fake-s3

a fake s3 server for testing purposes.

This is a zero dependency implementation that stores all objects in memory

Example

const FakeS3 = require('fake-s3');
const AWS = require('aws-sdk')
 
const server = new FakeS3({
  buckets: ['my-bucket'],
  prefix: 'files-i-care-about/'
})
 
// starts the server on specified port
await server.bootstrap()
 
// this field now exists and contains the actual hostPort
server.hostPort
 
// Create an S3 client connected to it
const s3 = new AWS.S3({
  endpoint: `http://${server.hostPort}`
  sslEnabled: false,
  accessKeyId: '123',
  secretAccessKey: 'abc',
  s3ForcePathStyle: true
})
 
// can wait for files
const files = await server.waitForFiles('my-bucket', 2)
// will yield you back when two files have been uploaded
 
// shutdown server
await server.close()

Support

The following aws-sdk methods are supported

• s3.listBuckets()
• s3.listObjectsV2()
• s3.upload()

Features

Currently the fake-s3 module supports two different ways of getting data in & out of it.

One where you just set up the fake-s3 server and use the s3 api to upload and list files.

The second is to use the populateFromCache() method to load a bunch of fixtures of disk into memory.

Recommended local approach

I recommend copying the script/cache-from-prod.js script into your application and using it to download production data onto your laptop so that it can be used for offline development.

Docs

var server = new FakeS3(options)

• options.prefix : prefix for getFiles() and waitForFiles() ; necessary to support multi part uploads, otherwise waitForFiles() will return too early when N parts have been uploaded.
• options.buckets : an array of buckets to create.

server.hostPort

This is the hostPort that the server is listening on, this will be non-null after bootstrap() finishes.

await server.bootstrap()

starts the server

await getFiles(bucket)

gets all files in a bucket

await waitForFiles(bucket, count)

this will wait for file uploads to finish and calls getFiles() and returns them once it's finished.

This is useful if your application does background uploads and you want to be notified when they are finished.

await server.close()

closes the HTTP server.

await server.populateFromCache(cacheDir)

This will have the server fetch buckets & objects from a cache on disk. This can be useful for writing tests with fixtures or starting a local server with fixtures.

It's recommended you use cacheBucketsToDisk() and cacheObjectsToDisk() to create the fixtures directory.

await server.cacheBucketsToDisk(cacheDir, accessKeyId, data)

Calling this will write buckets to the disk cache. The data parameter is the response of s3.listBuckets().

The accessKeyId is the name of the AWS account you are writing to. If you only use one account you can just specify 'default' otherwise you can get it from the S3 client instance.

await server.cacheObjectsToDisk(cacheDir, accessKeyId, bucketName, data)

Calling this will write objects to the disk cache. The data parameter is the response of s3.listObjectsV2()

The accessKeyId is the name of the AWS account you are writing to. If you only use one account you can just specify 'default' otherwise you can get it from the S3 client instance.

Installation

npm install fake-s3

Tests

npm test

Contributors

• Raynos

MIT Licensed