As part of my test suite I run pylint and PEP8 over my code, and when using factory_boy factories, I run into issues with the linting of lines such as:

foo = FooFactory() # Factory for Foo() models, which include a method called 'bar(self)'
foo.field = foo.bar()

Which is flagged as:

E1101: instance of 'FooFactory' has no 'bar' member. 

Is there a way to avoid this other than per line flags telling pylint to ignore E1101 which seems rather inelegant?

Hi,
I am sorry for my bad english
With the factory.Iterator and cycle=False I can't reset the iterator.
For exemple :

ExempleFactory(factory.DjangoModelFactory):
    id = factory.Iterator([1,2,3], cycle=False)

ExempleFactory.reset(('id',))

ExempleFactory(reset=('id',)

Thanks you

Not a bug or anything. I don't know any other places to ask, so here I go:

My first glance tells me factory_boy is awesome. But as I am writing down my unittest for my custom User model, I didn't really know what to test.

1. Basically, factory_boy builds reusable fixtures that is local to test code (which adds extra doc quality...)
2. I don't want real database interaction, I will use build instead.
3. Say I have a User model. If I define a UserFactory, as point 1 above, I would define some dummy values and in my test I might supply UserFactory(username='yeukhon', hashed_pwd='hashed_value').build() and assert attribute hashed_pwd has the value hashed_value?

I guess the confusion is whether doing point 2 actually means testing my models (unittest). In the old days I would mock.patch out the model when I am interacting with them.

Maybe someone can reassure me.

Thanks,

This is more a question than an issue since I didn't find a mailing list where I could ask this question. Apologize if this isn't the place to ask.

I'm trying to use factory_boy without an ORM or Django. I've no experience in either.

So I've got a library in python which is used as a test framework. Over time the test data that was managed using dictionaries has grown in size that it has become unweildy to manage. Much of this test data needs to be passed to my functional tests. So, I decided to use factory_boy for defining this test-data as a representation of factory_boy managed objects.

My object model is similar to the examples shown in factory_boy's wiki page:

AccountFactory (inherit) AdminAccountFactory (inherit) UserAccountFactory

Similarly
NetworkFactory (inherit) IsolatedNetworkFactory (inherit) SharedNetworkFactory and so on...

So my questions are as follows

a) Is factory_boy the right tool for addressing this problem?
b) If yes, can it be used without dependency on Django, Mojo and/or any ORM?
c) If no, what (to your knowledge) would best solve my test management problem?

Thanks in advance,

is it possible to get rid of the ABSTRACT_FACTORY attribute?

At the moment, not providing a FACTORY_FOR raises the following error, even if the factory itself isn't used outside of its child classes:

AssociatedClassError: Could not determine the class associated with BaseFactory. Use the FACTORY_FOR attribute to specify an associated class.

One issue I've come up with is that I need a RelatedFactory that works like CircularSubFactory does: where I can pass in a class name to get a late import. However, rather than adding CircularRelatedFactory, doesn't it make sense to get rid of the idea of a CircularSubFactory and just allow users to pass in either a Class or a string to a class into SubFactory?

The init method could then just check if it's a callable or a string.

Assuming you use a single string to define the class/module (which admittedly maybe isn't backwards compatible) it would be:

class MyFactory
    employee1 = factory.SubFactory('tests.factories.EmployeeFactory')
    employee2 = factory.SubFactory(EmployeeFactory)
    employee3 = factory.RelatedFactory('tests.factories.EmployeeFactory', 'employee')
    employee4 = factory.RelatedFactory(EmployeeFactory, 'employee')

Hi,

i'm trying to use factory boy with some nested factories but i get a PRIMARY KEY error (.. must be unique).

these are my factories

class CategoryParentFactory(factory.Factory):
    FACTORY_FOR = Category
    parent = None
    name = "foo"

class CategoryFactory(factory.Factory):
    FACTORY_FOR = Category
    parent = factory.SubFactory(CategoryParentFactory)
    name = "bar"

class ProductFactory(factory.Factory):
    FACTORY_FOR = Product
    category = factory.SubFactory(CategoryFactory)
    name = "fuz"

if i try something like

cp = CategoryParentFactory.create()
c = CategoryFactory.create()

it works, but

p = ProductFactory.create()

gives me "IntegrityError: PRIMARY KEY must be unique"

How can i debug/understand what happens?

