#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
交付物阻断式核查脚本（对齐软件改造部署外包交付规范）。

检查项：
  A. 7 项交付物：源码 / 当前产品截图 / 全功能录屏 / 网盘归档 / README / Docker 部署 / 依赖与注意事项
  B. README 12 项必备内容（按标题关键词启发式匹配）
  C. 痕迹/来源清理：对用户提供的商业标识与来源关键词做全文检索（交付物中不得命中）
  D. 交付清单 .Codex/delivery/manifest.md 存在且 7 项条目齐全

退出码：0 全部通过；1 存在阻断项。
用法见 --help。
"""
from __future__ import annotations

import argparse
import json
import os
import re
import sys
from datetime import datetime
from pathlib import Path

README_KEYWORDS = {
    "一句话简介": ["简介", "项目介绍"],
    "详细功能描述": ["功能", "功能特性", "功能特点"],
    "行业痛点": ["痛点"],
    "技术优势": ["技术优势", "核心优势"],
    "应用场景": ["应用场景", "使用场景", "场景"],
    "目标行业与人群": ["目标行业", "适用行业", "适配行业", "目标用户", "适用人群"],
    "商业应用价值": ["商业价值", "应用价值"],
    "技术架构(含拓扑图)": ["技术架构", "架构"],
    "环境前置依赖": ["环境要求", "前置依赖", "依赖要求", "环境依赖"],
    "快速开始": ["快速开始", "快速上手", "quick start"],
    "安装部署教程": ["部署", "安装"],
    "版权与使用说明": ["版权", "使用说明"],
}

SCAN_SUFFIXES = {
    ".ts", ".tsx", ".js", ".jsx", ".vue", ".java", ".py", ".go", ".json",
    ".md", ".html", ".yml", ".yaml", ".less", ".css", ".scss", ".xml",
}

IMAGE_SUFFIXES = {".png", ".jpg", ".jpeg", ".webp"}
VIDEO_SUFFIXES = {".mp4", ".mkv", ".mov", ".avi", ".webm"}

SKIP_DIRS = {
    "node_modules", ".git", "dist", "build", "target", ".next", ".umi",
    "__pycache__", ".venv", "venv",
}


def iter_files(root: Path):
    # os.walk 自顶向下并在目录层剪枝，避免遍历 node_modules/.git 等巨型目录
    for dirpath, dirnames, filenames in os.walk(root):
        dirnames[:] = [
            d for d in dirnames
            if d not in SKIP_DIRS and not d.lower().startswith(".codex")
        ]
        for name in filenames:
            yield Path(dirpath) / name


def find_files(root: Path, suffixes: set[str]) -> list[Path]:
    return [p for p in iter_files(root) if p.suffix.lower() in suffixes]


def main() -> int:
    ap = argparse.ArgumentParser(description="开源改造项目交付物阻断式核查")
    ap.add_argument("--project-dir", required=True, help="项目根目录")
    ap.add_argument("--readme", default="README.md", help="README 相对路径")
    ap.add_argument("--delivery-dir", default=".Codex/delivery", help="交付目录（相对项目）")
    ap.add_argument("--assets-dir", default=".Codex/delivery/assets",
                    help="截图/录屏素材目录（相对项目）")
    ap.add_argument("--trace-keywords", default="",
                    help="逗号分隔的商业标识与来源宣传关键词（原公司名、外部仓库、项目代号等）")
    ap.add_argument("--allow-no-docker", action="store_true",
                    help="项目确实不适配 Docker 时，要求存在书面说明而非 Dockerfile")
    ap.add_argument("--netdisk-link", default="",
                    help="网盘归档链接（manifest 中也可提供，二选一）")
    ap.add_argument("--out", default=".Codex/reports/deliverables-check.json")
    args = ap.parse_args()

    root = Path(args.project_dir).resolve()
    violations: list[str] = []
    warnings: list[str] = []
    evidence: dict = {}

    # ---------- A1. 源码可构建产物线索 ----------
    # 多模块项目（如 frontend/backend 分离）清单可能在子目录，递归查找
    manifest_names = {
        "package.json", "pom.xml", "build.gradle", "go.mod",
        "requirements.txt", "pyproject.toml", "Cargo.toml",
    }
    manifest_hits = [p for p in iter_files(root) if p.name in manifest_names]
    if not manifest_hits:
        violations.append("未发现项目清单文件（package.json/pom.xml/go.mod 等），无法确认源码完整性")
    else:
        evidence["manifests"] = sorted(str(p.relative_to(root)) for p in manifest_hits)

    # ---------- A2/A3. 截图与录屏 ----------
    assets_dir = root / args.assets_dir
    alt_dirs = [root / args.delivery_dir / "screenshots", root / args.delivery_dir / "recordings"]
    images: list[Path] = []
    videos: list[Path] = []
    search_dirs = [d for d in [assets_dir, *alt_dirs] if d.exists()]
    for d in search_dirs:
        images.extend(find_files(d, IMAGE_SUFFIXES))
        videos.extend(find_files(d, VIDEO_SUFFIXES))
    # 对外只提供当前产品界面截图；交付素材目录不得含改造前/before 旧界面
    leaked_before = [p for p in images if re.search(r"(before|改造前|原版)", p.name, re.I)]
    if leaked_before:
        violations.append("交付素材中包含改造前/原版截图（不应随包交付）：%s"
                          % "、".join(p.name for p in leaked_before[:8]))
    if not images:
        violations.append("缺少产品界面截图（应放在 %s 或交付目录）" % args.assets_dir)
    else:
        evidence["screenshots"] = len(images)
    if not videos:
        violations.append("缺少全功能操作演示录屏（mp4/mov 等）")
    else:
        evidence["recordings"] = len(videos)

    # ---------- A5. README ----------
    readme_path = root / args.readme
    readme_text = ""
    if not readme_path.exists():
        violations.append("缺少 README 文档：%s" % args.readme)
    else:
        readme_text = readme_path.read_text(encoding="utf-8", errors="ignore")
        missing = []
        low = readme_text.lower()
        for item, kws in README_KEYWORDS.items():
            if not any(k.lower() in low for k in kws):
                missing.append(item)
        if missing:
            violations.append("README 缺少必备内容：%s" % "、".join(missing))
        # 架构图线索：mermaid 或图片引用
        if not re.search(r"```mermaid|!\[.*\]\(.*\)|架构图|拓扑图", readme_text):
            warnings.append("README 未发现架构/拓扑图（mermaid 或图片）")
        # 截图图文对应
        if not re.search(r"!\[.*\]\(.*\)", readme_text):
            warnings.append("README 无图片引用，合同要求功能逐条对应截图")

    # ---------- A6. Docker / 书面说明 ----------
    docker_files = [p for p in [root / "Dockerfile", root / "docker-compose.yml", root / "docker-compose.yaml"] if p.exists()]
    if not docker_files:
        if args.allow_no_docker:
            note = root / args.delivery_dir / "no-docker说明.md"
            if not note.exists():
                violations.append("已声明不适配 Docker，但缺少书面说明（%s）" % note.relative_to(root))
        else:
            violations.append("缺少 Dockerfile / docker-compose（不适配 Docker 请加 --allow-no-docker 并提供书面说明）")

    # ---------- A7. 依赖清单与部署注意事项 ----------
    dep_docs = list((root / args.delivery_dir).glob("*依赖*")) + list((root / args.delivery_dir).glob("*注意事项*"))
    dep_docs += [root / "依赖清单.md", root / "部署注意事项.md"]
    dep_docs = [p for p in dep_docs if p.exists()]
    if not dep_docs:
        # 若 README 已含环境依赖/注意事项章节，降级为警告
        if re.search(r"环境(要求|依赖|前置)|注意事项", readme_text):
            warnings.append("未单独提供依赖清单/部署注意事项文档（README 中有相关章节，建议单独输出）")
        else:
            violations.append("缺少环境依赖清单与部署注意事项说明文档")

    # ---------- A4. 网盘链接 + D. manifest ----------
    manifest_path = root / args.delivery_dir / "manifest.md"
    manifest_text = manifest_path.read_text(encoding="utf-8", errors="ignore") if manifest_path.exists() else ""
    netdisk = args.netdisk_link or (re.search(r"https?://\S*(?:pan\.baidu|aliyundrive|alipan|123pan|cloud\.189)\S*", manifest_text) or [""])[0]
    if not netdisk:
        if not manifest_text:
            violations.append("缺少网盘归档链接，且未提供交付清单 manifest.md")
        else:
            violations.append("未提供网盘素材归档链接（命令行 --netdisk-link 或写入 manifest）")
    if not manifest_path.exists():
        violations.append("缺少交付清单：%s" % manifest_path.relative_to(root))
    else:
        required_items = ["源码", "截图", "录屏", "网盘", "README", "Docker", "依赖"]
        miss_items = [w for w in required_items if w not in manifest_text]
        if miss_items:
            violations.append("manifest.md 未覆盖交付物条目：%s" % "、".join(miss_items))

    # ---------- C. 痕迹清理检索 ----------
    kws = [k.strip() for k in args.trace_keywords.split(",") if k.strip()]
    traces: list[str] = []
    if kws:
        # before/after 需要按独立词识别，避免 beforeEach、::before、::after 等代码语法误报。
        exact_keywords = [k for k in kws if k.lower() not in {"before", "after"}]
        exact_pat = (
            re.compile("|".join(re.escape(k) for k in exact_keywords), re.I)
            if exact_keywords else None
        )
        boundary_pats = {
            "before": re.compile(r"(?<![A-Za-z_:-])before(?![A-Za-z_:-])", re.I),
            "after": re.compile(r"(?<![A-Za-z_:-])after(?![A-Za-z_:-])", re.I),
        }
        active_boundary_pats = [
            boundary_pats[k.lower()] for k in kws if k.lower() in boundary_pats
        ]
        for file_path in iter_files(root):
            if file_path.suffix.lower() not in SCAN_SUFFIXES:
                continue
            rel = str(file_path.relative_to(root))
            if re.search(r"license|copying", rel, re.I):
                continue
            try:
                file_text = file_path.read_text(encoding="utf-8", errors="ignore")
            except OSError:
                continue
            matched = bool(exact_pat and exact_pat.search(file_text))
            if not matched:
                matched = any(pat.search(file_text) for pat in active_boundary_pats)
            if matched:
                traces.append(rel)
        if traces:
            violations.append("痕迹清理不彻底，以下文件命中原标识关键词：%s" % "、".join(sorted(set(traces))[:20]))

    # ---------- 输出 ----------
    result = {
        "timestamp": datetime.now().isoformat(),
        "project_dir": str(root),
        "violations": violations,
        "warnings": warnings,
        "evidence": evidence,
        "passed": not violations,
    }
    out_path = root / args.out
    out_path.parent.mkdir(parents=True, exist_ok=True)
    out_path.write_text(json.dumps(result, ensure_ascii=False, indent=2), encoding="utf-8")

    print("交付物核查：%d 个阻断项，%d 个警告" % (len(violations), len(warnings)))
    for v in violations:
        print("  [FAIL]", v)
    for w in warnings:
        print("  [WARN]", w)
    print("report=%s" % out_path)
    return 0 if not violations else 1


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