Skip to content

Patch fetch/XMLHttpRequest to fake a REST API server in the browser, based on JSON data.

License

Notifications You must be signed in to change notification settings

marmelab/FakeRest

Repository files navigation

FakeRest

A browser library that intercepts AJAX calls to mock a REST server based on JSON data.

Use it in conjunction with MSW, fetch-mock, or Sinon.js to test JavaScript REST clients on the client side (e.g. single page apps) without a server.

See it in action in the react-admin demo (source code).

Installation

npm install fakerest --save-dev

Usage

FakeRest lets you create a handler function that you can pass to an API mocking library. FakeRest supports MSW, fetch-mock, and Sinon. If you have the choice, we recommend using MSW, as it will allow you to inspect requests as you usually do in the dev tools network tab.

MSW

Install MSW and initialize it:

npm install msw@latest --save-dev
npx msw init <PUBLIC_DIR> # eg: public

Then configure an MSW worker:

// in ./src/fakeServer.js
import { http } from 'msw';
import { setupWorker } from "msw/browser";
import { getMswHandler } from "fakerest";

const handler = getMswHandler({
    baseUrl: 'http://localhost:3000',
    data: {
        'authors': [
            { id: 0, first_name: 'Leo', last_name: 'Tolstoi' },
            { id: 1, first_name: 'Jane', last_name: 'Austen' }
        ],
        'books': [
            { id: 0, author_id: 0, title: 'Anna Karenina' },
            { id: 1, author_id: 0, title: 'War and Peace' },
            { id: 2, author_id: 1, title: 'Pride and Prejudice' },
            { id: 3, author_id: 1, title: 'Sense and Sensibility' }
        ],
        'settings': {
            language: 'english',
            preferred_format: 'hardback',
        }
    }
});
export const worker = setupWorker(
    // Make sure you use a RegExp to target all calls to the API
    http.all(/http:\/\/localhost:3000/, handler)
);

Finally, call the worker.start() method before rendering your application. For instance, in a Vite React application:

import React from "react";
import ReactDom from "react-dom";
import { App } from "./App";
import { worker } from "./fakeServer";

worker.start({
    quiet: true, // Instruct MSW to not log requests in the console
    onUnhandledRequest: 'bypass', // Instruct MSW to ignore requests we don't handle
}).then(() => {
  ReactDom.render(<App />, document.getElementById("root"));
});

FakeRest will now intercept every fetch request to the REST server.

fetch-mock

Install fetch-mock:

npm install fetch-mock --save-dev

You can then create a handler and pass it to fetch-mock:

import fetchMock from 'fetch-mock';
import { getFetchMockHandler } from "fakerest";

const handler = getFetchMockHandler({
    baseUrl: 'http://localhost:3000',
    data: {
        'authors': [
            { id: 0, first_name: 'Leo', last_name: 'Tolstoi' },
            { id: 1, first_name: 'Jane', last_name: 'Austen' }
        ],
        'books': [
            { id: 0, author_id: 0, title: 'Anna Karenina' },
            { id: 1, author_id: 0, title: 'War and Peace' },
            { id: 2, author_id: 1, title: 'Pride and Prejudice' },
            { id: 3, author_id: 1, title: 'Sense and Sensibility' }
        ],
        'settings': {
            language: 'english',
            preferred_format: 'hardback',
        }
    }
});

fetchMock.mock('begin:http://localhost:3000', handler);

FakeRest will now intercept every fetch request to the REST server.

Sinon

Install Sinon:

npm install sinon --save-dev

Then, configure a Sinon server:

import sinon from 'sinon';
import { getSinonHandler } from "fakerest";

const handler = getSinonHandler({
    baseUrl: 'http://localhost:3000',
    data: {
        'authors': [
            { id: 0, first_name: 'Leo', last_name: 'Tolstoi' },
            { id: 1, first_name: 'Jane', last_name: 'Austen' }
        ],
        'books': [
            { id: 0, author_id: 0, title: 'Anna Karenina' },
            { id: 1, author_id: 0, title: 'War and Peace' },
            { id: 2, author_id: 1, title: 'Pride and Prejudice' },
            { id: 3, author_id: 1, title: 'Sense and Sensibility' }
        ],
        'settings': {
            language: 'english',
            preferred_format: 'hardback',
        }
    },
});

// use sinon.js to monkey-patch XmlHttpRequest
const sinonServer = sinon.fakeServer.create();
// this is required when doing asynchronous XmlHttpRequest
sinonServer.autoRespond = true;
sinonServer.respondWith(handler);

FakeRest will now intercept every XMLHttpRequest request to the REST server.

