This is a ultra light-weight synchronous fetch mock for unit testing with Jest.
It can be used to mock the native fetch or 3rd party libraries such as unfetch.
Module verion 2.0.0 was intended to make this mock more in-line with the official fetch API.
This however brings some braking changes:
fetchfunction now has only two arguments (resourceandinit) instead of three (url,data,config)mockResultno longer acceptsdataparam - you now need to provide implementation to some of the usuall response methods (i.e.text,json, etc)
Please update your code accordingly.
Installation is simple - just run:
npm i --save-dev jest-mock-fetch
Create a setupJest.js file to setup the mock with the folloging content:
// setupJest.js
global.fetch = require('jest-mock-fetch');Edit the jest.config.js file and add the following:
"jest": {
"automock": false,
"setupFiles": [
"./setupJest.js"
]
}If you are using a polyfill/ponyfill library (i.e. unfetch) which implement fetch API then you need to use a different setup procedure.
Here's an example for unfetch:
- We need to setup a manual Jest mock
- create
__mocks__directory in your project root - inside this new directory create a files named
unfetch.js - copy & past the following snippets to
unfetch.jsfile// ./__mocks__/unfetch.js import unfetch from 'jest-mock-fetch'; export default unfetch;
- create
Let's consider that we want to test a component which uses fetch. This component returns a promise, which will be resolved after fetch is done communicating with the server.
Here's a Jest snippet, which explains how we would test this component:
// ./test/UppercaseProxy.spec.js
import UppercaseProxy from '../src/UppercaseProxy';
import fetch from 'jest-mock-fetch';
afterEach(() => {
// cleaning up the mess left behind the previous test
fetch.reset();
});
it('UppercaseProxy should get data from the server and convert it to UPPERCASE', () => {
let catchFn = jest.fn(),
thenFn = jest.fn();
// using the component, which should make a server response
let clientMessage = 'client is saying hello!';
UppercaseProxy(clientMessage)
.then(thenFn)
.catch(catchFn);
// since `post` method is a spy, we can check if the server request was correct
// a) the correct method was used (post)
// b) went to the correct web service URL ('/web-service-url/')
// c) if the payload was correct ('client is saying hello!')
expect(fetch).toHaveBeenCalledWith("/web-service-url/", {
body: clientMessage,
});
// simulating a server response
fetch.mockResponse({
text: () => 'server says hello!'
});
// checking the `then` spy has been called and if the
// response from the server was converted to upper case
expect(thenFn).toHaveBeenCalledWith('SERVER SAYS HELLO!');
// catch should not have been called
expect(catchFn).not.toHaveBeenCalled();
});To make this example complete and easier to understand, let's have a look at a (verbose) implementation of component we are testing:
// ./src/UppercaseProxy.js
const UppercaseProxy = (clientMessage) => {
return(
// requesting data from server
fetch('/web-service-url/', { body: clientMessage })
// get the response text
.then(response => response.text())
// convert text to uppercase
.then(text => text.toUpperCase())
);
};
export default UppercaseProxy;At the bottom of this page you can find additional examples.
In addition to mock fetch itself being a spy, it also has additional public methods, which are intended to facilitate mocking:
mockResponse- simulates a server (web service) responsemockError- simulates a (network/server) errorlastReqGet- returns extended info about the most recent requestlastPromiseGet- returns promise created when the most recent request was madereset- resets thefetchmock object - prepare it for the next test (typically used inafterEach)
After a request has been made to the server (web service), this method resolves that request by simulating a server response. Status meaning is ignored, i.e. 400 will still resolve fetch promise. Use mockError for non-2xx responses.
NOTE: This method should be called after the fetch call in your test for the promise to resolve properly. After all remember that this mock works synchronously.
The first argument of this method is the a response object returned by the server, with a structure illustrated by the snippet below. All the properties are optional, meaning that if a property is ommitted it will be replaced by a default value (defaults are shown in the snippet).
response = {
body: new PassThrough(),
headers: new Headers(),
status: 200,
statusText: "OK",
ok: true,
url: resource as string,
arrayBuffer: () => new ArrayBuffer(0),
blob: () => new Blob(),
clone: jest.fn(),
error: jest.fn(),
formData: () => new FormData(),
json: () => ({ }),
redirect: jest.fn(),
text: () => "dummy text"
}The given response object will get passed to then even handler function.
The second argument enables us to pinpoint an exact server request we wish to resolve. This can be useful if we're making multiple server requests and are planing to resolve them in a different order from the one in which they were made.
We can supply two different objects:
- an extended request info object, which can be accessed by calling
lastReqGetmethod - a
promiseobject, which is can be accessed by calling thelastPromiseGetmethod
If ommited this argument defaults to the latest request made (internally the lastReqGet method is called).
At the end of this document you can find an example which demonstrates how this parameter can be used.
Both mockResponse and mockError will throw an error if there's no pending request to resolve/reject.
You can change this behavior by passing true as third argument, activating the so-called silentMode. With silentMode activated, the methods will just do nothing.
This method simulates an error while making a server request (network error, server error, etc ...).
