Skip to main content

Plugin for nose or py.test that automatically reruns flaky tests.

Project description

http://opensource.box.com/badges/active.svg https://travis-ci.org/box/flaky.png?branch=master https://img.shields.io/pypi/v/flaky.svg https://img.shields.io/pypi/dm/flaky.svg

About

Flaky is a plugin for nose or py.test that automatically reruns flaky tests.

Ideally, tests reliably pass or fail, but sometimes test fixtures must rely on components that aren’t 100% reliable. With flaky, instead of removing those tests or marking them to @skip, they can be automatically retried.

For more information about flaky, see this presentation.

Marking tests flaky

To mark a test as flaky, simply import flaky and decorate the test with @flaky:

from flaky import flaky
@flaky
def test_something_that_usually_passes(self):
    value_to_double = 21
    result = get_result_from_flaky_doubler(value_to_double)
    self.assertEqual(result, value_to_double * 2, 'Result doubled incorrectly.')

By default, flaky will retry a failing test once, but that behavior can be overridden by passing values to the flaky decorator. It accepts two parameters: max_runs, and min_passes; flaky will run tests up to max_runs times, until it has succeeded min_passes times. Once a test passes min_passes times, it’s considered a success; once it has been run max_runs times without passing min_passes times, it’s considered a failure.

@flaky(max_runs=3, min_passes=2)
def test_something_that_usually_passes(self):
    """This test must pass twice, and it can be run up to three times."""
    value_to_double = 21
    result = get_result_from_flaky_doubler(value_to_double)
    self.assertEqual(result, value_to_double * 2, 'Result doubled incorrectly.')

Marking a class flaky

In addition to marking a single test flaky, entire test cases can be marked flaky:

@flaky
class TestMultipliers(TestCase):
    def test_flaky_doubler(self):
        value_to_double = 21
        result = get_result_from_flaky_doubler(value_to_double)
        self.assertEqual(result, value_to_double * 2, 'Result doubled incorrectly.')

    @flaky(max_runs=3)
    def test_flaky_tripler(self):
        value_to_triple = 14
        result = get_result_from_flaky_tripler(value_to_triple)
        self.assertEqual(result, value_to_triple * 3, 'Result tripled incorrectly.')

The @flaky class decorator will mark test_flaky_doubler as flaky, but it won’t override the 3 max_runs for test_flaky_tripler (from the decorator on that test method).

Don’t rerun certain types of failures

Depending on your tests, some failures are obviously not due to flakiness. Instead of rerunning after those failures, you can specify a filter function that can tell flaky to fail the test right away.

def is_not_crash(err, *args):
    return not issubclass(err[0], ProductCrashedError)

@flaky
def test_something():
    raise ProductCrashedError

@flaky(rerun_filter=is_not_crash)
def test_something_else():
    raise ProductCrashedError

Flaky will run test_something twice, but will only run test_something_else once.

Activating the plugin

Like any nose plugin, flaky can be activated via the command line:

nosetests --with-flaky

With py.test, flaky will automatically run. It can, however be disabled via the command line:

py.test -p no:flaky

Command line arguments

No Flaky Report

Pass --no-flaky-report to suppress the report at the end of the run detailing flaky test results.

Shorter Flaky Report

Pass --no-success-flaky-report to suppress information about successful flaky tests.

Force Flaky

Pass --force-flaky to treat all tests as flaky.

Pass --max-runs=MAX_RUNS and/or --min-passes=MIN_PASSES to control the behavior of flaky if --force-flaky is specified. Flaky decorators on individual tests will override these defaults.

Additional usage examples are in the code - see test/test_example.py

Installation

To install, simply:

pip install flaky

Compatibility

Flaky is tested with the following test runners and options:

  • Nosetests. Doctests cannot be marked flaky.

  • Py.test. Works with pytest-xdist but not with the –boxed option. Doctests cannot be marked flaky.

Contributing

See CONTRIBUTING.rst.

Setup

Create a virtual environment and install packages -

mkvirtualenv flaky
pip install -r requirements-dev.txt

Testing

Run all tests using -

tox

The tox tests include code style checks via pep8 and pylint.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

flaky-2.4.1.tar.gz (27.4 kB view hashes)

Uploaded Source

Built Distribution

flaky-2.4.1-py2-none-any.whl (40.2 kB view hashes)

Uploaded Python 2

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page