REST Syntax

FakeRest uses a simple REST syntax described below.

Get A Collection of records

GET /[name] returns an array of records in the name collection. It accepts 4 query parameters: filter, sort, range, and embed. It responds with a status 200 if there is no pagination, or 206 if the list of items is paginated. The response mentions the total count in the Content-Range header.

GET /books?filter={"author_id":1}&embed=["author"]&sort=["title","desc"]&range=[0-9]

HTTP 1.1 200 OK
Content-Range: items 0-1/2
Content-Type: application/json
[
    { "id": 3, "author_id": 1, "title": "Sense and Sensibility", "author": { "id": 1, "first_name": "Jane", "last_name": "Austen" } },
    { "id": 2, "author_id": 1, "title": "Pride and Prejudice", "author": { "id": 1, "first_name": "Jane", "last_name": "Austen" } }
]

The filter param must be a serialized object literal describing the criteria to apply to the search query. See the supported filters for more details.

GET /books?filter={"author_id":1} // return books where author_id is equal to 1
HTTP 1.1 200 OK
Content-Range: items 0-1/2
Content-Type: application/json
[
    { "id": 2, "author_id": 1, "title": "Pride and Prejudice" },
    { "id": 3, "author_id": 1, "title": "Sense and Sensibility" }
]

// array values are possible
GET /books?filter={"id":[2,3]} // return books where id is in [2,3]
HTTP 1.1 200 OK
Content-Range: items 0-1/2
Content-Type: application/json
[
    { "id": 2, "author_id": 1, "title": "Pride and Prejudice" },
    { "id": 3, "author_id": 1, "title": "Sense and Sensibility" }
]

// use the special "q" filter to make a full-text search on all text fields
GET /books?filter={"q":"and"} // return books where any of the book properties contains the string 'and'

HTTP 1.1 200 OK
Content-Range: items 0-2/3
Content-Type: application/json
[
    { "id": 1, "author_id": 0, "title": "War and Peace" },
    { "id": 2, "author_id": 1, "title": "Pride and Prejudice" },
    { "id": 3, "author_id": 1, "title": "Sense and Sensibility" }
]

// use _gt, _gte, _lte, _lt, or _neq suffix on filter names to make range queries
GET /books?filter={"price_lte":20} // return books where the price is less than or equal to 20
GET /books?filter={"price_gt":20} // return books where the price is greater than 20

// when the filter object contains more than one property, the criteria combine with an AND logic
GET /books?filter={"published_at_gte":"2015-06-12","published_at_lte":"2015-06-15"} // return books published between two dates

The sort param must be a serialized array literal defining first the property used for sorting, then the sorting direction.

GET /author?sort=["date_of_birth","asc"]  // return authors, the oldest first
GET /author?sort=["date_of_birth","desc"]  // return authors, the youngest first

The range param defines the number of results by specifying the rank of the first and last results. The first result is #0.

GET /books?range=[0-9] // return the first 10 books
GET /books?range=[10-19] // return the 10 next books

The embed param sets the related objects or collections to be embedded in the response.

