fix(trend): remove inferred DCA fill prices across all exchanges

Unify display on trend_leg_display_price: use recorded fills or grid triggers only; last done row avg comes from live entry, never back-solve trigger prices.

Co-authored-by: Cursor <cursoragent@cursor.com>

strategy_trend_lib.py

@@ -359,48 +359,46 @@ def append_leg_fill_price_json(existing_json: str | None, fill_px: float) -> str
     return json.dumps(fills, ensure_ascii=False, separators=(",", ":"))
 
 
-def _trend_leg_contracts(
-    leg_idx: int, first_amt: float, leg_amounts: list[float]
-) -> float:
+def trend_leg_grid_price(plan: dict, leg_idx: int) -> Optional[float]:
+    """补仓 leg_idx(1..N) 的计划网格触发价；首仓返回 None。"""
+    if leg_idx <= 0:
+        return None
+    try:
+        grid = [float(x) for x in json.loads((plan or {}).get("grid_prices_json") or "[]")]
+    except Exception:
+        grid = []
+    gi = leg_idx - 1
+    if 0 <= gi < len(grid):
+        return float(grid[gi])
+    return None
+
+
+def trend_leg_display_price(plan: dict, leg_idx: int) -> Optional[float]:
+    """
+    四所统一：单档展示价 = leg_fill_prices_json 实际记录，否则计划网格（首仓用均价/参考价）。
+    禁止为凑均价反推虚构成交价。
+    """
+    p = plan or {}
+    fills = parse_leg_fill_prices(p)
+    if len(fills) > leg_idx:
+        return float(fills[leg_idx])
     if leg_idx == 0:
-        return float(first_amt)
-    li = leg_idx - 1
-    if 0 <= li < len(leg_amounts):
-        return float(leg_amounts[li])
-    return 0.0
-
-
-def _infer_trend_fill_from_target_avg(
-    leg_idx: int,
-    prices: list[float],
-    *,
-    first_amt: float,
-    leg_amounts: list[float],
-    legs_done: int,
-    target_avg: float,
-) -> float:
-    """已知其余档位成交价时，反推单档成交价使加权均价等于 target_avg。"""
-    total_contracts = 0.0
-    known_cost = 0.0
-    unknown_amt = _trend_leg_contracts(leg_idx, first_amt, leg_amounts)
-    for i in range(legs_done + 1):
-        amt = _trend_leg_contracts(i, first_amt, leg_amounts)
-        if amt <= 0:
-            continue
-        total_contracts += amt
-        if i == leg_idx:
-            continue
-        known_cost += float(prices[i]) * amt
-    if unknown_amt <= 0 or total_contracts <= 0:
-        return float(prices[leg_idx])
-    return (float(target_avg) * total_contracts - known_cost) / unknown_amt
+        try:
+            return float(p.get("avg_entry_price"))
+        except (TypeError, ValueError):
+            pass
+        try:
+            ref = p.get("live_price_ref")
+            if ref not in (None, ""):
+                return float(ref)
+        except (TypeError, ValueError):
+            pass
+        return None
+    return trend_leg_grid_price(p, leg_idx)
 
 
 def reconcile_trend_leg_fill_prices(plan: dict) -> list[float]:
-    """
-    首仓(0)+已补仓(1..legs_done) 成交价。
-    优先 leg_fill_prices_json；缺口用计划网格价；再对齐 avg_entry_price。
-    """
+    """首仓(0)+已补仓(1..legs_done) 展示价列表（四所共用 trend_leg_display_price）。"""
     p = plan or {}
     if int(p.get("first_order_done") or 0) == 0:
         return []
@@ -408,64 +406,10 @@ def reconcile_trend_leg_fill_prices(plan: dict) -> list[float]:
         legs_done = int(p.get("legs_done") or 0)
     except (TypeError, ValueError):
         legs_done = 0
-    try:
-        first_amt = float(p.get("first_order_amount"))
-    except (TypeError, ValueError):
-        first_amt = 0.0
-    try:
-        target_avg = float(p.get("avg_entry_price"))
-    except (TypeError, ValueError):
-        target_avg = None
-
-    fills = parse_leg_fill_prices(p)
-    try:
-        grid = [float(x) for x in json.loads(p.get("grid_prices_json") or "[]")]
-    except Exception:
-        grid = []
-    try:
-        leg_amounts = [float(x) for x in json.loads(p.get("leg_amounts_json") or "[]")]
-    except Exception:
-        leg_amounts = []
-
-    def _default_px(leg_idx: int) -> float:
-        if leg_idx == 0:
-            if target_avg is not None and legs_done == 0:
-                return target_avg
-            try:
-                return float(p.get("avg_entry_price"))
-            except (TypeError, ValueError):
-                pass
-            try:
-                ref = p.get("live_price_ref")
-                if ref not in (None, ""):
-                    return float(ref)
-            except (TypeError, ValueError):
-                pass
-            return 0.0
-        gi = leg_idx - 1
-        if 0 <= gi < len(grid):
-            return float(grid[gi])
-        return _default_px(0)
-
     result: list[float] = []
-    estimated: list[int] = []
     for leg_idx in range(legs_done + 1):
-        if len(fills) > leg_idx:
-            result.append(float(fills[leg_idx]))
-        else:
-            result.append(_default_px(leg_idx))
-            estimated.append(leg_idx)
-
-    if target_avg is not None and estimated:
-        adjust_idx = estimated[0] if len(estimated) == 1 else estimated[-1]
-        result[adjust_idx] = _infer_trend_fill_from_target_avg(
-            adjust_idx,
-            result,
-            first_amt=first_amt,
-            leg_amounts=leg_amounts,
-            legs_done=legs_done,
-            target_avg=target_avg,
-        )
+        px = trend_leg_display_price(p, leg_idx)
+        result.append(float(px) if px is not None else 0.0)
     return result
 
