Skip to content

Latest commit

 

History

History
106 lines (85 loc) · 3.51 KB

README.md

File metadata and controls

106 lines (85 loc) · 3.51 KB
Raw

Generative AI Application Builder on AWS - WebApp

This project is the web interface (UI Application) that provides the front-end experience. The application is based on Reactjs framework and uses components from the AWS Cloudscape Design System

Local Configuration Setup

To build and run the application locally, the setup requires

Follow the below steps before building the web app for local execution

  • The backend infrastructure stacks from source/infrastructure are deployed in your AWS account
  • Create a file source/ui-chat/public/runtimeConfig.json (if one does not exist) by executing
mkdir -p source/ui-chat/public
touch source/ui-chat/public/runtimeConfig.json
  • From the AWS CloudFormation console, navigate to the Outputs tab of the main/ parent stack deployed and copy the Value of the Key named WebConfigKey.
  • Navigate to AWS Systems Manager Parameter Store in the AWS web console, search for the Key in the previous step and copy the contents into the file created in the previous (step #2) steps (source/ui-chat/public/runtimeConfig.json)

For reference, the string in the Parameter Store should look something like the below:

{
    "ModelFamilyParams": {
        "amazon": <Prompt>,
        "anthropic": <Prompt>,
        "ai21": <Prompt>
    },
    "UserPoolId": "us-east-1_mCkSfPSu6",
    "SupportedModels": [
        "ai21.j2-ultra",
        "ai21.j2-mid",
        "amazon.titan-text-express-v1",
        "anthropic.claude-v1",
        "anthropic.claude-v2",
        "anthropic.claude-instant-v1",
        "meta.llama2-13b-chat-v1",
        "meta.llama2-70b-chat-v1",
        "cohere.command-text-v14",
        "cohere.command-light-text-v14"
    ],
    "AllowsStreaming": "true",
    "IsInternalUser": "true",
    "UserPoolClientId": "bs147rk5tak12cnq080ajsieh",
    "ModelProviderName": "Bedrock",
    "ApiEndpoint": "wss://4g3lbfhl3k.execute-api.us-east-1.amazonaws.com",
    "SocketURL": "wss://4g3lbfhl3k.execute-api.us-east-1.amazonaws.com/prod",
    "AwsRegion": "us-east-1",
    "UseCaseConfig": {
        "UseCaseName": "demo-bedrock-aws-docs",
        "ConversationMemoryType": "DynamoDB",
        "KnowledgeBaseType": "Kendra",
        "KnowledgeBaseParams": {
            "NumberOfDocs": 5,
            "ReturnSourceDocs": false
        },
        "LlmParams": {
            "ModelProvider": "Bedrock",
            "ModelId": "anthropic.claude-v1",
            "ModelParams": {},
            "PromptTemplate": "",
            "Streaming": true,
            "Verbose": false,
            "Temperature": 0.1,
            "RAGEnabled": true
        }
    }
}

After completing the above steps, you can run the web application locally.

Build and Run the App Locally

  1. From the project root directory, change directory to source/ui-chat
    cd source/ui-chat
  1. Install the library modules if building the project for the first time
    npm install
  1. Building the project with the below command will generate a build folder which contains the compiled components and minified javascript files.
    npm run build
  1. Executing the following command will run a local server on port 3000 (http://localhost:3000)
    npm start

You should now be able to log in with the User Id and Password for which you should have received an email during deployment. You can also create additional users in the Amazon Cognito User Pool from the AWS web console.