pytest¶

py.test hacks to support XFAIL/XPASS

sympy.testing.pytest.SKIP(reason)[source]¶: Similar to skip(), but this is a decorator.

sympy.testing.pytest.nocache_fail(func)[source]¶: Dummy decorator for marking tests that fail when cache is disabled

sympy.testing.pytest.raises(expectedException, code=None)[source]¶

Tests that code raises the exception expectedException.

code may be a callable, such as a lambda expression or function name.

If code is not given or None, raises will return a context manager for use in with statements; the code to execute then comes from the scope of the with.

raises() does nothing if the callable raises the expected exception, otherwise it raises an AssertionError.

Examples

>>> from sympy.testing.pytest import raises

>>> raises(ZeroDivisionError, lambda: 1/0)
<ExceptionInfo ZeroDivisionError(...)>
>>> raises(ZeroDivisionError, lambda: 1/2)
Traceback (most recent call last):
...
Failed: DID NOT RAISE

>>> with raises(ZeroDivisionError):
...     n = 1/0
>>> with raises(ZeroDivisionError):
...     n = 1/2
Traceback (most recent call last):
...
Failed: DID NOT RAISE

Note that you cannot test multiple statements via with raises:

>>> with raises(ZeroDivisionError):
...     n = 1/0    # will execute and raise, aborting the ``with``
...     n = 9999/0 # never executed

This is just what with is supposed to do: abort the contained statement sequence at the first exception and let the context manager deal with the exception.

To test multiple statements, you’ll need a separate with for each:

>>> with raises(ZeroDivisionError):
...     n = 1/0    # will execute and raise
>>> with raises(ZeroDivisionError):
...     n = 9999/0 # will also execute and raise

sympy.testing.pytest.skip_under_pyodide(message)[source]¶: Decorator to skip a test if running under pyodide.

sympy.testing.pytest.warns(warningcls, *, match='', test_stacklevel=True)[source]¶

Like raises but tests that warnings are emitted.

>>> from sympy.testing.pytest import warns
>>> import warnings

>>> with warns(UserWarning):
...     warnings.warn('deprecated', UserWarning, stacklevel=2)

>>> with warns(UserWarning):
...     pass
Traceback (most recent call last):
...
Failed: DID NOT WARN. No warnings of type UserWarning    was emitted. The list of emitted warnings is: [].

test_stacklevel makes it check that the stacklevel parameter to warn() is set so that the warning shows the user line of code (the code under the warns() context manager). Set this to False if this is ambiguous or if the context manager does not test the direct user code that emits the warning.

If the warning is a SymPyDeprecationWarning, this additionally tests that the active_deprecations_target is a real target in the active-deprecations.md file.

sympy.testing.pytest.warns_deprecated_sympy()[source]¶

Shorthand for warns(SymPyDeprecationWarning)

This is the recommended way to test that SymPyDeprecationWarning is emitted for deprecated features in SymPy. To test for other warnings use warns. To suppress warnings without asserting that they are emitted use ignore_warnings.

Note

warns_deprecated_sympy() is only intended for internal use in the SymPy test suite to test that a deprecation warning triggers properly. All other code in the SymPy codebase, including documentation examples, should not use deprecated behavior.

If you are a user of SymPy and you want to disable SymPyDeprecationWarnings, use warnings filters (see Silencing SymPy Deprecation Warnings).

>>> from sympy.testing.pytest import warns_deprecated_sympy
>>> from sympy.utilities.exceptions import sympy_deprecation_warning
>>> with warns_deprecated_sympy():
...     sympy_deprecation_warning("Don't use",
...        deprecated_since_version="1.0",
...        active_deprecations_target="active-deprecations")

>>> with warns_deprecated_sympy():
...     pass
Traceback (most recent call last):
...
Failed: DID NOT WARN. No warnings of type     SymPyDeprecationWarning was emitted. The list of emitted warnings is: [].

Note

Sometimes the stacklevel test will fail because the same warning is emitted multiple times. In this case, you can use sympy.utilities.exceptions.ignore_warnings() in the code to prevent the SymPyDeprecationWarning from being emitted again recursively. In rare cases it is impossible to have a consistent stacklevel for deprecation warnings because different ways of calling a function will produce different call stacks.. In those cases, use warns(SymPyDeprecationWarning) instead.