DEV Community

Cover image for AWS Amplify Video
Alex Patterson for AWS Community Builders

Posted on • Originally published at codingcat.dev on

AWS Amplify Video

AWS Amplify Video

AWS Account is required for this lesson. You may (and probably will) incur fees associated with these activities. Livestreaming is VERY expensive so be careful!

Startup

Please make sure that you see the prerequisites, the are Node, NPM and Git.

https://docs.amplify.aws/start/getting-started/installation/q/integration/react

Installing Amplify

Make sure you install amplify globally. This will allow you to execute commands within your project.

npm install @aws-amplify/cli@latest -g
Enter fullscreen mode Exit fullscreen mode

Configuring Amplify with AWS

Run the below command to add configuration to for AWS.

amplify configure
Enter fullscreen mode Exit fullscreen mode

You will need to finish adding the required role for your amplify CLI to connect with AWS..

Initializing Project

If you prefer to follow the guide see the link in the React section.

amplify init
Enter fullscreen mode Exit fullscreen mode

This is the setup that I provided for using VSCode and creating a React

TODO_UPLOAD

React setup

I would recommend this guide for fully setting up a React project.

https://docs.amplify.aws/start/getting-started/setup/q/integration/react#initialize-a-new-backend

npx is an npm command that will fetch the necessary dependencies and create a React Application.

npx create-react-app react-amplified
cd react-amplified
Enter fullscreen mode Exit fullscreen mode

Amplify Add Video

At the time of writing this lesson there are two options that you can use for video, Livestream and Video on Demand (VOD). VOD is currently still in beta but it still produces a great streaming setup for VOD.

Livestream

I just want to repeat, Livestream is a cool experiement but SUPER expensive to run!

For adding Livestream you can accept all of the defaults (I used a custom name for the project).

TODO_UPLOAD

Whenever you add new services to Amplify you then need to push them to AWS. Amplify uses Cloudformation to create all of the required services. The key item that is created for Livestream is AWS Elemental MediaLive.

TODO_UPLOAD

You will be able to find a great amount of detail from this link https://aws.amazon.com/medialive/

This image represents exactly how we are utilizing the service. Our input source is OBS instead of a camera like depicted on the left.

TODO_UPLOAD

Video on Demand (VOD)

Probably the best explantion for what is going to happen for this part of the lesson is described in the GitHub Issue for VOD. https://github.com/awslabs/amplify-video/issues/4

TODO_UPLOAD

This is very similar to the larger image diagram that AWS has listed, however it is much simpler to follow.

https://aws.amazon.com/solutions/implementations/video-on-demand-on-aws/

TODO_UPLOAD

Adding VOD to Amplify

We will first use amplify to add VOD by running the below command

amplify video add
Enter fullscreen mode Exit fullscreen mode

In the video I choose to add the GraphQL option, but if you would rather keep things simple you can choose these options to ignore it.

TODO_UPLOAD

Now that VOD resources have been added to the project you can push them to aws.

amplify push
Enter fullscreen mode Exit fullscreen mode

You should see a similar message (depending on the changes you have made it could add api).

TODO_UPLOAD

Once this operation completes you should receive a message like below where your S3 buckets are listed for testing out the process. Select yes to overwrite, you will use this URL from the configuration in your React app.

TODO_UPLOAD

Showing Videos In React

Now that you have locations for both Livestreaming and VOD, we can update App.js to include a package called video.js.

npm i video.js
Enter fullscreen mode Exit fullscreen mode

Once this package is installed we can use it within our React app.

import React from "react";
import logo from "./logo.svg";
import "./App.css";
import videojs from "video.js";
import "video.js/dist/video-js.css";
import awsvideoconfig from "./aws-video-exports";

class VideoPlayer extends React.Component {
  componentDidMount() {
    this.player = videojs(this.videoNode, this.props);
  }

  componentWillUnmount() {
    if (this.player) {
      this.player.dispose();
    }
  }

  render() {
    return (
      <div>
        <div data-vjs-player>
          <video
            ref={(node) => {
              this.videoNode = node;
            }}
            className="video-js"
          />
        </div>
      </div>
    );
  }
}

const videoJsOptions = {
  autoplay: true,
  controls: true,
  sources: [
    {
      src: awsvideoconfig.awsOutputLiveLL,
    },
  ],
};

const videoOnDemandJsOptions = {
  autoplay: true,
  controls: true,
  sources: [
    {
      src: `https://${awsvideoconfig.awsOutputVideo}/mediaConvertTest/mediaConvertTest.m3u8`,
    },
  ],
};

function App() {
  return (
    <div className="App">
      <header className="App-header">
        <VideoPlayer {...videoJsOptions} />
        <VideoPlayer {...videoOnDemandJsOptions} />
      </header>
    </div>
  );
}

export default App;

Enter fullscreen mode Exit fullscreen mode

Above you will notice two configruations, videoJsOptions for our Livestream and videOnDemanJsOptions for our VOD. Because I was manually uploading a test file called mediaConvertTest I hardcoded this file directly in the VOD.

