django-fakery 4.1.1

Installation

Install with:

$ pip install django-fakery

QuickStart

from django_fakery import factory
from myapp.models import MyModel

factory.m(MyModel)(field='value')

If you’re having issues with circular imports, you can also reference a model by using the M utility function:

from django_fakery import factory, M

factory.m(M("myapp.MyModel"))(field="value")

If you really don’t want to import things, you could also just reference a model by using the <app_label>.<ModelName> syntax. This is not encouraged, as it will likely break type-hinting:

from django_fakery import factory

factory.m("myapp.MyModel")(field="value")

If you use pytest, you can use the fakery and fakery_shortcuts` fixtures (requires ``pytest and pytest-django):

import pytest
from myapp.models import MyModel

@pytest.mark.django_db
def test_mymodel(fakery, fakery_shortcuts):
    fakery.m(MyModel)(field=fakery_shortcuts.future_datetime())

If you’d rather, you can use a more wordy API:

from django_fakery import factory
from myapp.models import MyModel

factory.make(
    MyModel,
    fields={
        'field': 'value',
    }
)

We will use the short API thorough the documentation.

The value of a field can be any python object, a callable, or a lambda:

from django.utils import timezone
from django_fakery import factory
from myapp.models import MyModel

factory.m(MyModel)(created=timezone.now)

When using a lambda, it will receive two arguments: n is the iteration number, and f is an instance of faker:

from django.contrib.auth.models import User

user = factory.m(User)(
    username=lambda n, f: 'user_{}'.format(n),
)

django-fakery includes some pre-built lambdas for common needs. See shortcuts for more info.

You can create multiple objects by using the quantity parameter:

from django_fakery import factory
from django.contrib.auth.models import User

factory.m(User, quantity=4)

For convenience, when the value of a field is a string, it will be interpolated with the iteration number:

from myapp.models import MyModel

user = factory.m(User, quantity=4)(
    username='user_{}',
)

Custom fields

You can add support for custom fields by adding your custom field class and a function in factory.field_types:

from django_fakery import factory

from my_fields import CustomField

def func(faker, field, count, *args, **kwargs):
    return 43


factory.field_types.add(
    CustomField, (func, [], {})
)

As a shortcut, you can specified any Faker function by its name:

from django_fakery import factory

from my_fields import CustomField


factory.field_types.add(
    CustomField, ("random_int", [], {"min": 0, "max": 60})
)

Foreign keys

Non-nullable ForeignKey s create related objects automatically.

If you want to explicitly create a related object, you can pass a factory like any other value:

from django.contrib.auth.models import User
from food.models import Pizza

pizza = factory.m(Pizza)(
    chef=factory.m(User)(username='Gusteau'),
)

If you’d rather not create related objects and reuse the same value for a foreign key, you can use the special value django_fakery.rels.SELECT:

from django_fakery import factory, rels
from food.models import Pizza

pizza = factory.m(Pizza, quantity=5)(
    chef=rels.SELECT,
)

django-fakery will always use the first instance of the related model, creating one if necessary.

ManyToManies

Because ManyToManyField s are implicitly nullable (ie: they’re always allowed to have their .count() equal to 0), related objects on those fields are not automatically created for you.

If you want to explicitly create a related objects, you can pass a list as the field’s value:

from food.models import Pizza, Topping

pizza = factory.m(Pizza)(
    toppings=[
        factory.m(Topping)(name='Anchovies')
    ],
)

You can also pass a factory, to create multiple objects:

from food.models import Pizza, Topping

pizza = factory.m(Pizza)(
    toppings=factory.m(Topping, quantity=5),
)

Shortcuts

django-fakery includes some shortcut functions to generate commonly needed values.

from django_fakery import factory, shortcuts
from myapp.models import MyModel

factory.m(MyModel)(field=shortcuts.future_datetime('+1w'))

Lazies

You can refer to the created instance’s own attributes or method by using Lazy objects.

For example, if you’d like to create user with email as username, and have them always match, you could do:

from django_fakery import factory, Lazy
from django.contrib.auth.models import User

factory.m(auth.User)(
    username=Lazy('email'),
)

If you want to assign a value returned by a method on the instance, you can pass the method’s arguments to the Lazy object:

from django_fakery import factory, Lazy
from myapp.models import MyModel

factory.m(MyModel)(
    myfield=Lazy('model_method', 'argument', keyword='keyword value'),
)

Pre-save and Post-save hooks

You can define functions to be called right before the instance is saved or right after:

from django.contrib.auth.models import User
from django_fakery import factory

factory.m(
    User,
    pre_save=[
        lambda u: u.set_password('password')
    ],
)(username='username')

Since settings a user’s password is such a common case, we special-cased that scenario, so you can just pass it as a field:

from django.contrib.auth.models import User
from django_fakery import factory

factory.m(User)(
    username='username',
    password='password',
)

Get or Make

You can check for existance of a model instance and create it if necessary by using the g_m (short for get_or_make) method:

from myapp.models import MyModel

myinstance, created = factory.g_m(
    MyModel,
    lookup={
        'myfield': 'myvalue',
    }
)(myotherfield='somevalue')

If you’re looking for a more explicit API, you can use the .get_or_make() method:

from myapp.models import MyModel

myinstance, created = factory.get_or_make(
    MyModel,
    lookup={
        'myfield': 'myvalue',
    },
    fields={
        'myotherfield': 'somevalue',
    },
)

Get or Update

You can check for existence of a model instance and update it by using the g_u (short for get_or_update) method:

from myapp.models import MyModel

myinstance, created = factory.g_u(
    MyModel,
    lookup={
        'myfield': 'myvalue',
    }
)(myotherfield='somevalue')

If you’re looking for a more explicit API, you can use the .get_or_update() method:

from myapp.models import MyModel

myinstance, created = factory.get_or_update(
    MyModel,
    lookup={
        'myfield': 'myvalue',
    },
    fields={
        'myotherfield': 'somevalue',
    },
)

Non-persistent instances

You can build instances that are not saved to the database by using the .b() method, just like you’d use .m():

from django_fakery import factory
from myapp.models import MyModel

factory.b(MyModel)(
    field='value',
)

Note that since the instance is not saved to the database, .build() does not support ManyToManies or post-save hooks.

If you’re looking for a more explicit API, you can use the .build() method:

from django_fakery import factory
from myapp.models import MyModel

factory.build(
    MyModel,
    fields={
        'field': 'value',
    }
)

Blueprints

Use a blueprint:

from django.contrib.auth.models import User
from django_fakery import factory

user = factory.blueprint(User)

user.make(quantity=10)

Blueprints can refer other blueprints:

from food.models import Pizza

pizza = factory.blueprint(Pizza).fields(
        chef=user,
    )
)

You can also override the field values you previously specified:

from food.models import Pizza

pizza = factory.blueprint(Pizza).fields(
        chef=user,
        thickness=1
    )
)

pizza.m(quantity=10)(thickness=2)

Or, if you’d rather use the explicit api:

from food.models import Pizza

pizza = factory.blueprint(Pizza).fields(
        chef=user,
        thickness=1
    )
)

thicker_pizza = pizza.fields(thickness=2)
thicker_pizza.make(quantity=10)

Seeding the faker

from django.contrib.auth.models import User
from django_fakery import factory

factory.m(User, seed=1234, quantity=4)(
    username='regularuser_{}'
)

Credits

The API is heavily inspired by model_mommy.

License

This software is released under the MIT License.