LoopBack Test Lab
A collection of test utilities we use to write LoopBack tests.
Overview
Test utilities to help writing LoopBack 4 tests:
expect
- behavior-driven development (BDD) style assertionssinon
- test spies: functions recording arguments and other information for all of their calls
- stubs: functions (spies) with pre-programmed behavior
- mocks: fake methods (like spies) with pre-programmed behavior (like stubs) as well as pre-programmed expectations
- Helpers for creating
supertest
clients for LoopBack applications - HTTP request/response stubs for writing tests without a listening HTTP server
- Swagger/OpenAPI spec validation
- Test sandbox
Installation
npm install --save-dev @loopback/testlab
This package is typically used in tests, save it to devDependencies
via
--save-dev
.
Basic use
import {expect} from '@loopback/testlab';
describe('Basic assertions', => {
it('asserts equal values', => {
expect({key: 'value'}).to.deepEqual({key: 'value'});
expect.exists(1);
});
});
API documentation
Table of contents:
- expect - Better assertions.
- sinon - Mocks, stubs and more.
- shot - HTTP Request/Response stubs.
- validateApiSpec - Open API Spec validator.
- skipIf - Skip tests when a condition is met.
- skipOnTravis - Skip tests on Travis env.
- createRestAppClient - Create a supertest client connected to a running RestApplication.
- givenHttpServerConfig - Generate HTTP server config.
- httpGetAsync - Async wrapper for HTTP GET requests.
- httpsGetAsync - Async wrapper for HTTPS GET requests.
- toJSON - A helper to obtain JSON data representing a given object.
- createUnexpectedHttpErrorLogger - An error logger that only logs errors for unexpected HTTP statuses.
- TestSandbox - A sandbox directory for tests
expect
Should.js configured in “as-function” mode
(global Object.prototype
is left intact) with an extra chaining word to
.
sinon
Spies, mocks and stubs. Learn more at http://sinonjs.org/.
shot
Stub implementation of HTTP Request and Response objects, useful for unit tests.
Besides the API provided by shot
module (see
API Reference), we provide
additional APIs to better support async/await flow control and usage in
Express-based code.
There are three primary situations where you can leverage stub objects provided by Shot in your unit tests:
- Code parsing core HTTP Request
- Code modifying core HTTP Response, including full request/response handlers
- Code parsing Express HTTP Request or modifying Express HTTP Response
Test request parsing
Use the factory function stubServerRequest
to create a stub request that can
be passed to methods expecting core HTTP Request on input.
import {stubServerRequest, expect} from '@loopback/testlab';
describe('parseParams', () => {
it('parses query string arguments', () => {
const request = stubServerRequest({
method: 'GET',
url: '/api/products?count=10',
});
const args = parseParams(request, [
{name: 'count', in: 'query', type: 'number'},
]);
expect(args).to.eql([10]);
});
});
Test response producers
Use the factory function stubHandlerContext
to create request & response stubs
and a promise to observe the actual response as received by clients.
import {stubHandlerContext, expect} from '@loopback/testlab';
describe('app REST handler', () => {
it('returns 404 with JSON body when URL not found', async () => {
const app = express();
const context = stubHandlerContext({
method: 'GET',
url: '/path-does-not-exist',
});
// Invoke Express' request handler with stubbed request/response objects
app(context.request, context.response);
// Wait until Express finishes writing the response
const actualResponse = await context.result;
// Verify the response seen by clients
expect(actualResponse.statusCode).to.equal(404);
expect(JSON.parse(actualResponse.payload)).to.containEql({
error: {
statusCode: 404,
message: 'Not Found',
},
});
});
});
Test code expecting Express Request or Response
Express modifies core HTTP request and response objects with additional properties and methods, it also cross-links request with response and vice versa. As a result, it’s not possible to create Express Request object without the accompanying Response object.
Use the factory function stubExpressContext
to create Express-flavored request
& response stubs and a promise to observe the actual response as received by
clients.
If your tested function is expecting a request object only:
import {stubExpressContext, expect} from '@loopback/testlab';
describe('operationArgsParser', () => {
it('parses body parameter', async () => {
const req = givenRequest({
url: '/',
payload: {key: 'value'},
});
const spec = givenOperationWithRequestBody({
description: 'data',
content: {'application/json': {schema: {type: 'object'}}},
});
const route = givenResolvedRoute(spec);
const args = await parseOperationArgs(req, route);
expect(args).to.eql([{key: 'value'}]);
});
function givenRequest(options?: ShotRequestOptions): Request {
return stubExpressContext(options).request;
}
});
Tests verifying code producing HTTP response can await context.result
to
receive the response as returned to clients.
import {stubExpressContext, expect} from '@loopback/testlab';
describe('response writer', () => {
it('writes object result to response as JSON', async () => {
const context = stubExpressContext();
writeResultToResponse(context.response, {name: 'Joe'});
const result = await context.result;
expect(result.headers['content-type']).to.eql('application/json');
expect(result.payload).to.equal('{"name":"Joe"}');
});
});
skipIf
Helper function for skipping tests when a certain condition is met. Use this
helper together with it
or describe
.
skipIf(someCondition, it, 'does something', async () => {
// the test code
});
Unfortunately, type inference does not work well for describe
, you have to
help the compiler to figure out the correct types.
skipIf<[(this: Suite) => void], void>(
someCondition,
describe,
'some suite name',
() => {
// define the test cases
},
);
Under the hood, skipIf
invokes the provided test verb by default (e.g. it
).
