FactoryBoy

factory_boy: Create Python Test Data with Factories

factory_boy replaces static test fixtures with configurable factories that build Python objects for individual tests. It supports multiple creation strategies and integrations with popular ORMs, making it useful for projects that need flexible, repeatable test data.

View on GitHub Compare
Stars
3.8k
Forks
424
Stars, last 30 days
+0

Our take

Health score

34/100

How it is scored
Activity
19
Community
75
Issues
21
Pull requests
20

+0 stars in the last 30 days

  • More contributors than 78% of the projects we track

Project history

Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.

Selected snapshot

Stars

3,811

Forks

427

Change since Nov 14, 2026: 0 stars·+1 forks

Overview

factory_boy is a Python library for generating test objects from reusable factory definitions. Rather than maintaining large sets of static fixtures, developers can specify defaults and override only the fields relevant to a test, including values for associated objects.

It is intended for Python projects that need object setup for tests or demos. ORM-specific factory classes support Django, MongoDB libraries, and SQLAlchemy; the project can also be extended for other libraries.

Key Features

  • Declarative factory classes with defaults for model attributes.
  • Per-call keyword overrides for test-specific values.
  • Build unsaved objects, create saved ORM instances, or return stub objects.
  • Generate batches of objects and unique sequence-based values.
  • Compose related objects with subfactories and lazy attributes.
  • Generate realistic values with Faker and reseed randomness for reproducible tests.
  • ORM integrations for Django, Mogo, MongoEngine, and SQLAlchemy.

Use Cases

  • Python application teams can replace brittle, shared fixture files with object definitions tailored to each test.
  • Django, SQLAlchemy, or supported MongoDB users can prepare model instances using ORM-aware factories.
  • Test authors can create related object graphs, such as a post with an author, without repeating setup code.
  • Developers can generate batches of representative records for tests or demos, and seed random values when failures need to be reproduced.

Getting Started

Install from PyPI:

pip install factory_boy

See the README and documentation for setup and usage details.

Alternatives

  • mixer: Mixer generates and optionally saves model instances with inferred or supplied values, while factory_boy centers on explicit, reusable factory definitions and creation strategies.
ProjectLanguageLicenseStarsStatus
factory_boyPythonMIT3.8kNot checked yet
mixerPython–953Inactive

Considerations

  • ORM persistence requires using the appropriate ORM-specific factory class; the default factory can build objects without saving them.
  • The project supports active Python versions and PyPy3, and its documented ORM support policies vary by integration. Check the documentation for compatibility details.
  • Randomized values can make failures harder to reproduce unless random seeds are controlled.

Found this useful?

Share it with someone who would like factory_boy.

Comparisons