"""
政府采购网爬虫集成示例
演示如何爬取政府采购信息并生成分析报表
"""
import os
from dotenv import load_dotenv
from crawler.ccgp_crawler import CCGPCrawler
from rag import RAGProcessor
from report import ReportGenerator


def main():
    """主函数"""
    load_dotenv()

    # 获取配置
    qwen_api_key = os.getenv("QWEN_API_KEY")
    qwen_model = os.getenv("QWEN_MODEL", "qwen-max")
    data_dir = os.getenv("DATA_DIR", "./data")
    vector_db_dir = os.getenv("VECTOR_DB_DIR", "./vector_db")

    print("="*60)
    print("政府采购信息分析系统 - 信控/绿波")
    print("="*60)

    # 步骤1: 爬取政府采购信息
    print("\n[步骤 1/3] 爬取政府采购网信息...")
    ccgp_crawler = CCGPCrawler()

    # 搜索关键词
    keywords = ["信控", "绿波"]

    results = ccgp_crawler.crawl_by_keywords(
        keywords=keywords,
        max_per_keyword=20  # 每个关键词最多20条
    )

    if not results:
        print("未能获取到采购信息")
        return

    print(f"[成功] 爬取了 {len(results)} 条采购信息")

    # 步骤2: 保存到向量数据库
    print("\n[步骤 2/3] 将采购信息存入向量数据库...")
    rag_processor = RAGProcessor(vector_db_dir=vector_db_dir)

    # 转换格式并组合标题+内容
    news_list = []
    for item in results:
        if not item.get('content'):
            continue

        # 标题和内容一起嵌入向量数据库
        combined_content = f"标题: {item['title']}\n\n{item['content']}"

        news_item = {
            "title": item["title"],
            "content": combined_content,
            "url": item["url"],
            "source": item["source"],
            "std_timestamp": item.get("std_timestamp", 0),
        }
        news_list.append(news_item)

    # 使用upsert模式存储
    rag_processor.process_news(news_list, upsert=True)

    db_stats = rag_processor.get_database_stats()
    print(f"[成功] 向量数据库现有 {db_stats['unique_news']} 条新闻")

    # 步骤3: 生成分析报表
    print("\n[步骤 3/3] 生成采购信息分析报表...")
    report_generator = ReportGenerator(api_key=qwen_api_key, model_name=qwen_model)

    # 检索相关内容增强报表
    relevant_docs = rag_processor.search("信控 绿波 交通信号 采购", k=15)

    report = report_generator.generate_summary_report(news_list, relevant_docs)

    # 保存报表
    report_path = report_generator.save_report(
        report,
        report_type='ccgp_xinlv_lvbo',
        output_dir=data_dir
    )

    print(f"[成功] 报表已生成: {report_path}")

    # 打印报表预览
    print("\n" + "="*60)
    print("报表内容预览：")
    print("="*60)
    print(report)

    # 统计信息
    print("\n" + "="*60)
    print("统计信息：")
    print("="*60)
    print(f"总采购信息数: {len(results)}")

    # 按关键词统计
    keyword_stats = {}
    for item in results:
        kw = item.get('keyword', '未知')
        keyword_stats[kw] = keyword_stats.get(kw, 0) + 1

    print("\n按关键词统计:")
    for kw, count in keyword_stats.items():
        print(f"  - {kw}: {count} 条")

    # 按项目类型统计
    type_stats = {}
    for item in results:
        ptype = item.get('project_type', '未知')
        if ptype:
            type_stats[ptype] = type_stats.get(ptype, 0) + 1

    if type_stats:
        print("\n按项目类型统计:")
        for ptype, count in type_stats.items():
            print(f"  - {ptype}: {count} 条")

    print("\n" + "="*60)
    print("任务完成！")
    print("="*60)


if __name__ == "__main__":
    main()