Back to Pytest

How to capture warnings

doc/en/how-to/capture-warnings.rst

9.1.0.dev017.4 KB
Original Source

.. _warnings:

How to capture warnings

Starting from version 3.1, pytest now automatically catches warnings during test execution and displays them at the end of the session:

.. code-block:: python

# content of test_show_warnings.py
import warnings


def api_v1():
    warnings.warn(UserWarning("api v1, should use functions from v2"))
    return 1


def test_one():
    assert api_v1() == 1

Running pytest now produces this output:

.. code-block:: pytest

$ pytest test_show_warnings.py
=========================== test session starts ============================
platform linux -- Python 3.x.y, pytest-9.x.y, pluggy-1.x.y
rootdir: /home/sweet/project
collected 1 item

test_show_warnings.py .                                              [100%]

============================= warnings summary =============================
test_show_warnings.py::test_one
  /home/sweet/project/test_show_warnings.py:5: UserWarning: api v1, should use functions from v2
    warnings.warn(UserWarning("api v1, should use functions from v2"))

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
======================= 1 passed, 1 warning in 0.12s =======================

.. _controlling-warnings:

Controlling warnings

Similar to Python's warning filter_ and :option:-W option <python:-W> flag, pytest provides its own -W flag to control which warnings are ignored, displayed, or turned into errors. See the warning filter_ documentation for more advanced use-cases.

.. _warning filter: https://docs.python.org/3/library/warnings.html#warning-filter

This code sample shows how to treat any UserWarning category class of warning as an error:

.. code-block:: pytest

$ pytest -q test_show_warnings.py -W error::UserWarning
F                                                                    [100%]
================================= FAILURES =================================
_________________________________ test_one _________________________________

    def test_one():
>       assert api_v1() == 1
               ^^^^^^^^

test_show_warnings.py:10:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

    def api_v1():
>       warnings.warn(UserWarning("api v1, should use functions from v2"))
E       UserWarning: api v1, should use functions from v2

test_show_warnings.py:5: UserWarning
========================= short test summary info ==========================
FAILED test_show_warnings.py::test_one - UserWarning: api v1, should use ...
1 failed in 0.12s

The same option can be set in the configuration file using the :confval:filterwarnings configuration option. For example, the configuration below will ignore all user warnings and specific deprecation warnings matching a regex, but will transform all other warnings into errors.

.. tab:: toml

.. code-block:: toml

    [pytest]
    filterwarnings = [
        'error',
        'ignore::UserWarning',
        # Note the use of single quote below to denote "raw" strings in TOML.
        'ignore:function ham\(\) is deprecated:DeprecationWarning',
    ]

.. tab:: ini

.. code-block:: ini

    [pytest]
    filterwarnings =
        error
        ignore::UserWarning
        ignore:function ham\(\) is deprecated:DeprecationWarning

When a warning matches more than one option in the list, the action for the last matching option is performed.

.. note::

The ``-W`` flag and the :confval:`filterwarnings` configuration option use warning filters that are
similar in structure, but each configuration option interprets its filter
differently. For example, *message* in ``filterwarnings`` is a string containing a
regular expression that the start of the warning message must match,
case-insensitively, while *message* in ``-W`` is a literal string that the start of
the warning message must contain (case-insensitively), ignoring any whitespace at
the start or end of message. Consult the `warning filter`_ documentation for more
details.

.. _filterwarnings:

@pytest.mark.filterwarnings

You can use the :ref:@pytest.mark.filterwarnings <pytest.mark.filterwarnings ref> mark to add warning filters to specific test items, allowing you to have finer control of which warnings should be captured at test, class or even module level:

.. code-block:: python

import warnings


def api_v1():
    warnings.warn(UserWarning("api v1, should use functions from v2"))
    return 1


@pytest.mark.filterwarnings("ignore:api v1")
def test_one():
    assert api_v1() == 1

You can specify multiple filters with separate decorators:

.. code-block:: python

# Ignore "api v1" warnings, but fail on all other warnings
@pytest.mark.filterwarnings("ignore:api v1")
@pytest.mark.filterwarnings("error")
def test_one():
    assert api_v1() == 1

You can also pass multiple filters to a single mark by providing multiple arguments:

.. code-block:: python

# Later arguments take precedence, matching warnings.filterwarnings behavior.
@pytest.mark.filterwarnings("error", "ignore:api v1")
def test_one():
    assert api_v1() == 1

.. important::

Regarding decorator order and filter precedence:
it's important to remember that decorators are evaluated in reverse order,
so you have to list the warning filters in the reverse order
compared to traditional :py:func:`warnings.filterwarnings` and :option:`-W option <python:-W>` usage.
This means in practice that filters from earlier :ref:`@pytest.mark.filterwarnings <pytest.mark.filterwarnings ref>` decorators
take precedence over filters from later decorators, as illustrated in the example above.

