hyungi_document_server/config.yaml

# hyungi_Document_Server 설정

ai:
  gateway:
    endpoint: "http://ai-gateway:8080"

  models:
    # ─── 단일 generation 호스트 routing (2026-05-14 GPU LLM 제거) ───
    # GPU Ollama gemma4:e4b-it-q8_0 제거. Mac mini 26B-A4B 가 triage + primary + classifier 모두 흡수.
    # fallback 은 Claude Sonnet 4 API (Mac mini 다운 시 자동 trigger, premium 과 budget 공유).
    # plan: ~/.claude/plans/rosy-launching-otter.md §C/§D/§E

    # triage: 상시 분류·요약·근거 선별. Mac mini 26B (primary 와 동일 endpoint, 짧은 max_tokens).
    triage:
      endpoint: "http://100.76.254.116:8801/v1/chat/completions"
      model: "mlx-community/gemma-4-26b-a4b-it-8bit"
      max_tokens: 4096
      timeout: 30
      context_char_limit: 120000

    # primary: 에스컬레이션 전용. 26B MLX (맥미니 Semaphore(1) 보호 대상).
    primary:
      endpoint: "http://100.76.254.116:8801/v1/chat/completions"
      model: "mlx-community/gemma-4-26b-a4b-it-8bit"
      max_tokens: 8192
      timeout: 180
      context_char_limit: 260000

    # fallback: primary 장애 시 최후 방어선. Claude Sonnet 4 API (소액 한도, 자동 trigger).
    # 호출 빈도 낮음 가정 (Mac mini 가 거의 항상 up) → premium 과 budget 공유 OK.
    fallback:
      endpoint: "https://api.anthropic.com/v1/messages"
      model: "claude-sonnet-4-20250514"
      max_tokens: 4096
      daily_budget_usd: 5.00
      require_explicit_trigger: false
      timeout: 120

    premium:
      endpoint: "https://api.anthropic.com/v1/messages"
      model: "claude-sonnet-4-20250514"
      max_tokens: 8192
      daily_budget_usd: 5.00
      require_explicit_trigger: true

    embedding:
      endpoint: "http://ollama:11434/api/embeddings"
      model: "bge-m3"

    rerank:
      endpoint: "http://reranker:80/rerank"
      model: "bge-reranker-v2-m3"

    # Phase 3.5a answerability classifier. 2026-05-14 GPU LLM 제거 후 Mac mini 26B 로 swap.
    # classifier_service 가 hasattr 체크로 optional 이므로 이 섹션 제거 시 classifier gate 는 자동 skip (score-only).
    classifier:
      endpoint: "http://100.76.254.116:8801/v1/chat/completions"
      model: "mlx-community/gemma-4-26b-a4b-it-8bit"
      max_tokens: 512
      timeout: 30  # 2026-05-17: 15s 도 동시 부하 시 elapsed 14.4s 직전이라 tight — 30s 로 2x 마진 (Mac mini 26B concurrent load). classifier_service.LLM_TIMEOUT_MS=30000 와 align
    # 제거: vision (미사용)

  # ─── deep_summary enqueue 폭발 억제 (B-1 R2) ───
  # 초기 튜닝 전 deep_summary 큐에 soft escalate 가 과발생하면 MLX 26B 가 포화된다.
  # 아래 임계치 중 하나라도 초과하면 soft escalate (recommend_deep_summary 만) 를
  # suppress. hard escalate (long_context / triage_json_invalid / low_confidence)는
  # 절대 suppress 되지 않는다.
  deep_summary_backlog:
    ratio_threshold: 0.3      # 지난 window 의 deep_n/classify_n
    pending_threshold: 5      # deep_summary stage 의 pending+processing
    window_minutes: 30

