Unit Testing

As a final best practice, we want to consider unit testing. Among the different testing

approaches, unit testing can indeed be considered a best practice because it tests Python

code on a rather fundamental level — i.e., the single units. What it does not test, however,

is the integration of the single units. Typically, such units are functions, classes, or

methods of classes. As a pretty simple example of a Python function that is also easily

testable, consider the one in Example A-5.

Example A-5. A rather simple Python function

Simple function to calculate

the square of the square root

of a positive number

simple_function.py

from math import sqrt

def f(x):
”’ Function to calculate the square of the square root.

            Parameters

            ==========

            x   :   float   or  int

                            input   number



            Returns

            =======

            fx  :   float

                            square  of  the square  root,   i.e.    sqrt(x) **  2



            Raises

            ======

            TypeError

                            if  argument    is  neither float   nor integer

            ValueError

                            if  argument    is  negative



            Examples

            ========

            >>> f(1)

            1

            >>> f(10.5)

            10.5

            ”’

if type(x) != float and type(x) != int:
raise TypeError(“Input has not the right type.”)
if x < 0 :
raise ValueError(“Number negative.”)
fx = sqrt(x) ** 2
return fx

There are many tools available that help support unit tests. We will make use of nose in

what follows. Example A-6 contains a small test suite for the simple function f from

Example A-5.

Example A-6. A test suite for the function f

Test suite for simple function f

nose_test.py

import nose.tools as nt
from simple_function import f

def test_f_calculation():