Hermes CI Fix
6b6043536a
升级1: 宏观敏感性弹性历史校准 - 内置宏观历史数据(oil/usd/cpi 2026-01~07月度) - 变化率弹性: 同period匹配KPI历史vs因素历史算弹性 - 合理性校验: |弹性|超出[0.01,0.5]视为噪声回退规则(诚实标注) 升级2: 预测偏差告警闭环 - 新表 kpi_forecast_log(预测历史)+模型KpiForecastLog - 预测时落库(同KPI同预测期覆盖) - alert_rules 支持 rule_type=forecast_deviation(threshold_pct) - POST /alert-rules/run-forecast-deviation: 预测vs实际偏差>阈值生成预警(去重, 超2倍阈值红色) - 端到端验证: 模拟实际500vs预测399.55→偏差20.1%>5%→红色预警生成 回归: pytest 40 passed(predict+alerts)
530 lines
21 KiB
Python
530 lines
21 KiB
Python
"""KPI预测引擎 — 基于历史KPI值做趋势预测（预测性成本智能 MVP）
|
||
|
||
模型（MVP原则：简单可用，不上深度学习）：
|
||
- linear 线性回归（最小二乘 y = a + b·x），输出95%预测区间
|
||
- moving_average 简单移动平均（默认窗口3期），输出均值±波动区间
|
||
|
||
置信度诚实标注：基于历史数据量 + 拟合优度（R² / 波动率CV）综合打分，
|
||
数据不足时明确给出 low，不做虚假高置信。
|
||
|
||
复用 cash_forecast_engine.get_entity_kpi_history 取历史数据（不重复写查询）。
|
||
"""
|
||
import logging
|
||
import math
|
||
from typing import Optional
|
||
from sqlalchemy.orm import Session
|
||
|
||
from app.utils.cash_forecast_engine import get_entity_kpi_history, find_kpi
|
||
|
||
logger = logging.getLogger("cma.kpi_forecast")
|
||
|
||
MODELS = ("linear", "moving_average")
|
||
DEFAULT_PERIODS = 3
|
||
TREND_THRESHOLD_PCT = 3.0 # |趋势百分比| ≥ 3% 判定为有明确趋势方向
|
||
|
||
# 置信度档位
|
||
CONF_LEVELS = {3: "high", 2: "medium", 1: "low"}
|
||
# 中文映射（供 summary 使用）
|
||
TREND_CN = {"up": "上升", "down": "下降", "flat": "基本平稳"}
|
||
CONF_CN = {"high": "高", "medium": "中", "low": "低"}
|
||
|
||
|
||
def next_period(period: str, steps: int = 1) -> str:
|
||
"""期数递增："2026-05" + 1 → "2026-06"；解析失败时退化为 period+N"""
|
||
try:
|
||
y, m = str(period).split("-")
|
||
total = int(y) * 12 + (int(m) - 1) + steps
|
||
return f"{total // 12:04d}-{total % 12 + 1:02d}"
|
||
except Exception:
|
||
return f"{period}+{steps}"
|
||
|
||
|
||
def _t_crit(n: int) -> float:
|
||
"""95%双尾学生t临界值近似（小样本查表取保守值，大样本趋近1.96）"""
|
||
table = {
|
||
2: 12.71, 3: 4.30, 4: 3.18, 5: 2.78, 6: 2.57, 7: 2.45,
|
||
8: 2.31, 9: 2.26, 10: 2.23, 12: 2.18, 15: 2.13,
|
||
20: 2.09, 30: 2.04, 60: 2.00,
|
||
}
|
||
for k in sorted(table):
|
||
if n <= k:
|
||
return table[k]
|
||
return 1.96
|
||
|
||
|
||
def _std(values: list) -> float:
|
||
"""样本标准差（n>=2），n==1 返回0"""
|
||
n = len(values)
|
||
if n < 2:
|
||
return 0.0
|
||
mean = sum(values) / n
|
||
return math.sqrt(sum((v - mean) ** 2 for v in values) / (n - 1))
|
||
|
||
|
||
def _rel_trend_pct(values: list) -> float:
|
||
"""趋势百分比 = 线性回归斜率 / |均值| × 100（与 cash_forecast_engine.calc_trend 同口径）"""
|
||
n = len(values)
|
||
if n < 2:
|
||
return 0.0
|
||
xbar = (n - 1) / 2.0
|
||
ybar = sum(values) / n
|
||