NOTE: This method should be called after the fetch call in your test for the promise to resolve properly. After all remember that this mock works synchronously.
Error object will get passed to catch event handler function. If omitted it defaults to an empty object.
The second argument is a item object, which works the same way as described part about the mockResponse method.
The third argument is the silentMode flag, which works the same way as described part about the mockResponse method.
lastReqGet method returns extended info about the most recent request. The returned value can be used to pinpoint exact server request we wish to resolve (the value is passed as the second param of mockResponse or mockError methods).
The returned info contains all the data relevant to the request. It has the following structure (an example):
let requestInfo = {
// promise created while
promise: SimplePromise,
// URL passed to the fetch
resource: "https://github.com/",
// payload sent to server
init: { body: "this is payload sent to the server" },
}Additional examples at the end of this document illustrate how this method can be used.
NOTE: this is a sibling method to the lastPromiseGet (which returns only the promise portion of this the request object).
getReqByUrl() returns the same info about a specific request as lastReqGet (see above). Instead of returning the
most recent request, it returns the most recent request matching the given url.
The url to be matched. Must match exactly the url passed to fetch before.
lastPromiseGet method returns a promise given when the most recent server request was made. The returned value can be used to pinpoint exact server request we wish to resolve (the value is passed as the second param of mockResponse or mockError methods).
Additional examples at the end of this document illustrate how this method can be used.
NOTE: This is a sibling method to the lastReqGet, which in addition to promise returns object containing extended info about the request.
reset method clears state of the fetch mock to initial values. It should be called after each test, so that we can start fresh with our next test (i.e. from afterEach method).
Since fetchMock is relatively simple, most of its functionality was covered in basic example at the beginning of this document. In this section we'll explore features not covered by that initial example.
The following example illustrates the meaning of the values returned by lastReqGet and lastPromiseGet methods.
The first snippet shows a component which will be tested. The component makes a request to the server and stores the promise returned by fetch.
// ./src/MyComponent.js
class MyComponent {
CallServer () {
// making a request and storing the given promise
this.fetchPromise = fetch('/web-service-url/', { data: clientMessage });
}
}
export default MyComponent;In our spec file we will compare promise stored inside the MyComponent with values returned by lastReqGet and lastPromiseGet methods:
// ./test/MyComponent.spec.js
import MyComponent from '../src/SomeSourceFile';
let myComp = new MyComponent();
myComp.CallServer();
// getting the extended info about the most recent request
let lastReqInfo = fetch.lastReqGet();
// getting the promise made when the most recent request was made
let lastPromise = fetch.lastPromiseGet();
// the following expression will write `true` to the console
// > here we compare promise stored in the `MyComponent` to the one
// returned by the `lastPromiseGet` method
console.log(myComp.fetchPromise === lastPromise);
// the following expression will also write `true` to the console
// > here we compare promise stored in the `MyComponent`
// to the one in the request info, which was returned by the
// `lastReqGet` method
console.log(myComp.fetchPromise === lastReqInfo.promise);
// the following will also write "true" to console,
// since it't the same object
console.log(lastPromise === lastReqInfo.promise);In the following example we'll have a look at how to resolve requests at desired order by using lastReqGet method.
In this example we'll create two consecutive requests before simulating a server response to the first one.
it('when resolving a request an appropriate handler should be called', () => {
let thenFn1 = jest.fn(),
thenFn2 = jest.fn();
// creating the FIRST server request
UppercaseProxy('client is saying hello!').then(thenFn1);
// storing the request info - we'll need it later to pinpoint the request
let firstRequestInfo = fetch.lastReqGet();
// creating the SECOND server request
// BEFORE the first had chance to be resolved
UppercaseProxy('client says bye bye!').then(thenFn2);
// Simulating a server response to the FIRST request
// -> we're using request info object to pinpoint the request
// ... IF the info object is ommited, the method would automatically
// resolve to the newest request from the internal queue (the SECOND one)
const firstResponse = fetch.mockResponse({ text: () => 'server says hello!' }, firstRequestInfo);
// only the first handler should have been called
expect(thenFn1).toHaveBeenCalled();
expect(thenFn2).not.toHaveBeenCalled();
// Simulating a server response to the SECOND request
// NOTE: here we don't need to provide the request info,
// since there is only one unresolved request left
// -> `mockResponse` resolves the last request in the
// queue if request info is ommited
fetch.mockResponse({ data: 'server says bye bye!' });
// the first `then` handles should be called only once
expect(thenFn1).toHaveBeenCalledTimes(1);
// now the second `then` handler should be called
expect(thenFn2).toHaveBeenCalled();
});Although this might not be the most realistic use-case of this functionality, it does illustrate how lastReqGet method can be used to alter the default behaviour of the mockResponse method.
NOTE: the identical effect can be achieved by using the lastPromiseGet method. These two methods perform a similar task, as described in the corresponding documentation.
The magic which enables fetch mock to work synchronously is hidden away in jest-mock-promise, which enables promises to be settled in synchronous manner.
This mock is based on: jest-mock-axios
MIT License, http://www.opensource.org/licenses/MIT