// embed author in books
GET /books?embed=["author"]
HTTP 1.1 200 OK
Content-Range: items 0-3/4
Content-Type: application/json
[
    { "id": 0, "author_id": 0, "title": "Anna Karenina", "author": { "id": 0, "first_name": "Leo", "last_name": "Tolstoi" } },
    { "id": 1, "author_id": 0, "title": "War and Peace", "author": { "id": 0, "first_name": "Leo", "last_name": "Tolstoi" } },
    { "id": 2, "author_id": 1, "title": "Pride and Prejudice", "author": { "id": 1, "first_name": "Jane", "last_name": "Austen" } },
    { "id": 3, "author_id": 1, "title": "Sense and Sensibility", "author": { "id": 1, "first_name": "Jane", "last_name": "Austen" } }
]

// embed books in author
GET /authors?embed=["books"]
HTTP 1.1 200 OK
Content-Range: items 0-1/2
Content-Type: application/json
[
    { id: 0, first_name: 'Leo', last_name: 'Tolstoi', books: [{ id: 0, author_id: 0, title: 'Anna Karenina' }, { id: 1, author_id: 0, title: 'War and Peace' }] },
    { id: 1, first_name: 'Jane', last_name: 'Austen', books: [{ id: 2, author_id: 1, title: 'Pride and Prejudice' }, { id: 3, author_id: 1, title: 'Sense and Sensibility' }] }
]

// you can embed several objects
GET /authors?embed=["books","country"]

Get A Single Record

GET /[name]/:id returns a JSON object, and a status 200, unless the resource doesn't exist.

GET /books/2

HTTP 1.1 200 OK
Content-Type: application/json
{ "id": 2, "author_id": 1, "title": "Pride and Prejudice" }

The embed param sets the related objects or collections to be embedded in the response.

GET /books/2?embed=['author']

HTTP 1.1 200 OK
Content-Type: application/json
{ "id": 2, "author_id": 1, "title": "Pride and Prejudice", "author": { "id": 1, "first_name": "Jane", "last_name": "Austen" } }

Create A Record

POST /[name] returns a status 201 with a Location header for the newly created resource, and the new resource in the body.

POST /books
{ "author_id": 1, "title": "Emma" }

HTTP 1.1 201 Created
Location: /books/4
Content-Type: application/json
{ "author_id": 1, "title": "Emma", "id": 4 }

Update A Record

PUT /[name]/:id returns the modified JSON object, and a status 200, unless the resource doesn't exist.

PUT /books/2
{ "author_id": 1, "title": "Pride and Prejudice" }

HTTP 1.1 200 OK
Content-Type: application/json
{ "id": 2, "author_id": 1, "title": "Pride and Prejudice" }

Delete A Single Record

DELETE /[name]/:id returns the deleted JSON object, and a status 200, unless the resource doesn't exist.

DELETE /books/2

HTTP 1.1 200 OK
Content-Type: application/json
{ "id": 2, "author_id": 1, "title": "Pride and Prejudice" }

Supported Filters

Operators are specified as suffixes on each filtered field. For instance, applying the _lte operator on the price field for the books resource is done like this:

    GET /books?filter={"price_lte":20} // return books where the price is less than or equal to 20
  • _eq: check for equality on simple values:

      GET /books?filter={"price_eq":20} // return books where the price is equal to 20
    
  • _neq: check for inequality on simple values

      GET /books?filter={"price_neq":20} // return books where the price is not equal to 20
    
  • _eq_any: check for equality on any passed values

      GET /books?filter={"price_eq_any":[20, 30]} // return books where the price is equal to 20 or 30
    
  • _neq_any: check for inequality on any passed values

      GET /books?filter={"price_neq_any":[20, 30]} // return books where the price is not equal to 20 nor 30
    
  • _inc_any: check for items that include any of the passed values

      GET /books?filter={"authors_inc_any":['William Gibson', 'Pat Cadigan']} // return books where authors include either 'William Gibson' or 'Pat Cadigan' or both
    
  • _q: check for items that contain the provided text

      GET /books?filter={"author_q":['Gibson']} // return books where the author includes 'Gibson' not considering the other fields
    
  • _lt: check for items that have a value lower than the provided value

      GET /books?filter={"price_lte":100} // return books that have a price lower that 100
    
  • _lte: check for items that have a value lower than or equal to the provided value

      GET /books?filter={"price_lte":100} // return books that have a price lower or equal to 100
    
  • _gt: check for items that have a value greater than the provided value

      GET /books?filter={"price_gte":100} // return books that have a price greater that 100
    
  • _gte: check for items that have a value greater than or equal to the provided value

      GET /books?filter={"price_gte":100} // return books that have a price greater or equal to 100
    

Single Elements

FakeRest allows you to define a single element, such as a user profile or global settings, that can be fetched, updated, or deleted.