sxx = sum((i - xbar) ** 2 for i in range(n))
|
||
if sxx == 0:
|
||
return 0.0
|
||
slope = sum((i - xbar) * (values[i] - ybar) for i in range(n)) / sxx
|
||
return slope / max(abs(ybar), 1.0) * 100
|
||
|
||
|
||
def judge_trend(trend_pct: float, threshold: float = TREND_THRESHOLD_PCT) -> str:
|
||
"""趋势方向判定：up / down / flat"""
|
||
if trend_pct > threshold:
|
||
return "up"
|
||
if trend_pct < -threshold:
|
||
return "down"
|
||
return "flat"
|
||
|
||
|
||
def _compute_r2(values: list, pred_fn) -> float:
|
||
"""拟合优度 R²（0~1），数据无波动时视为完全拟合"""
|
||
ybar = sum(values) / len(values)
|
||
ss_tot = sum((v - ybar) ** 2 for v in values)
|
||
if ss_tot == 0:
|
||
return 1.0
|
||
ss_res = sum((v - pred_fn(i)) ** 2 for i, v in enumerate(values))
|
||
return max(0.0, 1.0 - ss_res / ss_tot)
|
||
|
||
|
||
def compute_confidence(n: int, model: str, r2: Optional[float] = None,
|
||
cv: Optional[float] = None) -> str:
|
||
"""置信度诚实标注：数据量基数 + 拟合优度修正
|
||
|
||
- 数据量：n>=12 → 3分；n>=6 → 2分；否则 1分
|
||
- linear：R²>=0.7 +1；R²<0.3 -1
|
||
- moving_average：CV<0.3 +1（低波动更可信）；CV>0.6 -1
|
||
"""
|
||
score = 3 if n >= 12 else (2 if n >= 6 else 1)
|
||
# 拟合度修正仅在样本量足够时生效：
|
||
# n<4 时 R² 无统计意义（2点直线必然R²=1.0），CV 也噪声大，不做上调，避免虚假高置信
|
||
if n >= 4:
|
||
if model == "linear" and r2 is not None:
|
||
if r2 >= 0.7:
|
||
score += 1
|
||
elif r2 < 0.3:
|
||
score -= 1
|
||
elif model == "moving_average" and cv is not None:
|
||
if cv < 0.3:
|
||
score += 1
|
||
elif cv > 0.6:
|
||
score -= 1
|
||
score = max(1, min(3, score))
|
||
return CONF_LEVELS[score]
|
||
|
||
|
||
def linear_forecast(values: list, periods: int = 3) -> dict:
|
||
"""线性回归预测 — 返回未来periods期预测值 + 95%预测区间 + 拟合统计量"""
|
||
n = len(values)
|
||
x = list(range(n))
|
||
xbar = (n - 1) / 2.0
|
||
ybar = sum(values) / n
|
||
sxx = sum((i - xbar) ** 2 for i in x)
|
||
slope = sum((i - xbar) * (values[i] - ybar) for i in x) / sxx if sxx else 0.0
|
||
intercept = ybar - slope * xbar
|
||
|
||
def pred(i: int) -> float:
|
||
return intercept + slope * i
|
||
|
||
# 残差标准误（n>=3 用 n-2 自由度；n==2 用样本标准差近似）
|
||
if n >= 3:
|
||
resid = [values[i] - pred(i) for i in x]
|
||
se = math.sqrt(sum(r * r for r in resid) / (n - 2))
|
||
else:
|
||
se = _std(values)
|
||
if se == 0:
|
||
se = max(abs(ybar) * 0.05, 1e-9) # 完全拟合时给最小带，避免零宽区间
|
||
|
||
t_crit = _t_crit(n)
|
||
forecast = []
|
||
for k in range(periods):
|
||
x0 = n + k
|
||
predicted = pred(x0)
|
||
se_pred = se * math.sqrt(1.0 + 1.0 / n + (x0 - xbar) ** 2 / max(sxx, 1e-9)) * t_crit
|
||
band = max(se_pred, abs(predicted) * 0.02)
|
||
forecast.append({
|
||
