Mock Firestore

Jest Mock for testing Google Cloud Firestore

A simple way to mock calls to Cloud Firestore, allowing you to assert that you are requesting data correctly.

This is not a pseudo-database -- it is only for testing you are interfacing with firebase/firestore the way you expect.

⚠️ WARNING ⚠️

This library is NOT feature complete with all available methods exposed by Firestore.

Small, easy to grok pull requests are welcome, but please note that there is no official roadmap for making this library fully featured.

Table of Contents

• Mock Firestore
  • ⚠️ WARNING ⚠️
  • Table of Contents
  • What's in the Box
  • Installation
  • Usage
    • mockFirebase
    • @google-cloud/firestore compatibility
    • Subcollections
    • What would you want to test?
    • Don't forget to reset your mocks
      • I wrote a where clause, but all the records were returned!
    • Additional options
      • includeIdsInData
      • mutable
      • simulateQueryFilters
    • Functions you can test
      • Firestore
      • Firestore.Query
      • Firestore.FieldValue
      • Firestore.Timestamp
      • Firestore.Transaction
      • Auth
  • Contributing
  • Code of Conduct
  • About Upstatement

What's in the Box

This library provides an easy to use mocked version of firestore.

Installation

With npm:

npm install --save-dev firestore-jest-mock

With yarn:

yarn add --dev firestore-jest-mock

Usage

mockFirebase

The default method to use is mockFirebase, which returns a jest mock, overwriting firebase and firebase-admin. It accepts an object with two pieces:

• database -- A mock of your collections
• currentUser -- (optional) overwrites the currently logged in user

Example usage:

const { mockFirebase } = require('firestore-jest-mock');

// Create a fake Firestore with a `users` and `posts` collection
mockFirebase({
  database: {
    users: [
      { id: 'abc123', name: 'Homer Simpson' },
      { id: 'abc456', name: 'Lisa Simpson' },
    ],
    posts: [{ id: '123abc', title: 'Really cool title' }],
  },
});

If you are using TypeScript, you can import mockFirebase using ES module syntax:

import { mockFirebase } from 'firestore-jest-mock';

This will populate a fake database with a users and posts collection. This database is read-only by default, meaning that any Firestore write calls will not actually persist across invocations.

Now you can write queries or requests for data just as you would with Firestore:

const { mockCollection } = require('firestore-jest-mock/mocks/firestore');

test('testing stuff', () => {
  const firebase = require('firebase'); // or import firebase from 'firebase';
  const db = firebase.firestore();

  return db
    .collection('users')
    .get()
    .then(userDocs => {
      // Assert that a collection ID was referenced
      expect(mockCollection).toHaveBeenCalledWith('users');

      // Write other assertions here
    });
});

In TypeScript, you would import mockCollection using ES module syntax:

import { mockCollection } from 'firestore-jest-mock/mocks/firestore';

The other mock functions may be imported similarly.

@google-cloud/firestore compatibility

If you use @google-cloud/firestore, use mockGoogleCloudFirestore instead of mockFirebase in all the documentation.

const { mockGoogleCloudFirestore } = require('firestore-jest-mock');

mockGoogleCloudFirestore({
  database: {
    users: [
      { id: 'abc123', name: 'Homer Simpson' },
      { id: 'abc456', name: 'Lisa Simpson' },
    ],
    posts: [{ id: '123abc', title: 'Really cool title' }],
  },
});

const { mockCollection } = require('firestore-jest-mock/mocks/firestore');

test('testing stuff', () => {
  const { Firestore } = require('@google-cloud/firestore');
  const firestore = new Firestore();

  return firestore
    .collection('users')
    .get()
    .then(userDocs => {
      expect(mockCollection).toHaveBeenCalledWith('users');
      expect(userDocs[0].name).toEqual('Homer Simpson');
    });
});

Note: Authentication with @google-cloud/firestore is not handled in the same way as with firebase. The Auth module is not available for @google-cloud/firestore compatibility.

Subcollections

A common Firestore use case is to store data in document subcollections. You can model these with firestore-jest-mock like so:

const { mockFirebase } = require('firestore-jest-mock');
// Using our fake Firestore from above:
mockFirebase({
  database: {
    users: [
      {
        id: 'abc123',
        name: 'Homer Simpson',
      },
      {
        id: 'abc456',
        name: 'Lisa Simpson',
        _collections: {
          notes: [
            {
              id: 'note123',
              text: 'This is a document in a subcollection!',
            },
          ],
        },
      },
    ],
    posts: [{ id: '123abc', title: 'Really cool title' }],
  },
});

Similar to how the id key defines a document object to firestore-jest-mock, the _collections key defines a subcollection. You model each subcollection structure in the same way that database is modeled above: an object keyed by collection IDs and populated with document arrays.

