Pytest is available for installation from the PyPI, therefore installation with pip is recommended. To install the latest version, execute from the command line:

That will install allure-pytest and allure-python-commons packages to produce report data compatible with Allure 2. If you are using a previous version of adapter for the first generation of Allure reports then you will need to uninstall it first.

To enable Allure listener to collect results during the test execution simply add --alluredir option and provide path to the folder where results should be stored. E.g.:

To see the actual report after your tests have finished, you need to use Allure commandline utility to generate report from the results.

This command will show you generated report in your default browser.

Your can see all default pytest statuses in the Allure report: only tests that were not succeeded due to one of the assertion errors will be marked as failed, any other exception will cause a test to have a broken status.

Some of the common Pytest features that the Allure report supports include xfails, fixtures and finalizers, marks, conditional skips and parametrization.

Xfail

@pytest.mark.xfail(condition=lambda: True, reason='this test is expecting failure')
def test_xfail_expected_failure():
    """this test is an xfail that will be marked as expected failure"""
    assert False


@pytest.mark.xfail(condition=lambda: True, reason='this test is expecting failure')
def test_xfail_unexpected_pass():
    """this test is an xfail that will be marked as unexpected success"""
    assert True

@pytest.mark.skipif('2 + 2 != 5', reason='This test is skipped by a triggered condition in @pytest.mark.skipif')
def test_skip_by_triggered_condition():
    pass

@pytest.fixture(params=[True, False], ids=['param_true', 'param_false'])
def function_scope_fixture_with_finalizer(request):
    if request.param:
        print('True')
    else:
        print('False')
    def function_scope_finalizer():
        function_scope_step()
    request.addfinalizer(function_scope_finalizer)


@pytest.fixture(scope='class')
def class_scope_fixture_with_finalizer(request):
    def class_finalizer_fixture():
        class_scope_step()
    request.addfinalizer(class_finalizer_fixture)


@pytest.fixture(scope='module')
def module_scope_fixture_with_finalizer(request):
    def module_finalizer_fixture():
        module_scope_step()
    request.addfinalizer(module_finalizer_fixture)


@pytest.fixture(scope='session')
def session_scope_fixture_with_finalizer(request):
    def session_finalizer_fixture():
        session_scope_step()
    request.addfinalizer(session_finalizer_fixture)


class TestClass(object):

    def test_with_scoped_finalizers(self,
                                    function_scope_fixture_with_finalizer,
                                    class_scope_fixture_with_finalizer,
                                    module_scope_fixture_with_finalizer,
                                    session_scope_fixture_with_finalizer):
        step_inside_test_body()

import pytest

@pytest.fixture
def skip_fixture():
    pytest.skip()


@pytest.fixture
def fail_fixture():
    assert False


@pytest.fixture
def broken_fixture():
    raise Exception("Sorry, it's broken.")


def test_with_pytest_skip_in_the_fixture(skip_fixture):
    pass


def test_with_failure_in_the_fixture(fail_fixture):
    pass


def test_with_broken_fixture(broken_fixture):
    pass

import allure
import pytest


@allure.step
def simple_step(step_param1, step_param2 = None):
    pass


@pytest.mark.parametrize('param1', [True, False], ids=['id explaining value 1', 'id explaining value 2'])
def test_parameterize_with_id(param1):
    simple_step(param1)


@pytest.mark.parametrize('param1', [True, False])
@pytest.mark.parametrize('param2', ['value 1', 'value 2'])
def test_parametrize_with_two_parameters(param1, param2):
    simple_step(param1, param2)


@pytest.mark.parametrize('param1', [True], ids=['boolean parameter id'])
@pytest.mark.parametrize('param2', ['value 1', 'value 2'])
@pytest.mark.parametrize('param3', [1])
def test_parameterize_with_uneven_value_sets(param1, param2, param3):
    simple_step(param1, param3)
    simple_step(param2)

The first and probably most important aspect of the Allure report is that it allows to get a very detailed step-by-step representation of every test invocation. This is made possible with @allure.step decorator that adds invocation of the annotated method or function with provided arguments to the report.

Methods annotated with @step can be stored aside from your tests and just imported when needed. Step methods can have an arbitrarily deep nested structure.

