This documentation is for an unreleased version of MPF!

This is the developer documentation for MPF 0.54, which is the “dev” (next) release of MPF that is a work-in-progress. Use the “Read the Docs” link in the lower left corner to view the developer docs for the version of MPF you’re using.

Testing Tutorial Step 2. Create your machine folder

Step 2 of the MPF Tutorial is where you create your machine folder and get MPF up and running with an empty config. Since it ends with MPF running and the attract mode being active, we can actually write a test for it!

Here are the steps to take:

1. Create a “tests” folder in your machine folder

First, create a folder called tests in your machine folder. This would be alongside the other folders in there, which will be “config” (created in the MPF tutorial), as well as “logs” and “data” which were created automatically by MPF the first time it ran.

2. Add an empty “__init__.py” file

Next, inside your new tests folder, create a blank file called __init__.py. (That’s two underscores, then the word “init”, then two more underscores, then “.py”.) This file should be totally blank. (It just needs to exist.) This file is needed to let the Python test runner find and load the tests from this folder.

3. Add a test file

Next you need to add a Python file which actually holds your tests. You can name this file whatever you want as long as it starts with “test”. (The reason for starting it with “test” is also so that the Python test runner knows that this file contains tests, allowing it to automatically find and run tests from it.)

For now let’s call it test_step_2.py.

Open that file and add the following lines to it: (If you are interested in what all this means, then read on below the file. Otherwise you can skip down to Step 4.)

So what’s this file actually doing?

The import line just imports the base class we use for MPF machine tests. (More details on that is covered in the Testing Class API page).

Our specific class name TestTutorialMachine can be whatever you want. Again just make sure it starts with “Test” in order for the test runner to find out.

Our specific method is called test_step_2_mpf_startup(). (Also it has to start with “test”). When the tests are run each method represents a separate “run” of MPF. The test runner will start up MPF and get it all up and running, and then it will move through the code in the test method, then it will cleanly shut down MPF when it’s done. If there are multiple test methods, then the test running will start and stop MPF multiple times. The key is that each test method is run against a “fresh” MPF copy.

These test methods will also load the machine config files (just like if the command mpf was run the regular way).

Anyway, in our test method, we have the only actual line that does anything:


This just tests (“asserts”) that a mode called “attract” is running. There are all sorts of MPF-specific assertion methods which we’ll cover in later steps of this tutorial.

4. Run your test

You can run your tests via the command prompt from your machine folder. (In other words, the same place where you run mpf to run your machine.)

The exact command to run is python -m unittest. This should produce output similar to the following:

C:\pinball\your_machine>python -m unittest
C:\Python34\lib\imp.py:32: PendingDeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
Ran 1 test in 0.734s



That warning about the deprecation can be ignored (if you even have it.. you might not). The important thing is the message towards the bottom: “Ran 1 test in 0.734s” and the “OK” below it. That means your test passed!

5. Check out a failed test

When you’re writing unit tests, you’ll end up dealing with failed tests a lot! So let’s purposefully change the test so it fails. In this case, change the line which asserts a mode called “attract” is running to look for a mode called “foo” instead, like this:


Save the file and rerun the tests and you should see results like this:

C:\pinball\your_machine>python -m unittest
C:\Python34\lib\imp.py:32: PendingDeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
FAIL: test_mpf_starts (tests.test_step_2.TestTutorialMachine)
Tests Step 2 of the tutorial
Traceback (most recent call last):
  File "C:\pinball\your_machine\tests\test_step_2.py", line 18, in test_mpf_starts
  File "C:\Python34\lib\site-packages\mpf\tests\MpfTestCase.py", line 576, in assertModeRunning
    raise AssertionError("Mode {} not known.".format(mode_name))
AssertionError: Mode foo not known.

Ran 1 test in 0.594s

FAILED (failures=1)


Note that we see the test run failed, with one failure, and that we can scroll up and see the specific name of the test that failed along with the line that failed, and information about the failure. (In this case it tells us that the mode “foo” is not known.)

So to get this test to work, you either need to change your MPF config to start a mode called “foo”, or you need to change the test back to looking for a mode called “attract”. :)

What if it didn’t work?

If the unit tests don’t work for you, there are a few things you can try.

If you get some kind of loading error or config error, make sure you’re running python -m unittest from your machine folder (not from the “tests” folder).

If you get a message about 0 tests run, make sure you have that empty __init__.py in your tests folder.

And if you get some weird error that you can’t figure out, then post a message to the MPF Google Group.