"predicted": round(predicted, 2),
|
||
"lower": round(predicted - band, 2),
|
||
"upper": round(predicted + band, 2),
|
||
})
|
||
|
||
r2 = _compute_r2(values, pred)
|
||
trend_pct = slope / max(abs(ybar), 1.0) * 100
|
||
return {
|
||
"forecast": forecast,
|
||
"slope": slope,
|
||
"intercept": intercept,
|
||
"r2": round(r2, 3),
|
||
"trend_pct": round(trend_pct, 2),
|
||
"se": round(se, 4),
|
||
}
|
||
|
||
|
||
def moving_average_forecast(values: list, periods: int = 3, window: int = 3) -> dict:
|
||
"""简单移动平均预测 — 未来各期预测值 = 最近window期均值；区间=均值±1.96×波动"""
|
||
n = len(values)
|
||
w = max(1, min(window, n))
|
||
base = sum(values[-w:]) / w
|
||
std = _std(values)
|
||
if std == 0:
|
||
std = max(abs(base) * 0.05, 1e-9)
|
||
band = max(1.96 * std, abs(base) * 0.02)
|
||
|
||
forecast = [{
|
||
"predicted": round(base, 2),
|
||
"lower": round(base - band, 2),
|
||
"upper": round(base + band, 2),
|
||
} for _ in range(periods)]
|
||
|
||
cv = std / abs(base) if base else 0.0
|
||
trend_pct = _rel_trend_pct(values)
|
||
return {
|
||
"forecast": forecast,
|
||
"window": w,
|
||
"mean": round(base, 2),
|
||
"std": round(std, 4),
|
||
"cv": round(cv, 3),
|
||
"trend_pct": round(trend_pct, 2),
|
||
}
|
||
|
||
|
||
def build_summary(kpi_name: str, unit: str, trend: str, next_target: Optional[float],
|
||
periods: int, n_history: int, confidence: str, model: str) -> str:
|
||
"""中文一句话解读"""
|
||
trend_cn = TREND_CN.get(trend, trend)
|
||
conf_cn = CONF_CN.get(confidence, confidence)
|
||
unit_txt = unit or ""
|
||
if periods <= 0:
|
||
return f"基于{n_history}期历史数据，{kpi_name}当前趋势{trend_cn}（模型:{model}，置信度:{conf_cn}），未请求未来期数预测"
|
||
target_txt = f"{next_target:,.2f}{unit_txt}" if next_target is not None else "—"
|
||
return (
|
||
f"基于{n_history}期历史数据，{kpi_name}未来{periods}期预计{trend_cn}，"
|
||
f"下一期预测值约{target_txt}（模型:{model}，置信度:{conf_cn}）"
|
||
)
|
||
|
||
|
||
def forecast_kpi(entity_id: int, kpi_code: str, db: Session,
|
||
periods: int = DEFAULT_PERIODS, model: str = "linear") -> Optional[dict]:
|
||
"""单个KPI预测（多租户隔离：历史数据通过 entity_id 维度查询）
|
||
|
||
返回 None 表示 KPI 不存在或历史数据不足（<2条）。
|
||
"""
|
||
if model not in MODELS:
|
||
model = "linear"
|
||
|
||
history = get_entity_kpi_history(entity_id, kpi_code, db, limit_months=120)
|
||
if not history:
|
||
return None
|
||
hist_asc = list(reversed(history)) # 按 period 升序
|
||
values = [float(v.actual_value) for v in hist_asc if v.actual_value is not None]
|
||
if len(values) < 2:
|
||
return None
|
||
|
||
kpi_def = find_kpi(db, entity_id, [kpi_code])
|
||
kpi_name = str(kpi_def.kpi_name) if kpi_def else kpi_code