Thanks,

ps: i'm trying to get many products with a single category, so this

p1 = ProductFactory.create(category=c)
p1 = ProductFactory.create(category=c)

gives me the same error

Alessandro

For my model I'm using custom strategies for _build and _create. My _build strategy creates a plain python object which has necessary attributes to make a web server request. The _create strategy makes the actual request and includes additional attributes into the model from the response received. All goes well until I have factories that relate to each other.

I would like to ensure that when a factory X is 'created' then the factory Y (y contains x as RelatedFactory) also gets created. To create X: I do

class X(factory.Factory):
     FACTORY_FOR = EntityX

     EntityY = factory.RelatedFactory(Y)


X.create(webserverConnection)

but webserverConnection isn't passed when the RelatedFactory Y is created. How can I achieve this?

I am trying to work out how to make factoryboy play nice with neo4django

If I have a simple model defined as so:

from neo4django.db import models

class Country(models.NodeModel):
    code = models.StringProperty(indexed=True)

And then define a factory for the model like so:

import factory
from my.models import Country

class CountryFactory(factory.Factory):
    FACTORY_FOR = Country
    code = 'DE'

CountryFactory.default_strategy = factory.CREATE_STRATEGY

In the very simple example below I get an empty queryset instead of it containing the country I just created.

 import factory
 from .factories import CountryFactory                                                                               
 from my.models import Country

  def test_country():
      c1 = CountryFactory.create()
      pp(c1) # <Country> as expected
      country = Country.objects.all()
      pp(countries) # []

I have spent an afternoon, comparing the example above with the django orm version (which works), tracing each through the debugger to no avail. I was wondering if you would maybe able to point me in the direction of the right bits of code. Happy to supply patches if I get anything working.

I'm just starting, sorry if i'm wrong. In the Subfactory docs:

http://readthedocs.org/docs/factoryboy/en/latest/subfactory.html

The first code example is something similar to this:

# A standard factory
class UserFactory(factory.Factory):
    FACTORY_FOR = User

    # Various fields
    first_name = 'John'
    last_name = factory.Sequence(lambda n: 'D%se' % (o * n))
    email = factory.LazyAttribute(lambda o: '%s.%[email protected]' % (o.first_name.lower(), o.last_name.lower()))

In this piece of code, I don't see o defined anywhere:

    # Where's "o" defined?
    last_name = factory.Sequence(lambda n: 'D%se' % (o * n))

As suggested by B. Randel, the documentation lacks on such subjects.

The new section should also cover "through" m2m relationships.

It doesn't seem possible to use list comprehension in argument of InfiniteIterator().

As a result, the errors like this take place

TypeError: unbound method _create() must be called with <Factory> instance as first argument (got ModelBase instance instead)

Offending code is here:

    @classmethod
    def _setup_next_sequence(cls):
        """Compute the next available PK, based on the 'pk' database field."""

        model = cls._associated_class  # pylint: disable=E1101
        manager = cls._get_manager(model)

        try:
            return 1 + manager.values_list('pk', flat=True
                ).order_by('-pk')[0]
        except IndexError:
            return 1

This problem didn't exist in factory_boy 1.3. My field that was using a non-integer PK is using a sequence, which worked fine previously:

code = factory.Sequence(lambda n: str(n).zfill(3))

I haven't dug into the code enough to know much about the changes that caused this problem, but ultimately I'd like to be able to use the sequence for the field again.

What does everyone think about adding a load/dump API that takes/consumes lists of dictionaries? This would allow for easy consuming and producing JSON / YAML fixtures.

Something like the following.

import factory
from . import models

class UserFactory(factory.Factory):
    FACTORY_FOR = models.User

    first_name = 'John'
    last_name = 'Doe'
    admin = False

# Deserialize an iterable of dictionaries using the factory.
UserFactory.load([{...}, {...}])

# Serialize an iterable of targets to an list of dictionaries
# factoring out attributes provided by factories, etc. 
data = UserFactory.dump([models.User(first_name="Bob"), 
                         models.User()])

assert 'first_name' not in data[1]

Hi There,

This is for Django. We have a basic User model. When a User object is created a trigger is fired and UserProfile is automatically created.

