How can I stop the python process from terminating after a failed unit test so that I can call pdb.pm

Question

This is a bit of a newbie question about basic python workflow. I have a library of python files, and a collection of associated unit tests. One of the tests is failing and I'd like to inject pdb at the spot of the failure.

My workflow uses elpy and emacs, which I otherwise quite like and would prefer to stick with. I load the test file in emacs, start a python shell, and then run the test code in the python shell. Something throws an exception, which causes the test to fail. The python process ends and prints a stack trace.

If the process were still alive I'd call pdb.pm() to start the debugger at the source of the error so that I could investigate the stack interactively, and that is what I'm trying to achieve. The basic structure of the unit test file is

import unittest
import other_stuff

class MyTest(unittest.TestCase):
    def test_other_stuff(self):
        self.assertTrue(other_stuff.function_call() == expected_result)

    def test_other_other_stuff(self):
        pass # you get the idea

if __name__ == "__main__":
    unittest.main(verbosity=2)