This is an extraction of the S3 client files from aws-sdk node package

The goal here is to have a smaller package (in bytes) that contains only the files necessary to interact with AWS S3 in a way so a zipped version of the application is smaller than 5MB and we are able to see our code when deployed to lambda function using custom layers

• In order for you to be able to see your code on the AWS Lambda Function console your zip deployed package has to be not larger than 5MB
• If we use a custom layer for a version of the nodejs that is newer than the ones provided by AWS the aws-sdk package is not provided
• So, when using a custom layer, we need to install the aws-sdk library to interact with S3 with our lambda function
• When we install the aws-sdk package our zip-file gets larger than 5MB, which takes from us the ability to see our code on the AWS Lambda Function console, and we don't want that

npm install --save @everymundo/aws-s3-client

Using Async functions your code can be as simple as this:

// We recommend you using the following approach just so you can easily stub ```lib.S3``` on your tests
const lib = require('@everymundo/aws-s3-client')

const getS3File = async (Bucket, Key) => {
  const s3client = new lib.S3()

  return s3client.getObject({Bucket, Key}).promise()
})

If you prefer the callback approach you can do this:

// We recommend you using the following approach just so you can easily stub ```lib.S3``` on your tests
const lib = require('@everymundo/aws-s3-client')

const getS3File = (Bucket, Key) => new Promise((resolve, reject) => {
  const s3client = new lib.S3()

  s3client.getObject({Bucket, Key}, (err, response) => err ? reject(err) : resolve(response))
})

if you don't care for stubbing you can go ahead and import the S3 directly as a constant

const {S3} = require('@everymundo/aws-s3-client')