Ultimately I want to be able to do this:

user = BasicUserFactory(
    username="Foo", userprofile__title="Bar", 
    userprofile__company__name="Co")

How can I create a UserFactory which creates and sets attributes for that new UserProfile? Right now it creates it (but I think this is b/c of the trigger) but I can't seem to set any attributes presumably b/c it already exists.

# models.py
class Company(models.Model):
    name = models.CharField(max_length=32, null=True)

class UserProfile(models.Model):
    user = models.OneToOneField(User)
    company = models.ForeignKey('company.Company', related_name="users")
    title = models.CharField(max_length=32, null=True)
    department = models.CharField(max_length=16, blank=True)

# class User == contrib.auth.models.User

#factories.py
class BaseUserProfileFactory(factory.DjangoModelFactory):
    FACTORY_FOR = UserProfile
    FACTORY_DJANGO_GET_OR_CREATE = ('user', )
    company = factory.SubFactory(
       'apps.company.tests.factories.GeneralOrganizationFactory',
        is_customer=True, name="GeneralCo")
    title = "King of the world.."

class BasicUserFactory(factory.DjangoModelFactory):
    """A base user factory"""
    FACTORY_FOR = User
    FACTORY_DJANGO_GET_OR_CREATE = ('username', )
    username = factory.Sequence(lambda n: 'user_{0}'.format(n))
    first_name = factory.Sequence(lambda n: 'first_{0}'.format(n))
    last_name = factory.Sequence(lambda n: 'last_{0}'.format(n))
    email = factory.Sequence(lambda n: 'user_{0}@home.com'.format(n))
    is_active = True
    is_staff = False
    is_superuser = False
    password = factory.PostGenerationMethodCall('set_password', 'password')

    userprofile = factory.RelatedFactory(
BaseUserProfileFactory, 'user', )

There is no way to avoid the creation of an object via RelatedFactory even when this related param is set to None via kwargs in factory instantiation. Example:

** Factory file **

MyFactory(factory.DjangoModelFactory):
param1 = factory.RelatedFactory('model_file.MyModel')

** Test file **

obj = MyFactory(param1=None) # param1 exists in db

Regards,

I'm having fun using factories, but there are some "initial" - "always the same" - "needed by the app" data that i load with fixtures (and it's not so funny to recreate this with factories).

Now, what i want to achieve is this:

class FooFactory(factory.Factory):
    FACTORY_FOR = Foo
    bar = Bar.objects.get(id=1)

i have tried but i get this error executing the test
django.db.utils.DatabaseError: no such table: bar

I presume that it happens because the factories import execute before the database creation.

Any hint? Can i use the LazyAttribute in some way?

Thanks

PS. I ask here because i don't think there is a dedicated google groups, right?

As discussed today on IRC, we need a simple way to manage circular dependencies:

A.b --> B
B.a --> A

Thanks!

If you define a Factory with ABSTRACT_FACTORY set to True, subclasses do not inherit any declarations from the abstract factory. Is this by design?

Hi there,

Oh factory_boy where have you been all of my life. This thing is AWESOME!!

I can't figure out how to perform an action on a SelfAttribute. I want to transform it a bit..

class RaterUserFactory(NonCompanyProfile):
    """A factory for a rater user"""
    FACTORY_FOR = UserProfile
    FACTORY_DJANGO_GET_OR_CREATE = ('user', )
    user = factory.SubFactory(BaseUserNoProfileFactory, username="rater")
    company = factory.SubFactory('apps.company.tests.RaterOrganizationFactory', is_customer=True,
                                 name=factory.SelfAttribute('..user.username').capitalize() + "Co")```

This is example from documentation

class MyFactory(factory.Factory):
    blah = factory.PostGeneration(lambda obj, create, extracted, **kwargs: 42)

MyFactory(
    blah=42,  # Passed in the 'extracted' argument of the lambda
    blah__foo=1,  # Passed in kwargs as 'foo': 1
    blah__baz=2,  # Passed in kwargs as 'baz': 2
    blah_bar=3,  # Not passed
)

But in real, 'foo' and 'baz' will not occurs in kwargs dictionary when blah argument is set.

