ewoooc/services/code_review_pipeline_service.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
services/code_review_pipeline_service.py
Post-Deploy AI Agent Code Review Pipeline

觸發時機：CD 健康檢查通過後，由 Gitea Action webhook 呼叫
Pipeline：
  Step 1  system          讀取變更檔案內容
  Step 2  Hermes          程式碼掃描（bugs / security / performance）
  Step 3  OpenClaw        架構品質評估（Gemini 2.5 Flash）
  Step 4  ElephantAlpha   決策協調（severity 判定 + auto-fix 裁量）
  Step 5  NemoTron        行動派遣（action_plans 寫入 + AiderHeal 觸發）

結果輸出：
  - ai_insights（type='code_review_result'）
  - action_plans（type='code_review_fix'）
  - Telegram 告警（啟動 / 完成 / 錯誤）
  - 前端 /code-review/ 即時 polling 狀態
"""

import json
import logging
import os
import re
import threading
from datetime import datetime
from typing import Any, Dict, List, Optional

from database.manager import get_session
from sqlalchemy import text
# ADR-027 Phase 2 N3：HERMES_MODEL 仍 import（純常數），HERMES_URL 改 lazy
# 每次 _hermes_scan 才透過 get_hermes_url() 取最新解析（GCP 優先 / 111 備援），
# 避免 import-time freeze 導致主機切換不生效。
from services.hermes_analyst_service import HERMES_MODEL as _HERMES_MODEL
from config import get_hermes_url
from services.ai_call_logger import log_ai_call  # Operation Ollama-First v5.0 P1

logger = logging.getLogger(__name__)

# ── Pipeline 全域狀態（供前端 polling）─────────────────────────────────────
_current_pipeline: Dict[str, Any] = {}
_pipeline_lock = threading.Lock()

GEMINI_API_KEY = os.getenv("GEMINI_API_KEY", "")
# LOCKED-GEMINI: Code Review 全 repo diff 可達 100K+ tokens，超過 Ollama 32K context
# Phase 7 升級：CODE_REVIEW_USE_CLAUDE=true 時改走 Claude Opus 4.7（200K context, Arena code Elo 1548）
# 預設 OFF，行為與 Phase 6 完全相同；ADR-028 鎖定場景 #5
REVIEW_MODEL   = os.getenv("OPENCLAW_MODEL", "gemini-2.5-flash")
INTERNAL_TOKEN = os.getenv("INTERNAL_WEBHOOK_TOKEN", "")
AUTO_FIX_ENABLED = os.getenv("CODE_REVIEW_AUTO_FIX_ENABLED", "true").lower() == "true"
ALLOW_INSECURE_WEBHOOK = os.getenv("MOMO_ALLOW_INSECURE_INTERNAL_WEBHOOK_FOR_DEV", "").lower() == "true"

# Phase 7 Frontier 升級 feature flag — 預設 OFF，啟用後 _openclaw_assess 改走 Claude Opus 4.7
CODE_REVIEW_USE_CLAUDE = os.getenv("CODE_REVIEW_USE_CLAUDE", "false").lower() == "true"
CLAUDE_REVIEW_MODEL = os.getenv("CLAUDE_MODEL", "claude-opus-4-7")


# ═══════════════════════════════════════════════════════════════════════════════
# Pipeline Class
# ═══════════════════════════════════════════════════════════════════════════════

class CodeReviewPipeline:
    """
    5-Step post-deploy code review pipeline.
    Call pipeline.run() inside a daemon thread.
    """

    def __init__(self, commit_sha: str, changed_files: List[str],
                 branch: str = "main", deploy_type: str = "sync"):
        self.commit_sha   = commit_sha
        self.branch       = branch
        self.deploy_type  = deploy_type
        self.started_at   = datetime.now()
        self.pipeline_id  = f"cr_{commit_sha[:8]}_{self.started_at.strftime('%Y%m%d_%H%M%S')}"

        # 只 review Python + YAML 檔（跳過靜態資源）
        self.changed_files = [
            f for f in changed_files
            if f.endswith(('.py', '.yaml', '.yml', '.json'))
            and not f.startswith(('node_modules/', '.git/'))
        ]

        self.state: Dict[str, Any] = {
            "pipeline_id":       self.pipeline_id,
            "commit_sha":        commit_sha,
            "branch":            branch,
            "changed_files":     self.changed_files,
            "status":            "running",
            "current_step":      0,
            "total_steps":       5,
            "steps":             [],
            "findings":          [],
            "severity_summary":  {"critical": 0, "high": 0, "medium": 0, "low": 0},
            "openclaw_report":   "",
            "ea_decision":       {},
            "auto_fix_triggered": False,
            "started_at":        self.started_at.isoformat(),
            "completed_at":      None,
            "message":           "",
        }
        self._sync_global()

