Running a single test from unittest.TestCase via the command line

It can work well as you guess

python testMyCase.py MyCase.testItIsHot

And there is another way to just test testItIsHot:

    suite = unittest.TestSuite()
    suite.addTest(MyCase("testItIsHot"))
    runner = unittest.TextTestRunner()
    runner.run(suite)

If you organize your test cases, that is, follow the same organization like the actual code and also use relative imports for modules in the same package, you can also use the following command format:

python -m unittest mypkg.tests.test_module.TestClass.test_method

# In your case, this would be:
python -m unittest testMyCase.MyCase.testItIsHot

Python 3 documentation for this: Command-Line Interface

If you check out the help of the unittest module it tells you about several combinations that allow you to run test case classes from a module and test methods from a test case class.

python3 -m unittest -h

[...]

Examples:
  python3 -m unittest test_module               - run tests from test_module
  python3 -m unittest module.TestClass          - run tests from module.TestClass
  python3 -m unittest module.Class.test_method  - run specified test method
```lang-none

It does not require you to define a `unittest.main()` as the default behaviour of your module.

Inspired by yarkee, I combined it with some of the code I already got. You can also call this from another script, just by calling the function run_unit_tests() without requiring to use the command line, or just call it from the command line with python3 my_test_file.py.

import my_test_file
my_test_file.run_unit_tests()

Sadly this only works for Python 3.3 or above:

import unittest

class LineBalancingUnitTests(unittest.TestCase):

    @classmethod
    def setUp(self):
        self.maxDiff = None

    def test_it_is_sunny(self):
        self.assertTrue("a" == "a")

    def test_it_is_hot(self):
        self.assertTrue("a" != "b")

Runner code:

#! /usr/bin/env python3
# -*- coding: utf-8 -*-
import unittest
from .somewhere import LineBalancingUnitTests

def create_suite(classes, unit_tests_to_run):
    suite = unittest.TestSuite()
    unit_tests_to_run_count = len( unit_tests_to_run )

    for _class in classes:
        _object = _class()
        for function_name in dir( _object ):
            if function_name.lower().startswith( "test" ):
                if unit_tests_to_run_count > 0 
                        and function_name not in unit_tests_to_run:
                    continue
                suite.addTest( _class( function_name ) )
    return suite

def run_unit_tests():
    runner = unittest.TextTestRunner()
    classes =  [
        LineBalancingUnitTests,
    ]

    # Comment all the tests names on this list, to run all Unit Tests
    unit_tests_to_run =  [
        "test_it_is_sunny",
        # "test_it_is_hot",
    ]
    runner.run( create_suite( classes, unit_tests_to_run ) )

if __name__ == "__main__":
    print( "nn" )
    run_unit_tests()

Editing the code a little, you can pass an array with all unit tests you would like to call:

...
def run_unit_tests(unit_tests_to_run):
    runner = unittest.TextTestRunner()

    classes = 
    [
        LineBalancingUnitTests,
    ]

    runner.run( suite( classes, unit_tests_to_run ) )
...

And another file:

import my_test_file

# Comment all the tests names on this list, to run all unit tests
unit_tests_to_run = 
[
    "test_it_is_sunny",
    # "test_it_is_hot",
]

my_test_file.run_unit_tests( unit_tests_to_run )

---

Alternatively, you can use load_tests Protocol and define the following method in your test module/file:

def load_tests(loader, standard_tests, pattern):
    suite = unittest.TestSuite()

    # To add a single test from this file
    suite.addTest( LineBalancingUnitTests( 'test_it_is_sunny' ) )

    # To add a single test class from this file
    suite.addTests( unittest.TestLoader().loadTestsFromTestCase( LineBalancingUnitTests ) )

    return suite

If you want to limit the execution to one single test file, you just need to set the test discovery pattern to the only file where you defined the load_tests() function.

#! /usr/bin/env python3
# -*- coding: utf-8 -*-
import os
import sys
import unittest

test_pattern = 'mytest/module/name.py'
PACKAGE_ROOT_DIRECTORY = os.path.dirname( os.path.realpath( __file__ ) )

loader = unittest.TestLoader()
start_dir = os.path.join( PACKAGE_ROOT_DIRECTORY, 'testing' )

suite = loader.discover( start_dir, test_pattern )
runner = unittest.TextTestRunner( verbosity=2 )
results = runner.run( suite )

print( "results: %s" % results )
print( "results.wasSuccessful: %s" % results.wasSuccessful() )

sys.exit( not results.wasSuccessful() )

References:

1. Problem with sys.argv[1] when unittest module is in a script
2. Is there a way to loop through and execute all of the functions in a Python class?
3. looping over all member variables of a class in python

---

Alternatively, to the last main program example, I came up with the following variation after reading the unittest.main() method implementation:

1. https://github.com/python/cpython/blob/master/Lib/unittest/main.py#L65

#! /usr/bin/env python3
# -*- coding: utf-8 -*-

import os
import sys
import unittest

PACKAGE_ROOT_DIRECTORY = os.path.dirname( os.path.realpath( __file__ ) )
start_dir = os.path.join( PACKAGE_ROOT_DIRECTORY, 'testing' )

from testing_package import main_unit_tests_module
testNames = ["TestCaseClassName.test_nameHelloWorld"]

loader = unittest.TestLoader()
suite = loader.loadTestsFromNames( testNames, main_unit_tests_module )

runner = unittest.TextTestRunner(verbosity=2)
results = runner.run( suite )

print( "results: %s" % results )
print( "results.wasSuccessful: %s" % results.wasSuccessful() )
sys.exit( not results.wasSuccessful() )

In case you want to run only tests from a specific class:

if __name__ == "__main__":
    unittest.main(MyCase())

It works for me in Python 3.6.

TL;DR: This would very likely work:

python mypkg/tests/test_module.py MyCase.testItIsHot

The explanation:

• The convenient way

    python mypkg/tests/test_module.py MyCase.testItIsHot
  

  would work, but its unspoken assumption is you already have this conventional code snippet inside (typically at the end of) your test file.

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

• The inconvenient way

    python -m unittest mypkg.tests.test_module.TestClass.test_method
  

  would always work, without requiring you to have that if __name__ == "__main__": unittest.main() code snippet in your test source file.

So why is the second method considered inconvenient? Because it would be a pain in the <insert one of your body parts here> to type that long, dot-delimited path by hand. While in the first method, the mypkg/tests/test_module.py part can be auto-completed, either by a modern shell, or by your editor.

If you want to run the test directly from a script (for example, from a jupyter notebook), you can do this to run just one test:

from testMyCase import MyCase
unittest.main(argv=['ignored', '-v', 'MyCase.testItIsHot'], exit=False)

What worked for me was:

cd project_dir
python -m unittest -v pathtotesttestMyCase.py -k my_test_name

-v is for unittest verbose log output.