Filters applied using a mark take precedence over filters passed on the command line or configured by the :confval:filterwarnings configuration option.

You may apply a filter to all tests of a class by using the :ref:filterwarnings <pytest.mark.filterwarnings ref> mark as a class decorator or to all tests in a module by setting the :globalvar:pytestmark variable:

.. code-block:: python

# turns all warnings into errors for this module
pytestmark = pytest.mark.filterwarnings("error")

.. note::

If you want to apply multiple filters
(by assigning a list of :ref:`filterwarnings <pytest.mark.filterwarnings ref>` mark to :globalvar:`pytestmark`),
you must use the traditional :py:func:`warnings.filterwarnings` ordering approach (later filters take precedence),
which is the reverse of the decorator approach mentioned above.

Credits go to Florian Schulze for the reference implementation in the pytest-warnings_ plugin.

.. _pytest-warnings: https://github.com/fschulze/pytest-warnings

Disabling warnings summary

Although not recommended, you can use the :option:--disable-warnings command-line option to suppress the warning summary entirely from the test run output.

Disabling warning capture entirely

This plugin is enabled by default but can be disabled entirely in your configuration file with:

.. tab:: toml

.. code-block:: toml

    [pytest]
    addopts = ["-p", "no:warnings"]

.. tab:: ini

.. code-block:: ini

    [pytest]
    addopts = -p no:warnings

Or passing -p no:warnings in the command-line. This might be useful if your test suite handles warnings using an external system.

.. _deprecation-warnings:

DeprecationWarning and PendingDeprecationWarning

By default pytest will display DeprecationWarning and PendingDeprecationWarning warnings from user code and third-party libraries, as recommended by :pep:565. This helps users keep their code modern and avoid breakages when deprecated warnings are effectively removed.

However, in the specific case where users capture any type of warnings in their test, either with :func:pytest.warns, :func:pytest.deprecated_call or using the :fixture:recwarn fixture, no warning will be displayed at all.

Sometimes it is useful to hide some specific deprecation warnings that happen in code that you have no control over (such as third-party libraries), in which case you might use the warning filters options (configuration or marks) to ignore those warnings.

For example:

.. tab:: toml

.. code-block:: toml

    [pytest]
    filterwarnings = [
        'ignore:.*U.*mode is deprecated:DeprecationWarning',
    ]

.. tab:: ini

.. code-block:: ini

    [pytest]
    filterwarnings =
        ignore:.*U.*mode is deprecated:DeprecationWarning

This will ignore all warnings of type DeprecationWarning where the start of the message matches the regular expression ".*U.*mode is deprecated".

See :ref:@pytest.mark.filterwarnings <filterwarnings> and :ref:Controlling warnings <controlling-warnings> for more examples.

.. note::

If warnings are configured at the interpreter level, using
the :envvar:`python:PYTHONWARNINGS` environment variable or the
``-W`` command-line option, pytest will not configure any filters by default.

Also pytest doesn't follow :pep:`565` suggestion of resetting all warning filters because
it might break test suites that configure warning filters themselves
by calling :func:`warnings.simplefilter` (see :issue:`2430` for an example of that).

.. _ensuring a function triggers a deprecation warning:

.. _ensuring_function_triggers:

Ensuring code triggers a deprecation warning

You can also use :func:pytest.deprecated_call for checking that a certain function call triggers a DeprecationWarning, PendingDeprecationWarning or FutureWarning:

.. code-block:: python

import pytest


def test_myfunction_deprecated():
    with pytest.deprecated_call():
        myfunction(17)

This test will fail if myfunction does not issue a deprecation warning when called with a 17 argument.

.. _asserting warnings:

.. _assertwarnings:

.. _asserting warnings with the warns function:

.. _warns:

Asserting warnings with the warns function

You can check that code raises a particular warning using :func:pytest.warns, which works in a similar manner to :ref:raises <assertraises> (except that :ref:raises <assertraises> does not capture all exceptions, only the expected_exception):

.. code-block:: python

import warnings

import pytest


def test_warning():
    with pytest.warns(UserWarning):
        warnings.warn("my warning", UserWarning)

The test will fail if the warning in question is not raised. Use the keyword argument match to assert that the warning matches a text or regex. To match a literal string that may contain regular expression metacharacters like ( or ., the pattern can first be escaped with re.escape.

Some examples:

.. code-block:: pycon

>>> with warns(UserWarning, match="must be 0 or None"):
...     warnings.warn("value must be 0 or None", UserWarning)
...

>>> with warns(UserWarning, match=r"must be \d+$"):
...     warnings.warn("value must be 42", UserWarning)
...

>>> with warns(UserWarning, match=r"must be \d+$"):
...     warnings.warn("this is not here", UserWarning)
...
Traceback (most recent call last):
  ...
Failed: DID NOT WARN. No warnings of type ...UserWarning... were emitted...

>>> with warns(UserWarning, match=re.escape("issue with foo() func")):
...     warnings.warn("issue with foo() func")
...