# ─── /api/search/ask backend dispatcher (PR-MacBook-RAG-Backend-1) ───
# backend 미지정 (default) → Gemma Mac mini (settings.ai.primary 경로 그대로, 변동 0).
# backend="qwen-macbook" 명시 opt-in → MacBook M5 Max mlx-vlm.server. unavailable 시 503.
# 자동 fallback 없음 ([[macbook-inference-endpoint-role]] Invariant 1).
search:
  ask:
    backend:
      macmini_url: "http://100.76.254.116:8801"  # Gemma 경로 = settings.ai.primary 가 권위, 본 키는 spec 일관성 + 변경 추적용
      macbook_url: "http://100.118.112.84:8810"  # MacBook M5 Max Tailscale interface bind
      macbook_model: "mlx-community/Qwen3.6-27B-8bit"
      timeout_connect_s: 1   # MacBook sleep/wake 빠른 감지 (자동 fallback 부재 → 빠른 503)
      timeout_read_s: 30     # synthesis_service.LLM_TIMEOUT_MS=30000 와 align
    # PR-DocSrv-Ask-ToolCalling-ReAct-1: /api/search/ask/react ReAct loop (qwen-macbook only)
    react:
      enabled: true
      max_tool_rounds: 2     # G0-2: LLM 호출 최대 3회 (tool round 2 + final 1), search 실행 최대 2회
      search_tool_limit: 5
      search_tool_mode: "hybrid"

nas:
  mount_path: "/documents"
  pkm_root: "/documents/PKM"

# ─── 업로드 한도 정책 (authoritative) ───
# 프록시(home-caddy 등) request_body 한도는 max_bytes * content_length_slack_ratio 이상 유지.
upload:
  max_bytes: 100000000              # 100 MB (SI). 업로드 실제 제한의 단일 진실 공급원.
  content_length_slack_ratio: 1.05  # multipart form 오버헤드(헤더/바운더리) 여유.
  stream_chunk_bytes: 1048576       # 1 MiB 단위 스트리밍 read/write.

# ─── 문서 분류 체계 ───
taxonomy:
  Philosophy:
    Ethics: []
    Metaphysics: []
    Epistemology: []
    Logic: []
    Aesthetics: []
    Eastern_Philosophy: []
    Western_Philosophy: []
  Language:
    Korean: []
    English: []
    Japanese: []
    Translation: []
    Linguistics: []
  Engineering:
    Mechanical: [Piping, HVAC, Equipment]
    Electrical: [Power, Instrumentation]
    Chemical: [Process, Material]
    Civil: []
    Network: [Server, Security, Infrastructure]
  Industrial_Safety:
    Legislation: [Act, Decree, Foreign_Law, Korea_Law_Archive, Enforcement_Rule, Public_Notice, SAPA]
    Theory: [Industrial_Safety_General, Safety_Health_Fundamentals]
    Academic_Papers: [Safety_General, Risk_Assessment_Research]
    Cases: [Domestic, International]
    Practice: [Checklist, Contractor_Management, Safety_Education, Emergency_Plan, Patrol_Inspection, Permit_to_Work, PPE, Safety_Plan]
    Risk_Assessment: [KRAS, JSA, Checklist_Method]
    Safety_Manager: [Appointment, Duty_Record, Improvement, Inspection, Meeting]
    Health_Manager: [Appointment, Duty_Record, Ergonomics, Health_Checkup, Mental_Health, MSDS, Work_Environment]
  Programming:
    Programming_Language: [Python, JavaScript, Go, Rust]
    Framework: [FastAPI, SvelteKit, React]
    DevOps: [Docker, CI_CD, Linux_Administration]
    AI_ML: [Large_Language_Model, Computer_Vision, Data_Science]
    Database: []
    Software_Architecture: []
  General:
    Reading_Notes: []
    Self_Development: []
    Business: []
    Science: []
    History: []

document_types:
  - Reference
  - Standard
  - Manual
  - Drawing
  - Template
  - Note
  - Academic_Paper
  - Law_Document
  - Report
  - Memo
  - Checklist
  - Meeting_Minutes
  - Specification
  - 발주서
  - 세금계산서
  - 명세표
  - 도면
  - 증명서
  - 계획서
  - 시방서

schedule:
  law_monitor: "07:00"
  mailplus_archive: ["07:00", "18:00"]
  daily_digest: "20:00"
  file_watcher_interval_minutes: 5
  queue_consumer_interval_minutes: 10