Inspired by model-bakery, this project aims to simplify the process of creating MongoDB documents for testing purposes. The goal is to deliver a maintainable, intuitive, and developer-friendly API specifically designed for MongoDB. By streamlining the generation of test data, this tool will empower developers to efficiently create realistic document structures, enhancing the testing workflow for applications that rely on MongoDB as their primary database.
- To have the conveniences of model_bakery (from the Django world) in Flask with MongoEngine.
- We want more context within the test itself (instead of having fixtures in conftest where we don't know which fields are populated).
- We don't want to create a Factory for every Document in the application.
uv add mongo-bakerymongo_bakery fills in a MongoEngine Document (or EmbeddedDocument) with realistic fake data so you don't have to
hand-write every field in every test. All the examples below assume:
from mongoengine import BooleanField, Document, IntField, StringField
from mongo_bakery import baker
class Customer(Document):
name = StringField(required=True)
email = StringField(required=True)
company = StringField(required=True)
phone_number = StringField(required=True)
loyalty_points = IntField(required=True)
notes = StringField(required=False)
newsletter_opt_in = BooleanField(required=False)baker.make instantiates and saves the document, automatically generating values for every required field:
customer = baker.make(Customer)Pass _quantity to create and save several instances at once. baker.make returns a list when _quantity > 1:
customers = baker.make(Customer, _quantity=5)
len(customers) # 5Optional fields (required=False) are not filled in automatically — baker.make only generates data for
required fields, leaving optional ones at their MongoEngine default (e.g. None). Pass any field, required or not,
as a keyword argument to set it explicitly:
customer = baker.make(Customer, notes="VIP client", newsletter_opt_in=True)Keyword arguments always win over generated data, so this also works to override a required field with a specific value.
For StringFields, mongo_bakery checks whether Faker has a provider method whose
name matches the field's name, and uses it when available. That's why, on the Customer example above,
email, company and phone_number come out looking like real data instead of a random word:
customer = baker.make(Customer)
customer.email # e.g. "jean23@example.com" instead of a random word
customer.company # e.g. "Smith, Doe and Partners"
customer.phone_number # e.g. "+1-555-019-2837"Name your fields after a Faker provider (address, city,
job, url, ...) to get more meaningful fake data for free. Fields without a matching provider fall back to a
random word.
When a field declares choices, baker.make always picks one of the allowed values, so the generated document
passes MongoEngine's validation:
class Order(Document):
status = StringField(required=True, choices=["pending", "shipped", "delivered"])
order = baker.make(Order)
order.status in ["pending", "shipped", "delivered"] # always TruePass baker.seq(value) as a kwarg to get a different, incrementing value on each instance instead of the same
value repeated across _quantity instances. It supports str, int, float, date and datetime base values:
customers = baker.make(Customer, name=baker.seq("Chad"), _quantity=3)
[customer.name for customer in customers] # ["Chad1", "Chad2", "Chad3"]Use increment_by to control the step (a timedelta for date/datetime values) and start to control the
first value of the sequence (defaults to increment_by):
customers = baker.make(Customer, loyalty_points=baker.seq(0, increment_by=10, start=100), _quantity=3)
[customer.loyalty_points for customer in customers] # [100, 110, 120]Call baker.seed(value) to seed Faker's random generator, so baker.make produces the same mock data across runs
— useful for debugging a flaky test or reproducing a specific failure:
baker.seed(1234)
customer = baker.make(Customer) # always the same field values for this seedEmbeddedDocumentField and ReferenceField are resolved recursively with baker.make, so nested documents are
created for you as well.
baker.make keeps track of every instance it saved. Call baker.cleanup() (e.g. in a test teardown/fixture) to
delete them all:
baker.cleanup()mongo_bakery also ships as a pytest plugin, registered automatically once it's installed. Use the baker fixture
instead to get this cleanup for free after every test:
def test_something(baker):
customer = baker.make(Customer)
...
# cleanup() is called automatically once the test finishesSee the API Reference for the full Baker interface.
We welcome contributions to the mongo_bakery project! In short: fork the repo, install dependencies with
uv sync, make your change on a branch, and open a pull request against main.
See CONTRIBUTING.md for the full guide, including code style, testing conventions, commit message format, and the pull request checklist.
Made with contrib.rocks.