streamlit-llm-chat-aws

Streamlit LLM Chat App on AWS

reserve

create a user pool

create a Cognito user pool and a user pool client.

aws cloudformation deploy --template-file cloudformation/userpool.yaml --stack-name streamlit-chatgpt-userpool

The following three values are obtained.

• userpool id
• app client id
• client secret

1. userpool id:
   To obtain a list of user pools, execute the following command.

$ aws cognito-idp list-user-pools --max-results 20 | jq ".UserPools[] | {Id, Name}"

The value of --max-results can be adjusted accordingly to define the maximum number of user pools to be displayed.
The results of the run will show a list of available Cognito user pools, including the ID of each.

2. app client id:
   Retrieve a list of app clients in the user pool using the retrieved user pool ID.

$ aws cognito-idp list-user-pool-clients --user-pool-id <userPoolId>

Replace userPoolId with the user pool ID obtained earlier. This command will list the ID for each client.

3. client secret:
   Once you know the app client ID, use the following command to obtain the client secret.

$ aws cognito-idp describe-user-pool-client --user-pool-id <userPoolId> --client-id <AppClientId>

Replace userPoolId with the user pool ID and AppClientId with the app client ID. That is the client secret.

create a user

• Run the aws cognito-idp admin-create-user command with the following parameters:
  • –user-pool-id: The ID of the user pool that you want to add a user to.
  • –username: The user name for the user.
  • –user-attributes: The user attributes for the user, such as email, phone number, name, etc. You can specify multiple attributes by using the syntax Name=value,Name=value,…
  • –message-action: The option to send a verification email or SMS message to the user. You can choose SUPPRESS to suppress the message, or RESEND to resend the message if the user already exists.

environment setting

Write the OpenAPI access key, the model to be used, the user pool ID you just created, the app client ID, and the client secret in the .env file.

OPENAI_API_KEY=XX-XXXXX...
OPENAI_MODEL=gpt-4-1106-preview
POOL_ID=us-east-1_XXXXXX
APP_CLIENT_ID=XXXXX
APP_CLIENT_SECRET=XXXXXXX

execute

build a container

docker build ./ -t streamlit-app

deploy on local

docker compose up -d

deploy on AWS

create a repository in ECR

aws ecr create-repository --repository-name streamlit-app --region us-east-1

aws ecr get-login-password --region us-east-1 | docker login --username AWS --password-stdin xxxxxxxx.dkr.ecr.us-east-1.amazonaws.com/streamlit-app