DataOps-platform/scripts/generate_frontend_removal_inventory.py

#!/usr/bin/env python3
"""Generate the auditable frontend deletion inventory from the working tree."""

from __future__ import annotations

import subprocess
from collections import Counter, defaultdict
from pathlib import Path


ROOT = Path(__file__).resolve().parents[1]
OUTPUT = ROOT / "docs" / "architecture" / "FRONTEND_REMOVAL_INVENTORY.md"


def deleted_files() -> list[str]:
    result = subprocess.run(
        ["git", "diff", "--name-only", "--diff-filter=D", "--", "frontend"],
        cwd=ROOT,
        check=True,
        capture_output=True,
        text=True,
    )
    return sorted(line for line in result.stdout.splitlines() if line)


def group_name(path: str) -> str:
    parts = Path(path).parts
    if len(parts) >= 5 and parts[:3] == ("frontend", "src", "views"):
        return f"views/{parts[3]}"
    if len(parts) >= 4 and parts[:3] == ("frontend", "src", "api"):
        return "api"
    if len(parts) >= 4 and parts[:3] == ("frontend", "src", "components"):
        return "components"
    if len(parts) >= 4 and parts[:3] == ("frontend", "src", "layout"):
        return "layout"
    return "/".join(parts[1:3])


def render(files: list[str]) -> str:
    groups: dict[str, list[str]] = defaultdict(list)
    for path in files:
        groups[group_name(path)].append(path)
    counts = Counter({name: len(paths) for name, paths in groups.items()})

    lines = [
        "# 前端历史代码删除清单",
        "",
        "> 生成方式：`python scripts/generate_frontend_removal_inventory.py`。本文件基于当前未提交删除集生成；合并前应再次生成并复核。",
        "",
        "## 1. 基线与恢复方式",
        "",
        "- 原解析报告识别出 348 个前端未匹配调用，产品确认其主体为已下线后端服务遗留；本轮不重建这些旧接口。",
        f"- 当前 Git 删除集包含 {len(files)} 个前端文件；一个文件可包含多个调用，因此文件数不等于 348 个调用数。",
        "- 原始代码备份：`/Users/maxiaolong/Documents/Codex/2026-07-06/g-i/DataOps-platform-backup-before-cleanup-20260716-170321.tar.gz`。",
        "- SHA-256：`866c6c4aff127f88497a5ddf39ac8d8da93008cc784ae6f276f80a7cd17d4ef0`。",
        "- 恢复优先级：先从 Git 历史恢复单文件；若原始工作区未入库内容缺失，再从上述只读备份提取。禁止将整份备份直接覆盖当前工作区。",
        "",
        "## 2. 删除批次",
        "",
        "| 批次 | 范围 | 处置依据 |",
        "|---|---|---|",
        "| C1 | Vanna、Airflow、模型训练、问卷薪酬、图表定制 | 产品确认整体下线 |",
        "| C2 | 数据资源、数据模型、数据指标和旧数据资产原型 | 静态引用与路由复核后确认没有保留链路 |",
        "| C3 | 旧数据工厂非 n8n 页面、执行/调度/节点管理 | 执行能力统一由 n8n 提供 |",
        "| C4 | 旧系统管理、自助注册、租户、升级管理 | 后续以管理员/编辑者/查看者重建 |",
        "| C5 | 不可达公共组件、API 封装和演示页 | 无活跃路由或 import 消费方 |",
        "",
        "## 3. 明确保留例外",
        "",
        "- n8n 适配页面和 `/api/datafactory/*` 后端代理接口；Workflow 保留为 n8n 原生概念。",
        "- DataFlow 治理定义页面；后续新增一个 DataFlow 对多个 n8n Workflow 版本的映射。",
        "- 数据治理知识库入口和工作台；当前为验收占位，下一阶段补齐真实能力。",
        "- 数据安全入口；当前不宣称已实现安全治理能力。",
        "- HOPMS 数据导入工具及其验收样例。",
        "",
        "## 4. 按目录统计",
        "",
        "| 分组 | 删除文件数 |",
        "|---|---:|",
    ]
    for name, count in sorted(counts.items(), key=lambda item: (-item[1], item[0])):
        lines.append(f"| `{name}` | {count} |")

    lines.extend(
        [
            "",
            "## 5. 合并门禁",
            "",
            "1. `npm --prefix frontend run build` 成功。",
            "2. 路由表中不再出现已下线模块，活跃页面无失效 import。",
            "3. `tests/test_cleanup_contract.py`、Docker 静态隔离测试与本地冒烟测试全部通过。",
            "4. 数据安全、治理知识库、工作台、DataFlow 和 n8n 保留入口可访问。",
            "5. 生产数据库不执行删表；仅提交代码和建库脚本清理。",
            "",
            "## 6. 完整删除文件清单",
            "",
        ]
    )
    for name in sorted(groups):
        lines.append(f"### {name}")
        lines.append("")
        lines.extend(f"- `{path}`" for path in groups[name])
        lines.append("")
    return "\n".join(lines).rstrip() + "\n"


def main() -> None:
    files = deleted_files()
    if not files:
        raise SystemExit("No deleted frontend files found in the working tree")
    OUTPUT.parent.mkdir(parents=True, exist_ok=True)
    OUTPUT.write_text(render(files), encoding="utf-8")
    print(f"Generated {OUTPUT.relative_to(ROOT)} with {len(files)} deleted files")


if __name__ == "__main__":
    main()