GET /settings

HTTP 1.1 200 OK
Content-Type: application/json
{ "language": "english", "preferred_format": "hardback" }

PUT /settings
{ "language": "french", "preferred_format": "paperback" }

HTTP 1.1 200 OK
Content-Type: application/json
{ "language": "french", "preferred_format": "paperback" }

DELETE /settings

HTTP 1.1 200 OK
Content-Type: application/json
{ "language": "french", "preferred_format": "paperback" }

Middlewares

Middlewares let you intercept requests and simulate server features such as:

  • authentication checks
  • server-side validation
  • server dynamically generated values
  • simulate response delays

You can define middlewares on all handlers, by passing a middlewares option:

import { getMswHandler } from 'fakerest';
import { data } from './data';

const handler = getMswHandler({
    baseUrl: 'http://my.custom.domain',
    data,
    middlewares: [
        async (context, next) => {
            if (context.headers.Authorization === undefined) {
                return {
                    status: 401,
                    headers: {},
                };
            }

            return next(context);
        },
        withDelay(300),
    ],
});

A middleware is a function that receives 2 parameters:

  • The FakeRest context, an object containing the data extracted from the request that FakeRest uses to build the response. It has the following properties:
  • method: The request method as a string (GET, POST, PATCH or PUT)
  • url: The request URL as a string
  • headers: The request headers as an object where keys are header names
  • requestBody: The parsed request data if any
  • params: The request parameters from the URL search (e.g. the identifier of the requested record)
  • collection: The name of the targeted collection (e.g. posts)
  • single: The name of the targeted single (e.g. settings)
  • A next function to call the next middleware in the chain, to which you must pass the context

A middleware must return a FakeRest response either by returning the result of the next function or by returning its own response. A FakeRest response is an object with the following properties:

  • status: The response status as a number (e.g. 200)
  • headers: The response HTTP headers as an object where keys are header names
  • body: The response body which will be stringified

Authentication Checks

Here's how to implement an authentication check:

const handler = getMswHandler({
    baseUrl: 'http://my.custom.domain',
    data,
    middlewares: [
        async (context, next) => {
            if (context.headers.Authorization === undefined) {
                return { status: 401, headers: {} };
            }
            return next(context);
        }
    ]
});

Server-Side Validation

Here's how to implement server-side validation:

const handler = getMswHandler({
    baseUrl: 'http://my.custom.domain',
    data,
    middlewares: [
        async (context, next) => {
            if (
                context.collection === "books" &&
                request.method === "POST" &&
 !context.requestBody?.title
            ) {
                return {
                    status: 400,
                    headers: {},
                    body: {
                        errors: {
                            title: 'An article with this title already exists. The title must be unique.',
                        },
                    },
                };
            }

            return next(context);
        }
    ]
});

Dynamically Generated Values

Here's how to implement dynamically generated values on creation:

const handler = getMswHandler({
    baseUrl: 'http://my.custom.domain',
    data,
    middlewares: [
        async (context, next) => {
            if (
                context.collection === 'books' &&
                context.method === 'POST'
            ) {
                const response = await next(context);
                response.body.updatedAt = new Date().toISOString();
                return response;
            }

            return next(context);
        }
    ]
});

Simulate Response Delays

Here's how to simulate response delays:

const handler = getMswHandler({
    baseUrl: 'http://my.custom.domain',
    data,
    middlewares: [
        async (context, next) => {
            return new Promise((resolve) => {
                setTimeout(() => {
                    resolve(next(context));
                }, 500);
            });
        }
    ]
});

This is so common FakeRest provides the withDelay function for that:

import { getMswHandler, withDelay } from 'fakerest';

const handler = getMswHandler({
    baseUrl: 'http://my.custom.domain',
    data,
    middlewares: [
        withDelay(500), // delay in ms
    ]
});

Configuration

All handlers can be customized to accommodate your API structure.