I have added this file to the GitHub repo for your testing as well, it is located here: TestVideo/mediaConvertTest.mov

Running VOD conversion

In order to start the process of running Media Convert you will need to upload the test file to your S3 bucket it should have the word input within S3.

TODO_UPLOAD

It will then trigger a Lambda to start Media Convert, you can see what events are subscribed on the properties tab in S3.

TODO_UPLOAD

This Lambda should have the -dev-inputWatcher

TODO_UPLOAD

You can see in the image above how the trigger is creating a new MediaConvert Job. Below is the full code for the Lambda.

// Load the AWS SDK for Node.js
// eslint-disable-next-line import/no-extraneous-dependencies
/* eslint-disable */
const AWS = require('aws-sdk');
/* eslint-enable */
const jobSettings = require('./settings.json');
// Set the region

exports.handler = async (event) => {
  AWS.config.update({ region: event.awsRegion });
  console.log(event);
  if (event.Records[0].eventName.includes('ObjectCreated')) {
    await createJob(event.Records[0].s3);
    const response = {
      statusCode: 200,
      body: JSON.stringify(`Transcoding your file: ${event.Records[0].s3.object.key}`),
    };
    return response;
  }
};

// Function to submit job to Elemental MediaConvert
async function createJob(eventObject) {
  let mcClient = new AWS.MediaConvert();
  if (!AWS.config.mediaconvert) {
    try {
      const endpoints = await mcClient.describeEndpoints().promise();
      AWS.config.mediaconvert = { endpoint: endpoints.Endpoints[0].Url };
      // Override so config applies
      mcClient = new AWS.MediaConvert();
    } catch (e) {
      console.log(e);
      return;
    }
  }

  const queueParams = {
    Name: 'Default', /* required */
  };
  const AddedKey = eventObject.object.key;
  // Get the name of the file passed in the event without the extension
  const FileName = AddedKey.split('.').slice(0, -1).join('.');
  const Bucket = eventObject.bucket.name;
  const outputBucketName = process.env.OUTPUT_BUCKET;

  // Set the output to have the filename (without extension) as a folder
  jobSettings.OutputGroups[0].OutputGroupSettings.HlsGroupSettings.Destination = `s3://${outputBucketName}/${FileName}/`;
  jobSettings.Inputs[0].FileInput = `s3://${Bucket}/${AddedKey}`;

  let queueARN = '';
  if (process.env.QUEUE_ARN) {
    queueARN = process.env.QUEUE_ARN;
  } else {
    const q = await mcClient.getQueue(queueParams, (err, data) => {
      if (err) console.log(err, err.stack); // an error occurred
      else console.log(data);
    }).promise();
    queueARN = q.Queue.Arn;
  }

  const jobParams = {
    JobTemplate: process.env.ARN_TEMPLATE,
    Queue: queueARN,
    UserMetadata: {},
    Role: process.env.MC_ROLE,
    Settings: jobSettings,
  };
  await mcClient.createJob(jobParams).promise();
}

Enter fullscreen mode Exit fullscreen mode

Within Media Convert you should see a job has started, and hopefully completed.

TODO_UPLOAD

If you select one of those jobs you will find the input and outputs in the summary.

TODO_UPLOAD

Now you will find that exact structure in your output S3 bucket.

TODO_UPLOAD

If you open that "folder" you can see that there are several files listed that the conversion created.

TODO_UPLOAD

The URL to the top one is what we will be using for the example.

Showing VOD

Now that you have a VOD file in S3 you will view this file being served through Cloud Front which is Amazon's Global CDN. The URL can be found in src/aws-video-exports.js. With the full solution in the App.js file for React you will now see the VOD file playing (you may need to refresh).

TODO_UPLOAD

Final Thoughts

I have a message out to the labs team on GitHub about the next triggers. I believe what they are trying to accomplish next is to take the Lambda form the output file and store all of the meta-data into DynamoDB which can then be served through our GraphQL API. I know this is for sure possible and if I don't hear back I will implement my own for the full app. So far this beta version is really awesome though!

https://github.com/awslabs/amplify-video/issues/4

Top comments (0)