Status of every step is shown in a small icon on the right from the name. Nested steps are organized in a tree-like collapsible structure.

import allure

@allure.step('Step with placeholders in the title, positional: "{0}", keyword: "{key}"')
def step_with_title_placeholders(arg1, key=None):
    pass


def test_steps_with_placeholders():
    step_with_title_placeholders(1, key='something')
    step_with_title_placeholders(2)
    step_with_title_placeholders(3, 'anything')

Steps are supported in fixtures as well. Here is an example of a test using a fixture defined in conftest.py module (such fixtures will be resolved by Pytest even when not directly imported):

import allure
import pytest


@allure.step('step in conftest.py')
def conftest_step():
    pass


@pytest.fixture
def fixture_with_conftest_step():
    conftest_step()

import allure

from .steps import imported_step


@allure.step
def passing_step():
    pass


def test_with_step_in_fixture_from_conftest(fixture_with_conftest_step):
    passing_step()

import allure
import pytest


@pytest.fixture
def attach_file_in_module_scope_fixture_with_finalizer(request):
    allure.attach('A text attacment in module scope fixture', 'blah blah blah', allure.attachment_type.TEXT)
    def finalizer_module_scope_fixture():
        allure.attach('A text attacment in module scope finalizer', 'blah blah blah blah',
                      allure.attachment_type.TEXT)
    request.addfinalizer(finalizer_module_scope_fixture)


def test_with_attacments_in_fixture_and_finalizer(attach_file_in_module_scope_finalizer):
    pass


def test_multiple_attachments():
    allure.attach.file('./data/totally_open_source_kitten.png', attachment_type=allure.attachment_type.PNG)
    allure.attach('<head></head><body> a page </body>', 'Attach with HTML type', allure.attachment_type.HTML)

import allure

@allure.description_html("""
<h1>Test with some complicated html description</h1>
<table style="width:100%">
  <tr>
    <th>Firstname</th>
    <th>Lastname</th>
    <th>Age</th>
  </tr>
  <tr align="center">
    <td>William</td>
    <td>Smith</td>
    <td>50</td>
  </tr>
  <tr align="center">
    <td>Vasya</td>
    <td>Jackson</td>
    <td>94</td>
  </tr>
</table>
""")
def test_html_description():
    assert True


@allure.description("""
Multiline test description.
That comes from the allure.description decorator.

Nothing special about it.
""")
def test_description_from_decorator():
    assert 42 == int(6 * 7)


def test_unicode_in_docstring_description():
    """Unicode in description.

    Этот тест проверяет юникод.

    你好伙计.
    """
    assert 42 == int(6 * 7)

Also descriptions can be dynamically updated from within test body using allure.dynamic.description.

Test titles can be made more readable with special @allure.title decorator. Titles support placeholders for arguments and support dynamic replacement.

import allure

TEST_CASE_LINK = 'your any link'


@allure.link('your any link')
def test_with_link():
    pass


@allure.link('your any link', name='Click me')
def test_with_named_link():
    pass


@allure.issue('140', 'Pytest-flaky test retries shows like test steps')
def test_with_issue_link():
    pass


@allure.testcase(TEST_CASE_LINK, 'Test case title')
def test_with_testcase_link():
    pass

$ pytest directory_with_tests/ --alluredir=/tmp/my_allure_report \
 --allure-link-pattern=issue:your_any_link{}

import allure
import random
import time


@allure.step
def passing_step():
    pass


@allure.step
def flaky_broken_step():
    if random.randint(1, 5) != 1:
        raise Exception('Broken!')


def test_broken_with_randomized_time():
    passing_step()
    time.sleep(random.randint(1, 3))
    flaky_broken_step()

There are two decorators: @allure.feature and @allure.story to mark your tests according to Feature/Story breakdown specific to your project (for background see BDD article on Wikipedia). To mark that some feature or story belong to an epic, use a name that starts with epic_ prefix.

You can use following commandline options to specify different sets of tests to execute passing a list of comma-separated values:

1. --allure-epics

2. --allure-features

3. --allure-stories

for example:

To mark your tests by severity level you can use @allure.severity decorator. It takes a allure.severity_level enum value as an argument.

Severity decorator can be applied to functions, methods or entire classes.

By using --allure-severities commandline option with a list of comma-separated severity levels only tests with corresponding severities will be run.