    # ── State helpers ─────────────────────────────────────────────────────────

    def _sync_global(self):
        global _current_pipeline
        with _pipeline_lock:
            _current_pipeline = dict(self.state)

    def _step_start(self, num: int, name: str, agent: str):
        self.state["current_step"] = num
        self.state["steps"].append({
            "step":         num,
            "name":         name,
            "agent":        agent,
            "status":       "running",
            "started_at":   datetime.now().isoformat(),
            "completed_at": None,
            "summary":      "",
        })
        self._sync_global()
        logger.info("[CodeReview] ▶ Step %d/5 %s (%s)", num, name, agent)

    def _step_done(self, num: int, summary: str, ok: bool = True):
        for s in self.state["steps"]:
            if s["step"] == num:
                s["status"]       = "ok" if ok else "error"
                s["completed_at"] = datetime.now().isoformat()
                s["summary"]      = summary[:300]
        self._sync_global()
        logger.info("[CodeReview] %s Step %d — %s", "✓" if ok else "✗", num, summary[:100])

    def _finish(self, status: str, message: str):
        self.state["status"]       = status
        self.state["completed_at"] = datetime.now().isoformat()
        self.state["message"]      = message
        self._sync_global()

    # ── Main pipeline ─────────────────────────────────────────────────────────

    def run(self):
        """Execute full pipeline. Designed to run in daemon thread."""
        try:
            self._notify_start()

            # Step 1 ─ read files
            self._step_start(1, "讀取變更檔案", "system")
            file_contents = self._read_changed_files()
            if not file_contents:
                self._step_done(1, "無有效 Python/YAML 變更，跳過 Review")
                self._finish("skipped", "無有效變更檔案")
                return
            self._step_done(1, f"讀取 {len(file_contents)} 個檔案")

            # Step 2 ─ Hermes scan
            self._step_start(2, "Hermes 程式碼掃描", "Hermes")
            findings = self._hermes_scan(file_contents)
            cnt = self.state["severity_summary"]
            self._step_done(2, f"CRITICAL={cnt['critical']} HIGH={cnt['high']} MEDIUM={cnt['medium']} LOW={cnt['low']}")

            # Step 3 ─ OpenClaw assessment
            self._step_start(3, "OpenClaw 架構品質評估", "OpenClaw")
            openclaw_report = self._openclaw_assess(file_contents, findings)
            self.state["openclaw_report"] = openclaw_report
            self._step_done(3, openclaw_report[:120] if openclaw_report else "（Gemini 未回應）")

            # Step 4 ─ ElephantAlpha orchestration
            self._step_start(4, "ElephantAlpha 決策協調", "ElephantAlpha")
            ea = self._ea_orchestrate(findings, openclaw_report)
            self.state["ea_decision"] = ea
            self._step_done(4, f"優先度={ea.get('priority','?')} auto_fix={ea.get('auto_fix',False)}")

            # Step 5 ─ NemoTron dispatch
            self._step_start(5, "NemoTron 行動派遣", "NemoTron")
            dispatch = self._nemotron_dispatch(ea, findings)
            self._step_done(5, f"寫入 {dispatch['actions']} 筆 action_plans，自動修復={'是' if dispatch['auto_fix'] else '否'}")

            # Persist + notify
            self._save_to_db(findings, openclaw_report, ea)
            self._notify_complete(findings, openclaw_report, ea)
            self._finish("completed", "Pipeline 執行完成")

        except Exception as e:
            logger.error("[CodeReview] Pipeline 例外: %s", e, exc_info=True)
            self._finish("error", str(e)[:200])
            self._notify_error(str(e))

    # ── Step 1：讀取檔案 ───────────────────────────────────────────────────────

    def _read_changed_files(self) -> Dict[str, str]:
        project_root = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
        contents: Dict[str, str] = {}
        for rel_path in self.changed_files[:8]:          # 最多 8 個檔案
            abs_path = os.path.join(project_root, rel_path)
            try:
                with open(abs_path, encoding="utf-8", errors="ignore") as fh:
                    raw = fh.read()
                    contents[rel_path] = raw[:6000] + ("\n... (截斷)" if len(raw) > 6000 else "")
            except OSError:
                logger.debug("[CodeReview] 無法讀取 %s（部署路徑不同？）", rel_path)
        return contents

