Skip to content

Latest commit

 

History

History
55 lines (44 loc) · 2.38 KB

README.md

File metadata and controls

55 lines (44 loc) · 2.38 KB
Raw

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