|
||
unit = str(kpi_def.unit or "") if kpi_def else ""
|
||
|
||
if model == "moving_average":
|
||
res = moving_average_forecast(values, periods)
|
||
confidence = compute_confidence(len(values), model, cv=res["cv"])
|
||
else:
|
||
res = linear_forecast(values, periods)
|
||
confidence = compute_confidence(len(values), model, r2=res["r2"])
|
||
|
||
trend = judge_trend(res["trend_pct"])
|
||
|
||
# 未来期数（基于最近一期 period 递增）
|
||
last_period = hist_asc[-1].period
|
||
forecast = []
|
||
for k in range(periods):
|
||
fp = res["forecast"][k]
|
||
forecast.append({
|
||
"period": next_period(last_period, k + 1),
|
||
"predicted": fp["predicted"],
|
||
"lower": fp["lower"],
|
||
"upper": fp["upper"],
|
||
})
|
||
|
||
next_target = forecast[0]["predicted"] if forecast else None
|
||
summary = build_summary(kpi_name, unit, trend, next_target, periods,
|
||
len(values), confidence, model)
|
||
|
||
return {
|
||
"entity_id": entity_id,
|
||
"kpi": {"code": kpi_code, "name": kpi_name, "unit": unit},
|
||
"model": model,
|
||
"periods": periods,
|
||
"trend": trend,
|
||
"trend_pct": res["trend_pct"],
|
||
"confidence": confidence,
|
||
"history_count": len(values),
|
||
"history": [{"period": v.period, "value": round(float(v.actual_value), 2)} for v in hist_asc],
|
||
"forecast": forecast,
|
||
"next_target": next_target,
|
||
"summary": summary,
|
||
}
|
||
|
||
|
||
def forecast_finance_kpis(entity_id: int, db: Session,
|
||
periods: int = DEFAULT_PERIODS, model: str = "linear",
|
||
min_history: int = 3) -> list:
|
||
"""批量预测该企业全部财务维度KPI（历史≥min_history条），按可预测性排序"""
|
||
from app.models import KPIDefinition
|
||
kpis = db.query(KPIDefinition).filter(
|
||
KPIDefinition.entity_id == entity_id,
|
||
KPIDefinition.dimension == "finance",
|
||
KPIDefinition.status == "active",
|
||
).all()
|
||
|
||
results = []
|
||
for kpi in kpis:
|
||
r = forecast_kpi(entity_id, str(kpi.kpi_code), db, periods=periods, model=model)
|
||
if r and r["history_count"] >= min_history:
|
||
results.append(r)
|
||
|
||
# 可预测性排序：置信度(high=3/medium=2/low=1) 优先，其次历史数据量
|
||
score = {"high": 3, "medium": 2, "low": 1}
|
||
results.sort(key=lambda r: (score.get(r["confidence"], 0), r["history_count"]), reverse=True)
|
||
return results
|
||
|
||
|
||
def save_forecast_logs(entity_id: int, results: list, db: Session, model: str = "linear") -> int:
|
||
"""预测结果落库 kpi_forecast_log（预测偏差告警数据源, 2026-08-25 升级2a）
|
||
存每KPI的下一期预测；同KPI同预测期覆盖（保留最新）"""
|
||
from app.models import KpiForecastLog, KPIDefinition
|
||
# 预加载 KPI id 映射（返回结果里的 kpi 无 id 字段，需从DB查）
|
||
kpi_map = {k.kpi_code: k.id for k in db.query(KPIDefinition).filter(
|
||
KPIDefinition.entity_id == entity_id).all()}
|
||
saved = 0
|
||
for r in results:
|
||
kpi_info = r.get("kpi", {})
|
||
fc = r.get("forecast") or []
|
||
if not fc:
|
||
continue
|
||
first = fc[0]
|
||
period = first.get("period", "")
|
||
val = first.get("predicted") or first.get("value") or first.get("next_value")
|
||
if not period or val is None:
|
||
continue
|
||
kpi_code = kpi_info.get("code", "")
|
||
kpi_id = kpi_info.get("id") or kpi_map.get(kpi_code)
|
||
if not kpi_id:
|
||
continue
|
||
# 同KPI同预测期覆盖
|
||
existing = db.query(KpiForecastLog).filter(
|
||
KpiForecastLog.entity_id == entity_id,
|
||
KpiForecastLog.kpi_id == kpi_id,
|
||
KpiForecastLog.period == period,
|
||
).first()
|
||
if existing:
|
||
existing.forecast_value = float(val)
|
||
existing.model = model
|
||
existing.confidence = r.get("confidence")
|
||
existing.trend = r.get("trend")
|
||
else:
|
||
db.add(KpiForecastLog(
|
||
entity_id=entity_id, kpi_id=kpi_id, kpi_code=kpi_code,
|
||
period=period, forecast_value=float(val),
|
||
model=model, confidence=r.get("confidence"), trend=r.get("trend"),
|
||
))
|
||
saved += 1
|
||
db.commit()
|
||
return saved
|
||
|
||
|
||
# ════════════════════════════════════════════════════════════
|
||
# 宏观敏感性因素联动（IMA 2026.7 Predictive Cost Intelligence 完整版）
|
||
# 内置宏观因素 → 按KPI类型推断弹性系数 → 调整预测值
|
||
# MVP：弹性系数为规则推断+可调，非历史回归（诚实标注"模型弹性"）
|
||
# v2(2026-08-25): 内置宏观历史数据 → 变化率弹性校准（有数据用回归，无数据回退规则）
|
||
# ════════════════════════════════════════════════════════════
|
||
|
||
MACRO_FACTORS = [
|
||
{"key": "oil", "name": "原油价格", "unit": "美元/桶",
|
||
"desc": "油价↑ → 运输/能源成本↑ → 成本类KPI↑、利润类KPI↓"},
|
||
{"key": "usd", "name": "美元汇率", "unit": "USD/CNY",
|
||
"desc": "美元↑ → 进口成本↑(成本类↑)、出口收入↑(营收类↑)"},
|
||
{"key": "cpi", "name": "CPI通胀率", "unit": "%",
|
||
"desc": "CPI↑ → 成本↑、名义营收↑"},
|
||
]
|
||
|
||
# 内置宏观因素历史数据（月度，2026-01 ~ 2026-07，供变化率弹性校准）
|
||
MACRO_FACTOR_HISTORY = {
|
||
"oil": [
|
||
{"period": "2026-01", "value": 74.0}, {"period": "2026-02", "value": 78.0},
|
||
{"period": "2026-03", "value": 76.0}, {"period": "2026-04", "value": 82.0},
|
||
{"period": "2026-05", "value": 79.0}, {"period": "2026-06", "value": 85.0},
|
||
{"period": "2026-07", "value": 88.0},
|
||
],
|
||
"usd": [
|
||
{"period": "2026-01", "value": 7.05}, {"period": "2026-02", "value": 7.08},
|
||
{"period": "2026-03", "value": 7.06}, {"period": "2026-04", "value": 7.10},
|
||
{"period": "2026-05", "value": 7.12}, {"period": "2026-06", "value": 7.15},
|
||
{"period": "2026-07", "value": 7.18},
|
||
],
|
||
"cpi": [
|
||
{"period": "2026-01", "value": 1.8}, {"period": "2026-02", "value": 1.9},
|
||
{"period": "2026-03", "value": 1.9}, {"period": "2026-04", "value": 2.0},
|
||
{"period": "2026-05", "value": 2.1}, {"period": "2026-06", "value": 2.1},
|
||
{"period": "2026-07", "value": 2.2},
|
||
],
|
||
}
|
||
|
||
# KPI 类别关键词 → 因素方向/弹性 (direction: +因素涨KPI涨, -因素涨KPI跌)
|
||
FACTOR_RULES = {
|
||
"cost": { # 成本/费用类: 宏观涨 → 成本涨
|
||
"oil": {"direction": "+", "elasticity": 0.15},
|
||
"usd": {"direction": "+", "elasticity": 0.10},
|
||
"cpi": {"direction": "+", "elasticity": 0.10},
|
||
},
|
||
"revenue": { # 营收类: 通胀涨→名义营收涨
|
||
"oil": {"direction": "-", "elasticity": 0.05},
|
||
"usd": {"direction": "+", "elasticity": 0.08},
|
||
"cpi": {"direction": "+", "elasticity": 0.08},
|
||
},
|
||
"profit": { # 利润类: 宏观涨 → 成本挤压利润
|
||
"oil": {"direction": "-", "elasticity": 0.12},
|
||
"usd": {"direction": "-", "elasticity": 0.08},
|
||
"cpi": {"direction": "-", "elasticity": 0.08},
|
||
},
|
||
"cash": { # 现金流类
|
||
"oil": {"direction": "-", "elasticity": 0.06},
|
||
"usd": {"direction": "-", "elasticity": 0.04},
|
||
"cpi": {"direction": "-", "elasticity": 0.05},
|
||
},
|
||
}
|
||
|
||
# 类别关键词匹配（长词优先）
|
||
CATEGORY_KEYWORDS = [
|
||
("profit", ["净利润", "净利", "利润", "毛利", "ROE", "ROI", "EVA", "收益率", "报酬率"]),
|
||
("revenue", ["营收", "收入", "销售额", "销售", "产值", "客单"]),
|
||
("cost", ["费用率", "成本率", "费用", "成本", "费率", "应付", "返利", "渠补", "税"]),
|
||
("cash", ["现金流", "现金", "回款", "FCF", "资金"]),
|
||
]
|
||
|
||
|
||
def infer_kpi_category(kpi_name: str, kpi_code: str = "") -> str:
|
||
"""按KPI名称/编码推断类别: profit/revenue/cost/cash，兜底 profit（保守）"""
|
||
n = (kpi_name or "") + " " + (kpi_code or "")
|
||
for cat, kws in CATEGORY_KEYWORDS:
|
||
if any(kw in n for kw in kws):
|
||
return cat
|
||
return "profit"
|
||
|
||
|
||
def factor_sensitivity_for_kpi(kpi_name: str, kpi_code: str = "") -> list:
|
||
"""返回该KPI对3个宏观因素的敏感性（方向+弹性）— 规则推断版"""
|
||
cat = infer_kpi_category(kpi_name, kpi_code)
|
||
rules = FACTOR_RULES.get(cat, FACTOR_RULES["profit"])
|
||
out = []
|
||
for f in MACRO_FACTORS:
|
||
r = rules.get(f["key"], {"direction": "-", "elasticity": 0.05})
|
||
out.append({
|
||
"factor_key": f["key"],
|
||
"factor_name": f["name"],
|
||
"factor_unit": f["unit"],
|
||
"factor_desc": f["desc"],
|
||
"direction": r["direction"],
|
||
"elasticity": r["elasticity"],
|
||
"category": cat,
|
||
"elasticity_source": "rule",
|
||
})
|
||
return out
|
||
|
||
|
||
def _rate_of_change(series: list) -> list:
|
||
"""相邻期变化率列表 [(period, pct), ...]"""
|
||
out = []
|
||
for i in range(1, len(series)):
|
||
prev, cur = series[i - 1], series[i]
|
||
if prev and prev.get("value"):
|
||
pct = (cur["value"] - prev["value"]) / prev["value"] * 100
|
||
out.append((cur["period"], pct))
|
||
return out
|
||
|
||
|
||
def elasticity_from_history(kpi_history: list, factor_key: str,
|
||
direction: str) -> Optional[dict]:
|
||
"""变化率弹性校准：KPI历史 vs 宏观因素历史（同period匹配）
|
||
弹性 = mean(KPI变化率 / 因素变化率)（符号由实际数据决定）
|
||
匹配期数 < 2 或无因素数据 → 返回 None（回退规则）
|
||
"""
|
||
factor_hist = MACRO_FACTOR_HISTORY.get(factor_key)
|
||
if not factor_hist or not kpi_history:
|
||
return None
|
||
kpi_by_period = {h.get("period"): h.get("value") for h in kpi_history if h.get("value") is not None}
|
||
ratios = []
|
||
# 因素相邻期变化率
|
||
for i in range(1, len(factor_hist)):
|
||
fp = factor_hist[i]["period"]
|
||
fv = factor_hist[i]["value"]
|
||
fv_prev = factor_hist[i - 1]["value"]
|
||
if not fv_prev:
|
||
continue
|
||
f_chg = (fv - fv_prev) / fv_prev * 100
|
||
# KPI 同期值（以及上一期，用于算KPI变化）
|
||
k_cur = kpi_by_period.get(fp)
|
||
# KPI 在因素上一期的值（模糊匹配上一月度）
|
||
k_prev = kpi_by_period.get(factor_hist[i - 1]["period"])
|
||
if k_cur is not None and k_prev not in (None, 0) and abs(f_chg) > 0.01:
|
||
k_chg = (k_cur - k_prev) / k_prev * 100
|
||
ratios.append(k_chg / f_chg)
|
||
if len(ratios) < 2:
|
||
return None
|
||
import statistics
|
||
raw_elasticity = statistics.median(ratios)
|
||
# 弹性合理性校验: |弹性| 超出 [0.01, 0.5] 视为数据噪声 → 回退规则推断（诚实标注，不用失真校准）
|
||
if not (0.01 <= abs(raw_elasticity) <= 0.5):
|
||
return None
|
||
elasticity = round(raw_elasticity, 4)
|
||
# 方向由数据符号决定；数据符号与规则方向冲突时以数据为准（标注）
|
||
data_direction = "+" if elasticity >= 0 else "-"
|
||
return {
|
||
"elasticity": abs(elasticity),
|
||
"direction": data_direction,
|
||
"matched_periods": len(ratios),
|
||
"elasticity_source": "history",
|
||
"rule_direction": direction,
|
||
}
|
||
|
||
|
||
def factor_sensitivity_with_history(kpi_name: str, kpi_code: str = "",
|
||
kpi_history: Optional[list] = None) -> list:
|
||
"""增强版敏感性：有历史数据用变化率弹性校准，无数据回退规则推断"""
|
||
base = factor_sensitivity_for_kpi(kpi_name, kpi_code)
|
||
out = []
|
||
for s in base:
|
||
hist_el = elasticity_from_history(kpi_history or [], s["factor_key"], s["direction"]) if kpi_history else None
|
||
if hist_el:
|
||
out.append({
|
||
**s,
|
||
"elasticity": hist_el["elasticity"],
|
||
"direction": hist_el["direction"],
|
||
"elasticity_source": hist_el["elasticity_source"],
|
||
"matched_periods": hist_el["matched_periods"],
|
||
"rule_direction": hist_el["rule_direction"],
|
||
})
|
||
else:
|
||
out.append(s)
|
||
return out
|
||
|
||
|
||
def adjusted_next_with_factor(next_target: Optional[float], pct: float,
|
||
direction: str, elasticity: float) -> Optional[float]:
|
||
"""因素变动 pct% → 调整后预测值: 方向+ 因素涨预测涨; 方向- 因素涨预测跌
|
||
负值KPI(亏损)方向反转: 方向- 时因素涨 → 更亏(更负)"""
|
||
if next_target is None:
|
||
return None
|
||
factor_change = pct * 0.01 # ±5% → 0.05
|
||
sign = 1.0 if direction == "+" else -1.0
|
||
if next_target < 0:
|
||
sign = -sign # 负值(亏损): 因素涨 → 更亏
|
||
return round(next_target * (1 + sign * factor_change * elasticity), 2)
|