erlangshen 0.1.0

package/src/agents/equity.py

@@ -0,0 +1,136 @@
+"""
+股票分析师 Agent
+"""
+
+from typing import Optional, Dict, Any, List
+from src.agents.base import BaseAgent
+from src.brain import Brain
+from src.mcp.registry import MCPRegistry
+
+
+class EquityAgent(BaseAgent):
+    """
+    股票分析师
+    
+    专注于A股、港股、美股的基本面和技术面分析
+    """
+    
+    def __init__(self, brain: Brain, mcp: MCPRegistry):
+        super().__init__(brain, mcp)
+        self.role = "股票分析师"
+    
+    def _build_system_prompt(self) -> str:
+        return """你是一位资深股票分析师，专注于：
+- A股、港股、美股基本面分析
+- 公司财务分析 (营收、利润、现金流)
+- 行业研究和竞争格局分析
+- 估值分析 (PE, PB, DCF)
+- 技术分析 (趋势、支撑阻力、量价关系)
+- 个股和行业配置建议
+
+分析原则：
+1. 价值投资：注重企业内在价值和长期回报
+2. 风险收益：平衡收益预期和风险控制
+3. 基本面为本：技术面作为辅助参考
+4. 分散配置：避免过度集中单一标的
+
+输出格式：
+## 核心观点
+(1-2句话概括投资逻辑)
+
+## 基本面分析
+(公司概况、竞争优势、财务数据)
+
+## 估值分析
+(当前估值、历史分位、合理区间)
+
+## 技术面
+(近期走势、关键价位、量价配合)
+
+## 风险因素
+(主要风险点)
+
+## 投资建议
+(买入/持有/卖出建议，目标价位)
+"""
+    
+    async def process(self, query: str, **kwargs) -> str:
+        """
+        处理股票分析查询
+        
+        Args:
+            query: 分析查询
+            **kwargs: 其他参数 (symbol, symbols等)
+        """
+        symbol = kwargs.get("symbol")
+        symbols = kwargs.get("symbols", [])
+        days = kwargs.get("days", 30)
+        
+        data = {}
+        
+        # 获取股票数据
+        if symbol:
+            symbols = [symbol]
+        
+        if symbols:
+            quotes = await self.call_mcp_tool("get_realtime_quotes", symbols=symbols)
+            if isinstance(quotes, list):
+                for quote in quotes:
+                    symbol = quote.get("symbol")
+                    if symbol:
+                        data[symbol] = quote
+                        # 获取历史数据
+                        history = await self.call_mcp_tool(
+                            "get_stock_history",
+                            symbol=symbol,
+                            days=days
+                        )
+                        if isinstance(history, list):
+                            data[f"{symbol}_history"] = history[-5:]  # 最近5天
+        
+        # 执行分析
+        return await self.analyze(query, data=data)
+    
+    async def analyze_stock(self, symbol: str, angle: str = "综合") -> str:
+        """
+        分析单只股票
+        
+        Args:
+            symbol: 股票代码
+            angle: 分析角度 (综合/基本面/技术面/估值)
+        """
+        # 获取数据
+        quote = await self.call_mcp_tool("get_stock_price", symbol=symbol)
+        history = await self.call_mcp_tool("get_stock_history", symbol=symbol, days=60)
+        
+        data = {
+            "quote": quote,
+            "history": history[-10:] if isinstance(history, list) else []
+        }
+        
+        prompt = f"请从{angle}角度分析股票 {symbol}"
+        return await self.analyze(prompt, data=data)
+    
+    async def compare_stocks(self, symbols: List[str]) -> str:
+        """
+        比较股票
+        
+        Args:
+            symbols: 股票代码列表
+        """
+        quotes = await self.call_mcp_tool("get_realtime_quotes", symbols=symbols)
+        
+        data = {"quotes": quotes}
+        
+        prompt = f"请比较以下股票：{', '.join(symbols)}"
+        return await self.analyze(prompt, framework="比较分析", data=data)
+    
+    async def analyze_sector(self, sector: str) -> str:
+        """
+        分析行业板块
+        
+        Args:
+            sector: 行业名称
+        """
+        prompt = f"请分析{sector}行业的：\n1. 当前景气度\n2. 竞争格局\n3. 龙头公司\n4. 投资机会"
+        return await self.think(prompt)

