Mongo database mocking with fixtures

mocking, mongodb, pytest, python
pip install humongous==1.3.0


What is this?

This is a pytest plugin, that enables you to test your code that relies on a database connection to a MongoDB and expects certain data to be present. It allows you to specify fixtures for database collections in JSON/BSON or YAML format. Under the hood we use the mongomock library, that you should consult for documentation on how to use MongoDB mock objects. If suitable you can also use a real MongoDB server.

Note: This project has been renamed from humongous to pytest-mongodb in order to conform to the pytest plugin naming convention and to be easier to find on the Python package index. See the migration section for more information.


If you don't want to put your database fixtures on the top-level directory of your package you have to specify a directory where pytest-mongodb looks for your data definitions.

To do so put a line like the following under the pytest section of your pytest.ini-file put a

mongodb_fixture_dir =

pytest-mongodb would then look for files ending in .yaml or .json in that directory.

If you want only a subset of the available fixtures to be loaded, you can use the mongodb_fixtures config option. It takes a list of collection file-names without the file-extension. E.g.:

mongodb_fixtures =

In this case only the collections "players" and "championships" will be loaded.

You can also choose to use a real MongoDB server for your tests. In that case you might also want to configure the hostname and/or the credentials if you don't want to stick with the default (localhost and no credentials). Use the following configuration values in your pytest.ini to adapt the settings to your needs:

mongodb_engine = pymongo
mongodb_host = mongodb://user:passwd@server.tld
mongodb_dbname = mydbname

For Mac users, who installed mongodb using homebrew, you can configure the executable to be picked up from /usr/local/bin/mongod instead of /usr/local/bin/mongod by using mongo_exec = /usr/local/bin/mongod in the pytest.ini file.

Basic usage

After you configured pytest-mongodb so that it can find your fixtures you're ready to specify some data. Regardless of the markup language you choose, the data is provided as a list of documents (dicts). The collection that these documents are being inserted into is given by the filename of your fixture-file. E.g.: If you had a file named players.yaml with the following content:

- name: Mario
  surname: Götze
  position: striker

- name: Manuel
  surname: Neuer
  position: keeper

you'd end up with a collection players that has the above player definitions inserted. If your fixture file is in JSON/BSON format you can also use BSON specific types like $oid, $date, etc.

You get ahold of the database in your test-function by using the mongodb fixture like so:

def test_players(mongodb):
    assert 'players' in mongodb.list_collection_names()
    manuel = mongodb.players.find_one({'name': 'Manuel'})
    assert manuel['surname'] == 'Neuer'

For further information refer to the mongomock documentation.

If you want to skip specific tests if the engine is ie. a mongomock engine you could do that like so:

from pytest_mongodb.plugin import mongo_engine
from pytest import mark

@mark.skipif(mongo_engine() == 'mongomock', reason="mongomock does not support that")
def test_players(mongodb):
    assert 'players' in mongodb.list_collection_names()
    manuel = mongodb.players.find_one({'name': 'Manuel'})
    assert manuel['surname'] == 'Neuer'

Migration from humongous

In the course of migrating the package name from humongous to pytest-mongodb most configuration values which previously were prefixed with humongous_ have been renamed to a mongodb_-prefixed counterpart. The only notable exception is the humongous_basedir config value, which now is named mongodb_fixture_dir. Additionally the commandline options have been unified, in a way that multi-word option names are now consistently separated with dashes instead of underscores.