msw-auto-mock

A cli tool to generate random mock data from OpenAPI descriptions for msw.

Why

We already have all the type definitions from OpenAPI spec so hand-writing every response resolver is completely unnecessary.

Generative AI Support

Since v0.19.0, msw-auto-mock support using generative AI to generate the mock data instead of fakerjs. To enable this feature, you need to setup the related config in your package.json:

1{
2  "msw-auto-mock": {
3    "ai": {
4      "enable": true,
5      "provider": "openai",
6      "openai": {
7        "apiKey": "process.env.OPENAI_API_KEY",
8      }
9    }
10  }
11}

Currently, only openai, azure, anthropic are supported. The Configuration is like below:

1interface Config {
2  ai?: {
3    enable?: boolean;
4    provider: 'openai' | 'azure' | 'anthropic';
5    openai?: {
6      baseURL?: string;
7      apiKey?: string;
8      model?: string;
9    };
10    azure?: {
11      apiKey?: string;
12      resource?: string;
13      deployment?: string;
14    };
15    anthropic?: {
16      apiKey?: string;
17      model?: string;
18    };
19  };
20};

[!IMPORTANT] For security issue, it is recommended to put your api keys in the .env files, and only leave the env key in the settings, for example if you are using vite, the setting should be "apiKey": "import.meta.env.VITE_OPENAI_API_KEY" since this is the way how vite loaded env variables, for Next.js, you could just use "apiKey": "process.env.PUBLIC_OPENAI_API_KEY". If you want to use plain value in the setting, make sure they are quoted like "model": "'gpt-4o'"

Usage

This tool also requires @faker-js/faker >= 8 and msw >= 2.

Install:

1yarn add msw-auto-mock @faker-js/faker -D

Read from your OpenAPI descriptions and output generated code:

1# can be http url or a file path on your machine, support both yaml and json.
2npx msw-auto-mock http://your_openapi.json -o ./mock

Integrate with msw, see Mock Service Worker's doc for more detail:

1# Install msw
2yarn add msw --dev
3
4# Init service worker
5npx msw init public/ --save

Then import those mock definitions in you app entry:

1import { worker } from './mock/browser.js';
2
3await worker.start();

For conditional mocking:

1async function enableMocking() {
2  if (process.env.NODE_ENV !== 'development') {
3    return
4  }
5  const { worker } = await import('./mock/browser');
6  // `worker.start()` returns a Promise that resolves
7  // once the Service Worker is up and ready to intercept requests.
8  return worker.start();
9}
10
11function mountApp() {
12  const root = createRoot(document.getElementById('root'));
13  root.render(<App />);
14}
15
16enableMocking().then(mountApp);

For Node.js integration, you can import from your_output/node.js:

1import { worker } from './mock/node.js';

For React Native integration, you can import from your_output/native.js:

1import { worker } from './mock/native.js';

Run you app then you'll see a successful activation message from Mock Service Worker in your browser's console.

Options

• -o, --output: specify output file path or output to stdout.
• -m, --max-array-length <number>: specify max array length in response, default value is 20, it'll cost some time if you want to generate a huge chunk of random data.
• -t, --includes <keywords>: specify keywords to match if you want to generate mock data only for certain requests, multiple keywords can be seperated with comma.
• -e, --excludes <keywords>: specify keywords to exclude, multiple keywords can be seperated with comma.
• --base-url: output code with specified base url or fallback to server host specified in OpenAPI.
• --static: By default it will generate dynamic mocks, use this flag if you need it to be static.
• -c, --codes <keywords>: comma separated list of status codes to generate responses for
• -h, --help: show help info.