple scripts.

Within both languages we will write a simple add_func function which takes in two parameters and simply adds them together. With this function, we will also write a simple test to demonstrate how we can prepare ourselves for continuous integration.

Python

Using the Python language, I will make use of the nosetest package. Firstly create a folder structure similar to below:

.
|   sample.py
|
\---tests
        sample_tests.py

Where the contents of sample.py is:

def add_func(x,y):
    return x+y

sample_tests.py

from nose.tools import *

from sample import add_func

def test_add():
    assert_equal(add_func(1,2), 3)

Within nosetest:

  1. Test files go into tests/ and should be names *_tests.py.
  2. Within *_tests.py, all tests start with test_*

Running Python tests would be via:

nosetests

If the files are not picked up in *.nix then this might be because the tests are labeled as “executable”. To examine this, run:

nosetests -vv --collect-only

If you wish to force nose to consider executables then run:

nosetests --exe

R

Within R, I will make use of two libraries testthat and import. Import is used to “simulate” a script to a packge. The folder structure for R will be similar to Python:

.
|   sample.R
|
\---tests
        test_sample.R

Where the contents of sample.R is

add_func <- function(x, y) {
  #' Add function
  #'
  #' @param x numeric number to be added
  #' @param y numeric number to be added
  #'
  #' @examples
  #' add_func(1,2) # 3
  return(x+y)
}

and test_sample.R:

setwd("../")
import::from(sample.R, add_func)

test_that("simple add function works", {
  expect_equal(add_func(1,2), 3)
})

R running tests:

Rscript -e "library(testthat);test_dir(\"./tests\")"

In the code test_sample.R above, import is from the import library and setwd("../") will ensure that the relative directory location is picked up correctly from the import library.

Using testthat:

  • All tests should go into tests
  • All tests should have the pattern test_*.R