MimicQL

The problem

You want to write maintainable tests in a codebase that interacts with a graphql API and you need mock data. You need mocked data to match the structure of your document which can be a pain given the dynamic and nested nature of graphql queries. You want to have sensible data generated for you by default and the ability to customize on a case by case basis.

The solution

MimicQL is a small but robust solution for generating mocked graphql data. It provides functions for generating mock data that will match your query, mutation, and fragment definition structures. It gives you the ability to define the way your graphql schema should resolve and execute queries, mutations, AND fragments against it.

Getting Started

Installation

This module is distributed via npm which is bundled with node and should be installed as one of your project's devDependencies:

npm install --save-dev mimicQL

or

for installation via yarn

yarn add --dev mimicQL

Create Mocker

First create a file where your mocker instance will live. To initialize your mocker you need to pass a JSON schema into the mock factory generator exposed from mimicQL.

// mocker.js
import MockFactory from "mimicQL"
import schemaJson from "path/to/schema.json"

export default new MockFactory(schemaJson)

At this point you can generate mock data but without supplying default mock resolvers but your data will consist of only default values for each type. You will more than likely want to have some sensible defaults.

Create Default Mock Resolvers

To do that we'll want to define mock resolvers. The documentation on what constitutes a mock resolver can be found here.

// resolvers/Rocket.js
import faker from "faker"

const MockRocket = () => {
  return {
    id: () => `rocket-${faker.random.uuid()}`,
    name: () => `rocket-${faker.random.word()}`,
    type: () => `rocket-${faker.random.word()}`,
  }
}

export default MockRocket

Generally you only want to define values for the shallow properties on an object. You'll leave value definition for nested types to the resolver in charge of that type. For example if a Launch has a rocket: Rocket field you should leave the definition for rocket to the Rocket resolver.

Setup Mock Server

Head back to the mocker and add the new mock Rocket resolver so it can use it for its defaults.

// mocker.js
import MockFactory from 'mimicQL'
import schemaJson from 'path/to/schema.json'
import Rocket from './resolvers/Rocket'

export default new MockFactory(
  schemaJson,
  { mocks: { Rocket }
)

Execute Query

Now with a valid graphql document we can generate some sensible mock data.

import gql from "graphql-tag"
import mocker from "./mocker"

const query = gql`
  query Rockets($id: ID!) {
    rockets {
      id
      name
      type
    }
  }
`

const mockedRocketsQuery = mocker.mockQuery(query)
/**
 * [
 *   {
 *     id: 'rocket-<uuid>',
 *     name: 'rocket-<word>',
 *     type: 'rocket-<word>',
 *     __typename: 'Rocket'
 *   },
 *   {
 *     id: 'rocket-<uuid>',
 *     name: 'rocket-<word>',
 *     type: 'rocket-<word>',
 *     __typename: 'Rocket'
 *   },
 * ]
 */

Using variables

Recipes & Examples

TODO