    # ── Step 2：Hermes 掃描 ───────────────────────────────────────────────────

    def _hermes_scan(self, files: Dict[str, str]) -> List[Dict]:
        """走 resolve_ollama_host() 三主機級聯：GCP-A → GCP-B → 111（ADR-027 Phase 2）"""
        try:
            import requests as _req

            files_text = "\n\n".join(
                f"### {name}\n```python\n{content}\n```"
                for name, content in list(files.items())[:4]
            )

            prompt = f"""你是資深程式碼審查工程師，請掃描以下程式碼並列出所有問題。

{files_text}

輸出格式：純 JSON 陣列，每項包含以下欄位：
- severity: "CRITICAL" | "HIGH" | "MEDIUM" | "LOW"
- type: "bug" | "security" | "performance" | "maintainability"
- file: 檔案名稱
- line_hint: 約略行號或函式名稱
- description: 問題說明（繁體中文，精簡一句）
- suggestion: 修復建議（繁體中文，精簡一句）

只輸出 JSON 陣列，不含其他文字。無問題時輸出 []"""

            # ADR-027 Phase 2 N3：lazy resolve Hermes 主機（GCP 優先 / 111 備援），
            # 避開 import-time freeze。provider 標籤跟著解析結果動態決定。
            hermes_url = get_hermes_url()
            provider_tag = (
                'gcp_ollama' if ('34.21.145.224' in hermes_url or '34.143.170.20' in hermes_url)
                else 'ollama_111' if '192.168.0.111' in hermes_url
                else 'ollama_other'
            )
            # Phase 1 v5.0: 包 ai_call_logger 追蹤 Code Review Hermes scan
            with log_ai_call(
                caller='code_review_hermes',
                provider=provider_tag,
                model=_HERMES_MODEL,
                request_id=f"cr-{self.commit_sha[:8]}",
                meta={'commit': self.commit_sha[:8], 'branch': self.branch,
                      'files': len(files), 'host': hermes_url},
            ) as _ctx:
                resp = _req.post(
                    f"{hermes_url}/api/generate",
                    json={"model": _HERMES_MODEL, "prompt": prompt,
                          "stream": False, "options": {"temperature": 0.1}},
                    timeout=120,
                )
                resp.raise_for_status()
                body = resp.json()
                _ctx.set_tokens(
                    input=body.get("prompt_eval_count", 0),
                    output=body.get("eval_count", 0),
                )
                raw = body.get("response", "").strip()

            match = re.search(r"\[.*\]", raw, re.DOTALL)
            if not match:
                logger.warning("[CodeReview] Hermes 回應無 JSON: %s", raw[:200])
                return []
            findings = json.loads(match.group())

            for f in findings:
                sev = f.get("severity", "LOW").lower()
                if sev in self.state["severity_summary"]:
                    self.state["severity_summary"][sev] += 1
            self.state["findings"] = findings
            self._sync_global()
            return findings

        except Exception as e:
            logger.warning("[CodeReview] Hermes 掃描失敗: %s", e)
            return []

    # ── Step 3：OpenClaw 評估 ──────────────────────────────────────────────────

    def _openclaw_assess(self, files: Dict[str, str], findings: List[Dict]) -> str:
        """
        路由優先序：
          L1 (Phase 7, flag CODE_REVIEW_USE_CLAUDE=true) → Claude Opus 4.7 (Arena code #1)
          L2 (預設)                                       → Gemini（GEMINI_API_KEY）
          L3 (降級)                                       → ElephantAlpha via OpenRouter
        feature flag 預設 OFF，行為與 Phase 6 完全相同。
        """
        sev = self.state["severity_summary"]
        findings_json = json.dumps(findings[:8], ensure_ascii=False, indent=2)
        files_list = "\n".join(f"- {k} ({len(v)} 字元)" for k, v in list(files.items())[:5])

        system = (
            "你是 OpenClaw 程式碼品質戰略分析師，以技術主管視角評估部署後程式碼。"
            "語言：繁體中文。風格：精準、數據導向、可執行建議。"
        )
        user_prompt = f"""【部署】Commit {self.commit_sha[:8]} @ {self.branch}
【變更檔案】
{files_list}

