ailine/backend/app/agent_subgraphs/news_analysis/api_client.py

"""
资讯子图API调用工具
News Analysis API Client
支持 async 和真实数据库缓存
"""

from typing import Dict, Any, Optional, List
import random
from datetime import datetime
from dataclasses import dataclass


@dataclass
class NewsAPIClient:
    """
    资讯API客户端 - 可扩展支持多种API和数据库缓存
    """
    
    # 可以配置多个API（如 NewsAPI, 今日头条, 百度新闻等）
    newsapi_key: Optional[str] = None
    
    # 数据库 Repository（可选，用于缓存新闻）
    news_repository: Optional[Any] = None
    
    async def query_news_db(self, user_id: str, keyword: str) -> Optional[List[Dict[str, Any]]]:
        """从数据库缓存查询新闻"""
        if not self.news_repository:
            return None
        try:
            entities = await self.news_repository.search_by_keywords(user_id, keyword)
            if entities:
                return [
                    {
                        "title": e.title,
                        "source": e.source,
                        "summary": e.content,
                        "keywords": e.keywords.split(",") if e.keywords else [],
                        "author": "",
                        "published_at": e.created_at
                    }
                    for e in entities
                ]
        except Exception as e:
            print(f"从数据库查询新闻失败：{e}")
        return None
    
    async def cache_news_db(self, user_id: str, news: Dict[str, Any]):
        """把新闻缓存到数据库"""
        if not self.news_repository:
            return
        try:
            from ...db.models import NewsEntity
            entity = NewsEntity(
                user_id=user_id,
                title=news.get("title", ""),
                content=news.get("summary", ""),
                url=news.get("url", ""),
                source=news.get("source", ""),
                keywords=",".join(news.get("keywords", []))
            )
            await self.news_repository.insert(entity)
        except Exception as e:
            print(f"缓存新闻到数据库失败：{e}")
    
    def query_news_mock(self, query: str) -> List[Dict[str, Any]]:
        """
        模拟查询资讯 - 目前用于演示
        """
        # 模拟资讯数据库
        mock_news = [
            {
                "title": "OpenAI发布GPT-5：智能再升级",
                "source": "Tech News",
                "summary": "最新消息，OpenAI刚刚发布了GPT-5模型，智能水平再次取得重大突破...",
                "keywords": ["AI", "GPT-5", "OpenAI"],
                "author": "AI Team",
                "published_at": datetime.now().isoformat()
            },
            {
                "title": "大模型在医疗领域的应用",
                "source": "Health Tech",
                "summary": "大模型AI技术正在医疗领域展现巨大潜力，从辅助诊断到药物研发...",
                "keywords": ["医疗", "大模型", "应用"],
                "author": "Medical Team",
                "published_at": datetime.now().isoformat()
            },
            {
                "title": "2026年AI行业发展趋势报告",
                "source": "Business Daily",
                "summary": "最新行业报告显示，AI行业将继续保持高速增长，企业数字化转型加速...",
                "keywords": ["趋势", "AI", "商业"],
                "author": "Business Team",
                "published_at": datetime.now().isoformat()
            }
        ]
        
        # 根据查询词简单过滤
        results = []
        query_lower = query.lower()
        
        for news in mock_news:
            if (query_lower in news["title"].lower() or 
                query_lower in news["summary"].lower() or 
                any(keyword.lower() in query_lower for keyword in news["keywords"])):
                results.append(news)
        
        # 如果没有匹配到，返回前两条
        if not results:
            results = mock_news[:2]
        
        return results
    
    def analyze_url_mock(self, url: str) -> Dict[str, Any]:
        """
        模拟URL分析 - 目前用于演示
        """
        return {
            "title": f"分析结果：{url}",
            "source": "URL Analyzer",
            "summary": "已完成对该URL的内容分析，包含文章摘要和情感倾向判断...",
            "keywords": ["News", "Analysis", url.split("/")[-1] if url else "unknown"]
        }
    
    def extract_keywords_mock(self, text: str) -> List[str]:
        """
        模拟关键词提取 - 目前用于演示
        """
        # 简单的关键词提取模拟
        common_keywords = ["AI", "大模型", "应用场景", "行业趋势", "创新", "技术"]
        result = []
        
        for keyword in common_keywords:
            if keyword.lower() in text.lower():
                result.append(keyword)
        
        # 如果没找到，返回默认关键词
        if not result:
            result = ["AI", "大模型", "应用场景", "行业趋势"]
        
        return result
    
    def generate_report_mock(self, query: str) -> str:
        """
        模拟报告生成 - 目前用于演示
        """
        report = f"""═══════════════════════════════════════════
📊 资讯分析报告
═══════════════════════════════════════════

主题：{query}

📋 摘要：
这是一份关于 {query} 的资讯分析综合报告，包含最新行业动态和趋势分析。

🔍 主要发现：
1. AI技术持续快速发展
2. 大模型应用场景不断拓展
3. 行业数字化转型加速

🏷️ 关键词：
- AI
- 大模型
- 数字化转型
- 创新

═══════════════════════════════════════════
💡 建议：继续关注行业动态，把握发展机遇！
"""
        return report
    
    # ========== 统一入口（优先查缓存） ==========
    async def query_news(self, user_id: str = "default", query: str = "", use_cache: bool = True) -> List[Dict[str, Any]]:
        """查询新闻（统一入口，优先查数据库缓存）"""
        # 1. 先查数据库缓存
        if use_cache:
            cached = await self.query_news_db(user_id, query)
            if cached:
                return cached
        
        # 2. 查第三方 API（暂未实现）
        # api_result = await self.query_news_api(query)
        # if api_result:
        #     for news in api_result:
        #         await self.cache_news_db(user_id, news)
        #     return api_result
        
        # 3. 用模拟数据（兜底）
        mock_result = self.query_news_mock(query)
        if use_cache:
            for news in mock_result:
                await self.cache_news_db(user_id, news)
        return mock_result


# 单例实例（模拟模式，保持向后兼容）
news_api = NewsAPIClient()