Identifiers

By default, FakeRest assumes all records have a unique id field. Some databases such as MongoDB use _id instead of id for collection identifiers. You can customize FakeRest to do the same by setting the identifierName option:

const handler = getMswHandler({
    baseUrl: 'http://my.custom.domain',
    data,
    identifierName: '_id'
});

You can also specify that on a per-collection basis:

import { MswAdapter, Collection } from 'fakerest';

const adapter = new MswAdapter({ baseUrl: 'http://my.custom.domain', data });
const authorsCollection = new Collection({ items: [], identifierName: '_id' });
adapter.server.addCollection('authors', authorsCollection);
const handler = adapter.getHandler();

Primary Keys

By default, FakeRest uses an auto-incremented sequence for the item identifiers. If you'd rather use UUIDs for instance but would like to avoid providing them when you insert new items, you can provide your own function:

import { getMswHandler } from 'fakerest';
import uuid from 'uuid';

const handler = new getMswHandler({
    baseUrl: 'http://my.custom.domain',
    data,
    getNewId: () => uuid.v5()
});

You can also specify that on a per-collection basis:

import { MswAdapter, Collection } from 'fakerest';
import uuid from 'uuid';

const adapter = new MswAdapter({ baseUrl: 'http://my.custom.domain', data });
const authorsCollection = new Collection({ items: [], getNewId: () => uuid.v5() });
adapter.server.addCollection('authors', authorsCollection);
const handler = adapter.getHandler();

Default Queries

Some APIs might enforce some parameters on queries. For instance, an API might always include an embed or enforce a query filter. You can simulate this using the defaultQuery parameter:

import { getMswHandler } from 'fakerest';
import uuid from 'uuid';

const handler = getMswHandler({
    baseUrl: 'http://my.custom.domain',
    data,
    defaultQuery: (collection) => {
        if (resourceName == 'authors') return { embed: ['books'] }
        if (resourceName == 'books') return { filter: { published: true } }
        return {};
    }
});

Architecture

Behind a simple API (getXXXHandler), FakeRest uses a modular architecture that lets you combine different components to build a fake REST server that fits your needs.

Mocking Adapter

getXXXHandler is a shortcut to an object-oriented API of adapter classes:

export const getMswHandler = (options: MswAdapterOptions) => {
    const server = new MswAdapter(options);
    return server.getHandler();
};

FakeRest provides 3 adapter classes:

  • MswAdapter: Based on MSW
  • FetchMockAdapter: Based on fetch-mock
  • SinonAdapter: Based on Sinon

You can use the adapter class directly, e.g. if you want to make the adapter instance available in the global scope for debugging purposes:

import { MsWAdapter } from 'fakerest';

const adapter = new MswAdapter({
    baseUrl: 'http://my.custom.domain',
    data: {
        'authors': [
            { id: 0, first_name: 'Leo', last_name: 'Tolstoi' },
            { id: 1, first_name: 'Jane', last_name: 'Austen' }
        ],
        'books': [
            { id: 0, author_id: 0, title: 'Anna Karenina' },
            { id: 1, author_id: 0, title: 'War and Peace' },
            { id: 2, author_id: 1, title: 'Pride and Prejudice' },
            { id: 3, author_id: 1, title: 'Sense and Sensibility' }
        ],
        'settings': {
            language: 'english',
            preferred_format: 'hardback',
        }
    }
});
window.fakerest = adapter;
const handler = adapter.getHandler();

REST Server

Adapters transform requests to a normalized format, pass them to a server object, and transform the normalized server response into the format expected by the mocking library.

The server object implements the REST syntax. It takes a normalized request and exposes a handle method that returns a normalized response. FakeRest currently provides only one server implementation: SimpleRestServer.

