aria-code 4.1.3__py3-none-any.whl

agents/financial/macro.py

@@ -0,0 +1,99 @@
+"""
+agents/financial/macro.py — 宏观环境 Agent
+===========================================
+分析：利率环境、汇率、行业周期、政策面、大盘趋势。
+"""
+from __future__ import annotations
+from typing import Any, Dict, List
+from ..base import BaseAgent, AgentResult
+
+
+class MacroAgent(BaseAgent):
+    name        = "macro"
+    description = "宏观分析：利率/汇率/政策/大盘趋势"
+
+    _SYSTEM = (
+        "You are a macro strategist covering China and global markets. "
+        "Focus on: interest rate environment, USD/CNY trend, sector cycle, "
+        "regulatory policy, and market sentiment. Be concise and data-driven. "
+        "End with: TAILWIND / NEUTRAL / HEADWIND for the given stock."
+    )
+
+    async def fetch_data(self, symbol: str) -> Dict[str, Any]:
+        data = await super().fetch_data(symbol)
+        # 尝试拉取指数数据作为大盘参考
+        if self.data:
+            for idx in ["000001", "399006"]:  # 上证 + 创业板
+                try:
+                    q = self.data.quote(idx)
+                    if q:
+                        data.setdefault("indices", {})[idx] = q.to_dict()
+                except Exception:
+                    pass
+        return data
+
+    async def analyze(self, symbol: str, data: Dict[str, Any]) -> AgentResult:
+        quote   = data.get("quote", {})
+        indices = data.get("indices", {})
+
+        idx_summary = ""
+        for code, q in indices.items():
+            name = {"000001": "上证", "399006": "创业板"}.get(code, code)
+            idx_summary += f"  {name}: {q.get('price',0):.2f} {q.get('change_pct',0):+.2f}%\n"
+
+        prompt = (
+            f"Symbol: {symbol}\n"
+            f"Market Indices:\n{idx_summary or '  (unavailable)'}\n\n"
+            "Analyze the macroeconomic backdrop for this stock:\n"
+            "1. Current rate/liquidity environment (PBOC stance)\n"
+            "2. Sector regulatory environment (any recent policies)\n"
+            "3. Market sentiment and positioning\n"
+            "4. Key macro risks to watch\n"
+            "Conclusion: TAILWIND / NEUTRAL / HEADWIND"
+        )
+
+        analysis = await self._call_llm(self._SYSTEM, prompt, max_tokens=500, quote=quote)
+        if not analysis:
+            analysis = _template_macro(symbol, indices)
+
+        signal     = _extract_signal(analysis)
+        key_points = _extract_key_points(analysis)
+
+        return AgentResult(
+            agent=self.name, symbol=symbol,
+            analysis=analysis, confidence=0.65,
+            signal=signal, key_points=key_points,
+            data_used={"indices": indices},
+        )
+
+
+def _extract_signal(text: str) -> str:
+    t = text.upper()
+    if "TAILWIND" in t or "看多" in t:   return "BUY"
+    if "HEADWIND" in t or "看空" in t:   return "SELL"
+    return "HOLD"
+
+
+def _extract_key_points(text: str) -> List[str]:
+    points = []
+    for line in text.split("\n"):
+        line = line.strip()
+        if line.startswith(("1.", "2.", "3.", "4.", "•", "-", "·")) and len(line) > 5:
+            points.append(line.lstrip("1234567890.-•· "))
+    return points[:4]
+
+
+_IDX_NAMES = {"000001": "上证", "399006": "创业板"}
+
+def _template_macro(symbol: str, indices: Dict) -> str:
+    idx_str = "、".join(
+        f"{_IDX_NAMES.get(k, k)} {v.get('change_pct', 0):+.2f}%"
+        for k, v in indices.items()
+    ) if indices else "指数数据不可用"
+    return (
+        f"{symbol} 宏观分析（模板）:\n"
+        f"• 市场指数: {idx_str}\n"
+        "• 当前货币政策偏宽松，流动性充裕\n"
+        "• 建议关注政策面变化和外资动向\n"
+        "• 结论: NEUTRAL"
+    )

agents/financial/news.py

@@ -0,0 +1,207 @@
+"""
+agents/financial/news.py — 新闻与舆情分析 Agent
+================================================
+分析近期新闻标题、公告和媒体情绪，识别关键事件类型。
+数据源：yfinance.news（美股/港股）、akshare公告（A股，可选）
+"""
+
+from __future__ import annotations
+
+import logging
+import re
+from datetime import datetime, timezone
+from typing import Any, Dict, List
+
+from ..base import BaseAgent, AgentResult
+
+logger = logging.getLogger(__name__)
+
+_EVENT_KEYWORDS = {
+    "earnings":    ["earnings", "revenue", "profit", "EPS", "业绩", "净利润", "营收"],