In my tests:

MyFactory(blah=42,  blah__foo=1) # results in extracted=42 and kwargs={}
MyFactory(blah__foo=1) # results in extracted=None and kwargs={'foo'}

tested on latest master

The README says that I should use a LazyAttribute(lambda o: OtherFactory()) in order to have a "SubFactory" on one of my factories.

Readthedocs, however actually has a chapter about factory.SubFactory AND it has a chapter about factory.RelatedFactory

I am confused. Let's say I have a UserFactory and a UserProfileFactory (very commong for Django projects, I guess)... which of the three documented approaches would be the correct one?

Hello,

I need to get a full dump of a Factory as a dict so that I can serialize it to json. FactoryName.attributes() works great except when there is a SubField, in which case it evaluates it and returns an instantiated version of the object as opposed to a dictionary of key: value of the SubField attributes.

I tried to hack around this by traversing the Factory myself but it turns out that I'll have to duplicate a lot of code that already exists. Do you see an easy way to do this?

When using .build() to create an instance of a model (in this case, using Django), the resulting objects, even if attributes are different, compare as the same.

Example:

object1 = PersonFactory.build(name="Bob")
object2 = PersonFactory.build(name="Joe")
if object1 == object2:
    print "Oops."

If I create the objects, the comparison fails. If I build them (without hitting the DB), the comparison resolves as True.

During unit testing, it's often desirable to test with randomized values as opposed to hard-coded values, see Fuzzy Testing. While it's easy enough to add randomization with the random package, the problem is that evaluation of the randomized result is performed once at the class level. We would like to have randomization performed for each object created.

No matter how I try it. I can't seem to get a reference to DjangoModelFactory.

I tried the following like the tests do.

import factory.django

# or inheriting from 

factory.django.DjangoModelFactory

It keeps complaining about not being able to find django.

So we've got a set of SQLAlchemy models like so (don't ask why, it makes me sad).

class Test(Base):
    __tablename__ = "test"

    id = sqlalchemy.Column('id_', sqlalchemy.Integer, primary_key=True)

When using a factory for the class we get:

Traceback (most recent call last):
  File "<pyshell#0>", line 16, in <module>
    TestFactory.build()
  File "/usr/local/lib/python2.7/dist-packages/factory/base.py", line 434, in build
    attrs = cls.attributes(create=False, extra=kwargs)
  File "/usr/local/lib/python2.7/dist-packages/factory/base.py", line 316, in attributes
    force_sequence=force_sequence,
  File "/usr/local/lib/python2.7/dist-packages/factory/containers.py", line 263, in build
    sequence = self.factory._generate_next_sequence()
  File "/usr/local/lib/python2.7/dist-packages/factory/base.py", line 287, in _generate_next_sequence
    cls._next_sequence = cls._setup_next_sequence()
  File "/usr/local/lib/python2.7/dist-packages/factory/alchemy.py", line 37, in _setup_next_sequence
    pk = getattr(model, model.__mapper__.primary_key[0].name)
AttributeError: type object 'Test' has no attribute 'id_'

A full example:

from sqlalchemy.ext.declarative import declarative_base
import sqlalchemy
import factory.alchemy

Base = declarative_base()

class Test(Base):
    __tablename__ = "test"

    id = sqlalchemy.Column('id_', sqlalchemy.Integer, primary_key=True)

class TestFactory(factory.alchemy.SQLAlchemyModelFactory):
    FACTORY_FOR = Test
    FACTORY_SESSION = None

TestFactory.build()

Whenever the django.db.Model is not found in the inheritance chain the BUILD_STRATEGY should be used automatically instead of the CREATE_STRATEGY.

Through logging.debug, obviously.

This would trace important steps:

• Class mangling
• Call chains

I am using the setting FACTORY_DJANGO_GET_OR_CREATE. But in some cases, when I ask for an existing object with an existing SubFactory object, it creates an unused object despite this setting.