package/src/agents/equity_agent.py

@@ -0,0 +1,91 @@
+"""
+Equity Agent - 股票分析师
+专注于个股分析、行业研究、财务建模
+"""
+from typing import Optional
+from loguru import logger
+
+from .base_agent import BaseAgent
+
+
+class EquityAgent(BaseAgent):
+    """
+    股票分析师智能体
+
+    能力：
+    - 个股基本面分析
+    - 行业研究
+    - 财务建模
+    - 估值分析
+    """
+
+    def __init__(self, tools: Optional[dict] = None):
+        super().__init__(
+            name="股票分析师",
+            description="专注于个股基本面分析和行业研究",
+            tools=tools,
+        )
+
+    async def process(self, query: str, context: Optional[dict] = None) -> dict:
+        """
+        处理股票分析查询
+
+        Args:
+            query: 分析问题
+            context: 上下文
+
+        Returns:
+            dict 包含分析结论
+        """
+        logger.info(f"EquityAgent processing: {query[:80]}")
+        context = context or {}
+
+        # 1. 提取股票代码
+        symbol = context.get("symbol", self._extract_symbol(query))
+
+        # 2. 分类分析类型
+        analysis_type = self._classify_query(query)
+
+        # 3. 获取数据
+        data = {}
+        if symbol:
+            if "market_tools" in self.tools:
+                data["price"] = await self.call_tool("market_tools.get_stock_price", symbol=symbol)
+            if "search_tools" in self.tools:
+                data["news"] = await self.call_tool("search_tools.get_company_info", ticker=symbol)
+
+        # 4. 生成分析
+        analysis = {
+            "query": query,
+            "symbol": symbol,
+            "type": analysis_type,
+            "data": data,
+            "conclusion": f"基于{query}的分析结论",
+            "valuation": "估值：合理/偏高/偏低",
+            "recommendation": "建议：买入/持有/卖出",
+        }
+
+        return analysis
+
+    def _extract_symbol(self, query: str) -> str:
+        """从查询中提取股票代码"""
+        # 简单的代码提取逻辑
+        import re
+        match = re.search(r'\b\d{6}\b', query)
+        if match:
+            return match.group(0)
+        return ""
+
+    def _classify_query(self, query: str) -> str:
+        """分类查询类型"""
+        q = query.lower()
+        if any(k in q for k in ["估值", "价值", "贵", "便宜"]):
+            return "valuation"
+        elif any(k in q for k in ["业绩", "利润", "收入", "财报"]):
+            return "financial"
+        elif any(k in q for k in ["行业", "竞争", "市场份额"]):
+            return "industry"
+        elif any(k in q for k in ["技术", "走势", "趋势", "k线"]):
+            return "technical"
+        else:
+            return "general"

package/src/agents/erlang.py

@@ -0,0 +1,165 @@
+"""
+二郎神 - 二郎神主智能体
+全知全能的投资智能体，整合所有能力
+"""
+from typing import Optional
+from loguru import logger
+
+from .base_agent import BaseAgent
+from .macro_agent import MacroAgent
+from .equity_agent import EquityAgent
+from .multi_asset_agent import MultiAssetAgent
+
+
+class 二郎神(BaseAgent):
+    """
+    二郎神 - 主智能体
+
+    整合所有分析师能力，提供全品类、全方位的投资分析
+
+    架构：
+    - Brain: LLM大脑
+    - Cerebellum: 认知调度
+    - Agents: 专业分析师团队
+    - Tools: 工具集
+    - Memory: 记忆系统
+    """
+
+    def __init__(
+        self,
+        brain=None,
+        memory=None,
+        knowledge=None,
+        tools: Optional[dict] = None,
+    ):
+        super().__init__(
+            name="二郎神",
+            description="全知全能的投资智能体，具有天眼般全方位洞察能力",
+            tools=tools,
+        )
+
+        self.brain = brain
+        self.memory = memory
+        self.knowledge = knowledge
+
+        # 初始化专业分析师团队
+        self.macro_agent = MacroAgent(tools=tools)
+        self.equity_agent = EquityAgent(tools=tools)
+        self.multi_asset_agent = MultiAssetAgent(tools=tools)
+
+        logger.info("二郎神 (二郎神) fully initialized")
+
+    async def process(self, query: str, context: Optional[dict] = None) -> dict:
+        """
+        处理用户查询
+
+        Args:
+            query: 用户查询
+            context: 上下文
+
+        Returns:
+            dict 包含完整分析结果
+        """
+        logger.info(f"二郎神 processing: {query[:100]}")
+        context = context or {}
+
+        # 1. 确定查询类型
+        query_type = self._classify_query(query)
+
+        # 2. 路由到对应分析师
+        result = {}
+        if query_type == "macro":
+            result = await self.macro_agent.process(query, context)
+        elif query_type == "equity":
+            result = await self.equity_agent.process(query, context)
+        elif query_type == "multi_asset":
+            result = await self.multi_asset_agent.process(query, context)
+        else:
+            # 综合分析
+            result = await self._comprehensive_analysis(query, context)
+
+        # 3. 添加二郎神特色结论
+        result["erlangshen_insight"] = self._generate_insight(query, result)
+
+        # 4. 记录到记忆
+        if self.memory:
+            await self.memory.add_interaction(query, str(result.get("conclusion", "")))
+
+        # 5. 沉淀到知识库
+        if self.knowledge and result.get("conclusion"):
+            await self.knowledge.append_insight(
+                result["conclusion"],
+                tags=["analysis", query_type],
+            )
+
+        return result
+
+    async def _comprehensive_analysis(self, query: str, context: dict) -> dict:
+        """综合分析 - 多维度分析"""
+        logger.info("Running comprehensive analysis")
+
+        # 并行调用多个分析师
+        import asyncio
+        try:
+            results = await asyncio.gather(
+                self.macro_agent.process(query, context),
+                self.equity_agent.process(query, context),
+                self.multi_asset_agent.process(query, context),
+                return_exceptions=True,
+            )
+        except Exception:
+            results = []
+
+        return {
+            "query": query,
+            "type": "comprehensive",
+            "macro_analysis": results[0] if len(results) > 0 else {},
+            "equity_analysis": results[1] if len(results) > 1 else {},
+            "multi_asset_analysis": results[2] if len(results) > 2 else {},
+        }
+
+    def _classify_query(self, query: str) -> str:
+        """分类查询类型"""
+        q = query.lower()
+        if any(k in q for k in ["宏观", "经济", "gdp", "通胀", "利率", "政策"]):
+            return "macro"
+        elif any(k in q for k in ["股票", "个股", "茅台", "估值", "行业"]):
+            return "equity"
+        elif any(k in q for k in ["配置", "资产", "组合", "多资产", "黄金", "债券"]):
+            return "multi_asset"
+        else:
+            return "comprehensive"
+
+    def _generate_insight(self, query: str, result: dict) -> str:
+        """生成二郎神洞察"""
+        return (
+            "【二郎神天眼观察】"
+            f"针对「{query[:30]}...」的分析已完成。"
+            "建议结合宏观周期、资产配置和个人风险偏好综合决策。"
+        )
+
+    async def analyze_market(self, market: str = "A股") -> dict:
+        """分析市场整体状况"""
+        logger.info(f"Analyzing {market}")
+        return await self.process(f"分析当前{market}市场走势和投资机会")
+
+    async def analyze_stock(self, symbol: str) -> dict:
+        """分析个股"""
+        logger.info(f"Analyzing stock {symbol}")
+        return await self.process(f"分析股票{symbol}的投资价值", {"symbol": symbol})
+
+    async def analyze_macro(self, topic: str) -> dict:
+        """分析宏观经济"""
+        logger.info(f"Analyzing macro: {topic}")
+        return await self.process(f"分析{topic}对投资的影响")
+
+    async def generate_report(
+        self,
+        title: str,
+        content: str,
+    ) -> dict:
+        """生成报告并沉淀到知识库"""
+        if self.knowledge:
+            filepath = await self.knowledge.write_report(content, title)
+            return {"status": "success", "path": filepath}
+        return {"status": "skipped", "reason": "knowledge not available"}

package/src/agents/macro.py

@@ -0,0 +1,137 @@
+"""
+宏观分析师 Agent
+"""
+
+from typing import Optional, Dict, Any
+from src.agents.base import BaseAgent
+from src.brain import Brain
+from src.mcp.registry import MCPRegistry
+
+
+class MacroAgent(BaseAgent):
+    """
+    宏观分析师
+    
+    专注于宏观经济分析、政策解读、大类资产配置
+    """
+    
+    def __init__(self, brain: Brain, mcp: MCPRegistry):
+        super().__init__(brain, mcp)
+        self.role = "宏观分析师"
+    
+    def _build_system_prompt(self) -> str:
+        return """你是一位资深宏观经济分析师，专注于：
+- 宏观经济指标分析 (GDP, CPI, PPI, PMI等)
+- 货币政策解读 (利率、存款准备金率、LPR等)
+- 财政政策分析
+- 大类资产配置建议
+- 汇率走势分析
+- 全球经济联动
+
+分析原则：
+1. 数据驱动：基于宏观经济数据进行客观分析
+2. 政策导向：关注央行、财政部政策动向
+3. 全球视野：关注海外主要经济体动态
+4. 风险提示：明确指出潜在风险
+
+输出格式：
+## 核心观点
+(1-2句话概括)
+
+## 宏观环境
+(当前经济周期、货币财政政策)
+
+## 数据解读
+(关键指标数据及含义)
+
+## 资产配置建议
+(股票、债券、商品、外汇等)
+
+## 风险提示
+(主要风险点)
+
+## 近期关注
+(即将发布的重要数据、政策事件)
+"""
+    
+    async def process(self, query: str, **kwargs) -> str:
+        """
+        处理宏观分析查询
+        
+        Args:
+            query: 分析查询
+            **kwargs: 其他参数 (framework, indicators等)
+        """
+        framework = kwargs.get("framework")
+        indicators = kwargs.get("indicators", [])
+        
+        # 收集相关数据
+        data = {}
+        
+        # 获取请求的宏观指标
+        if indicators:
+            for indicator in indicators:
+                result = await self.call_mcp_tool("get_macro_indicator", indicator_code=indicator)
+                if isinstance(result, dict):
+                    data[indicator] = result
+        else:
+            # 默认获取主要指标
+            for indicator in ["GDP", "CPI", "PMI"]:
+                result = await self.call_mcp_tool("get_macro_indicator", indicator_code=indicator)
+                if isinstance(result, dict):
+                    data[indicator] = result
+            
+            # 获取利率数据
+            lpr_result = await self.call_mcp_tool("get_interest_rates", rate_type="LPR")
+            if isinstance(lpr_result, dict):
+                data["LPR"] = lpr_result
+            
+            # 获取汇率数据
+            currency_result = await self.call_mcp_tool("get_currency_rates", base="USD")
+            if isinstance(currency_result, dict):
+                data["USD_CNY"] = currency_result
+        
+        # 执行分析
+        if data:
+            return await self.analyze(query, framework=framework, data=data)
+        else:
+            return await self.analyze(query, framework=framework)
+    
+    async def analyze_policy(self, policy_type: str) -> str:
+        """
+        分析特定政策
+        
+        Args:
+            policy_type: 政策类型 (货币政策/财政政策/产业政策)
+        """
+        prompt = f"请详细分析当前{policy_type}的取向、力度和潜在影响"
+        return await self.think(prompt)
+    
+    async def compare_economies(
+        self,
+        countries: list,
+        indicators: list
+    ) -> str:
+        """
+        比较多国经济
+        
+        Args:
+            countries: 国家列表
+            indicators: 指标列表
+        """
+        data = {}
+        
+        for country in countries:
+            country_data = {}
+            for indicator in indicators:
+                result = await self.call_mcp_tool(
+                    "get_macro_indicator",
+                    indicator_code=indicator,
+                    country=country
+                )
+                if isinstance(result, dict):
+                    country_data[indicator] = result
+            data[country] = country_data
+        
+        prompt = f"请比较多国经济状况：{countries}"
+        return await self.analyze(prompt, framework="跨国比较", data=data)

package/src/agents/macro_agent.py

@@ -0,0 +1,81 @@
+"""
+Macro Agent - 宏观分析师
+专注于宏观经济分析、政策解读、大类资产配置
+"""
+from typing import Optional
+from loguru import logger
+
+from .base_agent import BaseAgent
+
+
+class MacroAgent(BaseAgent):
+    """
+    宏观分析师智能体
+
+    能力：
+    - 宏观经济指标分析
+    - 货币政策解读
+    - 大类资产配置建议
+    - 宏观风险识别
+    """
+
+    def __init__(self, tools: Optional[dict] = None):
+        super().__init__(
+            name="宏观分析师",
+            description="专注于宏观经济分析、政策解读和大类资产配置",
+            tools=tools,
+        )
+
+    async def process(self, query: str, context: Optional[dict] = None) -> dict:
+        """
+        处理宏观分析查询
+
+        Args:
+            query: 分析问题
+            context: 上下文
+
+        Returns:
+            dict 包含分析结论和建议
+        """
+        logger.info(f"MacroAgent processing: {query[:80]}")
+        context = context or {}
+
+        # 1. 解析分析类型
+        analysis_type = self._classify_query(query)
+
+        # 2. 获取相关数据
+        data = {}
+        if "indicator" in analysis_type:
+            # 获取宏观指标
+            indicator = context.get("indicator", "CPI")
+            if "macro_tools" in self.tools:
+                data["macro"] = await self.call_tool("macro_tools.get_macro_indicator", code=indicator)
+            if "search_tools" in self.tools:
+                data["news"] = await self.call_tool("search_tools.news_search", query=f"{indicator} 宏观 经济")
+
+        # 3. 生成分析
+        analysis = {
+            "query": query,
+            "type": analysis_type,
+            "data": data,
+            "conclusion": f"基于{query}的分析结论",
+            "recommendation": "建议配置...",
+        }
+
+        return analysis
+
+    def _classify_query(self, query: str) -> str:
+        """分类查询类型"""
+        q = query.lower()
+        if any(k in q for k in ["gdp", "增长", "经济"]):
+            return "economic_growth"
+        elif any(k in q for k in ["通胀", "cpi", "ppi", "物价"]):
+            return "inflation"
+        elif any(k in q for k in ["利率", "美联储", "央行", "货币政策"]):
+            return "monetary_policy"
+        elif any(k in q for k in ["汇率", "外汇", "人民币", "美元"]):
+            return "fx_rate"
+        elif any(k in q for k in ["配置", "资产"]):
+            return "asset_allocation"
+        else:
+            return "general_macro"