Web-based Python Application (Flask) to predict/recognize faces.

1. Clone project.

2. Create virtual environment for local development and install project dependencies:

   pip install -r requirements.txt.

3. Configure local dev by creating .env file and specifying project config (see the content of .env.example)

4. Run docker-compose -f docker-compose.yml -f docker-compose.override.yml -f docker-compose.access.yml up

5. Access service via browser 127.0.0.1:80 (localhost) or use curl to perform POST calls to respective endpoints:

   • /api/v1/predictions/ (eg. curl -F "file=@000323.png" http://127.0.0.1/api/v1/predictions/)
   • /api/v1/recognitions/

Project is configured to be deployed in Google Cloud Platform via Google Kubernetes Engine.

To deploy a new version of app just run ./deploy.sh.

Note: Make sure that kubectl and gcloud CLI tools are installed and properly configured.

• Add batch image processing support for face detection.
• Implement face recognition endpoint - /api/v1/recognitions/.
• Add authorization flow.
• Add tests.