For example, in a brand new project, I typed:

# models.py
from django.db import models

class A(models.Model):
    name = models.CharField(max_length=10)

class B(models.Model):
    name = models.CharField(max_length=10)
    a = models.ForeignKey(A)

And

# factories.py
import factory

from . import models

class AFactory(factory.DjangoModelFactory):
    FACTORY_FOR = models.A
    FACTORY_DJANGO_GET_OR_CREATE = ('name',)

    name = factory.Sequence(lambda n: 'A-{0}'.format(n))

class BFactory(factory.DjangoModelFactory):
    FACTORY_FOR = models.B
    FACTORY_DJANGO_GET_OR_CREATE = ('name',)

    name = factory.Sequence(lambda n: 'B-{0}'.format(n))
    a = factory.SubFactory(AFactory)

Then:

from factories import *

a = AFactory(name="Apple")
models.A.objects.all()
# one object
b = BFactory(a__name="Apple", name="Beetle")
models.B.objects.all()
models.A.objects.all()
# one A object, one B object
b = BFactory(name="Beetle")
models.B.objects.all()
models.A.objects.all()
# still one B object, but now a new, unused A object too

The final call to BFactory has brought into being a new object of class A, even though the B object with name Beetle already exists (and is not re-created). Why, and how do I stop this new A object being created?

(I know I can get around this by calling instead:

b = BFactory(name="Beetle", a__name="Apple")

but in my actual use case, I have several dependencies and levels of hierarchy, and it's messy to supply extra redundant parameters this way - and I can't seem to get the right combination of parameters.)

Thanks!

(I also asked this on StackOverflow here too.)

We have a model which has a M2M relation to the user model. So in our factory we do this:

@classmethod
def _prepare(cls, create, **kwargs):
    obj = super(TaskFactory, cls)._prepare(create, **kwargs)
    if create:
        obj.users.add(UserFactory())
    return obj

It's a bit annoying to do this. Wouldn't it be sexy if we could do something like this:

class MyFactory(factory.Factory):
    ...
    users = factory.M2MSubfactory(UserFactory, 2)

So that would mean: Please add two users. Do you think this would be possible? Would it be a difficult change? Does it make sense at all?

factory_boy can be used to replace django fixtures.

Fixtures

• create model instances using instance.save_base(raw=True, **kwargs). raw=True is passed to post_save handlers
• load data for parent and child model separately in case of multi-table inheritance (MTI)

DjangoModelFactory

• creates models instances using models manager's create method. post_save handlers are invoked with raw=False
• calls child save method in case of MTI, no need to save parent model separately

Issue objective

• support raw kwarg for DjangoModelFactory.create: MyModelDjangoFactory.create(raw=True, **kwargs). When raw is True create method calls save_base instance method
• handle MTI case and call save_base for all models from hierarchy for given model

Could you explain what is the best practice to use factory boy to load initial data to a development database ?

Suppose the following setup:

class RegionFactory(Factory):
    pass

class BuildingFactory(Factory):
    region = SubFactory(RegionFactory)

class PersonFactory(Factory):
   region = SubFactory(RegionFactory)

class RoomFactory(Factory):
   building = SubFactory(BuildingFactory)
   person = SubFactory(PersonFactory)

If I were to construct a RoomFactory, this would result in two RegionFactorys being created (room.building.region, room.person.region).

What I've thus far been doing to solve this problem is the following:

region = RegionFactory()
room = RoomFactory(
    building__region=region,
    person__region=region,
)

This is a common pattern. Is there anyway to accomplish this within a factory definition, so I do not have to repeat this pattern?

Thank you!

I'm using LazyAttribute to check if a foreign key exists and return it or else create it. Currently there isn't a way in that function to know the build strategy so it can know if it should build or create the model it is returning.

I have factory

import factory
class GroupFactory(factory.Factory):
    FACTORY_FOR = Group    
    remote_id = factory.Sequence(lambda n: n)

Method create doesn't save model, I should call it especially:

In [1]: from vkontakte_groups.factories import GroupFactory    
In [2]: group = GroupFactory.create()    
In [3]: group.id, group.pk
Out[3]: (None, None)    

In [4]: group.save()    
In [5]: group.id, group.pk
Out[5]: (890, 890)    

In [6]: import factory    
In [7]: factory.__version__
Out[7]: '2.1.0-dev'

That is very strange. Before it was behaviour with saving by default and I don't know what was changed.. Should I post here more environment settings?

Primary key of model is not necessary to be named as id.
pk would be better.

https://github.com/rbarrois/factory_boy/blob/master/factory/base.py#L579

I had to dig throguh the code and docs to figure out what I should do.
Mention that use_strategy is a decorator and show useage.

