hyungi_document_server/docker-compose.yml

services:
  postgres:
    image: pgvector/pgvector:pg16
    volumes:
      - pgdata:/var/lib/postgresql/data
      - ./migrations:/docker-entrypoint-initdb.d
    environment:
      POSTGRES_DB: pkm
      POSTGRES_USER: pkm
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD}
    ports:
      - "100.110.63.63:15432:5432"
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U pkm"]
      interval: 5s
      timeout: 5s
      retries: 5
    restart: unless-stopped

  kordoc-service:
    build: ./services/kordoc
    ports:
      - "127.0.0.1:3100:3100"
    volumes:
      - ${NAS_NFS_PATH:-/mnt/nas/Document_Server}:/documents:ro
    mem_limit: 4g
    memswap_limit: 4g
    healthcheck:
      test: ["CMD", "node", "-e", "fetch('http://localhost:3100/health').then(r=>{process.exit(r.ok?0:1)}).catch(()=>process.exit(1))"]
      interval: 10s
      timeout: 5s
      retries: 3
    restart: unless-stopped

  ocr-service:
    build: ./services/ocr
    expose:
      - "3200"
    volumes:
      - ${NAS_NFS_PATH:-/mnt/nas/Document_Server}:/documents:ro
      - ocr_models:/root/.cache
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
    healthcheck:
      test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:3200/health')"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 180s
    restart: unless-stopped

  # Phase 1B (2026-05-01): PDF → markdown 변환. ocr-service 와 별도 컨테이너 (deps 충돌 회피).
  marker-service:
    build: ./services/marker
    ports:
      - "127.0.0.1:3300:3300"
    expose:
      - "3300"
    environment:
      - HF_HOME=/models/huggingface
      - TORCH_HOME=/models/torch
    volumes:
      - ${NAS_NFS_PATH:-/mnt/nas/Document_Server}:/documents:ro
      - marker_models:/models
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:3300/ready"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 300s
    restart: unless-stopped

  stt-service:
    # 2026-05-08 (D9 Track B revised): GPU is canonical STT owner.
    # 정책: Mac mini = Gemma 26B 전용 우선이므로 STT/Whisper 는 호출량 무관 GPU 서버 소유.
    # 이전 "Mac mini 이전본" 주석은 trace 오인 기반이었고 본 revised 결정으로 폐기.
    # fastapi 의 STT_ENDPOINT 는 `http://stt-service:3300` (compose 내부 DNS) 사용.
    build: ./services/stt
    expose:
      - "3300"
    volumes:
      - ${NAS_NFS_PATH:-/mnt/nas/Document_Server}:/documents:ro
      - stt_models:/root/.cache
    environment:
      - WHISPER_MODEL=${WHISPER_MODEL:-large-v3}
      - WHISPER_DEVICE=${WHISPER_DEVICE:-cuda}
      - WHISPER_COMPUTE_TYPE=${WHISPER_COMPUTE_TYPE:-float16}
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
    healthcheck:
      # /ready: CUDA 디바이스 + 모델 적재 둘 다 확인. ready=true 만 healthy 처리.
      # /health 는 단순 liveness 라 모델 미적재 상태도 healthy 로 잡혀 운영 신호로 부적합.
      test: ["CMD", "python3", "-c", "import json,urllib.request,sys; r=urllib.request.urlopen('http://localhost:3300/ready'); sys.exit(0 if json.load(r).get('ready') else 1)"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 300s
    restart: unless-stopped

  ollama:
    image: ollama/ollama
    volumes:
      - ollama_data:/root/.ollama
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
    ports:
      - "127.0.0.1:11434:11434"
    restart: unless-stopped

  # Phase 1.3: bge-reranker-v2-m3 (TEI) — internal only, fastapi에서 reranker:80으로 호출
  # fastapi가 depends_on 안 함 → 단독 시작 가능, 없어도 fastapi 동작 (rerank=false fallback)
  reranker:
    image: ghcr.io/huggingface/text-embeddings-inference:1.7
    container_name: hyungi_document_server-reranker-1
    expose:
      - "80"
    environment:
      - MODEL_ID=BAAI/bge-reranker-v2-m3
      # PR-2Q-Rerank-Payload-Fix (2026-05-24): 2 env 변경 — 413 root cause 분리.
      #   (a) MAX_BATCH_TOKENS 8192 → 16384: 각 batch 의 token sum 한도
      #   (b) MAX_CLIENT_BATCH_SIZE 32 → 64: 1 request 안 entries 수 한도 (default 32).
      #       multi-query cap 60 chunks (×2 chunks_per_doc dedup) 가 batch entries 54~60
      #       → 32 한도 초과 → 413. 64 로 늘림.
      # GPU VRAM free 6199MiB 충분. baseline path (MAX_RERANK_INPUT=200) 영향 0.
      - MAX_BATCH_TOKENS=16384
      - MAX_CLIENT_BATCH_SIZE=64
      - MAX_CONCURRENT_REQUESTS=4
    volumes:
      - reranker_cache:/data
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
    healthcheck:
      test: ["CMD", "curl", "-fsS", "http://localhost/health"]
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 120s
    restart: unless-stopped