+    "upgrade":     ["upgrade", "outperform", "buy", "overweight", "上调", "买入"],
+    "downgrade":   ["downgrade", "underperform", "sell", "underweight", "下调", "卖出"],
+    "insider":     ["insider", "director", "CEO", "增持", "减持", "大股东"],
+    "regulatory":  ["SEC", "CSRC", "证监会", "regulatory", "investigation", "调查"],
+    "dividend":    ["dividend", "分红", "派息", "股息"],
+    "merger":      ["merger", "acquisition", "takeover", "合并", "收购"],
+}
+
+
+class NewsAgent(BaseAgent):
+
+    name        = "news"
+    description = "新闻与舆情分析 — 近期重大事件、公告、媒体情绪"
+
+    _SYSTEM = (
+        "You are a financial news analyst. Analyze the provided news headlines "
+        "and summaries for a stock. Identify the most impactful recent events, "
+        "classify sentiment (POSITIVE / NEUTRAL / NEGATIVE), and assess likely "
+        "price impact. Focus on material events: earnings, analyst changes, "
+        "regulatory actions, M&A, and insider activity. "
+        "Conclude with: POSITIVE / NEUTRAL / NEGATIVE"
+    )
+
+    async def fetch_data(self, symbol: str) -> Dict[str, Any]:
+        data = await super().fetch_data(symbol)
+        news_items: List[Dict] = []
+
+        # 1. yfinance news (works for US / HK stocks)
+        try:
+            import yfinance as yf
+            ticker = yf.Ticker(symbol)
+            raw = ticker.news or []
+            now = datetime.now(timezone.utc).timestamp()
+            for item in raw[:12]:
+                pub = item.get("providerPublishTime", 0)
+                age_days = (now - pub) / 86400 if pub else 999
+                news_items.append({
+                    "title":    item.get("title", ""),
+                    "summary":  item.get("summary", "") or item.get("description", ""),
+                    "source":   item.get("publisher", ""),
+                    "age_days": round(age_days, 1),
+                })
+        except Exception as e:
+            logger.debug("[news] yfinance fetch failed for %s: %s", symbol, e)
+
+        # 2. akshare A-share announcements (optional)
+        if not news_items and re.match(r"^[036]\d{5}$", symbol):
+            try:
+                import akshare as ak
+                df = ak.stock_news_em(symbol=symbol)
+                if df is not None and not df.empty:
+                    for _, row in df.head(10).iterrows():
+                        news_items.append({
+                            "title":    str(row.get("新闻标题", "")),
+                            "summary":  str(row.get("新闻内容", ""))[:200],
+                            "source":   str(row.get("新闻来源", "")),
+                            "age_days": 0,
+                        })
+            except Exception as e:
+                logger.debug("[news] akshare fetch failed for %s: %s", symbol, e)
+
+        data["news"] = news_items
+        return data
+
+    async def analyze(self, symbol: str, data: Dict[str, Any]) -> AgentResult:
+        news  = data.get("news", [])
+        quote = data.get("quote", {})
+        price = quote.get("price", 0)
+
+        if not news:
+            return AgentResult(
+                agent=self.name, symbol=symbol,
+                analysis=f"{symbol}: 未获取到近期新闻数据。",
+                confidence=0.3, signal="HOLD",
+                key_points=["无近期新闻数据"],
+            )
+
+        events = _classify_events(news)
+        news_block = _format_news(news[:8])
+
+        prompt = (
+            f"Stock: {symbol}  Price: {price}\n\n"
+            f"Recent News ({len(news)} items):\n{news_block}\n\n"
+            "Tasks:\n"
+            "1. Identify the 2-3 most impactful recent events\n"
+            "2. Classify overall media sentiment\n"
+            "3. Assess potential price catalyst (short-term, 1-5 days)\n"
+            "4. Conclude with: POSITIVE / NEUTRAL / NEGATIVE"
+        )
+
+        analysis = await self._call_llm(self._SYSTEM, prompt, max_tokens=500)
+        if not analysis:
+            analysis = _template_analysis(symbol, news, events)
+
+        signal     = _extract_signal(analysis, events)
+        confidence = _estimate_confidence(news, events)
+        key_points = _build_key_points(news, events)
+
+        return AgentResult(
+            agent=self.name, symbol=symbol,
+            analysis=analysis,
+            confidence=confidence,
+            signal=signal,
+            key_points=key_points,
+            data_used={"news_count": len(news), "events": list(events.keys())},
+        )
+
+
+# ── Helpers ───────────────────────────────────────────────────────────────────
+
+def _classify_events(news: List[Dict]) -> Dict[str, int]:
+    counts: Dict[str, int] = {}
+    for item in news:
+        text = (item.get("title", "") + " " + item.get("summary", "")).lower()
+        for event_type, keywords in _EVENT_KEYWORDS.items():
+            if any(kw.lower() in text for kw in keywords):
+                counts[event_type] = counts.get(event_type, 0) + 1
+    return counts
+
+
+def _format_news(news: List[Dict]) -> str:
+    lines = []
+    for i, item in enumerate(news, 1):
+        age = f"{item['age_days']:.0f}d ago" if item.get("age_days", 999) < 30 else ""
+        src = item.get("source", "")
+        title = item.get("title", "").strip()
+        lines.append(f"{i}. [{src}] {title} {age}".strip())
+    return "\n".join(lines)
+
+
+def _extract_signal(analysis: str, events: Dict[str, int]) -> str:
+    text = analysis.upper()
+    if "POSITIVE" in text:
+        if events.get("upgrade", 0) > 0 or "STRONG" in text:
+            return "BUY"
+        return "BUY"
+    if "NEGATIVE" in text:
+        if events.get("downgrade", 0) > 0 or events.get("regulatory", 0) > 0:
+            return "SELL"
+        return "SELL"
+    return "HOLD"
+
+
+def _estimate_confidence(news: List[Dict], events: Dict[str, int]) -> float:
+    base = 0.45
+    recent = sum(1 for n in news if n.get("age_days", 999) <= 3)
+    base += min(recent * 0.05, 0.15)
+    if events.get("earnings") or events.get("upgrade") or events.get("downgrade"):
+        base += 0.1
+    return min(round(base, 2), 0.75)
+
+
+def _build_key_points(news: List[Dict], events: Dict[str, int]) -> List[str]:
+    points = []
+    recent = [n for n in news if n.get("age_days", 999) <= 3]
+    if recent:
+        points.append(f"近3日 {len(recent)} 条新鲜新闻")
+    for etype, count in events.items():
+        label = {
+            "earnings": "业绩/财报相关",
+            "upgrade": "分析师升级评级",
+            "downgrade": "分析师降级评级",
+            "insider": "内部人士交易",
+            "regulatory": "监管/调查事件",
+            "dividend": "股息/分红公告",
+            "merger": "并购/重组消息",
+        }.get(etype, etype)
+        points.append(f"{label} × {count}")
+    return points[:5]
+
+
+def _template_analysis(symbol: str, news: List[Dict], events: Dict[str, int]) -> str:
+    recent = [n for n in news if n.get("age_days", 999) <= 7]
+    sentiment = "中性"
+    if events.get("upgrade", 0) > events.get("downgrade", 0):
+        sentiment = "偏正面"
+    elif events.get("downgrade", 0) > 0 or events.get("regulatory", 0) > 0:
+        sentiment = "偏负面"
+
+    titles = "\n".join(f"  • {n['title'][:60]}" for n in recent[:3])
+    return (
+        f"{symbol} 近期新闻情绪：{sentiment}\n"
+        f"近7日 {len(recent)} 条相关新闻\n"
+        f"主要标题：\n{titles or '  （无最新标题）'}\n"
+        f"事件分类：{', '.join(events.keys()) or '无特定事件'}\n"
+        f"结论：{'POSITIVE' if sentiment == '偏正面' else ('NEGATIVE' if sentiment == '偏负面' else 'NEUTRAL')}"
+    )

agents/financial/risk.py

@@ -0,0 +1,132 @@
+"""
+agents/financial/risk.py — 风险评估 Agent
+==========================================
+分析：波动率、最大回撤、与大盘相关性、仓位建议。
+"""
+from __future__ import annotations
+from typing import Any, Dict, List
+from ..base import BaseAgent, AgentResult
+
+try:
+    import numpy as np
+    _HAS_NP = True
+except ImportError:
+    _HAS_NP = False
+    np = None  # type: ignore[assignment]
+
+
+class RiskAgent(BaseAgent):
+    name        = "risk"
+    description = "风险评估：波动率/回撤/相关性/仓位建议"
+
+    _SYSTEM = (
+        "You are a risk manager. Assess: historical volatility, max drawdown, "
+        "liquidity risk, correlation to market, and position sizing recommendation. "
+        "Output: risk score 1-10 (1=very low, 10=very high) and "
+        "POSITION: <percentage> (e.g. POSITION: 10%)."
+    )
+
+    async def fetch_data(self, symbol: str) -> Dict[str, Any]:
+        data = await super().fetch_data(symbol)
+        if self.data:
+            h = self.data.history(symbol, days=252)  # 1年
+            if h and h.data is not None:
+                data["risk_metrics"] = _compute_risk(h.data)
+        return data
+
+    async def analyze(self, symbol: str, data: Dict[str, Any]) -> AgentResult:
+        quote   = data.get("quote", {})
+        metrics = data.get("risk_metrics", {})
+        price   = quote.get("price", 0)
+
+        risk_str = (
+            f"  Ann. Volatility: {metrics.get('ann_vol',0):.1f}%\n"
+            f"  Max Drawdown:    {metrics.get('max_dd',0):.1f}%\n"
+            f"  Sharpe (1Y):     {metrics.get('sharpe',0):.2f}\n"
+            f"  Beta est.:       {metrics.get('beta',1.0):.2f}\n"
+            f"  Avg Daily Vol:   {metrics.get('avg_vol_usd',0):,.0f}"
+        ) if metrics else "  (risk metrics unavailable)"
+
+        prompt = (
+            f"Stock: {symbol}  Price: {price}\n"
+            f"Risk Metrics:\n{risk_str}\n\n"
+            "Assess:\n"
+            "1. Volatility profile (high/medium/low)\n"
+            "2. Drawdown history and recovery\n"
+            "3. Liquidity adequacy\n"
+            "4. Tail risk and black swan exposure\n"
+            "5. Recommended position size for a diversified portfolio\n"
+            "Output format: Risk Score: X/10\nPOSITION: Y%"
+        )
+
+        analysis  = await self._call_llm(self._SYSTEM, prompt, max_tokens=400, quote=quote)
+        if not analysis:
+            analysis = _template_risk(symbol, metrics)
+
+        signal     = _risk_to_signal(metrics)
+        confidence = 0.70
+        key_points = [
+            f"年化波动率 {metrics.get('ann_vol',0):.1f}%",
+            f"最大回撤 {metrics.get('max_dd',0):.1f}%",
+            f"夏普比率 {metrics.get('sharpe',0):.2f}",
+        ] if metrics else ["风险数据不可用"]
+
+        return AgentResult(
+            agent=self.name, symbol=symbol,
+            analysis=analysis, confidence=confidence,
+            signal=signal, key_points=key_points,
+            data_used=metrics,
+        )
+
+
+def _compute_risk(df) -> Dict:
+    if not _HAS_NP:
+        return {}
+    try:
+        close = df["close"].values if "close" in df.columns else df.iloc[:, -1].values
+        vol   = df["volume"].values if "volume" in df.columns else np.ones(len(close))
+        rets  = np.diff(close) / close[:-1]
+
+        ann_vol   = float(np.std(rets) * np.sqrt(252) * 100)
+        ann_ret   = float(np.mean(rets) * 252 * 100)
+        sharpe    = ann_ret / max(ann_vol, 0.01)
+        cum       = np.cumprod(1 + rets)
+        peaks     = np.maximum.accumulate(cum)
+        drawdowns = (cum - peaks) / peaks
+        max_dd    = float(drawdowns.min() * 100)
+        avg_vol   = float(np.mean(vol[-20:]) * close[-1]) if len(close) > 20 else 0
+
+        return {
+            "ann_vol":    round(ann_vol, 1),
+            "ann_ret":    round(ann_ret, 1),
+            "sharpe":     round(sharpe, 2),
+            "max_dd":     round(max_dd, 1),
+            "beta":       1.0,      # 需要指数数据对比
+            "avg_vol_usd": round(avg_vol, 0),
+        }
+    except Exception:
+        return {}
+
+
+def _risk_to_signal(metrics: Dict) -> str:
+    vol = metrics.get("ann_vol", 30)
+    dd  = metrics.get("max_dd", -20)
+    if vol < 20 and dd > -15:
+        return "BUY"
+    if vol > 50 or dd < -40:
+        return "REDUCE"
+    return "HOLD"
+
+
+def _template_risk(symbol: str, m: Dict) -> str:
+    vol = m.get("ann_vol", 0)
+    dd  = m.get("max_dd",  0)
+    risk_level = "高" if vol > 40 else ("中" if vol > 20 else "低")
+    return (
+        f"{symbol} 风险分析（模板）:\n"
+        f"• 年化波动率: {vol:.1f}%（风险水平: {risk_level}）\n"
+        f"• 最大回撤: {dd:.1f}%\n"
+        f"• 夏普比率: {m.get('sharpe',0):.2f}\n"
+        f"• 建议仓位: {'5-10%' if vol>40 else '10-15%' if vol>20 else '15-20%'}\n"
+        f"• Risk Score: {min(10, int(vol/5))}/10"
+    )