When the provided condition was true, then it calls .skip
instead (e.g.
it.skip
).
skipOnTravis
Helper function for skipping tests on Travis environment. If you need to skip
testing on Travis for any reason, use this helper together with it
or
describe
.
skipOnTravis(it, 'does something when some condition', async () => {
// the test code
});
Under the hood, skipOnTravis
invokes the provided test verb by default (e.g.
it
). When the helper detects Travis CI environment variables, then it calls
.skip
instead (e.g. it.skip
).
createRestAppClient
Helper function to create a supertest
client connected to a running
RestApplication. It is the responsibility of the caller to ensure that the app
is running and to stop the application after all tests are done.
Example use:
import {Client, createRestAppClient} from '@loopback/testlab';
describe('My application', () => {
app: MyApplication; // extends RestApplication
client: Client;
before(givenRunningApplication);
before(() => {
client = createRestAppClient(app);
});
after(() => app.stop());
it('invokes GET /ping', async () => {
await client.get('/ping?msg=world').expect(200);
});
});
givenHttpServerConfig
Helper function for generating Travis-friendly host (127.0.0.1). This is required because Travis is not able to handle IPv6 addresses.
httpGetAsync
Async wrapper for making HTTP GET requests.
import {httpGetAsync} from '@loopback/testlab';
const response = await httpGetAsync('http://example.com');
httpsGetAsync
Async wrapper for making HTTPS GET requests.
import {httpsGetAsync} from '@loopback/testlab';
const response = await httpsGetAsync('https://example.com');
toJSON
JSON encoding does not preserve properties that are undefined. As a result,
deepEqual
checks fail because the expected model value contains these
undefined property values, while the actual result returned by REST API does
not. Use this function to convert a model instance into a data object as
returned by REST API.
import {createClientForHandler, toJSON} from '@loopback/testlab';
it('gets a todo by ID', () => {
return client
.get(`/todos/${persistedTodo.id}`)
.expect(200, toJSON(persistedTodo));
});
validateApiSpec
Verify that your application API specification is a valid OpenAPI spec document.
import {validateApiSpec} from '@loopback/testlab';
import {RestServer} from '@loopback/rest';
describe('MyApp', () => {
it('has valid spec', async () => {
const app = new MyApp();
const server = await app.getServer(RestServer);
await validateApiSpec(server.getApiSpec());
});
});
createUnexpectedHttpErrorLogger
An error logger that logs the error only when the HTTP status code is not the expected HTTP status code. This is useful when writing tests for error responses:
-
When we don’t want any error messages printed to the console when the server responds with the expected error and the test passes.
-
When something else goes wrong and the server returns an unexpected error status code, and we do want an error message to be printed to the console so that we have enough information to troubleshoot the failing test.
import {createUnexpectedHttpErrorLogger} from '@loopback/testlab';
import {RestApplication} from '@loopback/rest';
describe('MyApp', () => {
it('does not log a known 401 error to console', async () => {
const app = new RestApplication();
const errorLogger = createUnexpectedHttpErrorLogger(401);
// binds the custom error logger
app.bind(SequenceActions.LOG_ERROR).to(errorLogger);
const spec = {
responses: {
/*...*/
},
};
function throwUnauthorizedError() {
throw new HttpErrors.Unauthorized('Unauthorized!');
}
app.route('get', '/', spec, throwUnauthorizedError);
await app.start();
// make `GET /` request, assert that 401 is returned
});
});
TestSandbox
Many tests need use a temporary directory as the sandbox to mimic a tree of
files. The TestSandbox
class provides such facilities to create and manage a
sandbox on the file system.
Create a sandbox
// Create a sandbox as a unique temporary subdirectory under the rootPath
const sandbox = new TestSandbox(rootPath);
const sandbox = new TestSandbox(rootPath, {subdir: true});
// Create a sandbox in the root path directly
// This is same as the old behavior
const sandbox = new TestSandbox(rootPath, {subdir: false});
// Create a sandbox in the `test1` subdirectory of the root path
const sandbox = new TestSandbox(rootPath, {subdir: 'test1'});
// To access the target directory of a sandbox
console.log(sandbox.path);
Reset a sandbox
All files inside a sandbox will be removed when the sandbox is reset. We also
try to remove cache from require
.
await sandbox.reset();
Delete a sandbox
Removes all files and mark the sandbox unusable.
await sandbox.delete();
Create a directory
Recursively creates a directory within the sandbox.
await sandbox.mkdir(dir);
Copy a file
Copies a file from src to the TestSandbox. If copying a .js
file which has an
accompanying .js.map
file in the src file location, the dest file will have
its sourceMappingURL updated to point to the original file as an absolute path
so you don’t need to copy the map file.
await sandbox.copyFile(src, dest);
Write a json file
Creates a new file and writes the given data serialized as JSON.
await sandbox.writeJsonFile(dest, data);
Write a file
Creates a new file and writes the given data as a UTF-8-encoded text.
await sandbox.writeFile(dest, data);
Related resources
For more info about supertest
, please refer to
supertest
Contributions
Tests
Run npm test
from the root folder.
Contributors
See all contributors.
License
MIT