stock-analyzer-skill 1.1.0 → 1.2.0

package/experts/chaogu_yangjia.md

@@ -123,3 +123,53 @@
 ### 调用方式
 
 debate 模式中重点关注：情绪周期阶段——用全市场涨停家数/跌停家数/炸板率/昨涨停今日溢价四条数据定位（主升/震荡/退潮/冰点）。养家的核心判断不在个股，而在"此刻市场情绪处于什么阶段"。
+
+**代码示例：**
+
+```python
+from data import get_quote
+from common import to_float
+
+# 基础数据
+quote = get_quote("sh600989")
+
+# 情绪维度：情绪周期判断（需结合市场整体数据）
+# 示例市场数据（实际需从市场接口获取）
+market_up_count = 65    # 全市场涨停家数
+market_down_count = 15  # 全市场跌停家数
+break_rate = 0.25       # 炸板率
+
+# 判断情绪周期阶段
+if market_down_count > 50 and break_rate > 0.6:
+    # 冰点期：绝望中蕴含机会
+    phase = "冰点"
+    sentiment_score = 100
+elif market_up_count > 80:
+    # 主升期：赚钱效应扩散
+    phase = "主升"
+    sentiment_score = 80
+elif 40 <= market_up_count <= 60:
+    # 震荡期：分化
+    phase = "震荡"
+    sentiment_score = 50
+else:
+    # 退潮期
+    phase = "退潮"
+    sentiment_score = 0
+
+# 技术面维度：市场温度计
+if market_up_count > 80:
+    tech_score = 100
+elif market_up_count >= 40:
+    tech_score = 50
+else:
+    tech_score = 0
+
+# 风险维度
+if market_down_count < 10:
+    risk_score = 100
+elif market_down_count <= 30:
+    risk_score = 30
+else:
+    risk_score = 0
+```

package/experts/decide.md

@@ -1,5 +1,8 @@
 # 专家圆桌决策引擎
 
+> 版本：v1.1.0（2026-06-08）
+> 更新：适配新的 `scripts/data/` 和 `scripts/common/` 模块
+
 > 整合 8 位专家的独立评分，生成加权投票结果。用于 stock skill debate 模式。
 
 ## 一、投票机制
@@ -122,13 +125,62 @@
 
 ## 五、使用流程（stock skill debate 模式）
 
-1. **获取数据**：运行 `scripts/quote.py`、`scripts/finance.py`、`scripts/kline.py` 获取目标股票完整数据
-2. **获取大盘数据**：运行 `scripts/quote.py sh000001,sh510300 -j` 判断市场环境
+1. **获取数据**：使用 `scripts/data/` 获取目标股票完整数据
+2. **获取大盘数据**：使用 `get_quote("sh000001")` 或 `get_quote("sh510300")` 判断市场环境
 3. **专家打分**：每位专家按各自 `§九 评分矩阵` 对目标股票独立打分（使用步骤1的数据）
 4. **应用调整**：根据市场环境（步骤2）调整长线/短线权重
 5. **冲突解决**：按 §三 规则处理分歧，特别注意巴菲特否决权和养家情绪周期
 6. **输出结果**：按 §四 格式输出，含评分表、方向、风险、仓位
 
+**代码示例：**
+
+```python
+from data import get_quote, get_finance, get_kline
+import statistics
+
+# 1. 获取股票数据
+quote = get_quote("sh600989")
+fin = get_finance("sh600989")[0]
+bars = get_kline("sh600989", scale=240, datalen=30)
+
+# 2. 获取大盘数据判断市场环境
+index_quote = get_quote("sh000001")
+prev_close = index_quote.get("prev_close", 0)
+current_price = index_quote.get("price", 0)
+
+# 判断市场状态
+if prev_close > 0:
+    change_pct = (current_price / prev_close - 1) * 100
+    if change_pct > 3:
+        market_state = "牛市"
+        long_weight = 0.40
+        short_weight = 0.60
+    elif change_pct < -3:
+        market_state = "熊市"
+        long_weight = 0.65
+        short_weight = 0.35
+    else:
+        market_state = "震荡"
+        long_weight = 0.55
+        short_weight = 0.45
+
+# 3. 计算专家评分（以巴菲特为例）
+roe = float(fin.get("ROEJQ", 0))
+pe = float(quote.get("pe", 0))
+
+# 巴菲特评分矩阵
+buffett_scores = {
+    "基本面": 100 if roe >= 20 else (75 if roe >= 15 else (40 if roe >= 10 else 0)),
+    "估值": 100 if pe <= 15 else (60 if pe <= 25 else (25 if pe <= 40 else 0)),
+}
+
+# 4. 计算综合分
+# ... (按 decide.md 规则计算)
+
+# 5. 输出结果
+print(f"市场状态: {market_state} | 长线权重: {long_weight} | 短线权重: {short_weight}")
+```
+
 ---
 
