I made a Ruby container image and moved the Lambda function

@Kiitan of Increments Inc. will be in charge of the 21st day of Increments × cyma (Ateam Inc.) Advent Calendar 2020.

Introduction

Amazon Web Services (AWS) announced AWS Lambda container support at the annual event "AWS re: Invent 2020" that is being held online. In short, the image pushed to the ECR repository can now be used in Lambda functions. So I immediately touched it.

What to make

When a video is uploaded to S3, we will create a gif and create a Lambda function that will upload it to S3.

Implementation

The file structure this time is as follows.

├── Dockerfile
├── Gemfile
├── Gemfile.lock
└── app.rb

For app.rb, S3 just gets the video from, converts it to a gif, and uploads it again. When saving a file by Lambda's method, you need to be careful because you do not have write permission unless it is under tmp /.

app.rb

require 'aws-sdk-s3'
require 'streamio-ffmpeg'

module LambdaFunction
  class Handler
    def self.process(event:, context:)
      s3_client = Aws::S3::Client.new(
                           :region => ENV['REGION'],
                           :access_key_id => ENV['ACCESS_KEY'],
                           :secret_access_key => ENV['SECRET_ACCESS_KEY']
                          )
      input_bucket = event['Records'][0]['s3']['bucket']['name']
      input_key = event['Records'][0]['s3']['object']['key']
      file_name = input_key.split("/").last
      file_path = "/tmp/#{file_name}"

      #Get the video uploaded from S3/tmp/Save to
      movie_file = s3_client.get_object(:bucket => input_bucket, :key => input_key).body.read
      File.write(file_path, movie_file)

      #Load video with ffmpeg and convert it to gif
      movie = FFMPEG::Movie.new(file_path)
      #Cut the start 6 seconds and convert it to gif
      options = {frame_rate: '10', duration: '00:00:6.000'}
      movie.transcode("#{file_path}.gif", options)

      #S3 gif/Upload the gif created below
      movie_object = s3_client.put_object({ 
        :bucket => input_bucket,
        :key => "gif/#{file_name}.gif",
        :body => File.new("#{file_path}.gif"),
      })
    end
  end
end

source 'https://rubygems.org'

ruby '2.7.2'

gem 'streamio-ffmpeg'
gem 'aws-sdk'

Next, regarding Dockerfile, not all base images can be used, and the Lambda Runtime API for Lambda must be implemented.

<!-Container for Lambda ruby-> in ECR Public gallery

This time, we will use amazon/aws-lambda-ruby: 2.7 as a base image. There is nothing to be careful about other than the base image, it is a very simple Dockerfile that just downloads ffmpeg and installs gem.

FROM amazon/aws-lambda-ruby:2.7

RUN yum -y install wget tar xz xz-devel
RUN wget https://johnvansickle.com/ffmpeg/builds/ffmpeg-git-amd64-static.tar.xz \
      && tar Jxvf ./ffmpeg-git-amd64-static.tar.xz \
      && cp ./ffmpeg*amd64-static/ffmpeg /usr/local/bin/ \
      && cp ./ffmpeg*amd64-static/ffprobe /usr/local/bin/ 
COPY Gemfile Gemfile.lock ./
RUN bundle install --path vendor/bundle
COPY app.rb ./

CMD ["app.LambdaFunction::Handler.process"]

Push to ECR

Pushing to ECR is not difficult, and you will not get lost because the repository is created and the push command on the repository screen is carefully written on the display.

1. Obtain an authentication token and authenticate the Docker client to the registry. Use the AWS CLI.

aws ecr get-login-password --region REGION | docker login --username AWS --password-
stdin XXX.dkr.ecr.XXX.amazonaws.com

2. Build a Docker image using the following command.

docker build -t ffmpeg-lambda .

3. Once the build is complete, tag the image so that you can push it to this repository.

docker tag ffmpeg-lambda:latest XXX.dkr.ecr.XXX.amazonaws.com/ffmpeg-lambda:latest

4. Run the following command to push this image to your newly created AWS repository.

docker push XXX.dkr.ecr.XXX.amazonaws.com/ffmpeg-lambda:latest

Creating a Lambda function

Let's make Lambda. Select the container image and specify the container image you created earlier to complete the creation.

After that, create and set the trigger as usual

If you execute it as it is, it fails because the default timeout setting is 3 seconds. It was necessary to extend the timeout time from the basic settings.

If you upload the video to S3 and have a gif, you're done!

Impressions

As for the impression, the implementation was completed very easily.

In addition, uploading of the conventional zip package is now allowed up to 10GB of image size compared to the maximum of 250MB, making it more flexible than before. I think that the range of available scenes has expanded because it can be installed with yum etc.

Finally

The 22nd day of Increments x cyma (Ateam Inc.) Advent Calendar 2020 @ O93 of Ateam Co., Ltd. is in charge!