Mongo Bakery

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 empowers developers to efficiently create realistic document structures, enhancing the testing workflow for applications that rely on MongoDB as their primary database.

Motivation

• 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.

Installation

uv add mongo-bakery

Usage

mongo_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)

Creating a basic Document

baker.make instantiates and saves the document, automatically generating values for every required field:

customer = baker.make(Customer)

Creating multiple instances with _quantity

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)  # 5

Not-required (optional) fields

Optional 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.

Fields with realistic values

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.

Fields restricted with choices

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 True

Reproducible data with baker.seed

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 seed

Embedded and referenced Documents

EmbeddedDocumentField and ReferenceField are resolved recursively with baker.make, so nested documents are created for you as well.

Cleaning up

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 finishes

See the API Reference for the full Baker interface.

Contributing

Contributions are welcome! See the Contributing guide in the project README for how to get started.