 
@@ -618,7 +562,10 @@ def build_trend_preview_level_rows(preview: dict) -> tuple[dict, list[dict]]:
 
 
 def enrich_trend_dca_levels_with_tp(plan: dict, levels: list[dict]) -> list[dict]:
-    """运行中计划：补仓表按实际成交价重算触发价/均价/金额盈亏比；未补档仍用计划触发价预估。"""
+    """
+    四所统一补仓表 enrich（实例策略页 + 中控 monitor 共用）。
+    触发价：实际成交价或计划网格；末档加仓后均价用持仓均价；禁止反推虚构成交价。
+    """
     if not levels:
         return levels
     p = plan or {}
@@ -641,7 +588,10 @@ def enrich_trend_dca_levels_with_tp(plan: dict, levels: list[dict]) -> list[dict
     except (TypeError, ValueError):
         legs_done = 0
     first_done = int(p.get("first_order_done") or 0) != 0
-    reconciled_fills = reconcile_trend_leg_fill_prices(p)
+    try:
+        target_avg = float(p.get("avg_entry_price"))
+    except (TypeError, ValueError):
+        target_avg = None
 
     ref_raw = p.get("live_price_ref")
     if ref_raw in (None, ""):
@@ -671,9 +621,8 @@ def enrich_trend_dca_levels_with_tp(plan: dict, levels: list[dict]) -> list[dict
             except (TypeError, ValueError):
                 amt_f = first_amt
             if first_done:
-                if reconciled_fills:
-                    fill_px = float(reconciled_fills[0])
-                else:
+                fill_px = trend_leg_display_price(p, 0)
+                if fill_px is None:
                     try:
                         fill_px = float(p.get("avg_entry_price") or ref)
                     except (TypeError, ValueError):
@@ -681,8 +630,11 @@ def enrich_trend_dca_levels_with_tp(plan: dict, levels: list[dict]) -> list[dict
                 accumulated = [(float(fill_px), amt_f)]
                 cum_contracts = amt_f
                 row_cum = cum_contracts
-                row["avg_entry"] = float(fill_px)
                 row["price"] = fill_px
+                if target_avg is not None and legs_done == 0:
+                    row["avg_entry"] = target_avg
+                else:
+                    row["avg_entry"] = float(fill_px)
             else:
                 accumulated = [(ref, amt_f)]
                 cum_contracts = amt_f
@@ -704,19 +656,19 @@ def enrich_trend_dca_levels_with_tp(plan: dict, levels: list[dict]) -> list[dict
                 leg_contracts = 0.0
             done = row.get("status") == "done" or (leg_num > 0 and leg_num <= legs_done)
             if done and leg_contracts > 0:
-                if leg_num < len(reconciled_fills):
-                    fill_px = float(reconciled_fills[leg_num])
-                elif grid_trigger_f is not None:
-                    fill_px = grid_trigger_f
-                else:
-                    fill_px = ref
+                fill_px = trend_leg_display_price(p, leg_num)
+                if fill_px is None:
+                    fill_px = grid_trigger_f if grid_trigger_f is not None else ref
                 row["price"] = fill_px
                 accumulated.append((fill_px, leg_contracts))
                 cum_contracts += leg_contracts
                 row_cum = cum_contracts
-                avg = weighted_avg_entry(accumulated)
-                if avg is not None:
-                    row["avg_entry"] = avg
+                if leg_num == legs_done and target_avg is not None:
+                    row["avg_entry"] = target_avg
+                else:
+                    avg = weighted_avg_entry(accumulated)
+                    if avg is not None:
+                        row["avg_entry"] = avg
             elif grid_trigger_f is not None and leg_contracts > 0:
                 row["price"] = grid_trigger_f
                 projected = accumulated + [(grid_trigger_f, leg_contracts)]