# -*- coding: utf-8 -*-
"""DoD 验收清单文档（docs/DoD验收清单.md）完整性与指标一致性检查。

检查项：
1. 必备章节齐全（目的用法 / 内核验收 / 首模板验收 / NFR 核对 / 验收流程节点 / 风险登记 / 一致性说明）；
2. PRD §10 验收标准中的关键量化指标全部出现（P99≤1.8s / 丢失率≤0.02% / 可用性≥99.8% /
   路由准确率≥96.5% / 首屏≤2s / 质量预测准确率≥90% / 预警提前≥30分钟 / 误报率 / 本地首token≤1.5s /
   交接班报告≤2分钟）；
3. 状态口径四态（未开始 / 待验证 / 通过 / 阻塞）齐全；
4. 验收流程六节点（自测 / 内部评审 / 客户UAT / 上线 / 质保观察 / 闭环）齐全；
5. 引用的核心验收证据文件路径在仓库内真实存在。

用法：python _check_dod.py
"""
import os
import re
import sys

HERE = os.path.dirname(os.path.abspath(__file__))
REPO_ROOT = os.path.dirname(HERE)
DOC_PATH = os.path.join(HERE, "DoD验收清单.md")

REQUIRED_SECTIONS = [
    "## 1. 目的与用法",
    "## 2. iAOP-Core 内核验收",
    "## 3. iAOP-Template-Ti 一期首模板验收",
    "## 4. NFR 量化底线核对表",
    "## 5. 验收流程与节点",
    "## 6. 阻塞与风险登记",
    "## 7. 一致性说明",
]

# PRD §9/§10 关键量化指标（去空格后匹配，兼容全角/半角差异）
KEY_METRICS = [
    "1.8s",          # 采集 P99
    "0.02%",         # 丢失率
    "99.8%",         # 可用性
    "96.5%",         # 路由准确率 / DLP（路由口径）
    "2s",            # 首屏
    "90%",           # 质量预测准确率
    "30 分钟",       # 预警提前量
    "1.5s",          # 本地首 token
    "2 分钟",        # 交接班报告生成
    "100%",          # DLP 拦截率
]

STATUS_TOKENS = ["未开始", "待验证", "通过", "阻塞"]

# 状态表格中的符号（核对口径可勾选）
STATUS_SYMBOLS = ["☐", "◐", "✅", "⛔"]

FLOW_NODES = ["自测", "内部评审", "客户 UAT", "生产上线", "质保观察", "闭环"]

# 关键证据文件须真实存在（仓库内）
EVIDENCE_PATHS = [
    os.path.join(REPO_ROOT, "core", "data-bus", "scripts", "bench_write.py"),
    os.path.join(REPO_ROOT, "core", "edge-gateway", "scripts", "verify_resilience.py"),
    os.path.join(REPO_ROOT, "docs", "采集网只读隔离与网闸配置.md"),
    os.path.join(REPO_ROOT, "core", "edge-gateway", "config", "point_dict.example.csv"),
    os.path.join(REPO_ROOT, "core", "llm-gateway", "router.py"),
    os.path.join(REPO_ROOT, "core", "llm-gateway", "dlp.py"),
    os.path.join(REPO_ROOT, "deploy", "k8s", "helm", "iaop"),
    os.path.join(REPO_ROOT, "templates", "resin", "version.yaml"),
    os.path.join(REPO_ROOT, "templates", "ti-cl4", "point-dict", "point_dict.default.csv"),
    os.path.join(REPO_ROOT, "docs", "LIMS质检接口字段清单.md"),
    os.path.join(REPO_ROOT, "docs", "产品设计文档_iAOP.md"),
]


def _norm(s: str) -> str:
    """归一化用于指标比对：移除所有空白。"""
    return re.sub(r"\s+", "", s)


def main() -> int:
    failures = []
    if not os.path.isfile(DOC_PATH):
        print("FAIL")
        print(" -", "DoD 验收清单文档缺失：DoD验收清单.md")
        return 1

    text = open(DOC_PATH, "r", encoding="utf-8").read()

    # 1) 必备章节
    for sec in REQUIRED_SECTIONS:
        if sec not in text:
            failures.append(f"缺少必备章节：{sec}")

    # 2) 关键量化指标（归一化后比对）
    norm_text = _norm(text)
    for metric in KEY_METRICS:
        if _norm(metric) not in norm_text:
            failures.append(f"缺少关键量化指标：{metric}")

    # 3) 状态口径四态齐全
    for token in STATUS_TOKENS:
        if token not in text:
            failures.append(f"缺少状态口径说明：{token}")
    symbol_present = [sym for sym in STATUS_SYMBOLS if sym in text]
    if len(symbol_present) < 3:
        failures.append(f"状态符号至少覆盖 3 种，实际出现：{symbol_present}")

    # 4) 验收流程六节点
    for node in FLOW_NODES:
        if node not in text:
            failures.append(f"缺少验收流程节点：{node}")

    # 5) 引用证据文件真实存在
    for p in EVIDENCE_PATHS:
        if not os.path.exists(p):
            failures.append(f"引用证据路径不存在：{os.path.relpath(p, REPO_ROOT)}")

    if failures:
        print("FAIL")
        for f in failures:
            print(" -", f)
        return 1
    print("OK: DoD 清单章节/量化指标/状态口径/流程节点/证据路径校验通过")
    return 0


if __name__ == "__main__":
    sys.exit(main())