Amazon Chime Meeting Summary Creator demo

This Demo Generate a brief summary of a chime meeting and email that to the meeting inviter’s email.

We record the meeting using Amazon Chime SDK and using a lambda function save the audio file to an S3 bucket. With the help of a second lambda function, we use AWS Transcribe on the audio file and convert it into a text file and upload this text file to an S3 bucket. Using the third lambda function we call Amazon Sagemaker which has an image of a summariser which creates a summary from the text file using BERT. Further, we use AWS SNS to email this file which constitutes the summary of the meeting in text to the meeting inviter.

Architecture Diagram

Prerequisites

For this walkthrough, you should have the following prerequisites:

• An AWS account

Walkthrough

The sections as enumerated below will walk you through the process of creating and recording a Chime SDK meeting:

• Setup AWS Cloud9 environment and clone the demo application
• Create an Amazon Elastic Container Registry (ECR) and register a Docker container image
• Deploy the Amazon Chime SDK Recording template
• Start a Amazon Chime SDK meeting with the serverless demo application and join with multiple participants
• Start the meeting recording
• Stop the recording and view the recording artifact

Create an AWS Cloud9 environment

1. Log into the AWS console with your AWS credentials and go to the AWS Cloud9 Console.
2. If you have previously set up a AWS Cloud9 environment in your account you can use it and skip this step entirely.
3. Press the Create environment button or go here.
4. For the Name enter and press the Next step button.
5. For Environment Settings
   • For Instance Type, select Other instance type and select t3.medium from the dropdown.
   • For Platform, select Ubuntu Server 18.04 LTS
   • Press the Next step button.
6. Review the Environment name and settings and press the Create environment button. You can optionally use your own Amazon EC2 instance if you have configured.
7. Wait for the environment to start.

Create a Amazon ECR repository, build and push the docker image to Amazon ECR

1. In the AWS Cloud9 instance, run the below command to create a repository in Amazon ECR

   aws ecr create-repository --repository-name <repository-name>
   

   You will get a response similar to this

   {
       "repository": {
           "repositoryArn": "arn:aws:ecr:us-east-1:123456789012:repository/chime-sdk-recording-demo",
           "registryId": "123456789012",
           "repositoryName": "chime-sdk-recording-demo",
           "repositoryUri": "123456789012.dkr.ecr.us-east-1.amazonaws.com/chime-sdk-recording-demo",
           "createdAt": 1585247726.0,
           "imageTagMutability": "MUTABLE",
           "imageScanningConfiguration": {
               "scanOnPush": false
           }
       }
   }
   

2. In the AWS Cloud9 instance, execute the following commands to download the recording demo.

   git clone https://github.com/shivampanjeta/chime-meeting-recordings.git
   cd amazon-chime-sdk-recording-demo
   

3. Execute the following command with the value of repositoryUri from step 1 to build and push the docker image to Amazon ECR

   make ECR_REPO_URI=<repositoryUri>
   

   Once the above commands execute successfully you will see an entry for the image in Amazon ECR as follows:

   

   Note: If this command fails due to AWS CLI version 2.0 not available, you can follow the instructions given here: Installing the AWS CLI version 2 on Linux and try again.

Deploy an Amazon Chime SDK Recording AWS CloudFormation Template

1. Execute the following command to create a AWS CloudFormation stack containing an Amazon ECS cluster, Amazon ECS task definition, Amazon S3 bucket, Amazon Lambda and an Amazon API Gateway deployment along with IAM roles and networking resources required for the Amazon ECS Cluster including an Amazon VPC, subnets, security groups, and an auto-scaling group.

   node ./deploy.js -b <my-bucket> -s <my-stack> -i <my-docker-image-uri> -r <region>
   

   Here is an example:

   node ./deploy.js -b recording-demo-cnf-deploy-bucket -s recording-demo-cnf-stack -i 123456789012.dkr.ecr.us-east-1.amazonaws.com/recording-demo:latest -r us-east-1
   

The above step deploys a AWS CloudFormation stack that creates resources needed to run the recording service. It may take several minutes to complete. You will get an Amazon API Gateway invoke URL in the output.

Create an Endpoint for Bert Extractive Summarizer Model on Sagemaker

1. In the AWS Cloud9 instance, run the below command to create a repository in Amazon ECR

   aws ecr create-repository --repository-name <repository-name>
   

   You will get a response similar to this

   {
       "repository": {
           "repositoryArn": "arn:aws:ecr:us-east-1:123456789012:repository/bert-sagemaker-inference",
           "registryId": "123456789012",
           "repositoryName": "chime-sdk-recording-demo",
           "repositoryUri": "123456789012.dkr.ecr.us-east-1.amazonaws.com/bert-sagemaker-inference",
           "createdAt": 1585247726.0,
           "imageTagMutability": "MUTABLE",
           "imageScanningConfiguration": {
               "scanOnPush": false
           }
       }
   }
   

2. In the AWS Cloud9 instance, execute the following commands to download the recording demo.

   git clone https://github.com/shivampanjeta/sagemaker-inference.git
   cd sagemaker-inference
   

3. Execute the following command to create a Amazon ECR repository, build and push the docker image to Amazon ECR

   sh build_and_push.sh
   

4. Create a Model on AWS Sagemaker by referencing to the ECR URI created after the last step.
5. Use that model for creating an Endpoint Configuration in Sagemaker.
6. Create an Endpoint with the created Enpoint Configuration in the last step, this Endpoint will be invoked from the summarizer lambda

Start an Amazon Chime SDK meeting with our serverless demo and join with multiple participants

1. At this point you can use our Amazon Chime SDK demo application by executing the following

   cd ../
   git clone https://github.com/aws/amazon-chime-sdk-js
   cd amazon-chime-sdk-js/demos/serverless
   

2. Deploy the demo using:

   npm run deploy -- -r us-east-1 -b <my-bucket> -s <my-stack-name> -a meeting
   

   The script will create an Amazon S3 bucket and AWS CloudFormation stack with Amazon Lambda and Amazon API Gateway resources required to run this demo. After the script finishes, it will output a URL that can be opened in a browser.

3. Open the Amazon Chime SDK serverless demo application using the link which was obtained in the previous step in multiple tabs in any web browser supported by the Amazon Chime SDK to simulate several participant joins. Optionally enable video or content sharing modalities for each participant in addition to audio.

Start the meeting recording

There are multiple ways for invoking a REST API in Amazon API Gateway. In this example we will use the Postman app.

1. Follow the steps highlighted in this document to install the Postman app.

   You need to select “AWS Signature” and add your AccessKey, SecretKey & AWS Region in the Authorization tab.

   

2. Start recording by passing the recordingAction as "start" and a url encoded meetingURL for our demo application as query parameters in the POST request to API Gateway.

   

   At this point the entire web page is captured by FFmpeg at 1280 X 720 fidelity and automatically transcoded and uploaded to Amazon S3. The file created in Amazon S3 will remain hidden until the capture is complete. Our demo application has been modified to suppress prompts for input device permissions and display customized UI for the recording bot.

   

Stop the recording and view the recording artifact

1. To stop the recording, we need to pass the task ARN as taskId that was received in the API response to start the recording in addition to the recordingAction denoting "stop".

   

2. Once the recording stops, open the AWS Console and navigate to Amazon S3. You will find a recording in the bucket chime-meeting-sdk-<aws-account-id>-<region>-recording-artifacts, with a key name YYYY/MM/DD/HH/.mp4.

   

Cleaning up

To avoid incurring future charges, please delete any resources in your account that you are not using such as files in Amazon S3, Amazon ECS and Amazon Lambda instances, AWS Cloud9 environment and Amazon API Gateway entries.

License

This project is licensed under the Apache-2.0 License.

Disclaimer: You and your end users understand that recording Amazon Chime SDK meetings with this feature may be subject to laws or regulations regarding the recording of electronic communications, and that it is your and your end users’ responsibility to comply with all applicable laws regarding the recording, including properly notifying all participants in a recorded session or to a recorded communication that the session or communication is being recorded and obtain their consent.

Disclaimer: Deploying the Amazon Chime SDK demo applications contained in this repository will cause your AWS Account to be billed for services, including the Amazon Chime SDK, used by the application.