docker-compose.modal.yml

services:
  # FhirFlame with Modal L4 GPU integration + A2A API
  fhirflame-modal:
    build:
      context: .
      dockerfile: Dockerfile
    image: fhirflame-modal:latest
    container_name: fhirflame-modal
    ports:
      - "${GRADIO_PORT:-7860}:7860"  # Gradio UI
    environment:
      - PYTHONPATH=/app
      - GRADIO_SERVER_NAME=0.0.0.0
      - DEPLOYMENT_TARGET=modal
      # Modal Configuration
      - ENABLE_MODAL_SCALING=${ENABLE_MODAL_SCALING:-true}
      - MODAL_TOKEN_ID=${MODAL_TOKEN_ID}
      - MODAL_TOKEN_SECRET=${MODAL_TOKEN_SECRET}
      - MODAL_ENDPOINT_URL=${MODAL_ENDPOINT_URL}
      - MODAL_L4_HOURLY_RATE=${MODAL_L4_HOURLY_RATE:-0.73}
      - MODAL_PLATFORM_FEE=${MODAL_PLATFORM_FEE:-15}
      # Environment
      - FHIRFLAME_DEV_MODE=${FHIRFLAME_DEV_MODE:-false}
      - FHIR_VERSION=${FHIR_VERSION:-R4}
      - ENABLE_HIPAA_LOGGING=${ENABLE_HIPAA_LOGGING:-true}
      # API Keys (from .env)
      - HF_TOKEN=${HF_TOKEN}
      - MISTRAL_API_KEY=${MISTRAL_API_KEY}
      # Fallback Configuration
      - USE_MISTRAL_FALLBACK=${USE_MISTRAL_FALLBACK:-true}
      - USE_MULTIMODAL_FALLBACK=${USE_MULTIMODAL_FALLBACK:-true}
      # Auth0 for production (optional)
      - AUTH0_DOMAIN=${AUTH0_DOMAIN:-}
      - AUTH0_AUDIENCE=${AUTH0_AUDIENCE:-}
    volumes:
      - ./src:/app/src
      - ./tests:/app/tests
      - ./logs:/app/logs
      - ./.env:/app/.env
    networks:
      - fhirflame-modal
    command: python frontend_ui.py
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:7860"]
      interval: 30s
      timeout: 10s
      retries: 3

  # A2A API Server with Modal integration
  fhirflame-a2a-modal:
    build:
      context: .
      dockerfile: Dockerfile
    image: fhirflame-modal:latest
    container_name: fhirflame-a2a-modal
    ports:
      - "${A2A_API_PORT:-8000}:8000"  # A2A API
    environment:
      - PYTHONPATH=/app
      - FHIRFLAME_DEV_MODE=${FHIRFLAME_DEV_MODE:-false}
      - FHIRFLAME_API_KEY=${FHIRFLAME_API_KEY:-fhirflame-modal-key}
      - PORT=8000
      # Auth0 Configuration for production
      - AUTH0_DOMAIN=${AUTH0_DOMAIN:-}
      - AUTH0_AUDIENCE=${AUTH0_AUDIENCE:-}
      # Modal Integration
      - MODAL_TOKEN_ID=${MODAL_TOKEN_ID}
      - MODAL_TOKEN_SECRET=${MODAL_TOKEN_SECRET}
      - MODAL_ENDPOINT_URL=${MODAL_ENDPOINT_URL}
    volumes:
      - ./src:/app/src
      - ./.env:/app/.env
    networks:
      - fhirflame-modal
    command: python -c "from src.mcp_a2a_api import app; import uvicorn; uvicorn.run(app, host='0.0.0.0', port=8000)"
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 30s
      timeout: 10s
      retries: 3

  # Modal deployment service
  modal-deployer:
    build:
      context: .
      dockerfile: Dockerfile
    image: fhirflame-modal:latest
    container_name: modal-deployer
    environment:
      - PYTHONPATH=/app
      - MODAL_TOKEN_ID=${MODAL_TOKEN_ID}
      - MODAL_TOKEN_SECRET=${MODAL_TOKEN_SECRET}
    volumes:
      - ./modal:/app/modal
      - ./.env:/app/.env
    networks:
      - fhirflame-modal
    working_dir: /app
    command: >
      sh -c "
        echo '🚀 Deploying Modal L4 GPU functions...' &&
        python modal/deploy.py --a2a &&
        echo '✅ Modal deployment complete!'
      "
    profiles:
      - deploy

  # HuggingFace fallback service (local backup)
  hf-fallback:
    build:
      context: .
      dockerfile: Dockerfile
    image: fhirflame-modal:latest
    container_name: hf-fallback
    environment:
      - PYTHONPATH=/app
      - HF_TOKEN=${HF_TOKEN}
      - DEPLOYMENT_TARGET=huggingface
    volumes:
      - ./src:/app/src
      - ./.env:/app/.env
    networks:
      - fhirflame-modal
    command: python -c "print('HuggingFace fallback ready')"
    profiles:
      - fallback

  # Test runner for Modal integration
  test-modal:
    build:
      context: .
      dockerfile: Dockerfile
    image: fhirflame-modal:latest
    container_name: fhirflame-modal-tests
    environment:
      - PYTHONPATH=/app
      - MODAL_TOKEN_ID=${MODAL_TOKEN_ID}
      - MODAL_TOKEN_SECRET=${MODAL_TOKEN_SECRET}
      - FHIRFLAME_DEV_MODE=true
    volumes:
      - ./src:/app/src
      - ./tests:/app/tests
      - ./test_results:/app/test_results
      - ./.env:/app/.env
    networks:
      - fhirflame-modal
    depends_on:
      - fhirflame-a2a-modal
    command: python tests/test_modal_scaling.py
    profiles:
      - test

  # Langfuse Database for monitoring
  langfuse-db:
    image: postgres:15
    container_name: langfuse-db-modal
    environment:
      - POSTGRES_DB=langfuse
      - POSTGRES_USER=langfuse
      - POSTGRES_PASSWORD=langfuse
    volumes:
      - langfuse_db_data:/var/lib/postgresql/data
    networks:
      - fhirflame-modal
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U langfuse -d langfuse"]
      interval: 10s
      timeout: 5s
      retries: 5
      start_period: 10s

  # Langfuse for comprehensive monitoring
  langfuse:
    image: langfuse/langfuse:latest
    container_name: langfuse-modal
    depends_on:
      langfuse-db:
        condition: service_healthy
    ports:
      - "${LANGFUSE_PORT:-3000}:3000"
    environment:
      - DATABASE_URL=postgresql://langfuse:langfuse@langfuse-db:5432/langfuse
      - NEXTAUTH_SECRET=mysecret
      - SALT=mysalt
      - NEXTAUTH_URL=http://localhost:3000
      - TELEMETRY_ENABLED=${TELEMETRY_ENABLED:-true}
      - NEXT_PUBLIC_SIGN_UP_DISABLED=${NEXT_PUBLIC_SIGN_UP_DISABLED:-false}
      - LANGFUSE_ENABLE_EXPERIMENTAL_FEATURES=${LANGFUSE_ENABLE_EXPERIMENTAL_FEATURES:-false}
    networks:
      - fhirflame-modal
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:3000/api/public/health"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 60s

networks:
  fhirflame-modal:
    driver: bridge

volumes:
  langfuse_db_data: