ltcai 9.0.0 → 9.1.0

package/auto_setup.py

@@ -1,857 +1,11 @@
-"""
-Lattice AI — Zero-Config Auto Setup
-===================================
+"""Compatibility shim for :mod:`latticeai.setup.auto_setup`."""
 
-명세: ``lattice_ai_full_spec.pptx`` 슬라이드 16·17 (자동 환경 매트릭스 + 5단계 흐름)
-
-5단계
------
-①  PROBE       OS · CPU · GPU · RAM · 디스크 · 가속 SDK 감지
-②  RECOMMEND   사양 점수 → 최적 모델 / 런타임 / 양자화 자동 선택
-③  INSTALL     OS별 패키지 매니저 어댑터 호출 (winget · brew · apt · 스토어)
-④  VERIFY      추론 토큰/초 측정, 첫 응답 지연, 메모리 누수 점검
-⑤  PRESET      기본/고급 모드 분기 + 단축키·MCP·테마 적용
-
-원칙
-----
-- **표준 라이브러리 only.**  외부 패키지 import 는 모두 try/except 로 감싼다.
-- **변경하지 않는다, 추천만 한다.**  INSTALL 단계는 *실행 명령어* 를 생성하고
-  돌려보낼 뿐, ``--apply`` 플래그 없이는 시스템을 건드리지 않는다.
-- **모든 출력은 JSON-직렬화 가능**해야 UI(설치 마법사 화면) 에서 그대로 표시 가능.
-
-사용
-----
-```bash
-python3 auto_setup.py probe                  # 1단계만
-python3 auto_setup.py recommend              # 1+2단계
-python3 auto_setup.py plan                   # 1+2+3 (설치 계획 출력)
-python3 auto_setup.py plan --apply --confirm-token <token>
-python3 auto_setup.py verify                 # 4단계 단독
-python3 auto_setup.py preset                 # 5단계
-python3 auto_setup.py all                    # 전체 흐름
-```
-"""
-
-from __future__ import annotations
-
-import argparse
-import json
-import os
-import platform
-import re
-import shutil
-import subprocess
-import sys
-import time
-from dataclasses import asdict, dataclass, field
-from pathlib import Path
-from typing import Any, Dict, List, Optional, Tuple
-
-from latticeai.services.process_audit import (
-    CommandConfirmationError,
-    append_process_audit_event,
-    command_plan,
-    command_plan_for_commands,
-    require_command_confirmation,
-)
-from latticeai.services.setup_detection import (
-    detect_cuda,
-    detect_tools,
-    detect_wsl_from_text,
-    parse_windows_video_controllers as _parse_windows_video_controllers,
-)
-
-__all__ = [
-    "SystemProfile", "Recommendation", "InstallPlan",
-    "probe", "recommend", "plan", "verify", "preset", "run_all",
-]
-
-
-# ── 1. PROBE ────────────────────────────────────────────────────────────────
-@dataclass
-class GPUInfo:
-    vendor: str = "unknown"          # nvidia | amd | intel | apple | none
-    model: str = ""
-    vram_mb: int = 0
-    sdk: List[str] = field(default_factory=list)   # ['cuda', 'metal', 'mlx', ...]
-
-
-@dataclass
-class SystemProfile:
-    os: str = ""                     # windows | darwin | linux | ios | android
-    os_version: str = ""
-    arch: str = ""                   # x86_64 | arm64 | …
-    cpu_model: str = ""
-    cpu_cores: int = 0
-    cpu_logical_cores: int = 0
-    cpu_instructions: List[str] = field(default_factory=list)
-    ram_mb: int = 0
-    disk_free_mb: int = 0
-    gpu: GPUInfo = field(default_factory=GPUInfo)
-    package_manager: Optional[str] = None   # winget | brew | apt | dnf | pacman
-    has_internet: bool = True
-    python_version: str = ""
-    is_wsl: bool = False
-    wsl_version: str = ""
-    cuda_available: bool = False
-    cuda_version: str = ""
-    tools: Dict[str, str] = field(default_factory=dict)
-
-    def score(self) -> int:
-        """LLM 적합도 점수 (0..100). RECOMMEND 의 입력."""
-        s = 0
-        s += min(self.cpu_cores * 2, 24)
-        s += min(self.ram_mb // 1024 * 2, 40)
-        s += min(self.gpu.vram_mb // 1024 * 4, 36)
-        return min(s, 100)
-
-    def to_json(self) -> Dict[str, Any]:
-        d = asdict(self)
-        d["score"] = self.score()
-        return d
-
-
-def _read_text(path: str) -> str:
-    try:
-        return Path(path).read_text(encoding="utf-8", errors="replace")
-    except Exception:
-        return ""
-
-
-def _run(cmd: List[str], timeout: float = 4.0) -> str:
-    try:
-        out = subprocess.run(cmd, capture_output=True, text=True,
-                             timeout=timeout, check=False)
-        return (out.stdout or "") + (out.stderr or "")
-    except Exception:
-        return ""
-
-
-def _windows_candidate_paths(binary: str) -> List[str]:
-    local_appdata = os.environ.get("LOCALAPPDATA", "")
-    program_files = os.environ.get("ProgramFiles", r"C:\Program Files")
-    program_files_x86 = os.environ.get("ProgramFiles(x86)", r"C:\Program Files (x86)")
-    candidates = {
-        "ollama": [
-            str(Path(local_appdata) / "Programs" / "Ollama" / "ollama.exe") if local_appdata else "",
-            str(Path(program_files) / "Ollama" / "ollama.exe"),
-        ],
-        "lms": [
-            str(Path(local_appdata) / "Programs" / "LM Studio" / "resources" / "app" / ".webpack" / "lms.exe") if local_appdata else "",
-            str(Path(program_files) / "LM Studio" / "resources" / "app" / ".webpack" / "lms.exe"),
-        ],
-        "nvidia-smi": [
-            str(Path(program_files) / "NVIDIA Corporation" / "NVSMI" / "nvidia-smi.exe"),
-            str(Path(program_files_x86) / "NVIDIA Corporation" / "NVSMI" / "nvidia-smi.exe"),
-        ],
-    }
-    return [item for item in candidates.get(binary, []) if item]
-
-
-def _which(binary: str) -> Optional[str]:
-    found = shutil.which(binary)
-    if found:
-        return found
-    if platform.system() == "Windows":
-        for candidate in _windows_candidate_paths(binary):
-            if Path(candidate).exists():
-                return candidate
-    return None
-
-
-def _detect_gpu(prof_os: str, arch: str) -> GPUInfo:
-    """OS별 휴리스틱으로 GPU 감지. 외부 라이브러리 없이 가능한 만큼만."""
-    gpu = GPUInfo()
-
-    # NVIDIA
-    nvidia_smi = _which("nvidia-smi")
-    if nvidia_smi:
-        info = _run([nvidia_smi, "--query-gpu=name,memory.total",
-                     "--format=csv,noheader,nounits"])
-        if info.strip():
-            first = info.strip().splitlines()[0]
-            try:
-                name, mem = [x.strip() for x in first.split(",", 1)]
-                gpu.vendor = "nvidia"
-                gpu.model = name
-                gpu.vram_mb = int(float(mem))
-                gpu.sdk.append("cuda")
-            except ValueError:
-                pass
-
-    # Apple Silicon / Metal
-    if prof_os == "darwin":
-        sp = _run(["system_profiler", "SPDisplaysDataType"], timeout=6.0)
-        if "Apple" in sp and arch == "arm64":
-            gpu.vendor = "apple"
-            for line in sp.splitlines():
-                if "Chipset Model" in line:
-                    gpu.model = line.split(":", 1)[-1].strip()
-                    break
-            # Apple Silicon 의 GPU 메모리는 통합 메모리 = RAM. 별도 표기 안 함.
-            gpu.sdk.extend(["metal", "mlx" if _has_module("mlx") else ""])
-            gpu.sdk = [s for s in gpu.sdk if s]
-
-    # Windows
-    if prof_os == "windows" and gpu.vendor == "unknown":
-        shell = _which("powershell") or _which("pwsh")
-        info = ""
-        if shell:
-            info = _run([
-                shell, "-NoProfile", "-Command",
-                "Get-CimInstance Win32_VideoController | Select-Object Name,AdapterRAM | ConvertTo-Json -Compress",
-            ], timeout=8.0)
-        if not info:
-            info = _run(["wmic", "path", "win32_VideoController", "get",
-                         "Name,AdapterRAM", "/format:list"])
-        controllers = _parse_windows_video_controllers(info)
-        if controllers:
-            primary = max(controllers, key=lambda item: int(item.get("vram_mb") or 0))
-            name = str(primary.get("name") or "")
-            gpu.model = name
-            gpu.vram_mb = int(primary.get("vram_mb") or 0)
-            low = name.lower()
-            if "nvidia" in low or "rtx" in low or "geforce" in low:
-                gpu.vendor = "nvidia"; gpu.sdk.append("cuda")
-            elif "amd" in low or "radeon" in low:
-                gpu.vendor = "amd"; gpu.sdk.extend(["directml", "vulkan"])
-            elif "intel" in low or "arc" in low or "iris" in low:
-                gpu.vendor = "intel"; gpu.sdk.extend(["directml", "vulkan"])
-
-    # Linux (lspci)
-    if prof_os == "linux" and gpu.vendor == "unknown":
-        info = _run(["lspci"], timeout=3.0).lower()
-        if "nvidia" in info:
-            gpu.vendor = "nvidia"; gpu.sdk.append("cuda")
-        elif "amd/ati" in info or "advanced micro devices" in info:
-            gpu.vendor = "amd"; gpu.sdk.extend(["rocm", "vulkan"])
-        elif "intel corporation" in info and "vga" in info:
-            gpu.vendor = "intel"; gpu.sdk.append("vulkan")
-
-    return gpu
-
-
-def _detect_package_manager(prof_os: str) -> Optional[str]:
-    if prof_os == "windows":
-        return "winget" if _which("winget") else None
-    if prof_os == "darwin":
-        return "brew" if _which("brew") else None
-    if prof_os == "linux":
-        for pm in ("apt", "dnf", "pacman", "zypper", "apk"):
-            if _which(pm):
-                return pm
-    return None
-
-
-def _detect_tools() -> Dict[str, str]:
-    detected = detect_tools(_which, ("ollama", "lms", "nvidia-smi", "nvcc", "winget", "brew", "apt", "git", "node", "python", "python3"))
-    return {binary: path for binary, path in detected.items() if path}
-
-
-def _detect_wsl(prof_os: str) -> Tuple[bool, str]:
-    return detect_wsl_from_text(prof_os, _read_text("/proc/version"))
-
-
-def _detect_cuda() -> Tuple[bool, str]:
-    available, version, _nvidia_smi, _nvcc = detect_cuda(_which, lambda args: _run(args, timeout=4.0))
-    return available, version
-
-
-def _detect_cpu_details(prof_os: str) -> Tuple[str, int, int, List[str]]:
-    model = platform.processor() or ""
-    physical = os.cpu_count() or 0
-    logical = os.cpu_count() or 0
-    flags: List[str] = []
-    if prof_os == "darwin":
-        model = _run(["sysctl", "-n", "machdep.cpu.brand_string"]).strip() or model
-        try:
-            physical = int((_run(["sysctl", "-n", "hw.physicalcpu"]).strip() or physical))
-            logical = int((_run(["sysctl", "-n", "hw.logicalcpu"]).strip() or logical))
-        except ValueError:
-            pass
-        flags = [item.lower() for item in _run(["sysctl", "-n", "machdep.cpu.features"]).split()]
-    elif prof_os == "linux":
-        text = _read_text("/proc/cpuinfo")
-        for line in text.splitlines():
-            if line.lower().startswith("model name") and not model:
-                model = line.split(":", 1)[-1].strip()
-            if line.lower().startswith(("flags", "features")) and not flags:
-                flags = line.split(":", 1)[-1].strip().lower().split()
-    elif prof_os == "windows":
-        raw = _run(["wmic", "cpu", "get", "Name,NumberOfCores,NumberOfLogicalProcessors", "/format:list"])
-        for line in raw.splitlines():
-            key, _, value = line.partition("=")
-            if key == "Name" and value.strip():
-                model = value.strip()
-            elif key == "NumberOfCores" and value.strip():
-                try:
-                    physical = int(value.strip())
-                except ValueError:
-                    pass
-            elif key == "NumberOfLogicalProcessors" and value.strip():
-                try:
-                    logical = int(value.strip())
-                except ValueError:
-                    pass
-        try:
-            import ctypes
-            kernel32 = ctypes.windll.kernel32
-            feature_map = {6: "sse", 10: "sse2", 13: "sse3", 19: "neon", 28: "rdrand"}
-            flags.extend(name for code, name in feature_map.items() if kernel32.IsProcessorFeaturePresent(code))
-        except Exception:
-            pass
-    interesting = {"avx", "avx2", "avx512f", "fma", "neon", "sse4_2", "sse", "sse2", "sse3", "rdrand"}
-    return model, physical, logical, sorted({flag for flag in flags if flag in interesting})
-
-
-def _has_module(name: str) -> bool:
-    try:
-        __import__(name)
-        return True
-    except Exception:
-        return False
-
-
-def probe() -> SystemProfile:
-    """① PROBE — 외부 의존성 없이 가능한 만큼 환경을 감지한다."""
-    prof = SystemProfile()
-    prof.os = {"Darwin": "darwin", "Windows": "windows",
-               "Linux": "linux"}.get(platform.system(), platform.system().lower())
-    prof.os_version = platform.release()
-    prof.arch = platform.machine().lower()
-    cpu_model, cpu_cores, cpu_logical_cores, cpu_instructions = _detect_cpu_details(prof.os)
-    prof.cpu_model = cpu_model
-    prof.cpu_cores = cpu_cores
-    prof.cpu_logical_cores = cpu_logical_cores
-    prof.cpu_instructions = cpu_instructions
-    prof.python_version = platform.python_version()
-    prof.is_wsl, prof.wsl_version = _detect_wsl(prof.os)
-    prof.cuda_available, prof.cuda_version = _detect_cuda()
-    prof.tools = _detect_tools()
-
-    # RAM
-    try:
-        if prof.os == "linux":
-            for line in _read_text("/proc/meminfo").splitlines():
-                if line.startswith("MemTotal:"):
-                    prof.ram_mb = int(line.split()[1]) // 1024
-                    break
-        elif prof.os == "darwin":
-            out = _run(["sysctl", "-n", "hw.memsize"])
-            if out.strip():
-                try:
-                    prof.ram_mb = int(out.strip()) // (1024 * 1024)
-                except ValueError:
-                    prof.ram_mb = 0
-            if not prof.ram_mb:
-                profiler = _run(["system_profiler", "SPHardwareDataType"], timeout=8.0)
-                m = re.search(r"Memory:\s+([\d.]+)\s*(TB|GB|MB)", profiler, re.IGNORECASE)
-                if m:
-                    value = float(m.group(1))
-                    unit = m.group(2).lower()
-                    if unit == "tb":
-                        prof.ram_mb = int(value * 1024 * 1024)
-                    elif unit == "gb":
-                        prof.ram_mb = int(value * 1024)
-                    else:
-                        prof.ram_mb = int(value)
-                if not prof.ram_mb:
-                    hostinfo = _run(["hostinfo"])
-                    m = re.search(r"Primary memory available:\s+([\d.]+)\s+gigabytes", hostinfo, re.IGNORECASE)
-                    if m:
-                        prof.ram_mb = int(float(m.group(1)) * 1024)
-        elif prof.os == "windows":
-            out = _run(["wmic", "ComputerSystem", "get", "TotalPhysicalMemory",
-                        "/format:list"])
-            for line in out.splitlines():
-                if line.startswith("TotalPhysicalMemory="):
-                    prof.ram_mb = int(line.split("=", 1)[-1].strip()) // (1024 * 1024)
-                    break
-    except Exception:
-        pass
-
-    # Disk
-    try:
-        usage = shutil.disk_usage(Path.home())
-        prof.disk_free_mb = usage.free // (1024 * 1024)
-    except Exception:
-        pass
-
-    prof.gpu = _detect_gpu(prof.os, prof.arch)
-    prof.package_manager = _detect_package_manager(prof.os)
-    return prof
-
-
-# ── 2. RECOMMEND ────────────────────────────────────────────────────────────
-@dataclass
-class Recommendation:
-    runtime: str              # llama.cpp | mlx | vllm | mlc-llm | tflite
-    backend: str              # cuda | metal+mlx | directml | vulkan | rocm | cpu
-    model_id: str             # 추천 모델 (huggingface-like id)
-    quantization: str         # q4_K_M | q5_K_M | mxfp4 | f16
-    rationale: List[str]      # 왜 이걸 골랐는지 (UI에 표시)
-    estimated_tokens_per_sec: Optional[float] = None
-
-    def to_json(self) -> Dict[str, Any]:
-        return asdict(self)
-
-
-# 모델 카탈로그. PPT 슬라이드 16 의 "추천 모델" 열과 동기화.
-_MODEL_CATALOG: List[Dict[str, Any]] = [
-    # (min_ram_mb, min_vram_mb, model_id, quant, runtime_preference)
-    # OS 오버헤드(~4-6 GB) + KV 캐시 여유를 감안한 보수적 RAM 임계값
-    {"ram": 64 * 1024, "vram": 32 * 1024,
-     "id": "mlx-community/gemma-4-31b-it-4bit", "q": "4bit", "multimodal": True},
-    {"ram": 64 * 1024, "vram": 32 * 1024,
-     "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", "q": "q4_K_M", "multimodal": True},
-    {"ram": 48 * 1024, "vram": 24 * 1024,
-     "id": "mlx-community/gemma-4-31b-it-4bit", "q": "4bit", "multimodal": True},
-    {"ram": 32 * 1024, "vram": 16 * 1024,
-     "id": "mlx-community/gemma-4-26b-a4b-it-4bit", "q": "4bit", "multimodal": True},
-    {"ram": 48 * 1024, "vram": 24 * 1024,
-     "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", "q": "q4_K_M", "multimodal": True},
-    {"ram": 24 * 1024, "vram": 12 * 1024,
-     "id": "mlx-community/Llama-4-Scout-17B-16E-Instruct-4bit", "q": "4bit", "multimodal": True},
-    {"ram": 16 * 1024, "vram": 8 * 1024,
-     "id": "mlx-community/gemma-4-12b-it-4bit", "q": "4bit", "multimodal": True},
-    {"ram": 32 * 1024, "vram": 16 * 1024,
-     "id": "Qwen/Qwen3-VL-8B-Instruct", "q": "q5_K_M", "multimodal": True},
-    {"ram": 24 * 1024, "vram": 12 * 1024,
-     "id": "Qwen/Qwen3-VL-8B-Instruct", "q": "q4_K_M", "multimodal": True},
-    {"ram": 16 * 1024, "vram": 8 * 1024,
-     "id": "Qwen/Qwen3-VL-8B-Instruct", "q": "q4_K_M", "multimodal": True},
-    {"ram": 12 * 1024, "vram": 6 * 1024,
-     "id": "Qwen/Qwen3-VL-4B-Instruct", "q": "q4_K_M", "multimodal": True},
-    {"ram":  8 * 1024, "vram": 4 * 1024,
-     "id": "Qwen/Qwen3-VL-4B-Instruct", "q": "q4_K_M", "multimodal": True},
-    {"ram":  4 * 1024, "vram": 0,
-     "id": "Qwen/Qwen3-VL-4B-Instruct", "q": "q4_K_M", "multimodal": True},
-]
-
-
-def recommend(profile: SystemProfile) -> Recommendation:
-    """② RECOMMEND — 프로파일을 보고 런타임/모델/양자화를 결정한다."""
-    rationale: List[str] = []
-
-    # backend / runtime
-    if profile.os == "darwin" and profile.gpu.vendor == "apple":
-        backend = "metal+mlx"
-        runtime = "mlx" if _has_module("mlx_vlm") else "llama.cpp"
-        rationale.append("Apple Silicon → Metal + MLX-VLM")
-    elif profile.gpu.vendor == "nvidia" and profile.cuda_available and (profile.os == "linux" or profile.is_wsl):
-        backend = "cuda"
-        runtime = "vllm" if profile.gpu.vram_mb >= 12 * 1024 else "llama.cpp"
-        rationale.append(f"NVIDIA GPU {profile.gpu.vram_mb} MB VRAM + CUDA → {runtime}")
-    elif profile.gpu.vendor == "nvidia":
-        backend = "cuda" if profile.cuda_available else "vulkan"
-        runtime = "lmstudio" if profile.tools.get("lms") else ("ollama" if profile.tools.get("ollama") else "llama.cpp")
-        rationale.append("Windows NVIDIA는 LM Studio/Ollama 우선, vLLM은 WSL/Linux 권장")
-    elif profile.os == "windows" and profile.gpu.vendor in ("amd", "intel"):
-        backend = "directml/vulkan"
-        runtime = "lmstudio" if profile.tools.get("lms") else ("ollama" if profile.tools.get("ollama") else "llama.cpp")
-        rationale.append("Windows + AMD/Intel GPU → DirectML/Vulkan")
-    elif profile.os == "linux" and profile.gpu.vendor == "amd":
-        backend = "rocm" if "rocm" in profile.gpu.sdk else "vulkan"
-        runtime = "ollama" if profile.tools.get("ollama") else "llama.cpp"
-        rationale.append("Linux + AMD GPU → ROCm/Vulkan")
-    else:
-        backend = "cpu"
-        runtime = "ollama" if profile.tools.get("ollama") else "llama.cpp"
-        instruction_hint = ", ".join(profile.cpu_instructions) or "명령어 미감지"
-        rationale.append(f"GPU 가속이 없거나 미감지 → CPU 추론 ({profile.cpu_logical_cores or profile.cpu_cores} threads, {instruction_hint})")
-
-    # model size by RAM/VRAM
-    pick = _MODEL_CATALOG[-1]   # 가장 작은 모델 기본값
-    for entry in _MODEL_CATALOG:
-        if profile.ram_mb >= entry["ram"] and (
-            backend in {"cpu", "metal+mlx"} or profile.gpu.vram_mb >= entry["vram"]
-        ):
-            pick = entry
-            break
-    rationale.append(
-        f"RAM {profile.ram_mb} MB · VRAM {profile.gpu.vram_mb} MB → {pick['id']}"
-    )
-    if pick.get("multimodal"):
-        rationale.append("최신 멀티모달 모델을 우선 선택")
-
-    # 양자화: VRAM 충분 → 더 정밀한 양자화로 업그레이드
-    quant = pick["q"]
-    if profile.gpu.vram_mb >= 24 * 1024:
-        quant = "f16"
-        rationale.append("VRAM ≥ 24 GB → f16 풀 정밀도")
-
-    # 거친 tokens/sec 예측 (very rough)
-    est_tps = None
-    if backend == "cuda":
-        est_tps = max(8.0, profile.gpu.vram_mb / 800)
-    elif backend == "metal+mlx":
-        est_tps = max(6.0, (profile.ram_mb // 1024) * 0.7)
-    elif backend == "cpu":
-        est_tps = max(1.5, profile.cpu_cores * 0.6)
-
-    return Recommendation(
-        runtime=runtime, backend=backend,
-        model_id=pick["id"], quantization=quant,
-        rationale=rationale, estimated_tokens_per_sec=est_tps,
-    )
-
-
-# ── 3. INSTALL plan ─────────────────────────────────────────────────────────
-@dataclass
-class InstallStep:
-    name: str
-    why: str
-    command: List[str]
-    requires_admin: bool = False
-
-    def to_json(self) -> Dict[str, Any]:
-        return {
-            "name": self.name,
-            "why": self.why,
-            "command": self.command,
-            "requires_admin": self.requires_admin,
-            "command_plan": command_plan(
-                self.command,
-                name=self.name,
-                purpose="auto_setup_install",
-                requires_admin=self.requires_admin,
-                metadata={"why": self.why},
-            ),
-        }
-
-
-@dataclass
-class InstallPlan:
-    package_manager: Optional[str]
-    steps: List[InstallStep]
-    notes: List[str] = field(default_factory=list)
-
-    def to_json(self) -> Dict[str, Any]:
-        plan_summary = command_plan_for_commands(
-            [step.command for step in self.steps],
-            name="auto_setup_plan",
-            purpose="auto_setup_install",
-            metadata={"package_manager": self.package_manager},
-        )
-        return {
-            "package_manager": self.package_manager,
-            "steps": [s.to_json() for s in self.steps],
-            "notes": self.notes,
-            "command_plan": plan_summary,
-            "confirmation_token": plan_summary["confirmation_token"],
-        }
-
-
-# 패키지 카탈로그: 핵심 의존성을 OS별 명령으로 매핑
-_PKG_MAP: Dict[str, Dict[str, Tuple[str, ...]]] = {
-    # name : { pm : (cmd parts) }
-    "python3.11+": {
-        "winget": ("winget", "install", "-e", "--id", "Python.Python.3.11"),
-        "brew":   ("brew", "install", "python@3.11"),
-        "apt":    ("apt-get", "install", "-y", "python3.11"),
-        "dnf":    ("dnf", "install", "-y", "python3.11"),
-    },
-    "node20":      {
-        "winget": ("winget", "install", "-e", "--id", "OpenJS.NodeJS.LTS"),
-        "brew":   ("brew", "install", "node@20"),
-        "apt":    ("apt-get", "install", "-y", "nodejs"),
-        "dnf":    ("dnf", "install", "-y", "nodejs"),
-    },
-    "ollama":      {
-        "brew":   ("brew", "install", "ollama"),
-        "winget": ("winget", "install", "-e", "--id", "Ollama.Ollama"),
-        "apt":    ("sh", "-c", "curl -fsSL https://ollama.com/install.sh | sh"),
-    },
-    "huggingface-cli": {
-        "brew":   ("pip3", "install", "--upgrade", "huggingface_hub"),
-        "winget": ("pip", "install", "--upgrade", "huggingface_hub"),
-        "apt":    ("pip3", "install", "--upgrade", "huggingface_hub"),
-        "dnf":    ("pip3", "install", "--upgrade", "huggingface_hub"),
-    },
-}
-
-
-def plan(profile: SystemProfile, rec: Recommendation) -> InstallPlan:
-    """③ INSTALL — 추천을 만족시키는 *명령 계획* 을 만든다.  실행하지 않는다."""
-    pm = profile.package_manager
-    steps: List[InstallStep] = []
-    notes: List[str] = []
-
-    def need(name: str, why: str) -> None:
-        cmd_tuple = _PKG_MAP.get(name, {}).get(pm or "")
-        if cmd_tuple:
-            steps.append(InstallStep(
-                name=name, why=why,
-                command=list(cmd_tuple),
-                requires_admin=(cmd_tuple[0] in ("apt-get", "dnf", "pacman")),
-            ))
-        else:
-            notes.append(f"패키지 매니저 어댑터 없음: {name} ({pm}) — 수동 설치 필요")
-
-    if sys.version_info < (3, 11):
-        need("python3.11+", "Lattice AI 서버는 Python 3.11 이상이 필요합니다.")
-    if not _which("node"):
-        need("node20", "VSCode 확장 / npm CLI 부트스트랩에 필요")
-
-    # 런타임별 추가
-    if rec.runtime == "mlx" and not _has_module("mlx_vlm"):
-        steps.append(InstallStep(
-            name="mlx-vlm", why="Apple Silicon 멀티모달 추론",
-            command=["pip3", "install", "--upgrade", "mlx-vlm"],
-        ))
-    if rec.runtime in {"llama.cpp", "ollama"} and not _which("ollama"):
-        need("ollama", "llama.cpp 가중치를 가장 쉽게 받는 경로")
-    if rec.runtime == "lmstudio" and not _which("lms"):
-        notes.append("LM Studio CLI(lms)를 찾지 못했습니다. https://lmstudio.ai/download 에서 설치하면 Windows/macOS/Linux 모델 다운로드와 GPU 백엔드를 자동 감지합니다.")
-    if rec.runtime == "vllm" and not _has_module("vllm"):
-        steps.append(InstallStep(
-            name="vllm", why="NVIDIA CUDA/WSL/Linux 서버형 추론",
-            command=["pip3", "install", "--upgrade", "vllm", "huggingface_hub"],
-        ))
-    if profile.gpu.vendor == "nvidia" and not profile.cuda_available:
-        notes.append("NVIDIA GPU는 감지됐지만 CUDA/nvidia-smi를 찾지 못했습니다. Windows에서는 NVIDIA 드라이버와 CUDA Toolkit 설치 후 재검사를 권장합니다.")
-    if profile.os == "windows" and profile.gpu.vendor == "nvidia" and not profile.is_wsl:
-        notes.append("vLLM은 Windows native보다 WSL2/Linux에서 안정적입니다. Windows 데스크톱은 LM Studio 또는 Ollama GPU 경로를 먼저 권장합니다.")
-
-    if not _which("huggingface-cli"):
-        need("huggingface-cli", "추천 모델 가중치 다운로드용")
-
-    # 모델 가중치 풀
-    model_command = ["huggingface-cli", "download", rec.model_id, "--quiet"]
-    if rec.runtime == "ollama":
-        lower = rec.model_id.lower()
-        if "gemma-4-31b" in lower:
-            model_command = ["ollama", "pull", "hf.co/ggml-org/gemma-4-31B-it-GGUF:Q4_K_M"]
-        elif "gemma-4-12b" in lower:
-            model_command = ["ollama", "pull", "hf.co/ggml-org/gemma-4-12B-it-GGUF:Q4_K_M"]
-        elif "llama-4-scout" in lower:
-            model_command = ["ollama", "pull", "hf.co/ggml-org/Llama-4-Scout-17B-16E-Instruct-GGUF:Q4_K_M"]
-        elif "qwen3-vl-8b" in lower:
-            model_command = ["ollama", "pull", "qwen3-vl:8b"]
-        elif "qwen3-vl-4b" in lower:
-            model_command = ["ollama", "pull", "qwen3-vl:4b"]
-    elif rec.runtime == "lmstudio":
-        model_command = ["lms", "get", rec.model_id]
-    steps.append(InstallStep(
-        name=f"weights:{rec.model_id}",
-        why="추론에 사용할 모델 가중치",
-        command=model_command,
-    ))
-
-    return InstallPlan(package_manager=pm, steps=steps, notes=notes)
-
-
-def apply_plan(
-    plan_obj: InstallPlan,
-    *,
-    confirm: bool = False,
-    confirmation_token: Optional[str] = None,
-) -> List[Dict[str, Any]]:
-    """위험: 실제로 설치 명령을 실행한다. ``confirm`` and token are required."""
-    if not confirm:
-        raise RuntimeError("refuse to apply: pass confirm=True")
-    expected = plan_obj.to_json().get("confirmation_token")
-    if str(confirmation_token or "").strip() != expected:
-        raise CommandConfirmationError("refuse to apply: confirmation token mismatch")
-    results: List[Dict[str, Any]] = []
-    for step in plan_obj.steps:
-        step_plan = command_plan(
-            step.command,
-            name=step.name,
-            purpose="auto_setup_install",
-            requires_admin=step.requires_admin,
-            metadata={"why": step.why},
-        )
-        try:
-            require_command_confirmation(
-                step.command,
-                step_plan["confirmation_token"],
-                purpose="auto_setup_install",
-            )
-            append_process_audit_event("installer_command", plan=step_plan, status="started")
-            r = subprocess.run(step.command, capture_output=True, text=True,
-                               timeout=300, check=False)
-            append_process_audit_event(
-                "installer_command",
-                plan=step_plan,
-                status="finished",
-                returncode=r.returncode,
-                stdout=r.stdout,
-                stderr=r.stderr,
-            )
-            results.append({
-                "name": step.name,
-                "command_hash": step_plan["command_hash"],
-                "command_preview": step_plan["command_preview"],
-                "returncode": r.returncode,
-                "stdout_tail": (r.stdout or "")[-2000:],
-                "stderr_tail": (r.stderr or "")[-2000:],
-            })
-        except Exception as exc:
-            append_process_audit_event("installer_command", plan=step_plan, status="error", error=str(exc))
-            results.append({"name": step.name, "command_hash": step_plan["command_hash"], "error": str(exc)})
-    return results
-
-
-# ── 4. VERIFY ───────────────────────────────────────────────────────────────
-def verify(profile: SystemProfile, rec: Recommendation) -> Dict[str, Any]:
-    """④ VERIFY — 가벼운 sanity check.  실제 LLM 추론 벤치는 별도 도구로.
-    여기서는 ‘설치된 것들이 import 되는가’ + ‘디스크/RAM 여유’ 정도만 본다."""
-    checks: List[Dict[str, Any]] = []
-
-    def add(label: str, ok: bool, detail: str = "") -> None:
-        checks.append({"label": label, "ok": ok, "detail": detail})
-
-    add("Python 3.11+", sys.version_info >= (3, 11), platform.python_version())
-    add("RAM ≥ 4 GB", profile.ram_mb >= 4 * 1024, f"{profile.ram_mb} MB")
-    add("디스크 여유 ≥ 8 GB", profile.disk_free_mb >= 8 * 1024,
-        f"{profile.disk_free_mb} MB free")
-
-    if rec.runtime == "mlx":
-        add("mlx_vlm import", _has_module("mlx_vlm"), "Apple Silicon 멀티모달 런타임")
-    if rec.runtime in {"llama.cpp", "ollama"}:
-        add("ollama binary", _which("ollama") is not None,
-            _which("ollama") or "not found")
-    if rec.runtime == "lmstudio":
-        add("LM Studio CLI", _which("lms") is not None, _which("lms") or "not found")
-    if rec.backend == "cuda":
-        add("CUDA/nvidia-smi", profile.cuda_available, profile.cuda_version or "not found")
-
-    # CPU/메모리 잠깐 측정
-    t0 = time.perf_counter()
-    _ = sum(i * i for i in range(200_000))
-    cpu_ms = (time.perf_counter() - t0) * 1000
-    add("CPU latency sample", cpu_ms < 200, f"{cpu_ms:.1f} ms / 200k ops")
-
-    return {
-        "checks": checks,
-        "all_pass": all(c["ok"] for c in checks),
-    }
-
-
-# ── 5. PRESET ───────────────────────────────────────────────────────────────
-def preset(profile: SystemProfile, rec: Recommendation) -> Dict[str, Any]:
-    """⑤ PRESET — UX 분기 + 단축키 + 테마 + MCP 도구 기본값.
-
-    PPT 슬라이드 3 (모드 선택) · 17 (PRESET) 명세를 따른다.
-    """
-    # 기본 모드 vs 고급 모드 자동 선택 휴리스틱
-    advanced = (
-        profile.gpu.vendor in ("nvidia", "apple") or
-        profile.ram_mb >= 24 * 1024 or
-        "code" in (profile.cpu_model or "").lower()   # 개발자 머신 추정
-    )
-    mode = "advanced" if advanced else "basic"
-
-    # 단축키는 OS별 자연 컨벤션 따름
-    mod = "Cmd" if profile.os == "darwin" else "Ctrl"
-    shortcuts = {
-        "newChat":        f"{mod}+N",
-        "toggleSidebar":  f"{mod}+B",
-        "openGraph":      f"{mod}+G",
-        "search":         f"{mod}+K",
-        "toggleMode":     f"{mod}+Shift+M",
-        "submit":         "Enter",
-        "newline":        "Shift+Enter",
-    }
-
-    # 기본 MCP 도구 (PPT 슬라이드 11 의 기본 5종)
-    mcp_defaults = [
-        {"id": "filesystem",      "scope": "local",  "enabled": True},
-        {"id": "web-search",      "scope": "remote", "enabled": True},
-        {"id": "code-execute",    "scope": "local",  "enabled": True},
-        {"id": "browser-automation","scope": "remote","enabled": False},
-        {"id": "database",        "scope": "remote", "enabled": False},
-    ]
-
-    # 테마: OS 다크 모드 추종이 기본
-    theme = {"mode": "auto",                    # auto | light | dark
-             "accent": "#6E4AE6",               # PPT 슬라이드 19 토큰
-             "density": "comfortable" if mode == "basic" else "compact"}
-
-    # 다국어: OS locale 기반 추정
-    locale = os.environ.get("LANG", os.environ.get("LC_ALL", "ko_KR"))
-    lang = "ko" if locale.lower().startswith("ko") else (
-        "ja" if locale.lower().startswith("ja") else "en"
-    )
-
-    return {
-        "mode": mode,
-        "model": {"id": rec.model_id, "runtime": rec.runtime,
-                  "backend": rec.backend, "quantization": rec.quantization},
-        "shortcuts": shortcuts,
-        "mcp": mcp_defaults,
-        "theme": theme,
-        "language": lang,
-        "tips": (
-            ["기본 모드는 카드형 액션과 큰 입력창 위주. 언제든 '고급 모드' 로 전환할 수 있어요."]
-            if mode == "basic"
-            else ["고급 모드: 사이드바·디테일 패널·파이프라인 도구가 모두 활성화됩니다."]
-        ),
-    }
-
-
-# ── orchestrator ────────────────────────────────────────────────────────────
-def run_all(*, apply_install: bool = False, confirmation_token: Optional[str] = None) -> Dict[str, Any]:
-    p = probe()
-    r = recommend(p)
-    pl = plan(p, r)
-    install_results = None
-    if apply_install:
-        install_results = apply_plan(pl, confirm=True, confirmation_token=confirmation_token)
-    v = verify(p, r)
-    ps = preset(p, r)
-    return {
-        "probe":   p.to_json(),
-        "recommend": r.to_json(),
-        "plan":    pl.to_json(),
-        "install": install_results,
-        "verify":  v,
-        "preset":  ps,
-    }
-
-
-# ── CLI ────────────────────────────────────────────────────────────────────
-def _main() -> int:
-    parser = argparse.ArgumentParser(prog="auto_setup",
-                                     description="Lattice AI zero-config setup")
-    sub = parser.add_subparsers(dest="cmd", required=True)
-    sub.add_parser("probe")
-    sub.add_parser("recommend")
-    sp_plan = sub.add_parser("plan")
-    sp_plan.add_argument("--apply", action="store_true",
-                         help="actually run the install commands (DANGER)")
-    sp_plan.add_argument("--confirm-token", default=None,
-                         help="confirmation token from the dry-run plan output")
-    sub.add_parser("verify")
-    sub.add_parser("preset")
-    sub.add_parser("all")
-    args = parser.parse_args()
-
-    if args.cmd == "probe":
-        print(json.dumps(probe().to_json(), indent=2, ensure_ascii=False)); return 0
-    if args.cmd == "recommend":
-        p = probe(); r = recommend(p)
-        print(json.dumps({"probe": p.to_json(), "recommend": r.to_json()},
-                         indent=2, ensure_ascii=False))
-        return 0
-    if args.cmd == "plan":
-        p = probe(); r = recommend(p); pl = plan(p, r)
-        out: Dict[str, Any] = {"plan": pl.to_json()}
-        if args.apply:
-            out["install"] = apply_plan(pl, confirm=True, confirmation_token=args.confirm_token)
-        print(json.dumps(out, indent=2, ensure_ascii=False)); return 0
-    if args.cmd == "verify":
-        p = probe(); r = recommend(p)
-        print(json.dumps(verify(p, r), indent=2, ensure_ascii=False)); return 0
-    if args.cmd == "preset":
-        p = probe(); r = recommend(p)
-        print(json.dumps(preset(p, r), indent=2, ensure_ascii=False)); return 0
-    if args.cmd == "all":
-        print(json.dumps(run_all(apply_install=False), indent=2, ensure_ascii=False))
-        return 0
-    return 2
+from latticeai.setup import auto_setup as _impl
 
 
 if __name__ == "__main__":
-    raise SystemExit(_main())
+    raise SystemExit(_impl._main())
+else:
+    import sys
+
+    sys.modules[__name__] = _impl