 ## 六、信心指数计算

package/experts/duan_yongping.md

@@ -104,3 +104,48 @@
 ### 调用方式
 
 debate 模式中，先判断"生意模式"—商业模式好→所有维度分不减；一般→总分×0.8。估值用 PE 绝对值+PE 历史分位综合判断。A 股需对管理层可信度单独折扣 10-20%。
+
+**代码示例：**
+
+```python
+from data import get_quote, get_finance
+from common import to_float
+
+# 获取数据
+quote = get_quote("sh600989")
+fin = get_finance("sh600989")[0]
+
+# 估值维度：PE
+pe = to_float(quote.get("pe", 0))
+risk_free_rate = 0.03  # 无风险利率
+pe_threshold = 1 / risk_free_rate  # 约 33 倍
+
+if pe < pe_threshold * 0.75:  # < 25 倍
+    pe_score = 100
+elif pe < pe_threshold:  # < 33 倍
+    pe_score = 70
+elif pe < pe_threshold * 1.2:  # < 40 倍
+    pe_score = 40
+else:
+    pe_score = 0
+
+# 安全边际维度：自由现金流
+eps = to_float(fin.get("EPSJB", 0))
+ocf_per_share = to_float(fin.get("MGJYXJJE", 0))
+if eps > 0 and ocf_per_share > eps * 0.8:
+    fcf_score = 100  # FCF 持续为正
+elif eps > 0:
+    fcf_score = 50
+else:
+    fcf_score = 0
+
+# 商业模式判断（简化版：基于行业和ROE）
+industry = infer_industry(quote.get("name", ""), quote.get("code", ""))
+roe = to_float(fin.get("ROEJQ", 0))
+if roe >= 20:
+    business_score = 100  # 优秀商业模式
+elif roe >= 15:
+    business_score = 60
+else:
+    business_score = 20
+```

package/experts/lynch.md

@@ -124,4 +124,51 @@
 
 ### 调用方式
 
-debate 模式中，用 scripts/ 获取 PE 和净利增速计算 PEG，对照上表打分。增速取近 3 年 CAGR，若数据不足则用最新季度年化。
+debate 模式中，用 `scripts/data/` 获取 PE 和净利增速计算 PEG，对照上表打分。增速取近 3 年 CAGR，若数据不足则用最新季度年化。
+
+**代码示例：**
+
+```python
+from data import get_quote, get_finance
+from common import to_float
+
+# 获取数据
+quote = get_quote("sh600989")
+fin = get_finance("sh600989")[0]
+
+# 基本面维度：净利增速
+profit_growth = to_float(fin.get("PARENTNETPROFITTZ", 0))  # 净利润同比增长%
+if profit_growth >= 25:
+    base_score = 100
+elif profit_growth >= 20:
+    base_score = 80
+elif profit_growth >= 15:
+    base_score = 50
+else:
+    base_score = 20
+
+# 估值维度：PEG
+pe = to_float(quote.get("pe", 0))
+growth = to_float(fin.get("PARENTNETPROFITTZ", 0))
+peg = pe / max(growth, 1)  # 避免除零
+
+if peg <= 0.5:
+    peg_score = 100
+elif peg <= 1.0:
+    peg_score = 80
+elif peg <= 1.5:
+    peg_score = 50
+elif peg <= 2.0:
+    peg_score = 30
+else:
+    peg_score = 0
+
+# 风险维度：负债率
+debt_ratio = to_float(fin.get("ZCFZL", 0))
+if debt_ratio < 60:
+    risk_score = 100
+elif debt_ratio < 100:
+    risk_score = 60
+else:
+    risk_score = 0
+```

package/experts/soros.md

@@ -87,3 +87,46 @@ _简化案例_：某银行 PB 0.5 倍，市场担忧不良资产→股价跌至
 ### 调用方式
 
 debate 模式中，重点关注：趋势是否明确（从 kline.py 20日均线判断）、市场情绪是否一致（从涨停家数/炸板率判断）、是否有反身性强化信号。个股基本面只作参考背景。
+
+**代码示例：**
+
+```python
+from data import get_quote, get_kline
+from common import to_float
+import statistics
+
+# 获取K线数据判断趋势
+bars = get_kline("sh600989", scale=240, datalen=30)
+closes = [to_float(b["close"]) for b in bars if to_float(b.get("close", 0)) > 0]
+
+if len(closes) >= 20:
+    ma20 = statistics.mean(closes[-20:])
+    current_price = closes[-1]
+
+    if current_price > ma20:
+        trend_score = 100  # 上升趋势
+    else:
+        trend_score = 0    # 下降趋势
+
+# 技术面维度：趋势强度
+if len(closes) >= 10:
+    recent = closes[-10:]
+    if all(recent[i] < recent[i+1] for i in range(len(recent)-1)):
+        tech_score = 100  # 明确上升趋势
+    elif all(recent[i] > recent[i+1] for i in range(len(recent)-1)):
+        tech_score = 0    # 明确下降趋势
+    else:
+        tech_score = 40   # 震荡
+else:
+    tech_score = 40
+
+# 情绪/反身性维度（需结合市场数据）
+# 实际使用时从市场情绪接口获取涨停家数、炸板率等
+market_up_count = 80  # 示例：涨停家数
+if market_up_count > 80:
+    sentiment_score = 40   # 亢奋期，一致性过高
+elif market_up_count > 40:
+    sentiment_score = 60   # 正常
+else:
+    sentiment_score = 100  # 冰点，反向机会
+```

package/experts/xu_xiang.md

@@ -105,3 +105,47 @@
 ### 调用方式
 
 debate 模式中重点关注：涨停板基因（kline.py 近30日是否有9.5%+涨幅日）、板块联动（同板涨停数）、大盘趋势（所在指数是否站上20日线）。基本面只看EPS排雷，不深度分析。
+
+**代码示例：**
+
+```python
+from data import get_quote, get_kline
+from common import to_float
+import statistics
+
+# 基础数据
+quote = get_quote("sh600989")
+
+# 技术面维度：涨停板基因
+# 获取近30日K线，检查是否有接近涨停的交易日
+bars = get_kline("sh600989", scale=240, datalen=30)
+high_prices = [to_float(b.get("high", 0)) for b in bars]
+prev_close = to_float(quote.get("prev_close", 0))
+
+if prev_close > 0:
+    # 计算每日涨跌幅
+    limit_up_pct = 0.095 if "30" in quote.get("code", "") or "68" in quote.get("code", "") else 0.095
+    limit_up_count = sum(1 for hp in high_prices if hp / prev_close - 1 >= limit_up_pct * 0.9)
+
+    if limit_up_count >= 2:
+        tech_score = 100
+    elif limit_up_count >= 1:
+        tech_score = 50
+    else:
+        tech_score = 0
+
+# 情绪/题材维度：板块联动（需外部市场数据）
+# 示例：同板块涨停数量
+sector_limit_up_count = 3  # 实际从市场数据获取
+
+if sector_limit_up_count >= 3:
+    sentiment_score = 100
+elif sector_limit_up_count >= 2:
+    sentiment_score = 60
+else:
+    sentiment_score = 20
+
+# 基本面维度：仅排雷
+eps = to_float(quote.get("pe"))  # 这里用PE作为占位，实际不深入分析
+base_score = 60 if eps and eps > 0 else 0
+```

package/experts/zhao_laoge.md

@@ -141,3 +141,56 @@
 ### 调用方式
 
 debate 模式中重点关注：kline.py 30日K线判断均线排列形态（MA5/MA10/MA20是否多头）、题材在生命周期中位置（启动→高潮→加速→分歧→退潮）、量价是否健康（放量上涨+缩量回踩=最佳）。流通市值取 quote.py 的 circulating_cap 字段。
+
+**代码示例：**
+
+```python
+from data import get_quote, get_kline
+from common import to_float
+import statistics
+
+# 基础数据
+quote = get_quote("sh600989")
+circulating_cap = to_float(quote.get("circulating_cap", 0))
+
+# 估值维度：流通市值
+if 50 <= circulating_cap <= 300:
+    cap_score = 100
+elif 30 <= circulating_cap <= 500:
+    cap_score = 60
+else:
+    cap_score = 20
+
+# 技术面维度：均线系统
+bars = get_kline("sh600989", scale=240, datalen=30)
+closes = [to_float(b["close"]) for b in bars if to_float(b.get("close", 0)) > 0]
+
+if len(closes) >= 20:
+    ma5 = statistics.mean(closes[-5:])
+    ma10 = statistics.mean(closes[-10:])
+    ma20 = statistics.mean(closes[-20:])
+
+    if ma5 > ma10 > ma20:
+        tech_score = 100  # 多头排列
+    elif ma5 < ma10 < ma20:
+        tech_score = 0    # 空头排列
+    else:
+        tech_score = 70   # 均线粘合
+else:
+    tech_score = 40
+
+# 情绪/题材维度：量价配合
+volumes = [to_float(b.get("volume", 0)) for b in bars]
+if len(closes) >= 10 and len(volumes) >= 10:
+    recent_vol = statistics.mean(volumes[-5:])
+    prev_vol = statistics.mean(volumes[-10:-5])
+
+    if recent_vol > prev_vol * 1.2 and closes[-1] > statistics.mean(closes[-5:-1]):
+        sentiment_score = 100  # 放量上涨
+    elif recent_vol < prev_vol * 0.8:
+        sentiment_score = 60   # 缩量回调（健康）
+    else:
+        sentiment_score = 40
+else:
+    sentiment_score = 50
+```

package/experts/zuoshou_xinyi.md

@@ -142,3 +142,69 @@
 ### 调用方式
 
 debate 模式中重点关注：kline.py 30日K线判断是否强势股（有≥4板记录）、当前是否缩量调整（量能<主升期50%）、关键K线反转信号（锤子线/早晨之星/阳包阴/缩量十字星）。流通市值取 quote.py circulating_cap，止损位取前低或20日线。
+
+**代码示例：**
+
+```python
+from data import get_quote, get_kline
+from common import to_float
+import statistics
+
+# 基础数据
+quote = get_quote("sh600989")
+circulating_cap = to_float(quote.get("circulating_cap", 0))
+
+# 估值维度：流通市值 + 回调深度
+if 30 <= circulating_cap <= 200:
+    cap_score = 100
+elif 10 <= circulating_cap <= 300:
+    cap_score = 50
+else:
+    cap_score = 0
+
+# 技术面维度：K线反转形态 + 缩量程度
+bars = get_kline("sh600989", scale=240, datalen=30)
+closes = [to_float(b["close"]) for b in bars if to_float(b.get("close", 0)) > 0]
+volumes = [to_float(b.get("volume", 0)) for b in bars if to_float(b.get("volume", 0)) > 0]
+
+if len(closes) >= 10 and len(volumes) >= 10:
+    # 计算缩量程度
+    peak_vol = max(volumes[-10:]) if volumes else 1
+    recent_vol = statistics.mean(volumes[-3:]) if len(volumes) >= 3 else 1
+    vol_ratio = recent_vol / peak_vol if peak_vol > 0 else 1
+
+    # K线反转形态识别（简化版）
+    last3 = closes[-3:]
+    if len(closes) >= 4:
+        prev3 = closes[-4:-1]
+    else:
+        prev3 = closes[:-1] if len(closes) > 1 else [closes[0]] * 3
+
+    # 锤子线/早晨之星形态
+    is_hammer = last3[-1] < last3[-2] and closes[-1] < statistics.mean(last3)
+    is_engulfing = last3[-1] > last3[-2] and last3[-2] < last3[-3] and last3[-1] > last3[-3]
+
+    if vol_ratio <= 0.5 and (is_hammer or is_engulfing):
+        tech_score = 100
+    elif vol_ratio <= 0.5:
+        tech_score = 60
+    else:
+        tech_score = 20
+else:
+    tech_score = 40
+
+# 情绪维度：调整阶段
+if len(closes) >= 10:
+    # 计算调整天数（从最高点回调）
+    peak_idx = closes.index(max(closes))
+    adjust_days = len(closes) - peak_idx - 1
+
+    if 3 <= adjust_days <= 7:
+        sentiment_score = 100  # ���佳调整期
+    elif adjust_days < 3:
+        sentiment_score = 40   # 调整不足
+    else:
+        sentiment_score = 30   # 调整过长
+else:
+    sentiment_score = 50
+```