Automating testing of functions that use files

To test a function like calculate_VAT from the Process prices from a file and calculate the VAT exercise, which generates a file, you have to open the file it has generated to check that it has the content you expected. For example, if you used the file data1.txt as input, you would expect a file with name data1_VAT.txt that has the content as shown in the example above.

Doing that manually is tedious. Imagine doing it for 10 data files, from data1.txt to data10.txt. We would have to manually generate the 10 files from data1_VAT.txt to data10_VAT.txt, with calculate_VAT, and then open them one by one to check their content to see if it is what we expect.

We can automate it with pytest. However, the output of the function is not just a value as we have seen so far. The output of the calculate_VAT function is a file. So the expected output can also be a file, and the test simply compares that what comes out is equal to what we expect!

For that, we must create 10 files only once where we define the expected outputs of our tests: from to. Then we create a pytest like the one below:

@pytest.mark.parametrize("testcase, f_input, f_expected_output",[ (1, "data1.txt", "expected_output_data1_VAT.txt"), (2, "data2.txt", "expected_output_data2_VAT.txt"), (3, "data3.txt", "expected_output_data3_VAT.txt"), (4, "data4.txt", "expected_output_data4_VAT.txt"), (5, "data5.txt", "expected_output_data5_VAT.txt"), (6, "data6.txt", "expected_output_data6_VAT.txt"), (7, "data7.txt", "expected_output_data7_VAT.txt"), (8, "data8.txt", "expected_output_data8_VAT.txt"), (9, "data9.txt", "expected_output_data9_VAT.txt"), (10, "data10.txt", "expected_output_data10_VAT.txt"), ])

def test_calculate_VAT(testcase, f_input, f_expected_output):
    f_output = calculate_VAT(f_input)
    assert (open(f_output).read() == open(f_expected_output).read()), "case 0".format(testcase)

NOTE: To compare the files we have used read because the size of the files that we are handling here is very small.