Oven logo

Oven

Published

Allows you to use fixtures in @pytest.mark.parametrize.

pip install pytest-lazy-fixtures

Package Downloads

Weekly DownloadsMonthly Downloads

Requires Python

>=3.8

Dependencies

pytest-lazy-fixtures

codecov CI PyPI version PyPI downloads

Use your fixtures in @pytest.mark.parametrize.

This project was inspired by pytest-lazy-fixture.

Improvements that have been made in this project:

  1. You can use fixtures in any data structures
  2. You can access the attributes of fixtures
  3. You can use functions in fixtures
  4. It is compatible with pytest-deadfixtures

Installation

pip install pytest-lazy-fixtures

Usage

To use your fixtures inside @pytest.mark.parametrize you can use lf (lazy_fixture).

import pytest
from pytest_lazy_fixtures import lf

@pytest.fixture()
def one():
    return 1

@pytest.mark.parametrize('arg1,arg2', [('val1', lf('one'))])
def test_func(arg1, arg2):
    assert arg2 == 1

lf can be used with any data structures. For example, in the following example, lf is used in the dictionary:

import pytest
from pytest_lazy_fixtures import lf

@pytest.fixture()
def one():
    return 1

@pytest.mark.parametrize('arg1,arg2', [('val1', {"value": lf('one')})])
def test_func(arg1, arg2):
    assert arg2 == {"value": 1}

You can also specify getting an attribute through a dot:

import pytest
from pytest_lazy_fixtures import lf

class Entity:
    def __init__(self, value):
        self.value = value

@pytest.fixture()
def one():
    return Entity(1)

@pytest.mark.parametrize('arg1,arg2', [('val1', lf('one.value'))])
def test_func(arg1, arg2):
    assert arg2 == 1

And there is some useful wrapper called lfc (lazy_fixture_callable). It can work with any callable and your fixtures.

import pytest
from pytest_lazy_fixtures import lf, lfc

class Entity:
    def __init__(self, value):
        self.value = value

    def __str__(self) -> str:
        return str(self.value)

    def sum(self, value: int) -> int:
        return self.value + value

@pytest.fixture()
def entity():
    return Entity(1)

@pytest.fixture()
def two():
    return 2

@pytest.fixture()
def three():
    return 3

@pytest.fixture()
def entity_format():
    def _entity_format(entity: Entity):
        return {"value": entity.value}

    return _entity_format

# Recommended: implicit injection from the callable's parameter names
@pytest.mark.parametrize("as_text", [lfc(lambda entity: str(entity))])
def test_lfc_injection_basic(as_text):
    assert as_text == "1"

# Advanced: manual control over what to inject (explicit values into a callable)
@pytest.mark.parametrize(
    "message",
    [
        lfc(
            "There is two lazy fixture args, {} and {}! And one kwarg {field}! And also simple value {simple}".format,
            lf("entity"),
            lf("two"),
            field=lf("three"),
            simple="value",
        ),
    ],
)
def test_lazy_fixture_callable_with_func(message):
    assert message == "There is two lazy fixture args, 1 and 2! And one kwarg 3! And also simple value value"


@pytest.mark.parametrize("formatted", [lfc("entity_format", lf("entity"))])
def test_lazy_fixture_callable_with_lf(formatted, entity):
    assert formatted == {"value": entity.value}


@pytest.mark.parametrize("result", [lfc("entity.sum", lf("two"))])
def test_lazy_fixture_callable_with_attr_lf(result):
    assert result == 3

Injecting fixtures from a callable signature

lfc automatically injects fixtures based on the callable's parameter names. If a parameter name matches a fixture name, that fixture will be resolved and passed into the callable. This implicit injection is the standard, recommended behavior. You can still override any parameter explicitly using lf, either positionally or by name. Default values prevent implicit injection for that parameter.

Examples:

import pytest
from pytest_lazy_fixtures import lf, lfc

# Some fixtures used in the examples
@pytest.fixture()
def alpha():
    return 10

@pytest.fixture()
def beta():
    return 20

@pytest.fixture()
def gamma():
    return 30

# Simple implicit injection by parameter name
@pytest.mark.parametrize("result", [lfc(lambda alpha: alpha + 1)])
def test_signature_injection_basic(result):
    assert result == 11

# Override an implicitly matched parameter (positional or keyword)
@pytest.mark.parametrize(
    "result",
    [
        lfc(lambda alpha: alpha + 1, lf("beta")),      # positional override
        lfc(lambda alpha: alpha + 1, alpha=lf("beta")),  # keyword override
    ],
)
def test_signature_injection_override(result):
    assert result == 21

# Mix implicit and explicit params
@pytest.mark.parametrize(
    "pair",
    [
        lfc(lambda alpha, beta: (alpha, beta), beta=lf("gamma")),  # first implicit, second explicit
        lfc(lambda alpha, beta: (alpha, beta), lf("gamma")),       # first explicit, second implicit
    ],
)
def test_signature_injection_mixed(pair):
    assert pair in [(10, 30), (30, 20)]

# 4) Defaults prevent implicit injection for that parameter
@pytest.mark.parametrize("val", [lfc(lambda alpha=999: alpha)])
def test_signature_injection_defaults(val):
    assert val == 999

Notes:

  • Implicit injection only happens for parameters without an explicit value and without a default.
  • Explicit arguments passed to lfc (positional or keyword) always take precedence over implicit injection.
  • The mapping of positional explicit arguments follows the callable's parameter order.

Contributing

Contributions are very welcome. Tests can be run with pytest.

License

Distributed under the terms of the MIT license, pytest-lazy-fixtures is free and open source software

Issues

If you encounter any problems, please file an issue along with a detailed description.