docker-compose.yml - master - storeksfeed/RAG-Bot

docker-compose.yml
234 строки7 KB
Semyon Glazyrin
Updated some Docker-related files
01 май 2026, 13:41
01 май 2026, 13:41a2e9a74
services:
  # =============================================================================
  # DATABASE SERVICES
  # =============================================================================
  
  # PostgreSQL - for Django, KV and Doc Status storage
  postgres:
    image: postgres:15.15
    container_name: postgres
    ports:
      - "127.0.0.1:5432:5432"
    volumes:
      - postgres_data:/var/lib/postgresql/data
    environment:
      - POSTGRES_USER=${POSTGRES_USER}
      - POSTGRES_PASSWORD=${POSTGRES_PASSWORD}
      - POSTGRES_DB=${POSTGRES_DATABASE}
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER} -d ${POSTGRES_DATABASE}"]
      interval: 10s
      timeout: 5s
      retries: 5
      start_period: 30s
    restart: unless-stopped

  # Qdrant - for vector storage
  qdrant:
    image: qdrant/qdrant:v1.16
    container_name: qdrant
    ports:
      - "127.0.0.1:6333:6333"  # REST API
      - "127.0.0.1:6334:6334"  # gRPC API
    volumes:
      - qdrant_data:/qdrant/storage
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'exec 3<>/dev/tcp/127.0.0.1/6333 && echo -e \"GET /readyz HTTP/1.1\r\nHost: localhost\r\nConnection: close\r\n\r\n\" >&3 && grep -q \"HTTP/1.1 200\" <&3'"]
      interval: 10s
      timeout: 5s
      retries: 5
      start_period: 10s
    restart: unless-stopped

  # Neo4j - for graph storage
  neo4j:
    image: neo4j:5.15.0
    container_name: neo4j
    ports:
      - "127.0.0.1:7474:7474"  # HTTP
      - "127.0.0.1:7687:7687"  # Bolt
    volumes:
      - neo4j_data:/data
      - neo4j_logs:/logs
    environment:
      # Authentication: format is username/password
      - NEO4J_AUTH=${NEO4J_USERNAME}/${NEO4J_PASSWORD}
      # # Enable APOC plugin (Advanced Procedures On Cypher - utility library for Neo4j)
      # - NEO4J_PLUGINS=["apoc"]
      # # Allow APOC procedures to run without security restrictions
      # - NEO4J_dbms_security_procedures_unrestricted=apoc.*
      # # Whitelist APOC procedures (allows all apoc.* procedures)
      # - NEO4J_dbms_security_procedures_allowlist=apoc.*
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'exec 3<>/dev/tcp/127.0.0.1/7474 && echo -e \"GET / HTTP/1.1\r\nHost: localhost\r\nConnection: close\r\n\r\n\" >&3 && grep -q \"HTTP/1.1 200\" <&3'"]
      interval: 10s
      timeout: 5s
      retries: 5
      start_period: 30s
    restart: unless-stopped

  # =============================================================================
  # LLAMA.CPP SERVICES (GPU-accelerated)
  # =============================================================================
  
  # Embedding service
  llamacpp-embed:
    build:
      context: .
      dockerfile: Dockerfile.llamacpp
    image: llamacpp:latest
    container_name: llamacpp-embed
    ports:
      - "127.0.0.1:8080:8080"
    volumes:
      - ./models:/models
    command: >
      --model /models/user-bge-m3-q8_0.gguf
      --host 0.0.0.0
      --port 8080
      --embedding
      --pooling mean
      --ctx-size 8192
      --batch-size 2048
      --ubatch-size 2048
      --n-gpu-layers 99
      --log-disable
    # GPU access
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'exec 3<>/dev/tcp/127.0.0.1/8080 && echo -e \"GET /health HTTP/1.1\r\nHost: localhost\r\nConnection: close\r\n\r\n\" >&3 && grep -q \"HTTP/1.1 200\" <&3'"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 60s
    restart: unless-stopped

  # LLM service
  llamacpp-llm:
    image: llamacpp:latest
    container_name: llamacpp-llm
    ports:
      - "127.0.0.1:8081:8080"
    volumes:
      - ./models:/models
    command: >
      --model /models/qwen3-14b-q4_k_m.gguf
      --host 0.0.0.0
      --port 8080
      --ctx-size 32768
      --batch-size 512
      --ubatch-size 1024
      --n-gpu-layers 99
      --threads 12
      --cont-batching
      --log-disable
      --jinja
      --flash-attn auto
      --no-context-shift
    # GPU access
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]
    healthcheck:
      test: ["CMD-SHELL", "curl -f http://8080/health || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 60s
    restart: unless-stopped

  # =============================================================================
  # LIGHTRAG SERVICE
  # =============================================================================
  
  lightrag:
    build:
      context: .
      dockerfile: Dockerfile.lightrag
    container_name: lightrag
    ports:
      - "127.0.0.1:9621:9621"
    env_file:
      - .env
    # Wait for all required services to be healthy
    depends_on:
      postgres:
        condition: service_healthy
      qdrant:
        condition: service_healthy
      neo4j:
        condition: service_healthy
      llamacpp-embed:
        condition: service_healthy
      llamacpp-llm:
        condition: service_healthy
    healthcheck:
      test: ["CMD-SHELL", "bash -c 'exec 3<>/dev/tcp/127.0.0.1/9621 && echo -e \"GET /health HTTP/1.1\r\nHost: localhost\r\nConnection: close\r\n\r\n\" >&3 && grep -q \"HTTP/1.1 200\" <&3'"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 30s
    restart: unless-stopped

  # =============================================================================
  # APP SERVICES
  # =============================================================================
  
  # Django admin interface
  app-admin:
    build:
      context: .
      dockerfile: Dockerfile.app
    image: app:latest
    container_name: app-admin
    ports:
      - "127.0.0.1:8000:8000"
    env_file:
      - .env
    command: /bin/sh -c "python manage.py collectstatic --noinput && python manage.py runserver 0.0.0.0:8000"
    depends_on:
      postgres:
        condition: service_healthy
    healthcheck:
      test: ["CMD-SHELL", "pgrep -f 'python manage.py runserver' || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 30s
    restart: unless-stopped

  # Bot worker
  app-bot:
    image: app:latest
    container_name: app-bot
    env_file:
      - .env
    command: python manage.py runbot vk
    depends_on:
      postgres:
        condition: service_healthy
      lightrag:
        condition: service_healthy
    healthcheck:
      test: ["CMD-SHELL", "pgrep -f 'python manage.py runbot' || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 30s
    restart: unless-stopped

volumes:
  postgres_data:
  qdrant_data:
  neo4j_data:
  neo4j_logs: