AI Agent For Manufacturing Data Contextualization

Prerequisites AWS account with ECR permissions, CloudFormation permissions. 1 IAM role with VPC, EC2, S3, Secrets Manager permissions.

Step 1: Pull the CloudFormation template from Git

Pull the CloudFormation template from Git using below repository link: https://github.com/medt-pds/AIAgentForManufacturingDataContextualization_Cloudformation_Template.git 

Copy the CloudFormation template in a S3 bucket.

Step 2: Deploy the CloudFormation template

Login to AWS console.

Go to Cloudformation console.

Click on create stack and select with New resources Standard.

Specify template choose from the following 1. S3 URL

Configure stack details: Stack name.

Enter values for below parameters. InstanceType KeyName AvailabilityZone1 (e.g us-east-1a) AvailabilityZone2 (e.g us-east-1b) EnvironmentName Neo4jEC2AMIID RegionName LLMRegionName

Configure rollback triggers to revert changes in case of errors.

Set stack options: Tag Permissions Stack policy

Select both the options available under section Capabilities.

Review and deploy

Monitor stack creation.

Check the following services has been provisioned or not: 1 VPC, 1 EC2, 1 S3 bucket, 1 Secrets Manager

Step 3: S3 bucket structure for raw data

Before running the Docker image, ensure that the raw data is available in an S3 bucket following the required structure. You can download the sample S3 bucket structure from the following Git repository: https://github.com/medt-pds/AIAgentForManufacturingDataContextualization_S3_Structure.git 

Step 4: Subscribe and Pull Image

Subscribe to "AI Agent For Manufacturing Data Contextualization" on AWS Marketplace

Authenticate Docker to ECR: aws ecr get-login-password --region us-east-1 | docker login --username AWS --password-stdin <marketplace-ecr-uri>

Pull the image: docker pull <marketplace-ecr-uri>/709825985650.dkr.ecr.us-east-1.amazonaws.com/domain_aware_data_contextualization_agent_repo:domain_aware_data_contextualization_agent

Step 5: Deploy Container docker run -e AWS_ACCESS_KEY_ID=<access_key> -e AWS_SECRET_ACCESS_KEY=<secret_access_key> -e REGION_NAME=<region_name> -p 8080:8080 --name <container_name> <image_id>

Step 6: Verify Deployment

Check if the docker image is running using postman: - Request: Get method http://localhost:8080/ping  response: {"status":"ok", "timestamp": <current time in epoch format>}

Check if the Knowledge graph is created in Neo4j hosted on an EC2 using below link: - http://<public-ip-of-EC2-instance>:7474

AWS Bedrock AgentCore: aws bedrock-agentcore-control create-agent-runtime
--agent-runtime-name "DataContextualization"
--agent-runtime-artifact '{"containerConfiguration": {"containerUri": "<your-ecr-uri>"}}'
--environment-variables '{"ACCESS_KEY_ID": "", "AWS_SECRET_ACCESS_KEY": "", "REGION_NAME": "***"}'

Note: Before running the Docker image, initial configuration is required based on the type of input data. This configuration depends on the specific data sources such as SAP, Maintenance Manuals, SOPs, Operating Manuals, P&IDs, and RCAs.