Question

I have a long running simulation that I need to stop at a given time in the simulation and retrieve and observe some information and then allow the simulation to continue. I've recently started using a Test Driven Design method, unfortunately I have no idea how to unit test an application that drops into an interactive shell.

Here is the basic idea of what I am trying to do:

# peekaboo.py
from IPython import embed
from IPython.config.loader import Config

PAB_HEADER = 'Hello, my name is PAB. How may I help you?'
PAB_EXIT_MESSAGE = 'Goodbye, sir.'
PAB_PROMPT_IN_TEMPLATE = 'In [PAB \\#]: '
PAB_PROMPT_IN2_TEMPLATE = '   .\\D.: '
PAB_PROMPT_OUT_TEMPLATE = 'Out [PAB \\#]: '

def activate(**vars):
    """
    Activate PAB 0.1 by starting an interactive shell and putting
    the variables in the scope of the caller into the scope of this
    method.
    """
    # Add variables from caller to this scope
    locals().update(vars)
    cfg = None
    try:
        get_ipython
    except NameError:

        cfg = Config()
        prompt_config = cfg.PromptManager
        prompt_config.in_template = PAB_PROMPT_IN_TEMPLATE
        prompt_config.in2_template = PAB_PROMPT_IN2_TEMPLATE
        prompt_config.out_template = PAB_PROMPT_OUT_TEMPLATE

    embed(config=cfg, header=PAB_HEADER, exit_msg=PAB_EXIT_MESSAGE)

And here is an example of how it the peek_a_boo module might be used:

# long_running_app.py
import peek_a_boo
import datetime
import random

start_dt = datetime.datetime(2013,1,1)
datetimes = [start_dt + datetime.timedelta(days=i) for i in range(10)]
dt_of_interest = datetime.datetime(2013, 1, 8)
def long_running_process(dts):
    """
    simulate long running process
    """
    some_data = {}
    for dt in dts:
        some_data[dt] = random.random()
        if dt.date() == dt_of_interest.date():
            peek_a_boo.activate(**locals())

    return some_data

if __name__ == '__main__':
    data = long_running_process(datetimes)
    print data

My first inclination is to use mock and patch the embed method and verify that it has been called with the correct arguments, but I was wondering if anyone had other suggestions?

UPDATE:

So I am using nose for my unit test and I tried the following:

# test_peek_a_boo.py
import nose
import mock

class TestPeekABoo(object):
    def setup(self):
        pass

    def teardown(self):
        pass

    @mock.patch('IPython.embed')
    def test_activate(self, mock_embed):
        """
        Test that the activate method calls IPython embed with the correct arguments
        """
        import peek_a_boo
        a = 'Hello'
        b = 'World'
        peek_a_boo.activate(**locals())
        mock_embed.assert_called_once_with(header=peek_a_boo.PAB_HEADER, ...)

But when I run:

nosetests test_peek_a_boo.py

The process hangs. If I run:

nosetests test_peek_a_boo.py -s

I can see that I the process is dropping into the Interactive Shell.

UPDATE 2:

I was able to get the test above to run by importing peek_a_boo inside the test_method of the test class.

This tests that embed was actually call, but I'd like to be able to test that both a and b make it into the local scope of the activate method.

Was it helpful?

Solution

It seems like the solution that I came up with works, so I'll post it as a solution.

# test_peek_a_boo.py
import nose
import mock

class TestPeekABoo(object):
    def setup(self):
        pass

    def teardown(self):
        pass

    @mock.patch('IPython.embed')
    def test_activate(self, mock_embed):
        """
        Test that the activate method calls IPython embed with the correct arguments
        """
        import peek_a_boo
        a = 'Hello'
        b = 'World'
        peek_a_boo.activate(**locals())
        mock_embed.assert_called_once_with(header=peek_a_boo.PAB_HEADER, ...)
Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top