# -*- coding: utf-8 -*-
"""更新 pac_domains.json：从 v2fly/domain-list-community (geosite) 拉取境外服务
分类，与现有手工条目合并，生成新的强制走隧道域名后缀表。

用法:
    python update_pac_domains.py            # 拉取并更新（先自动备份）
    python update_pac_domains.py --dry-run  # 只统计，不写文件

geosite 数据格式（data/<分类>，每行一条规则）:
    include:xxx      引用其他分类（递归解析）
    example.com      匹配域名及其子域（与我们的后缀匹配一致）
    +example.com     同上（additive）
    .example.com     仅精确匹配（不匹配子域）-> 跳过
    -example.com     排除 -> 跳过
    /regex/          正则 -> 跳过
    example.com @cn  属性标签 -> 去掉标签
    无点的单标签（如 chrome）-> 无实际主机名会命中，跳过
"""
import argparse
import json
import sys
import urllib.request
from datetime import datetime
from pathlib import Path

BASE_URL = "https://raw.githubusercontent.com/v2fly/domain-list-community/master/data/"
# 需要强制走隧道的境外服务分类（geosite 分类名），可按需增删
CATEGORIES = ["google", "apple", "telegram", "openai", "twitter", "facebook",
              "github", "netflix", "discord", "spotify", "whatsapp", "paypal"]

SCRIPT_DIR = Path(__file__).resolve().parent
LITE_DIR = SCRIPT_DIR.parent.parent
PAC_FILE = LITE_DIR / "pac_domains.json"
BACKUP_DIR = LITE_DIR / "backups"
# 本地 pac_domains.json 已不再随项目保留：缺失时从 CDN 拉当前版本作合并基底
CDN_URL = "https://cdn.rrrosmontis.icu/RRPP/PAC/pac_domains.json"


def fetch(cat, tries=2):
    for i in range(tries):
        try:
            with urllib.request.urlopen(BASE_URL + cat, timeout=20) as r:
                return r.read().decode("utf-8", "replace").splitlines()
        except Exception:
            if i == tries - 1:
                raise


def resolve(cat, seen=None):
    """递归展开 include:，返回原始行列表；循环引用用 seen 防死循环。"""
    seen = seen or set()
    if cat in seen:
        return []
    seen.add(cat)
    lines = []
    try:
        for raw in fetch(cat):
            line = raw.strip()
            if line.startswith("include:"):
                lines += resolve(line[len("include:"):].strip(), seen)
            else:
                lines.append(line)
    except Exception as e:
        print(f"  [warn] 拉取 {cat} 失败: {e}")
    return lines


def parse_domains(lines):
    """过滤 geosite 行，只保留可映射为后缀匹配的纯域名。"""
    out = set()
    for line in lines:
        s = line.strip()
        if not s or s.startswith("#") or s.startswith("/"):
            continue
        if s.startswith("-") or s.startswith("."):
            continue
        s = s.lstrip("+").split("@")[0].strip()
        if "." not in s:
            continue
        out.add(s.lower())
    return out


def main():
    parser = argparse.ArgumentParser(description="更新 pac_domains.json")
    parser.add_argument("--dry-run", action="store_true", help="只统计，不写文件")
    args = parser.parse_args()

    if PAC_FILE.is_file():
        cur = set(json.loads(PAC_FILE.read_text("utf-8"))["domains"])
    else:
        # 项目目录已不保留该文件：从 CDN 拉当前版本作合并基底（保留其中的手工条目）
        print(f"本地未找到 {PAC_FILE}，从 CDN 拉取当前版本作为合并基底")
        try:
            with urllib.request.urlopen(CDN_URL, timeout=20) as r:
                cur = set(json.loads(r.read().decode("utf-8"))["domains"])
        except Exception as e:
            print(f"拉取失败: {e}")
            sys.exit(1)

    print(f"拉取 geosite 分类: {', '.join(CATEGORIES)}")
    merged = set()
    for cat in CATEGORIES:
        doms = parse_domains(resolve(cat))
        merged |= doms
        print(f"  {cat:<12} -> {len(doms):>5} 条")

    new = merged - cur
    kept = cur & merged
    only_cur = cur - merged
    print(f"\n合并后总数: {len(merged | cur)}（geosite {len(merged)} + 现有 {len(cur)}）")
    print(f"新增 {len(new)}，与现有重叠 {len(kept)}，现有手工条目 {len(only_cur)} 条全部保留")

    if args.dry_run:
        print("--dry-run: 未写文件")
        return

    final = sorted(merged | cur)
    now = datetime.now().strftime("%Y-%m-%d %H:%M")
    comment = (
        f"强制走代理隧道的域名后缀（IP 智能分流的补充规则）：命中这些域名的连接即使"
        f"解析到国内 IP 也强制走隧道。匹配规则：host == 条目 或 host 以 \".条目\" 结尾。"
        f"由 others/update_pac_domains/update_pac_domains.py 从 v2fly geosite "
        f"({', '.join(CATEGORIES)}) 生成于 {now}，手工条目已保留；"
        f"客户端按文件内 version/date 从 CDN 在线更新到 %LOCALAPPDATA%\\RRPP。"
    )
    out = {"_comment": comment,
           "version": datetime.now().strftime("%Y%m%d%H%M"),
           "date": datetime.now().strftime("%Y-%m-%d"),
           "domains": final}

    BACKUP_DIR.mkdir(exist_ok=True)
    bak = BACKUP_DIR / f"pac_domains.json.bak_{datetime.now().strftime('%Y%m%d_%H%M%S')}"
    bak.write_bytes(PAC_FILE.read_bytes())
    print(f"已备份: {bak}")

    PAC_FILE.write_text(json.dumps(out, ensure_ascii=False, indent=2) + "\n", "utf-8")
    size = PAC_FILE.stat().st_size / 1024
    print(f"已写入: {PAC_FILE}（{len(final)} 条，{size:.0f} KB）")


if __name__ == "__main__":
    main()