Using Amazon S3 in a Node.js application

If you're using Node.js to build "generic" web projects (like this one), chances are that you are going to need to process and store user-uploaded images at some point.

While ImageMagick works perfectly on most PaaS providers, usually the "local" filesystem space is volatile, so you have to come up with your own space/strategy to store images or other type of user-generated assets. There are, of course, fully-featured image management solutions in the cloud - Cloudinary being perhaps the best known at the moment.

But if you're after a really cheap (yet very reliable) alternative, you should take a look at the classic Amazon S3. There are quite a few Node.js modules out there and I've tried some of them, but in the end I ended up working with aws-sdk, the "official" one.

Although I was initially a bit put off by their verbose documentation style, in the end I found it to be quite easy to use.

You can configure it like this (IcedCoffeeScript code):

aws = require 'aws-sdk'
env = process.env

aws.config.update
  accessKeyId:     env.S3_KEY
  secretAccessKey: env.S3_SECRET
  region:          env.S3_REGION

s3 = new aws.S3

And here's a sample method to upload a local (temporary) file:

exports.upload = (tempFilePath, callback) ->
  fileName = path.basename tempFilePath
  await fs.readFile tempFilePath, defer err, data
  return callback err if err
  await
    s3.putObject
        Bucket:       env.S3_BUCKET
        Key:          "pictures/#{fileName}"
        ContentType:  'image/jpg'
        CacheControl: 'max-age=31536000' # 1 year
        Body:         data
      , defer err
    return callback err if err
  await fs.unlink tempFilePath, defer err
  callback err

Not exactly rocket-science, is it?...