Factory Boy 2.0.0

Given the following class:

class TagFactory(factory.DjangoModelFactory):
    FACTORY_FOR = Tag

    name = factory.Sequence(lambda n: 'tag_{}'.format(n))

all Tag objects have the same name. It works fine with factory.Factory

I have the following:

from factory import Factory
from factory.declarations import PostGenerationMethodCall
from django.contrib.auth.models import User

class UserFactory(Factory):
    FACTORY_FOR = User

    username = 'john'
    first_name = 'John'
    last_name = 'Doe'
    email = '[email protected]'

    password = PostGenerationMethodCall('set_password', password='123456')

and when only importing the file where that code is, I'm getting:

Traceback (most recent call last):
  File "<console>", line 1, in <module>
  File "C:\django-project\tests\__init__.py", line 15, in <module>
    class UserFactory(factory.Factory):
  File "C:\django-project\tests\__init__.py", line 25, in UserFactory
    method_args='123456')
  File "C:\Python27\lib\site-packages\factory\declarations.py", line 356, in __init__
    super(RelatedFactory, self).__init__(extract_prefix)
TypeError: super(type, obj): obj must be an instance or subtype of type

I was reading the documentation, and I found the PostGenerationMethodCall, which is exactly what I need, but I couldn't find that piece of code when installing factory_boy using pip install factory_boy, so I got the code directly from the repo.

When creating a factory for a model that has a FileField or ImageField, it would be nice to have some helper factories.

My current solution is this:

import StringIO

import factory
from PIL import Image

from django.core.files.uploadedfile import InMemoryUploadedFile

from apps.photos.models import Photo


def django_image(name, size=200, color='red'):
    thumb = Image.new('RGB', (size, size,), color)
    thumb_io = StringIO.StringIO()
    thumb.save(thumb_io, format='JPEG')
    thumb_io.seek(0)
    return InMemoryUploadedFile(thumb_io, None, name, 'image/jpeg', thumb_io.len, None)


class PhotoFactory(factory.DjangoModelFactory):
    FACTORY_FOR = Photo

    @factory.post_generation
    def image(self, create, extracted, **kwargs):
        if extracted:
            image_name, image = extracted
        else:
            image_name = 'image.jpg'
            image =django_image(image_name, **kwargs)
        self.image.save(image_name, image)

As moved over from dnerdy/factory_boy#9

@rbarrois said

By the way, I'm strongly against the proposed behavior:

• Django's way could change in the future, and requires a single model for each app
• factory_boy aims to handle more situations, ORMs and frameworks than just Django, and should keep a consistent behavior across all projects.

The convention of identifying model instances by app.ModelName is certainly not going anywhere in Django at present - I'm quite familiar with the app-refactor branch and this does not attempt to change this convention at all. As an example, consider the recently introduced AUTH_USER_MODEL setting which follows this convention.

I understand that factory_boy aims to support multiple ORMs etc, which is why the patch on the other project is applied only to DjangoModelFactory, which is already Django specific.

My main reason for doing this is that my factories.py have a large number of factories defined in them, and I generally do not require the model class for anything other than setting FACTORY_FOR - it results in a lot of imports to the file which aren't really needed.

If you are willing to reconsider your opposition to the proposed behaviour, then I can upgrade the patch apporpriately.

I have problem with upgrade from 1.3 to 2.1.1. This may be caused by my Django version (1.4.5). You should document what Django versions are supported.

It seems there is no easy way to create a User and a Profile using UserFactory (not the ProfileFactory) with factory_boy now. Creating User with Profile factory is not always a solution because there can be several profiles, and it is overall not nice to create main model through one of the side models with extra info.

What I've tried:

class ProfileFactory(DjangoModelFactory):
    FACTORY_FOR = Profile
    gender = 'F'

#   @factory.container_attribute
#   def user(self, cont):           
#      return cont[0]

#   @factory.container_attribute
#   def user_id(self, cont):           
#      return cont[0].id


class UserFactory(DjangoModelFactory):
    FACTORY_FOR = User        
    profile = SubFactory(ProfileFactory)

It is possible to use _prepare method to create the profile but this way 'profile__gender' kwarg won't work (and it is not easier than writing creation function without factory-boy).