You can specify the server to use in an adapter by passing the server option:

const server = new SimpleRestServer({
    baseUrl: 'http://my.custom.domain',
    data: {
        'authors': [
            { id: 0, first_name: 'Leo', last_name: 'Tolstoi' },
            { id: 1, first_name: 'Jane', last_name: 'Austen' }
        ],
        'books': [
            { id: 0, author_id: 0, title: 'Anna Karenina' },
            { id: 1, author_id: 0, title: 'War and Peace' },
            { id: 2, author_id: 1, title: 'Pride and Prejudice' },
            { id: 3, author_id: 1, title: 'Sense and Sensibility' }
        ],
        'settings': {
            language: 'english',
            preferred_format: 'hardback',
        }
    }
});
const adapter = new MswAdapter({ server });
const handler = adapter.getHandler();

You can provide an alternative server implementation. This class must implement the APIServer type:

export type APIServer = {
    baseUrl?: string;
    handle: (context: FakeRestContext) => Promise<BaseResponse>;
};

export type BaseResponse = {
    status: number;
    body?: Record<string, any> | Record<string, any>[];
    headers: { [key: string]: string };
};

export type FakeRestContext = {
    url?: string;
    headers?: Headers;
    method?: string;
    collection?: string;
    single?: string;
    requestBody: Record<string, any> | undefined;
    params: { [key: string]: any };
};

The FakerRestContext type describes the normalized request. It's usually the adapter's job to transform the request from the mocking library to this format.

Database

The querying logic is implemented in a class called Database, which is independent of the server. It contains collections and single.

You can specify the database used by a server by setting its database property:

const database = new Database({
    data: {
        'authors': [
            { id: 0, first_name: 'Leo', last_name: 'Tolstoi' },
            { id: 1, first_name: 'Jane', last_name: 'Austen' }
        ],
        'books': [
            { id: 0, author_id: 0, title: 'Anna Karenina' },
            { id: 1, author_id: 0, title: 'War and Peace' },
            { id: 2, author_id: 1, title: 'Pride and Prejudice' },
            { id: 3, author_id: 1, title: 'Sense and Sensibility' }
        ],
        'settings': {
            language: 'english',
            preferred_format: 'hardback',
        }
    }
});
const server = new SimpleRestServer({ baseUrl: 'http://my.custom.domain', database });

You can even use the database object if you want to manipulate the data:

database.updateOne('authors', 0, { first_name: 'Lev' });

Collections & Singles

The Database may contain collections and singles. In the following example, authors and books are collections, and settings is a single.

const handler = getMswHandler({
    baseUrl: 'http://localhost:3000',
    data: {
        'authors': [
            { id: 0, first_name: 'Leo', last_name: 'Tolstoi' },
            { id: 1, first_name: 'Jane', last_name: 'Austen' }
            ],
        'books': [
            { id: 0, author_id: 0, title: 'Anna Karenina' },
            { id: 1, author_id: 0, title: 'War and Peace' },
            { id: 2, author_id: 1, title: 'Pride and Prejudice' },
            { id: 3, author_id: 1, title: 'Sense and Sensibility' }
        ],
        'settings': {
            language: 'english',
            preferred_format: 'hardback',
        }
    }
});

A collection is the equivalent of a classic database table. It supports filtering and direct access to records by their identifier.

A single represents an API endpoint that returns a single entity. It's useful for things such as user profile routes (/me) or global settings (/settings).

Embeds

FakeRest supports embedding other resources in a main resource query result. For instance, embedding the author of a book.

GET /books/2?embed=['author']

HTTP 1.1 200 OK
Content-Type: application/json
{ "id": 2, "author_id": 1, "title": "Pride and Prejudice", "author": { "id": 1, "first_name": "Jane", "last_name": "Austen" } }

Embeds are defined by the query, they require no setup in the database.

Development

# Install dependencies
make install

# Run the demo with MSW
make run-msw

# Run the demo with fetch-mock
make run-fetch-mock

# Run the demo with sinon
make run-sinon

# Run tests
make test

# Build minified version
make build

You can sign-in to the demo with janedoe and password

License

FakeRest is licensed under the MIT License, sponsored by marmelab.