You can also call :func:pytest.warns on a function or code string:

.. code-block:: python

pytest.warns(expected_warning, func, *args, **kwargs)
pytest.warns(expected_warning, "func(*args, **kwargs)")

The function also returns a list of all raised warnings (as warnings.WarningMessage objects), which you can query for additional information:

.. code-block:: python

with pytest.warns(RuntimeWarning) as record:
    warnings.warn("another warning", RuntimeWarning)

# check that only one warning was raised
assert len(record) == 1
# check that the message matches
assert record[0].message.args[0] == "another warning"

Alternatively, you can examine raised warnings in detail using the :fixture:recwarn fixture (see :ref:below <recwarn>).

The :fixture:recwarn fixture automatically ensures to reset the warnings filter at the end of the test, so no global state is leaked.

.. _recording warnings:

.. _recwarn:

Recording warnings

You can record raised warnings either using the :func:pytest.warns context manager or with the :fixture:recwarn fixture.

To record with :func:pytest.warns without asserting anything about the warnings, pass no arguments as the expected warning type and it will default to a generic Warning:

.. code-block:: python

with pytest.warns() as record:
    warnings.warn("user", UserWarning)
    warnings.warn("runtime", RuntimeWarning)

assert len(record) == 2
assert str(record[0].message) == "user"
assert str(record[1].message) == "runtime"

The :fixture:recwarn fixture will record warnings for the whole function:

.. code-block:: python

import warnings


def test_hello(recwarn):
    warnings.warn("hello", UserWarning)
    assert len(recwarn) == 1
    w = recwarn.pop(UserWarning)
    assert issubclass(w.category, UserWarning)
    assert str(w.message) == "hello"
    assert w.filename
    assert w.lineno

Both the :fixture:recwarn fixture and the :func:pytest.warns context manager return the same interface for recorded warnings: a :class:~_pytest.recwarn.WarningsRecorder instance. To view the recorded warnings, you can iterate over this instance, call len on it to get the number of recorded warnings, or index into it to get a particular recorded warning.

.. _warns use cases:

Additional use cases of warnings in tests

Here are some use cases involving warnings that often come up in tests, and suggestions on how to deal with them:

  • To ensure that at least one of the indicated warnings is issued, use:

.. code-block:: python

def test_warning():
    with pytest.warns((RuntimeWarning, UserWarning)):
        ...
  • To ensure that only certain warnings are issued, use:

.. code-block:: python

def test_warning(recwarn):
    ...
    assert len(recwarn) == 1
    user_warning = recwarn.pop(UserWarning)
    assert issubclass(user_warning.category, UserWarning)
  • To ensure that no warnings are emitted, use:

.. code-block:: python

def test_warning():
    with warnings.catch_warnings():
        warnings.simplefilter("error")
        ...
  • To suppress warnings, use:

.. code-block:: python

with warnings.catch_warnings():
    warnings.simplefilter("ignore")
    ...

.. _custom_failure_messages:

Custom failure messages

Recording warnings provides an opportunity to produce custom test failure messages for when no warnings are issued or other conditions are met.

.. code-block:: python

def test():
    with pytest.warns(Warning) as record:
        f()
        if not record:
            pytest.fail("Expected a warning!")

If no warnings are issued when calling f, then not record will evaluate to True. You can then call :func:pytest.fail with a custom error message.

.. _internal-warnings:

Internal pytest warnings

pytest may generate its own warnings in some situations, such as improper usage or deprecated features.

For example, pytest will emit a warning if it encounters a class that matches :confval:python_classes but also defines an __init__ constructor, as this prevents the class from being instantiated:

.. code-block:: python

# content of test_pytest_warnings.py
class Test:
    def __init__(self):
        pass

    def test_foo(self):
        assert 1 == 1

.. code-block:: pytest

$ pytest test_pytest_warnings.py -q

============================= warnings summary =============================
test_pytest_warnings.py:1
  /home/sweet/project/test_pytest_warnings.py:1: PytestCollectionWarning: cannot collect test class 'Test' because it has a __init__ constructor (from: test_pytest_warnings.py)
    class Test:

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
1 warning in 0.12s

These warnings might be filtered using the same builtin mechanisms used to filter other types of warnings.

Please read our :ref:backwards-compatibility to learn how we proceed about deprecating and eventually removing features.

The full list of warnings is listed in :ref:the reference documentation <warnings ref>.

.. _resource-warnings:

Resource Warnings

Additional information of the source of a :class:ResourceWarning can be obtained when captured by pytest if :mod:tracemalloc module is enabled.

One convenient way to enable :mod:tracemalloc when running tests is to set the :envvar:PYTHONTRACEMALLOC to a large enough number of frames (say 20, but that number is application dependent).

For more information, consult the Python Development Mode <https://docs.python.org/3/library/devmode.html>__ section in the Python documentation.