Are there better ways to solve this problem or is it better to dig into sources and try to craft a pull request?

See https://gist.github.com/3002546 :

This should work:

class MyFactory(factory.Factory):
  foo = {
    'a': factory.InfiniteIterator([1, 2, 3]),
  }

A good implementation would be:

class MyFactory(factory.Factory):
  foo = factory.DictFactory(a=factory.InfiniteIterator([1, 2, 3]))

  bar = factory.DictFactory({
    'a': factory.InfiniteIterator([1, 2, 3]),
  })

Using for the most part the example code in RTD:

import factory
import random

#
# Object Classes
#
class Account(object):
    def __init__(self, username, email):
        self.username = username
        self.email = email

    def __str__(self):
        return '%s (%s)' % (self.username, self.email)


class Profile(object):

    GENDER_MALE = 'm'
    GENDER_FEMALE = 'f'
    GENDER_UNKNOWN = 'u'  # If the user refused to give it

    def __init__(self, account, gender, firstname, lastname, planet='Earth'):
        self.account = account
        self.gender = gender
        self.firstname = firstname
        self.lastname = lastname
        self.planet = planet

    def __unicode__(self):
        return u'%s %s (%s)' % (
            unicode(self.firstname),
            unicode(self.lastname),
            unicode(self.account.username),
        )


#
# Factories
#
class AccountFactory(factory.Factory):
    FACTORY_FOR = Account

    username = factory.Sequence(lambda n: 'john%s' % n)
    email = factory.LazyAttribute(lambda o: '%[email protected]' % o.username)


class ProfileFactory(factory.Factory):
    FACTORY_FOR = Profile

    account = factory.SubFactory(AccountFactory)
    gender = random.choice([Profile.GENDER_MALE, Profile.GENDER_FEMALE])
    firstname = u'John'
    lastname = u'Doe'

if __name__ == '__main__':
    accounts1 = [AccountFactory.build() for n in range(10)]
    profiles = [ProfileFactory.build() for n in range(10)]
    accounts2 = [AccountFactory.build() for n in range(10)]

    print 'Directly created Accounts through AccountFactory:'
    print '\n'.join([a.username for a in accounts1])
    print '\nIndirectly created Accounts through SubFactory within ProfileFactory:'
    print '\n'.join([p.account.username for p in profiles])
    print '\nMore directly created Accounts through AccountFactory:'
    print '\n'.join([a.username for a in accounts2])

Notice in __unicode__ method of Profile, self.account.accountname had to be changed to self.account.username.

When running the above, the output is:

Directly created Accounts through AccountFactory:
john0
john1
john2
john3
john4
john5
john6
john7
john8
john9

Indirectly created Accounts through SubFactory within ProfileFactory:
john10
john12
john14
john16
john18
john20
john22
john24
john26
john28

More directly created Accounts through AccountFactory:
john30
john31
john32
john33
john34
john35
john36
john37
john38
john39

Just looked at it quickly, but seems to be an issue with multiple calls of _generate_next_sequence within class hierarchy.

Hi there,

I took this example RelatedFactory and modified it with the "name" field.

Here is my source code:

# models.py
class User(models.Model):
    pass

class UserLog(models.Model):
    user = models.ForeignKey(User)
    name = models.CharField()
    action = models.CharField()


# factories.py
class UserFactory(factory.DjangoModelFactory):
    FACTORY_FOR = models.User

    log = factory.RelatedFactory(UserLogFactory, 'user', name='test', action='create')

This caused the following error:

TypeError: __init__() got multiple values for keyword argument 'name'

I have a ContactListFactory and a ContactFactory. Each ContactFactory has a related contact_list:

class ContactListFactory(factory.Factory):
    FACTORY_FOR = ContactList

    name = factory.LazyAttributeSequence(lambda _, n: 'Contact List {}'.format(n))


class ContactFactory(factory.Factory):
    FACTORY_FOR = Contact

    email = factory.LazyAttributeSequence(lambda _, n: '{}@example.com'.format(n))
    contact_list = factory.SubFactory(ContactListFactory)

Is it possible that when I create a ContactListFactory, to auto-create a Contact factory that has that list as it's contact_list?