fetch-vcr

fetch-vcr

Stop mocking HTTP Requests. Just record and then play them back. See vcr/vcr for the main idea.

Usage

After setting up (see below), the basics are:

1. set the VCR_MODE=cache environment variable before running your tests
2. run your tests

This will record (and load) all the HTTP responses into the ./_fixtures/ directory.

And when you run the steps again, viola! no network traffic happens.

What are the different modes?

• playback: (default) only uses the local fixture files
• cache: tries to use the recorded response and if not found then it is fetched and then saved (useful when adding new tests)
• record: forces HTTP requests and responses are saved to the filesystem (useful for regenerating all the fixtures)

How can I set the VCR mode?

You can set the mode either by:

• setting the VCR_MODE=record environment variable when running tests (NodeJS)
• explicitly running fetch.configure({mode: 'record'}) (NodeJS or browser)

How do I set this up?

There are separate examples for NodeJS, Jest, and in a browser (PhantomJS or Selenium)

NodeJS Setup

Here is how you would use it in a typical NodeJS app:

// import fetch from 'fetch';
import fetch from 'fetch-vcr';

// Configure where the recordings should be loaded/saved to.
// The path is relative to `process.cwd()` but can be absolute.
fetch.configure({
  fixturePath: './_fixtures',
  // mode: 'record'     <-- This is optional
})

// Use fetch like you would normally
fetch('http://openstax.org')
.then(response => {
  console.log(response.ok)
})

How do I ignore calls?

Here is how you would configure it to ignore certain request:

// import fetch from 'fetch';
import fetch from 'fetch-vcr';

// Configure where the recordings should be loaded/saved to.
// The path is relative to `process.cwd()` but can be absolute.
fetch.configure({
  fixturePath: './_fixtures',
  ignoreUrls: [/.+weedmaps\.com.+/] // <-- This is an array of Regular Expressions
  // mode: 'record'     <-- This is optional
})

fetch('https://weedmaps.com/sitemap') // <-- This will be ignored from vcr
.then(response => {
  console.log(response)
})

How can I tell if the response was from the cache?

You can check response.wasCached. It will be true if the response was loaded from the cache.

import fetch from 'fetch-vcr';

fetch('https://philschatz.com').then(response => {
  if (!response.wasCached) {
    sleep(1000) // wait before making another request
  }
})

Jest Setup

Just add the following to package.json:

  "jest": {
    "moduleNameMapper": {
      "hack-node-fetch": "node-fetch",
      "node-fetch": "fetch-vcr"
    }
  }

If you want to check which calls were made, you can use the following:

// Returns an array of {url, args, hash, bodyFilename, response, optionsFilename}
const allCalls = fetchVCR.getCalled()

// Clears the array of calls made
fetchVCR.clearCalled()

jsdom Setup

Many apps use jsdom for testing which makes it really easy to add fetch-vcr. Just replace the global fetch function with fetchVCR and you can record/play back the cassettes. See below for an example:

var fs = require('fs')
var jsdom = require('jsdom')
var fetchVCR = require('fetch-vcr')

// Configure the path to find cassettes
fetchVCR.configure({
  fixturePath: './_fixtures/'
})

var dom = new jsdom.JSDOM(fs.readFileSync('./jsdom-example.html'), {
  runScripts: 'dangerously',
  beforeParse: (window) => {
    // This changes the fetch global to be fetchVCR
    window.fetch = fetchVCR
  }
})

How can I use this in a browser?

It is easy to record HTTP requests in NodeJS and play them back in the browser.

To play them back in a browser, just run fetchVCR.configure({fixturePath: './path/to/_fixtures'}) and fetchVCR will use that path to load the files via AJAX requests.

To record HTTP requests in a browser you will need to do a little bit of work. Loading fixture files is relatively painless (using XMLHTTPRequest) but saving them to disk is non-trivial.

In order to save the fixture files to disk you will need to override fetchVCR.saveFile(rootPath, filename, contents) => Promise.

If you are using PhantomJS you will likely need to use the alert(msg) to get data out of PhantomJS and then save it to the filesystem (using fs.writeFile(...))