This lets you model and validate more complex document access:

const { mockCollection, mockDoc } = require('firestore-jest-mock/mocks/firestore');

test('testing stuff', () => {
  const firebase = require('firebase');
  const db = firebase.firestore();

  return db
    .collection('users')
    .doc('abc456')
    .collection('notes')
    .get()
    .then(noteDocs => {
      // Assert that a collection or document ID was referenced
      expect(mockCollection).toHaveBeenNthCalledWith(1, 'users');
      expect(mockDoc).toHaveBeenCalledWith('abc456');
      expect(mockCollection).toHaveBeenNthCalledWith(2, 'notes');

      // Write other assertions here
    });
});

What would you want to test?

The job of the this library is not to test Firestore, but to allow you to test your code without hitting Firebase servers or booting a local emulator. Since this package simulates most of the Firestore interface in plain JavaScript, unit tests can be quick and easy both to write and to execute.

Take this example:

function maybeGetUsersInState(state) {
  const query = firestore.collection('users');

  if (state) {
    query = query.where('state', '==', state);
  }

  return query.get();
}

We have a conditional query here. If you pass state to this function, we will query against it; otherwise, we just get all of the users. So, you may want to write a test that ensures you are querying correctly:

const { mockFirebase } = require('firestore-jest-mock');

// Import the mock versions of the functions you expect to be called
const { mockCollection, mockWhere } = require('firestore-jest-mock/mocks/firestore');
describe('we can query', () => {
  mockFirebase({
    database: {
      users: [
        {
          id: 'abc123',
          name: 'Homer Simpson',
          state: 'connecticut',
        },
        {
          id: 'abc456',
          name: 'Lisa Simpson',
          state: 'alabama',
        },
      ],
    },
  });

  test('query with state', async () => {
    await maybeGetUsersInState('alabama');

    // Assert that we call the correct Firestore methods
    expect(mockCollection).toHaveBeenCalledWith('users');
    expect(mockWhere).toHaveBeenCalledWith('state', '==', 'alabama');
  });

  test('no state', async () => {
    await maybeGetUsersInState();

    // Assert that we call the correct Firestore methods
    expect(mockCollection).toHaveBeenCalledWith('users');
    expect(mockWhere).not.toHaveBeenCalled();
  });
});

In this test, we don't necessarily care what gets returned from Firestore (it's not our job to test Firestore), but instead we try to assert that we built our query correctly.

If I pass a state to this function, does it properly query the users collection?

That's what we want to answer.

Don't forget to reset your mocks

In jest, mocks will not reset between test instances unless you specify them to. This can be an issue if you have two tests that make an asseration against something like mockCollection. If you call it in one test and assert that it was called in another test, you may get a false positive.

Luckily, jest offers a few methods to reset or clear your mocks.

• clearAllMocks() clears all the calls for all of your mocks. It's good to run this in a beforeEach to reset between each test

jest.clearAllMocks();

• mockClear() this resets one specific mock function

mockCollection.mockClear();

I wrote a where clause, but all the records were returned!

The where clause in the mocked Firestore will not actually filter the data at all. We are not recreating Firestore in this mock, just exposing an API that allows us to write assertions. It is also not the job of the developer (you) to test that Firestore filtered the data appropriately. Your application doesn't double-check Firestore's response -- it trusts that it's always correct!

Additional options

The default state of this mock is meant for basic testing that should cover most everyone.
However, you can pass an options object to the mock to overwrite some default behavior.

const options = {
  includeIdsInData: true,
  mutable: true,
  simulateQueryFilters: true,
};

mockFirebase(database, options);

includeIdsInData

By default, id's are not returned with the document's data. Although you can declare an id when setting up your fake database, it will not be returned with data() as that is not the default behavior of firebase. However, a common practice for firestore users is to manually write an id property to their documents, allowing them to query collectionGroup by id.

mutable

Warning: Thar be dragons

By default, the mock database you set up is immutable. This means it doesn't update, even when you call things like set or add, as the result isn't typically important for your tests. If you need your tests to update the mock database, you can set mutable to true when calling mockFirebase. Calling .set() on a document or collection would update the mock database you created. This can make your tests less predictable, as they may need to be run in the same order.

Use with caution.

Note: not all APIs that update the database are supported yet. PRs welcome!

simulateQueryFilters

By default, query filters (read: where clauses) pass through all mock Firestore data without applying the requested filters.

If you need your tests to perform where queries on mock database data, you can set simulateQueryFilters to true when calling mockFirebase.

Functions you can test

Firestore

Firestore.Query

Firestore.FieldValue

Firestore.Timestamp

Firestore.Transaction

Auth

Contributing

We welcome all contributions to our projects! Filing bugs, feature requests, code changes, docs changes, or anything else you'd like to contribute are all more than welcome! More information about contributing can be found in the contributing guidelines.

To get set up, simply clone this repository and npm install!

Code of Conduct

Upstatement strives to provide a welcoming, inclusive environment for all users. To hold ourselves accountable to that mission, we have a strictly-enforced code of conduct.

About Upstatement

Upstatement is a digital transformation studio headquartered in Boston, MA that imagines and builds exceptional digital experiences. Make sure to check out our services, work, and open positions!