docker-compose.yaml

services:
  train:
    build:
      context: .
    command: |
      python -m src.train_optuna_callbacks experiment=catdog_experiment ++task_name=train ++train=True ++test=False && \
      python -m src.create_artifacts && \
      touch ./checkpoints/train_done.flag
    volumes:
      - ./data:/app/data
      - ./checkpoints:/app/checkpoints
      - ./artifacts:/app/artifacts
      - ./logs:/app/logs
    environment:
      - PYTHONUNBUFFERED=1
      - PYTHONPATH=/app
    shm_size: '4g'
    networks:
      - default
    env_file:
      - .env
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]

  eval:
    build:
      context: .
    command: |
      sh -c 'while [ ! -f /app/checkpoints/train_done.flag ]; do sleep 10; done && python -m src.train_optuna_callbacks experiment=catdog_experiment ++task_name=test ++train=False ++test=True'
    volumes:
      - ./data:/app/data
      - ./checkpoints:/app/checkpoints
      - ./artifacts:/app/artifacts
      - ./logs:/app/logs
    environment:
      - PYTHONUNBUFFERED=1
      - PYTHONPATH=/app
    shm_size: '4g'
    networks:
      - default
    env_file:
      - .env
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
  

  server:
    build:
      context: .
    command: |
      sh -c 'while [ ! -f /app/checkpoints/train_done.flag ]; do sleep 10; done && python -m src.server'
    volumes:
      - ./data:/app/data
      - ./checkpoints:/app/checkpoints
      - ./artifacts:/app/artifacts
      - ./logs:/app/logs
    environment:
      - PYTHONUNBUFFERED=1
      - PYTHONPATH=/app
      - SERVER_URL=http://localhost:8080
    shm_size: '4g'
    networks:
      - default
    env_file:
      - .env
    ports:
      - "8080:8080"
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
      
  client:
    build:
      context: .
    command: |
      sh -c 'until curl -s http://server:8080/health; do echo "Waiting for server to be ready..."; sleep 5; done && \
      ./run_client.sh'
    volumes:
      - ./data:/app/data
      - ./checkpoints:/app/checkpoints
      - ./artifacts:/app/artifacts
      - ./logs:/app/logs
    environment:
      - PYTHONUNBUFFERED=1
      - PYTHONPATH=/app
      - SERVER_URL=http://server:8080
    shm_size: '4g'
    networks:
      - default
    env_file:
      - .env

    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]

  
volumes:
  data:
  checkpoints:
  artifacts:
  logs:

networks:
  default: