#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
人民日报搜索爬虫命令行工具
用法: python RmrbSearch.py --key "关键词" --total 10 --type 0
"""

import argparse
import json
import sys
from pathlib import Path

# Add parent directory to path to import crawler
sys.path.insert(0, str(Path(__file__).parent.parent))

from crawler.RmrbCrawler import RmrbCrawler
from loguru import logger


def main():
    """主函数"""
    parser = argparse.ArgumentParser(
        description='人民日报新闻搜索工具',
        formatter_class=argparse.RawDescriptionHelpFormatter,
        epilog="""
示例:
  python RmrbSearch.py --key "教育改革" --total 20
  python RmrbSearch.py -k "科技创新" -t 15 -n 1

新闻类型说明:
  0 - 所有类型 (默认)
  1 - 新闻
  2 - 互动
  3 - 报刊
  4 - 图片
  5 - 视频
        """
    )

    parser.add_argument(
        '--key', '-k',
        type=str,
        required=True,
        help='搜索关键词 (必需)'
    )

    parser.add_argument(
        '--total', '-t',
        type=int,
        default=10,
        help='获取新闻总数 (默认: 10)'
    )

    parser.add_argument(
        '--type', '-n',
        type=int,
        default=0,
        choices=[0, 1, 2, 3, 4, 5],
        help='新闻类型: 0=全部, 1=新闻, 2=互动, 3=报刊, 4=图片, 5=视频 (默认: 0)'
    )

    args = parser.parse_args()

    try:
        # 创建爬虫实例
        logger.info(f"开始搜索: 关键词='{args.key}', 数量={args.total}, 类型={args.type}")
        crawler = RmrbCrawler()

        # 执行搜索
        result = crawler.search(key=args.key, total=args.total, news_type=args.type)

        # 输出JSON结果
        output = {
            "code": result.code,
            "message": result.message,
            "success": result.success,
            "data": None,
            "dataList": [item.dict() for item in result.dataList] if result.dataList else []
        }

        print(json.dumps(output, ensure_ascii=False, indent=2))

        # 关闭爬虫
        crawler.close()

        # 退出码: 成功=0, 失败=1
        sys.exit(0 if result.success else 1)

    except Exception as e:
        logger.error(f"执行失败: {str(e)}")
        error_output = {
            "code": 500,
            "message": f"执行失败: {str(e)}",
            "success": False,
            "data": None,
            "dataList": []
        }
        print(json.dumps(error_output, ensure_ascii=False, indent=2))
        sys.exit(1)


if __name__ == "__main__":
    main()