arch-id-web

ArchitectureID.ai’s full stack web application.

For the ArchitectureID.ai model development repo, see arch-id-model.

Overview

This full stack web application serves ArchitectureID.ai’s data models with a end-user friendly UI. The application consists of three containerized microservices:

• models
  tensorflow/serving instance that hosts the tensorflow models generated by arch-id-model. Uses
• backend
  Browser-friendly python API to wrap the models service and handle image processing
• frontend
  User-friendly Typescript/React UI.

Configuration

Environment variables

• MODELS_AWS_ACCESS_KEY_ID (required):
  AWS access key for S3 bucket for models storage
• MODELS_AWS_SECRET_ACCESS_KEY (required):
  AWS secret key for S3 bucket for models storage
• MODELS_AWS_REGION (required):
  AWS region of S3 bucket for models storage
• MODELS_S3_ENDPOINT (required):
  AWS endpoint of S3 bucket for models storage (probably 's3.${MODELS_AWS_REGION}.amazonaws.com')
• EC2_HOST (required for deployment):
  EC2 instance host domain
• EC2_USER (required for deployment):
  EC2 instance user
• FRONTEND_AWS_ACCESS_KEY_ID (required for deployment):
  AWS secret key for S3 bucket for static site hosting
• FRONTEND_AWS_SECRET_ACCESS_KEY (required for deployment):
  AWS secret key for S3 bucket for static site hosting
• FRONTEND_AWS_DEFAULT_REGION (required for deployment):
  AWS region of S3 bucket for static site hosting
• FRONTEND_S3_BUCKET_NAME (required for deployment):
  AWS bucket name of S3 bucket for static site hosting

Development

To bring up a local development environment, use:

docker -f docker-compose.yml -f docker-compose.dev.yml up -d

Deployment

The included script ./scripts/deploy.sh will perform a quick single-node deployment to AWS. The script will and serve the models and backend services on the configured AWS EC2 instance at ports 8051 and 80, and build/push the frontend to a S3 bucket. The EC2 instance must have installations of docker and docker-compose V2 and