Python for Finance: Analyze Big Financial Data

”’ Tests if function f calculates correctly. ”’
nt.assert_equal(f(4.), 4.)
nt.assert_equal(f( 1000 ), 1000 )
nt.assert_equal(f(5000.5), 5000.5)

def test_f_type_error():
”’ Tests if type error is raised. ”’
nt.assert_raises(TypeError, f, ‘test string’)
nt.assert_raises(TypeError, f, [ 3 , ‘string’])

def test_f_value_error():
”’ Tests if value error is raised. ”’
nt.assert_raises(ValueError, f, - 1 )
nt.assert_raises(ValueError, f, -2.5)

def test_f_test_fails():
”’ Tests if function test fails. ”’
nt.assert_equal(f(5.), 10 )

Table A-2 describes the test functions that are implemented.

Table A-2. Test functions for simple function f

Function Description

test_f_calculation

Tests if the function generates correct results

test_f_type_error

Checks if the function raises a type error when expected

test_f_value_error

Checks if the function raises a value error when expected

test_f_test_fails

Tests if the calculation test fails as expected (for illustration)

From the command line/shell, you can run the following tests:

$   nosetests   nose_test.py

...F

======================================================================

FAIL:   Test    if  function    test    fails.

–––––––––––––––––––––––-

Traceback   (most   recent  call    last):

            File    “/Library/anaconda/lib/python2.7/site-packages/nose/case.py”,

            line    197,    in  runTest self.test(*self.arg)

            File    “//Users/yhilpisch/Documents/Work/Python4Finance/python/nose_test.py”,

            line    30, in  test_f_test_fails

                    nt.assert_equal(f(5.),  10)

AssertionError: 5.000000000000001   !=  10

–––––––––––––––––––––––-

Ran 4   tests   in  0.002s

FAILED  (failures=1)

$

Obviously, the first three tests are successful, while the last one fails as expected. Using

such tools — and more importantly, implementing a rigorous approach to unit testing —

may require more effort up front, but you and those working with your code will benefit in

the long run.