【Hermes 掃描摘要】CRITICAL={sev['critical']} HIGH={sev['high']} MEDIUM={sev['medium']} LOW={sev['low']}
【問題明細】
{findings_json}

請產出程式碼品質評估（使用 HTML <b> 標題，150字以內）：
<b>🔍 整體風險等級</b>（一句理由）
<b>⚠️ 最需關注問題</b>（TOP 2）
<b>💡 架構優化方向</b>（1條長期建議）
<b>✅ 本次部署亮點</b>"""

        # ── L1：Phase 7 Frontier — Claude Opus 4.7（程式碼能力 #1）────────────
        # feature flag 預設 OFF；ON 時優先走，失敗 fallback 到 L2 Gemini
        if CODE_REVIEW_USE_CLAUDE:
            try:
                from services.anthropic_service import anthropic_service
            except Exception as e:
                logger.warning("[CodeReview] Claude service import 失敗，退回 Gemini: %s", e)
                anthropic_service = None  # type: ignore

            if anthropic_service is not None and anthropic_service.is_available():
                with log_ai_call(
                    caller='code_review_openclaw',
                    provider='claude',
                    model=CLAUDE_REVIEW_MODEL,
                    request_id=f"cr-{self.commit_sha[:8]}",
                    meta={
                        'commit': self.commit_sha[:8],
                        'branch': self.branch,
                        'flag': 'CODE_REVIEW_USE_CLAUDE',
                    },
                ) as _ctx:
                    resp = anthropic_service.generate(
                        prompt=user_prompt,
                        system_prompt=system,   # ephemeral cache（5 分鐘 TTL，省 ~90% 成本）
                        model=CLAUDE_REVIEW_MODEL,
                        max_tokens=2048,
                        temperature=0.2,        # code review 要精確
                        cache_system=True,
                        timeout=120,
                    )
                    if resp.success:
                        _ctx.set_tokens(input=resp.input_tokens, output=resp.output_tokens)
                        _ctx.set_cache_hit(resp.cache_hit)
                        _ctx.add_meta('cache_creation_tokens', resp.cache_creation_tokens)
                        _ctx.add_meta('cache_read_tokens', resp.cache_read_tokens)
                        return resp.content or ""
                    # Claude 失敗 → fallback 到 Gemini（L2）
                    _ctx.set_error(resp.error or 'claude generate failed')
                    _ctx.fallback_to_caller('code_review_openclaw_gemini')
                    logger.warning(
                        "[CodeReview] Claude 失敗，降級 Gemini: %s", resp.error,
                    )
            else:
                logger.info(
                    "[CodeReview] CODE_REVIEW_USE_CLAUDE=true 但 Claude 不可用（缺 API key 或 SDK），退回 Gemini",
                )

        # ── L2：Gemini — Phase 1 v5.0 logger 追蹤 ────────────────────────────
        if GEMINI_API_KEY:
            with log_ai_call(
                caller='code_review_openclaw',
                provider='gemini',
                model=REVIEW_MODEL,
                request_id=f"cr-{self.commit_sha[:8]}",
                meta={'commit': self.commit_sha[:8], 'branch': self.branch},
            ) as _ctx:
                try:
                    import google.generativeai as genai
                    genai.configure(api_key=GEMINI_API_KEY)
                    model = genai.GenerativeModel(
                        model_name=REVIEW_MODEL,
                        generation_config=genai.types.GenerationConfig(
                            temperature=0.3, max_output_tokens=1500,
                        ),
                        system_instruction=system,
                    )
                    resp = model.generate_content(user_prompt, request_options={"timeout": 90})
                    try:
                        usage = getattr(resp, 'usage_metadata', None)
                        if usage is not None:
                            _ctx.set_tokens(
                                input=getattr(usage, 'prompt_token_count', 0) or 0,
                                output=getattr(usage, 'candidates_token_count', 0) or 0,
                            )
                    except Exception:
                        pass
                    return resp.text or ""
                except Exception as e:
                    logger.warning("[CodeReview] OpenClaw Gemini 失敗，降級 ElephantAlpha: %s", e)
                    _ctx.set_error(f"{type(e).__name__}: {e}")
                    _ctx.fallback_to_caller('code_review_elephant')

        # 降級：ElephantAlpha via OpenRouter（OPENROUTER_API_KEY 容器內一定有）
        # Phase 1 v5.0 logger 追蹤
        with log_ai_call(
            caller='code_review_elephant',
            provider='nim_via_elephant',
            model='nvidia/llama-3.3-nemotron-super-49b-v1.5',
            request_id=f"cr-{self.commit_sha[:8]}",
            meta={'commit': self.commit_sha[:8], 'branch': self.branch},
        ) as _ctx:
            try:
                from services.elephant_service import elephant_service
                resp = elephant_service.generate(
                    prompt=user_prompt,
                    system_prompt=system,
                    temperature=0.3,
                    timeout=90,
                )
                if resp.success:
                    # ElephantResponse 已含 input_tokens/output_tokens
                    _ctx.set_tokens(
                        input=getattr(resp, 'input_tokens', 0) or 0,
                        output=getattr(resp, 'output_tokens', 0) or 0,
                    )
                    return resp.content or ""
                else:
                    _ctx.set_error(getattr(resp, 'error', 'elephant generate failed'))
            except Exception as e:
                logger.warning("[CodeReview] OpenClaw ElephantAlpha 降級也失敗: %s", e)
                _ctx.set_error(f"{type(e).__name__}: {e}")

        return ""

    # ── Step 4：ElephantAlpha 決策 ─────────────────────────────────────────────

    def _ea_orchestrate(self, findings: List[Dict], openclaw_report: str) -> Dict:
        sev = self.state["severity_summary"]
        critical_n = sev["critical"]
        high_n     = sev["high"]

        # 嘗試呼叫 ElephantAlpha 做精細判斷
        try:
            from services.elephant_service import elephant_service

            top3 = json.dumps(
                [f for f in findings if f.get("severity") in ("CRITICAL", "HIGH")][:3],
                ensure_ascii=False,
            )
            prompt = f"""你是 Elephant Alpha，負責協調 Code Review 後的修復決策。

【部署】commit={self.commit_sha[:8]} branch={self.branch}
【問題統計】CRITICAL={critical_n} HIGH={high_n} MEDIUM={sev['medium']} LOW={sev['low']}
【Top 問題】{top3}
【OpenClaw評估摘要】{openclaw_report[:300]}

請以 JSON 回答（不含其他文字）：
{{
  "priority": "critical|high|medium|low",
  "auto_fix": true|false,
  "reasoning": "決策理由（繁體中文，一句話，需含具體數字）",
  "fix_files": ["需自動修復的檔案（最多5個，所有有問題的檔案）"],
  "human_review_needed": false
}}

規則（依 ADR-020 全自動修復政策，覆寫 ADR-012 L3 HITL 對 code review 的限制）：
- 任何 finding（不論 CRITICAL/HIGH/MEDIUM/LOW）→ auto_fix=true，human_review_needed=false
- 安全網是 Git revert + Gitea CI/CD 回滾，不依賴人工審查門檻
- priority 按最嚴重 severity 決定：CRITICAL>HIGH>MEDIUM>LOW
- fix_files 填入所有有問題的檔案（最多 5 個，AiderHeal 端會再限流）"""

            resp = elephant_service.generate(
                prompt=prompt,
                json_mode=True,
                temperature=0.1,
                timeout=60,
            )
            if resp.success:
                return self._guard_ea_decision(json.loads(resp.content), findings)
        except Exception as e:
            logger.warning("[CodeReview] ElephantAlpha 決策失敗，回退規則: %s", e)

        # 規則 fallback：ADR-020 全自動修復政策。任何 finding 一律 auto_fix=true。
        has_findings = len(findings) > 0
        priority = (
            "critical" if critical_n > 0 else
            "high"     if high_n     > 0 else
            "medium"   if sev["medium"] > 0 else
            "low"      if sev["low"]    > 0 else "low"
        )
        auto_fix = bool(has_findings and AUTO_FIX_ENABLED)
        fix_files = list({
            f.get("file", "") for f in findings if f.get("file")
        })[:5]

        return {
            "priority":            priority,
            "auto_fix":            auto_fix,
            "reasoning":           f"ADR-020 全自動修復：CRITICAL={critical_n} HIGH={high_n} MEDIUM={sev['medium']} LOW={sev['low']}，"
                                   + ("觸發 AiderHeal 自動修復（Git+CI/CD 為回滾安全網）" if auto_fix else "無 finding，無需修復"),
            "fix_files":           fix_files,
            "human_review_needed": False,
        }

    def _guard_ea_decision(self, decision: Dict, findings: List[Dict]) -> Dict:
        """ADR-020 全自動修復政策：有 finding 一律 auto_fix=true，僅受 AUTO_FIX_ENABLED 主開關控制。"""
        sev = self.state["severity_summary"]
        priority = (decision.get("priority") or "").lower() or (
            "critical" if sev["critical"] > 0 else
            "high"     if sev["high"] > 0 else
            "medium"   if sev["medium"] > 0 else
            "low"
        )
        has_findings = bool(findings)
        allowed_auto_fix = bool(has_findings and AUTO_FIX_ENABLED)
        if has_findings and not AUTO_FIX_ENABLED:
            logger.warning(
                "[CodeReview] auto_fix 被 CODE_REVIEW_AUTO_FIX_ENABLED=false 主開關擋下 priority=%s",
                priority,
            )

        decision["priority"] = priority
        decision["auto_fix"] = allowed_auto_fix
        decision["human_review_needed"] = False
        decision["reasoning"] = (
            f"{decision.get('reasoning', '')} "
            f"[ADR-020 全自動修復: auto_fix={'enabled' if allowed_auto_fix else 'flag_disabled'}, priority={priority}]"
        ).strip()
        return decision

    # ── Step 5：NemoTron 派遣 ──────────────────────────────────────────────────

    def _nemotron_dispatch(self, ea: Dict, findings: List[Dict]) -> Dict:
        auto_fix  = ea.get("auto_fix", False)
        fix_files = ea.get("fix_files", [])
        priority_map = {"critical": 1, "high": 2, "medium": 3, "low": 4}
        priority_num  = priority_map.get(ea.get("priority", "low"), 4)

        actions_created = 0
        session = get_session()
        try:
            # 每個需修復的檔案建立一筆 action_plan
            for fpath in fix_files[:3]:
                related = [f for f in findings if f.get("file") == fpath][:3]
                desc = f"Code Review 修復：{fpath}｜{', '.join(f.get('description','')[:40] for f in related)}"
                session.execute(text("""
                    INSERT INTO action_plans
                        (action_type, description, status, priority, metadata_json, created_at)
                    VALUES
                        ('code_review_fix', :desc, :status, :priority, :meta, NOW())
                """), {
                    "desc":     desc[:500],
                    "status":   "auto_pending" if auto_fix else "auto_disabled",
                    "priority": priority_num,
                    "meta":     json.dumps({
                        "pipeline_id":  self.pipeline_id,
                        "commit_sha":   self.commit_sha,
                        "file":         fpath,
                        "auto_fix":     auto_fix,
                        "ea_priority":  ea.get("priority"),
                        "findings":     related,
                    }, ensure_ascii=False),
                })
                actions_created += 1
            session.commit()
        except Exception as e:
            logger.warning("[CodeReview] action_plans 寫入失敗: %s", e)
            session.rollback()
        finally:
            session.close()

        # 觸發 AiderHeal（非阻塞）
        if auto_fix and fix_files:
            self.state["auto_fix_triggered"] = True
            self._sync_global()
            self._trigger_aider_heal(findings, fix_files)

        return {"actions": actions_created, "auto_fix": auto_fix}

    def _trigger_aider_heal(self, findings: List[Dict], fix_files: List[str]):
        """非阻塞觸發 AiderHeal 自動修復"""
        def _heal_worker():
            try:
                from services.aider_heal_executor import execute_code_fix
                for fpath in fix_files[:2]:             # 最多同時修 2 個檔案
                    related = [f for f in findings if f.get("file") == fpath]
                    if not related:
                        continue
                    worst = sorted(related, key=lambda x: {"CRITICAL":0,"HIGH":1,"MEDIUM":2,"LOW":3}.get(x.get("severity","LOW"),3))[0]
                    result = execute_code_fix(
                        error_type=f"code_review_{worst.get('type','bug')}",
                        error_message=worst.get("description", "Code Review 發現問題"),
                        target_file=fpath,
                        context={"suggestion": worst.get("suggestion", ""), "pipeline_id": self.pipeline_id},
                    )
                    logger.info("[CodeReview] AiderHeal %s → %s", fpath, result.get("message", ""))
            except Exception as e:
                logger.error("[CodeReview] AiderHeal 觸發失敗: %s", e)

        t = threading.Thread(target=_heal_worker, daemon=True, name=f"aider-heal-{self.commit_sha[:8]}")
        t.start()

    # ── DB 持久化 ──────────────────────────────────────────────────────────────

    def _save_to_db(self, findings: List[Dict], openclaw_report: str, ea: Dict):
        session = get_session()
        try:
            row = session.execute(text("""
                INSERT INTO ai_insights
                    (insight_type, content, confidence, created_by,
                     status, metadata_json, period, created_at)
                VALUES
                    ('code_review_result', :content, :conf, 'code_review_pipeline',
                     'active', :meta, :period, NOW())
                RETURNING id
            """), {
                "content": json.dumps({
                    "findings":         findings,
                    "openclaw_report":  openclaw_report,
                    "ea_decision":      ea,
                    "severity_summary": self.state["severity_summary"],
                }, ensure_ascii=False)[:8000],
                "conf":   0.90,
                "meta":   json.dumps({
                    "pipeline_id":       self.pipeline_id,
                    "commit_sha":        self.commit_sha,
                    "branch":            self.branch,
                    "changed_files":     self.changed_files,
                    "auto_fix_triggered": self.state["auto_fix_triggered"],
                }, ensure_ascii=False),
                "period": datetime.now().strftime("%Y-%m-%d"),
            }).fetchone()
            session.commit()
            if row:
                try:
                    from services.openclaw_learning_service import enqueue_insight_embedding
                    enqueue_insight_embedding(row[0], "code_review_result", json.dumps({
                        "findings": findings,
                        "openclaw_report": openclaw_report,
                        "ea_decision": ea,
                    }, ensure_ascii=False), datetime.now().strftime("%Y-%m-%d"))
                except Exception as embed_err:
                    logger.warning("[CodeReview] embedding queue enqueue failed: %s", embed_err)
            logger.info("[CodeReview] ai_insights 寫入成功 pipeline=%s", self.pipeline_id)
        except Exception as e:
            logger.error("[CodeReview] DB 寫入失敗: %s", e)
            session.rollback()
        finally:
            session.close()

    # ── Telegram 通知 ─────────────────────────────────────────────────────────

    def _notify_start(self):
        try:
            from services.telegram_templates import _send_telegram_raw
            files_list = "\n".join(f"  • {f}" for f in self.changed_files[:5])
            if len(self.changed_files) > 5:
                files_list += f"\n  （+{len(self.changed_files)-5} 個）"
            _send_telegram_raw(
                f"🔍 <b>Code Review 啟動</b>\n"
                f"══════════════════════════\n"
                f"📦 Commit　<code>{self.commit_sha[:8]}</code>  🌿 {self.branch}\n"
                f"📝 變更檔案：\n{files_list}\n"
                f"══════════════════════════\n"
                f"🤖 <i>Hermes → OpenClaw → Elephant Alpha → NemoTron</i>\n"
                f"📊 即時進度：https://mo.wooo.work/code-review/"
            )
        except Exception as e:
            logger.warning("[CodeReview] 啟動通知失敗: %s", e)

    def _notify_complete(self, findings: List[Dict], openclaw_report: str, ea: Dict):
        try:
            from services.telegram_templates import _send_telegram_raw
            sev      = self.state["severity_summary"]
            priority = ea.get("priority", "medium")
            auto_fix = ea.get("auto_fix", False)

            icon = {"critical": "🔴", "high": "🟠", "medium": "🟡", "low": "🟢"}.get(priority, "🟡")
            top_issues = [f for f in findings if f.get("severity") in ("CRITICAL", "HIGH")][:3]
            issues_lines = "\n".join(
                f"  {'🔴' if f['severity']=='CRITICAL' else '🟠'} [{f['severity']}] {f.get('description','')} — <i>{f.get('file','')}</i>"
                for f in top_issues
            ) or "  ✅ 無高風險問題"

            msg = (
                f"{icon} <b>Code Review 完成</b>  ·  <code>{self.commit_sha[:8]}</code>\n"
                f"══════════════════════════\n"
                f"🔴 CRITICAL　<b>{sev['critical']}</b>　"
                f"🟠 HIGH　<b>{sev['high']}</b>　"
                f"🟡 MEDIUM　<b>{sev['medium']}</b>　"
                f"🟢 LOW　<b>{sev['low']}</b>\n"
                f"══════════════════════════\n"
                f"<b>⚠️ 主要問題</b>\n{issues_lines}\n"
            )
            if openclaw_report:
                msg += f"\n{openclaw_report[:400]}\n"

            if auto_fix:
                fix_status = "🔧 已觸發自動修復（AiderHeal）"
            elif sev['critical'] + sev['high'] + sev['medium'] + sev['low'] == 0:
                fix_status = "✅ 無需修復動作"
            else:
                fix_status = "🛑 自動修復主開關關閉（CODE_REVIEW_AUTO_FIX_ENABLED=false）"
            msg += (
                f"══════════════════════════\n"
                f"🤖 Elephant Alpha：<b>{priority.upper()}</b>　{fix_status}\n"
                f"📊 完整報告：https://mo.wooo.work/code-review/"
            )
            _send_telegram_raw(msg)
        except Exception as e:
            logger.warning("[CodeReview] 完成通知失敗: %s", e)

    def _notify_error(self, error: str):
        try:
            from services.telegram_templates import _send_telegram_raw
            _send_telegram_raw(
                f"🚨 <b>Code Review Pipeline 失敗</b>\n"
                f"Commit：<code>{self.commit_sha[:8]}</code> @ {self.branch}\n"
                f"錯誤：{error[:200]}\n"
                f"📊 查看：https://mo.wooo.work/code-review/"
            )
        except Exception:
            pass


# ═══════════════════════════════════════════════════════════════════════════════
# 公開 API
# ═══════════════════════════════════════════════════════════════════════════════

def trigger_post_deploy_review(
    commit_sha:    str,
    changed_files: List[str],
    branch:        str = "main",
    deploy_type:   str = "sync",
) -> str:
    """
    啟動 Pipeline（後台 daemon thread）。
    回傳 pipeline_id。由 routes/code_review_routes.py 的 webhook 端點呼叫。
    """
    pipeline = CodeReviewPipeline(commit_sha, changed_files, branch, deploy_type)
    t = threading.Thread(
        target=pipeline.run,
        daemon=True,
        name=f"code-review-{commit_sha[:8]}",
    )
    t.start()
    logger.info("[CodeReview] 已派發 pipeline=%s files=%d", pipeline.pipeline_id, len(pipeline.changed_files))
    return pipeline.pipeline_id


def get_current_state() -> Dict[str, Any]:
    """前端 polling 用：取得目前 pipeline 即時狀態"""
    with _pipeline_lock:
        return dict(_current_pipeline)


def get_history(limit: int = 20) -> List[Dict]:
    """取得 ai_insights 中歷史 code_review_result 記錄"""
    session = get_session()
    try:
        rows = session.execute(text("""
            SELECT id, content, confidence, metadata_json, created_at, status
            FROM ai_insights
            WHERE insight_type = 'code_review_result'
            ORDER BY created_at DESC
            LIMIT :lim
        """), {"lim": limit}).fetchall()

        results = []
        for r in rows:
            meta, content = {}, {}
            try:
                meta    = json.loads(r[3]) if r[3] else {}
                content = json.loads(r[1]) if r[1] else {}
            except Exception:
                pass
            sev = content.get("severity_summary", {})
            results.append({
                "id":               r[0],
                "pipeline_id":      meta.get("pipeline_id", ""),
                "commit_sha":       meta.get("commit_sha", "")[:8],
                "branch":           meta.get("branch", ""),
                "changed_files":    meta.get("changed_files", []),
                "severity_summary": sev,
                "total_issues":     sum(sev.values()),
                "auto_fix":         meta.get("auto_fix_triggered", False),
                "findings":         content.get("findings", []),
                "openclaw_report":  content.get("openclaw_report", ""),
                "ea_decision":      content.get("ea_decision", {}),
                "created_at":       r[4].isoformat() if r[4] else "",
                "status":           r[5] or "active",
            })
        return results
    except Exception as e:
        logger.warning("[CodeReview] 歷史讀取失敗: %s", e)
        return []
    finally:
        session.close()


def verify_internal_token(request_token: str) -> bool:
    """驗證 CD webhook 來源 token。Production 預設必填，避免外部觸發 auto-review/fix 鏈。"""
    if not INTERNAL_TOKEN:
        if ALLOW_INSECURE_WEBHOOK:
            logger.warning("[CodeReview] INTERNAL_WEBHOOK_TOKEN 未設定，僅因 dev override 放行")
            return True
        logger.error("[CodeReview] INTERNAL_WEBHOOK_TOKEN 未設定，拒絕 webhook")
        return False
    return request_token == INTERNAL_TOKEN