docker-compose.yml

version: '3.8'

services:
  # Backend service
  backend:
    build:
      context: ./backend
      dockerfile: Dockerfile
    ports:
      - "5001:5001"
    volumes:
      - ./uploads:/app/uploads
      - ./documents:/app/documents
    environment:
      - FLASK_ENV=${FLASK_ENV:-production}
    networks:
      - app-network
    depends_on:
      - llm-service
    restart: unless-stopped
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:5001/api/health"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 40s

  # Frontend service
  frontend:
    build:
      context: ./frontend
      dockerfile: Dockerfile
    ports:
      - "3000:3000"
    depends_on:
      - backend
    networks:
      - app-network
    # Updates for production React environment configuration
    environment:
      - REACT_APP_API_URL=${REACT_APP_API_URL:-http://localhost:5001/api}
    restart: unless-stopped
    healthcheck:
      test: ["CMD", "wget", "--spider", "http://localhost:3000"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 40s

  # LLM service - assuming this is a local LLM server like Ollama
  llm-service:
    image: ollama/ollama:latest
    ports:
      - "1234:1234"
    volumes:
      - ollama_models:/root/.ollama
    environment:
      - OLLAMA_HOST=${OLLAMA_HOST:-0.0.0.0}
    networks:
      - app-network
    restart: unless-stopped
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:1234/api/tags"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 40s
    # Automatically pull models on startup
    entrypoint: >
      sh -c "
        echo 'Starting Ollama service...' &&
        ollama serve &
        sleep 10 &&
        echo 'Pulling required models...' &&
        ollama pull text-embedding-bge-m3 &&
        ollama pull deepseek-r1-distill-qwen-32b-mlx &&
        echo 'Models pulled successfully' &&
        wait
      "

# Shared network
networks:
  app-network:
    driver: bridge

# Volumes for persistent data
volumes:
  uploads:
  documents:
  ollama_models: