(This document is still work in progress)

This project sets up an AI photobooth system, leveraging AWS services including EC2, Lambda, S3, DynamoDB, and SQS. It deploys a machine learning model for generating images based on user input.

Project Structure

.
├── app
│   ├── api
│   └── frontend
└── iac
    ├── ai-photobooth.yaml  # CloudFormation template
    ├── functions.sh        # Helper functions for stack operations
    └── user-data.sh        # Startup script for the EC2 instance

Prerequisites

• AWS CLI installed and configured with appropriate IAM permissions.
• An existing EC2 Key Pair (EC2_KEY_PAIR_NAME environment variable).
• The EMAIL_FROM environment variable, containing the email address for sending notifications.

Environment Variables

Ensure the following environment variables are set:

• EC2_KEY_PAIR_NAME – Your EC2 Key Pair for SSH access.
• EMAIL_FROM – The email address to send notifications from.

AWS Resources Deployed

• EC2 Instance: Runs the Stable Diffusion WebUI API.
• S3 Bucket: Stores images and assets.
• DynamoDB Table: Manages metadata for image requests.
• SQS Queue: Manages requests for the Lambda function.
• Lambda Function: Processes image requests and interfaces with S3 and DynamoDB.
• Security Group: Controls SSH and HTTP access to the EC2 instance.

How to Deploy the CloudFormation Stack

Get all the helper functions first.

source

Create a Stack:

To create a new CloudFormation stack, run the following command:

create_stack stack-name

This command appends a random suffix to the stack-name to ensure uniqueness.

Monitor Stack Creation:

Run the following command to monitor the status of the stack creation:

list_stacks_active

Once the status of the stack is CREATE_COMPLETE, note down the full stack name with the appended suffix (e.g., stack-name-123abc).

Update the Lambda Function:

After the stack is created, update the Lambda function by passing the full stack name with the suffix:

update_lambda_function stack-name-123abc

Run the Stable Diffusion WebUI API:

To start the Stable Diffusion WebUI API on the EC2 instance:

run_sdxl_webui_api stack-name-123abc

This command may take some time during the initial run as it installs necessary packages. You might need to restart the server once the setup completes.

Get the Public IP of the EC2 Instance:

If you need to access the instance directly, retrieve its public IP address with:

get_instance_public_ip stack-name-123abc

Running the Frontend (Streamlit)

The frontend is built using Streamlit. To run it, you must first set up environment variables. The .env.example file contains the required variables:

Example .env Setup

Create a .env file in your project root directory and populate it with the following variables (using actual values from your stack):

export AWS_DEFAULT_REGION=your-region
export S3_BUCKET_NAME=your-s3-bucket-name
export DYNAMODB_TABLE_NAME=your-dynamodb-table-name
export SQS_QUEUE_URL=your-sqs-queue-url
export USER_EMAIL_DEMO=email-send-to@example.com

You can also use the provided .env.example file as a template.

Install Python Dependencies

Make sure to install the necessary dependencies. Run:

pip install -r app/frontend/requirements.txt

Run the Streamlit App

To start the frontend application, navigate to the app/frontend directory and run:

streamlit run main.py --server.port 8080

This will launch the Streamlit frontend on port 8080, allowing users to interact with the AI photobooth system.

Managing the Stack

List All Stacks:

To list all CloudFormation stacks, including deleted ones:

list_stacks

List Active Stacks:

To view only active stacks:

list_stacks_active

Delete a Stack:

To delete a stack (and clean up associated resources like S3 objects):

delete_stack stack-name-123abc

Notes:

The stack name passed during creation will automatically get a random suffix for uniqueness. Be sure to run list_stacks_active to retrieve the full name before further actions.