feat(backend): AI 服务层 — 千问流式 LLM + 月报解析 + 健康度计算 + 风险检测 + Copilot

- LLM 客户端：全部 SSE 流式输出，兼容 OpenAI 接口 - AI 月报解析：SSE 流式端点 POST /reports/{id}/parse - 健康度计算引擎：四维评分（财务/经营/AI商业化/AI成本） - 风险自动检测引擎：6 条规则自动检测指标越界 - AI Copilot：SSE 流式对话 POST /copilot/chat - 权限中间件：角色级 + 字段级权限控制 - 测试：21 个新测试（健康度 8 + 风险检测 8 + 权限 5），总计 64 passed
2026-07-18 22:16:40 +08:00
parent 94be6189e9
commit 7ec4fb0747
15 changed files with 1244 additions and 52 deletions
@@ -0,0 +1,228 @@
+"""健康度评分计算引擎。
+
+基于月报结构化数据，计算四维评分：
+- 财务健康度（financial_score）
+- 经营健康度（operational_score）
+- AI 商业化度（ai_commercial_score）
+- AI 成本效率（ai_cost_score）
+
+总分 = 加权平均，输出 0-100 分。
+"""
+
+import logging
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+# 权重配置
+WEIGHTS = {
+    "financial": 0.35,
+    "operational": 0.25,
+    "ai_commercial": 0.25,
+    "ai_cost": 0.15,
+}
+
+
+def _safe_float(value: Any, default: float = 0.0) -> float:
+    """安全转换为 float。"""
+    if value is None or value == "":
+        return default
+    try:
+        return float(value)
+    except (ValueError, TypeError):
+        return default
+
+
+def _calc_financial_score(data: dict[str, Any]) -> float:
+    """计算财务健康度。
+
+    指标：
+    - 现金跑道（runway_months）：>12 月=90+，6-12=60-90，<6=<60
+    - 营收同比增长（yoy_change）：正=加分，负=减分
+    - 烧钱率趋势（burn_rate trend）：down=加分，up=减分
+    """
+    score = 50.0  # 基础分
+
+    cash = data.get("cash_balance", {})
+    runway = _safe_float(cash.get("runway_months"))
+    if runway > 0:
+        if runway >= 12:
+            score += 30
+        elif runway >= 6:
+            score += 15
+        elif runway >= 3:
+            score -= 10
+        else:
+            score -= 30
+
+    revenue = data.get("revenue", {})
+    yoy = revenue.get("yoy_change", "")
+    if yoy:
+        yoy_val = _safe_float(str(yoy).replace("%", "").replace("+", ""))
+        if yoy_val > 0:
+            score += 15
+        elif yoy_val < 0:
+            score -= 15
+
+    burn = data.get("burn_rate", {})
+    trend = burn.get("trend", "")
+    if trend == "down":
+        score += 10
+    elif trend == "up":
+        score -= 10
+
+    return max(0, min(100, score))
+
+
+def _calc_operational_score(data: dict[str, Any]) -> float:
+    """计算经营健康度。
+
+    指标：
+    - 团队规模变化（headcount）：净增长=加分
+    - 关键指标达成情况
+    """
+    score = 60.0
+
+    headcount = data.get("headcount", {})
+    new_hires = _safe_float(headcount.get("new_hires"))
+    departures = _safe_float(headcount.get("departures"))
+    net_change = new_hires - departures
+    if net_change > 0:
+        score += 15
+    elif net_change < 0:
+        score -= 10
+    if departures > 5:
+        score -= 10  # 高流失率
+
+    key_metrics = data.get("key_metrics", [])
+    if key_metrics:
+        positive_count = sum(
+            1 for m in key_metrics
+            if _safe_float(str(m.get("change", "")).replace("%", "").replace("+", "")) > 0
+        )
+        score += (positive_count / len(key_metrics)) * 20
+
+    return max(0, min(100, score))
+
+
+def _calc_ai_commercial_score(data: dict[str, Any]) -> float:
+    """计算 AI 商业化度。
+
+    基于 key_metrics 中 AI 相关指标的达成情况。
+    """
+    score = 50.0
+
+    key_metrics = data.get("key_metrics", [])
+    ai_metrics = [
+        m for m in key_metrics
+        if "ai" in str(m.get("name", "")).lower()
+        or "模型" in str(m.get("name", ""))
+        or "推理" in str(m.get("name", ""))
+    ]
+
+    if ai_metrics:
+        for m in ai_metrics:
+            change = str(m.get("change", ""))
+            val = _safe_float(change.replace("%", "").replace("+", ""))
+            if val > 0:
+                score += 15
+            elif val < 0:
+                score -= 10
+    else:
+        # 无 AI 相关指标，给中等偏下分数
+        score = 40.0
+
+    return max(0, min(100, score))
+
+
+def _calc_ai_cost_score(data: dict[str, Any]) -> float:
+    """计算 AI 成本效率。
+
+    基于 burn_rate 和 AI 相关支出估算。
+    """
+    score = 55.0
+
+    burn = data.get("burn_rate", {})
+    trend = burn.get("trend", "")
+    if trend == "down":
+        score += 20
+    elif trend == "up":
+        score -= 15
+
+    # 如果有 key_metrics 中的成本相关指标
+    key_metrics = data.get("key_metrics", [])
+    cost_metrics = [
+        m for m in key_metrics
+        if "成本" in str(m.get("name", "")) or "cost" in str(m.get("name", "")).lower()
+    ]
+    for m in cost_metrics:
+        change = str(m.get("change", ""))
+        val = _safe_float(change.replace("%", "").replace("+", ""))
+        if val < 0:  # 成本下降是好事
+            score += 10
+        elif val > 0:
+            score -= 10
+
+    return max(0, min(100, score))
+
+
+def calculate_health_score(structured_data: dict[str, Any]) -> dict[str, float]:
+    """计算四维健康度评分。
+
+    Args:
+        structured_data: 月报 AI 解析后的结构化数据
+
+    Returns:
+        包含 total_score 和四个维度分数的字典
+    """
+    if not structured_data:
+        return {
+            "total_score": 0.0,
+            "financial_score": 0.0,
+            "operational_score": 0.0,
+            "ai_commercial_score": 0.0,
+            "ai_cost_score": 0.0,
+        }
+
+    financial = _calc_financial_score(structured_data)
+    operational = _calc_operational_score(structured_data)
+    ai_commercial = _calc_ai_commercial_score(structured_data)
+    ai_cost = _calc_ai_cost_score(structured_data)
+
+    total = (
+        financial * WEIGHTS["financial"]
+        + operational * WEIGHTS["operational"]
+        + ai_commercial * WEIGHTS["ai_commercial"]
+        + ai_cost * WEIGHTS["ai_cost"]
+    )
+
+    result = {
+        "total_score": round(total, 1),
+        "financial_score": round(financial, 1),
+        "operational_score": round(operational, 1),
+        "ai_commercial_score": round(ai_commercial, 1),
+        "ai_cost_score": round(ai_cost, 1),
+    }
+
+    logger.info("健康度评分计算完成: %s", result)
+    return result
+
+
+def determine_trend(current_score: float, previous_score: float | None) -> str:
+    """判断评分趋势。
+
+    Args:
+        current_score: 当前评分
+        previous_score: 上期评分（如有）
+
+    Returns:
+        "up" / "down" / "stable"
+    """
+    if previous_score is None:
+        return "stable"
+    diff = current_score - previous_score
+    if diff > 5:
+        return "up"
+    elif diff < -5:
+        return "down"
+    return "stable"