  ai-gateway:
    build: ./gpu-server/services/ai-gateway
    ports:
      - "127.0.0.1:8081:8080"
    environment:
      - PRIMARY_ENDPOINT=http://100.76.254.116:8801/v1/chat/completions
      - FALLBACK_ENDPOINT=http://ollama:11434/v1/chat/completions
      - CLAUDE_API_KEY=${CLAUDE_API_KEY:-}
      - DAILY_BUDGET_USD=${DAILY_BUDGET_USD:-5.00}
    depends_on:
      - ollama
    restart: unless-stopped

  fastapi:
    build: ./app
    ports:
      - "100.110.63.63:8000:8000"
    volumes:
      - ${NAS_NFS_PATH:-/mnt/nas/Document_Server}:/documents
      - ./config.yaml:/app/config.yaml:ro
      - ./domain_policy.yaml:/app/domain_policy.yaml:ro
      - ./scripts:/app/scripts:ro
      - ./logs:/app/logs
      - ./migrations:/app/migrations:ro
    depends_on:
      postgres:
        condition: service_healthy
      kordoc-service:
        condition: service_healthy
      marker-service:
        condition: service_healthy
    env_file:
      - credentials.env
    environment:
      - DATABASE_URL=postgresql+asyncpg://pkm:${POSTGRES_PASSWORD}@postgres:5432/pkm
      - KORDOC_ENDPOINT=http://kordoc-service:3100
      - OCR_ENDPOINT=http://ocr-service:3200
      - MARKER_ENDPOINT=http://marker-service:3300
      - MARKER_CONTAINER_PATH_PREFIX=/documents
      # 2026-05-08 (D9 Track B revised): GPU stt-service 정식 승격, 내부 DNS 사용.
      - STT_ENDPOINT=http://stt-service:3300
      # KGS Code 등 외부 학습 자료 추가 스캔 경로 (host .env 에서 주입). 빈 값이면 비활성.
      - ADDITIONAL_WATCH_TARGETS=${ADDITIONAL_WATCH_TARGETS:-}
      # PR-MacMini-Derived-Worker-1
      - STUDY_EXPLANATION_ENABLED=${STUDY_EXPLANATION_ENABLED:-true}
      - INTERNAL_WORKER_TOKEN=${INTERNAL_WORKER_TOKEN}
      # Voice Memo PoC v1 — bot 계정 한정 long-expiry access token. default false → 일반 운영 영향 0.
      # 활성화: host .env 에 VOICE_MEMO_BOT_TOKEN_ENABLED=true. plan: rosy-launching-otter.md
      - VOICE_MEMO_BOT_TOKEN_ENABLED=${VOICE_MEMO_BOT_TOKEN_ENABLED:-false}
      - VOICE_MEMO_BOT_USERNAME=${VOICE_MEMO_BOT_USERNAME:-voice-memo-bot}
      - VOICE_MEMO_BOT_TOKEN_EXPIRE_DAYS=${VOICE_MEMO_BOT_TOKEN_EXPIRE_DAYS:-365}
      # PR-Notebook-Client-1 — notebook-client-bot long-expiry token (laptop-worker-bot wrapper 재활용,
      # 1B Pilot dormant 후 username rename). default false → 운영 영향 0.
      - LAPTOP_WORKER_BOT_TOKEN_ENABLED=${LAPTOP_WORKER_BOT_TOKEN_ENABLED:-false}
      - LAPTOP_WORKER_BOT_USERNAME=${LAPTOP_WORKER_BOT_USERNAME:-laptop-worker-bot}
      - LAPTOP_WORKER_BOT_TOKEN_EXPIRE_DAYS=${LAPTOP_WORKER_BOT_TOKEN_EXPIRE_DAYS:-365}
      # PR-2 of DS AI routing policy (2026-05-23) — backends dispatcher via llm-router.
      # router_url default points at Mac mini Tailscale interface :8890 (PR-1).
      # DS_BACKENDS_VIA_ROUTER=false 로 legacy 직접 호출 path 즉시 복귀.
      - LLM_ROUTER_URL=${LLM_ROUTER_URL:-http://100.76.254.116:8890}
      - DS_BACKENDS_VIA_ROUTER=${DS_BACKENDS_VIA_ROUTER:-true}
    restart: unless-stopped

  frontend:
    build: ./frontend
    ports:
      - "127.0.0.1:3000:3000"
    depends_on:
      - fastapi
    restart: unless-stopped

  caddy:
    image: caddy:2
    ports:
      - "8080:80"
    volumes:
      - ./Caddyfile:/etc/caddy/Caddyfile
      - caddy_data:/data
    depends_on:
      - fastapi
      - frontend
    restart: unless-stopped

volumes:
  pgdata:
  caddy_data:
  ollama_data:
  reranker_cache:
  ocr_models:
  stt_models:
  marker_models: