Lambda download and cache gzipped file from s3

index.js

'use strict';

console.log('// loading function');

const aws = require('aws-sdk');
const s3 = new aws.S3({apiVersion: '2006-03-01'});
const gzip = require('zlib').createGunzip();
const fs = require('fs');


let isInitialised = false;


exports.handler = function(event, context, callback) {

  if (!isInitialised)
    return exports.init(event, context, callback);

  // Do stuff here

}



exports.init = function(event, context, callback) {

  const localFileName = '/tmp/chrome';
  const localFile = fs.createWriteStream(localFileName);


  // first check if the file have been downloaded before
  if (fs.existsSync(localFileName)) {
    isInitialised = true;
    return exports.handler(event, context, callback); // call handler function to resume process
  }


  s3.getObject({
    Bucket: process.env.BUCKET_NAME,
    Key: process.env.FILE_NAME,
  })
  .createReadStream()
  .pipe(gzip) // decompress compress file on the fly. No need for this line if file already decompressed
  .pipe(localFile) // write to local /tmp directory
  .on('end', function() {
    isInitialised = true;
    exports.handler(event, context, callback); // call handler function to resume process
  })
  .on('err', function(err) {
    callback(err);
  })

}

You need to define those environmental variables in your lambda function

• BUCKET_NAME: the name of the bucket were the file is hosted (example: puppeteer_chrome_test)
• FILE_NAME: the path to the file within the bucket (example: chrome_executable/file_name)

In the sample above the file will be downloaded to /tmp/chrome

Once the function run for the first time it will execute exports.init. Afterwards isInitialised will be cached in memory as true.

how does puppeteer know your chrome is in /tmp/chrome?

how does puppeteer know your chrome is in /tmp/chrome?

puppeteer.launch({executablePath: '/tmp/chrome'})