Files
NewQuant/src/execution_simulator.py

361 lines
18 KiB
Python
Raw Normal View History

2025-06-23 22:21:59 +08:00
# src/execution_simulator.py
from datetime import datetime
2025-06-18 10:25:05 +08:00
from typing import Dict, List, Optional
import pandas as pd
from .core_data import Order, Trade, Bar, PortfolioSnapshot
class ExecutionSimulator:
"""
模拟交易执行和管理账户资金持仓
"""
def __init__(self, initial_capital: float,
slippage_rate: float = 0.0001,
commission_rate: float = 0.0002,
initial_positions: Optional[Dict[str, int]] = None):
self.initial_capital = initial_capital
self.cash = initial_capital
self.positions: Dict[str, int] = initial_positions if initial_positions is not None else {}
self.average_costs: Dict[str, float] = {}
if initial_positions:
for symbol, qty in initial_positions.items():
2025-06-23 22:21:59 +08:00
self.average_costs[symbol] = 0.0
2025-06-18 10:25:05 +08:00
self.slippage_rate = slippage_rate
self.commission_rate = commission_rate
2025-06-23 22:21:59 +08:00
self.trade_log: List[Trade] = []
self.pending_orders: Dict[str, Order] = {}
self._current_time: Optional[datetime] = None
2025-06-18 10:25:05 +08:00
print(
f"模拟器初始化:初始资金={self.initial_capital:.2f}, 滑点率={self.slippage_rate}, 佣金率={self.commission_rate}")
if self.positions:
print(f"初始持仓:{self.positions}")
def update_time(self, current_time: datetime):
self._current_time = current_time
def get_current_time(self) -> datetime:
if self._current_time is None:
2025-06-23 22:21:59 +08:00
# 改进:如果时间未设置,可以抛出错误,防止策略在 on_init 阶段意外调用
# raise RuntimeError("Simulator time has not been set. Ensure update_time is called.")
return None
return self._current_time
2025-06-18 10:25:05 +08:00
def _calculate_fill_price(self, order: Order, current_bar: Bar) -> float:
"""
内部方法根据订单类型和滑点计算实际成交价格
2025-06-23 22:21:59 +08:00
撮合逻辑所有订单市价/限价都以当前K线的 **开盘价 (open)** 为基准进行撮合
2025-06-18 10:25:05 +08:00
"""
fill_price = -1.0 # 默认未成交
2025-06-23 22:21:59 +08:00
base_price = current_bar.open # 所有成交都以当前K线的开盘价为基准
2025-06-23 22:21:59 +08:00
if order.price_type == "MARKET":
# 市价单:直接以开盘价成交,考虑滑点
if order.direction == "BUY" or order.direction == "CLOSE_SHORT": # 买入/平空:向上偏离(多付)
fill_price = base_price * (1 + self.slippage_rate)
2025-06-23 22:21:59 +08:00
elif order.direction == "SELL" or order.direction == "CLOSE_LONG": # 卖出/平多:向下偏离(少收)
fill_price = base_price * (1 - self.slippage_rate)
2025-06-23 22:21:59 +08:00
else:
fill_price = base_price # 理论上不发生
elif order.price_type == "LIMIT" and order.limit_price is not None:
limit_price = order.limit_price
2025-06-23 22:21:59 +08:00
# 限价单:判断开盘价是否满足限价条件,如果满足,则以开盘价成交(考虑滑点)
if order.direction == "BUY" or order.direction == "CLOSE_SHORT": # 限价买入/平空
# 买单只有当开盘价低于或等于限价时才可能成交
# 即:我愿意出 limit_price 买,开盘价 open_price 更低或一样,当然买
if base_price <= limit_price:
fill_price = base_price * (1 + self.slippage_rate)
# else: 未满足限价条件,不成交
elif order.direction == "SELL" or order.direction == "CLOSE_LONG": # 限价卖出/平多
# 卖单只有当开盘价高于或等于限价时才可能成交
# 即:我愿意出 limit_price 卖,开盘价 open_price 更高或一样,当然卖
if base_price >= limit_price:
fill_price = base_price * (1 - self.slippage_rate)
# else: 未满足限价条件,不成交
# 最终检查成交价是否有效且合理大于0
if fill_price <= 0:
2025-06-23 22:21:59 +08:00
return -1.0 # 未成交或价格无效
2025-06-18 10:25:05 +08:00
return fill_price
2025-06-23 22:21:59 +08:00
def send_order_to_pending(self, order: Order) -> Optional[Order]:
2025-06-18 10:25:05 +08:00
"""
2025-06-23 22:21:59 +08:00
将订单添加到待处理队列 BacktestEngine Strategy 调用
此方法不进行撮合撮合由 process_pending_orders 统一处理
"""
if order.id in self.pending_orders:
# print(f"订单 {order.id} 已经存在于待处理队列。")
return None
self.pending_orders[order.id] = order
# print(f"订单 {order.id} 加入待处理队列。")
return order
def process_pending_orders(self, current_bar: Bar):
"""
处理所有待撮合的订单在每个K线数据到来时调用
"""
# 复制一份待处理订单的键,防止在迭代时修改字典
order_ids_to_process = list(self.pending_orders.keys())
for order_id in order_ids_to_process:
if order_id not in self.pending_orders: # 订单可能已被取消
continue
order = self.pending_orders[order_id]
# 只有当订单的symbol与当前bar的symbol一致时才尝试撮合
# 这样确保了在换月后,旧合约的挂单不会被尝试撮合 (尽管换月时会强制取消)
if order.symbol != current_bar.symbol:
# 这种情况理论上应该被换月逻辑清理掉的旧合约挂单,
# 如果因为某种原因漏掉了,这里直接跳过,避免异常。
continue
# 尝试成交订单
self._execute_single_order(order, current_bar)
def _execute_single_order(self, order: Order, current_bar: Bar) -> Optional[Trade]:
"""
内部方法尝试执行单个订单并处理资金和持仓变化
send_order process_pending_orders 调用
2025-06-18 10:25:05 +08:00
"""
# --- 处理撤单指令 ---
2025-06-23 22:21:59 +08:00
if order.direction == "CANCEL": # 策略主动发起撤单
2025-06-18 10:25:05 +08:00
success = self.cancel_order(order.id)
if success:
# print(f"[{current_bar.datetime}] 模拟器: 收到并成功处理撤单指令 for Order ID: {order.id}")
pass
2025-06-23 22:21:59 +08:00
return None # 撤单操作不返回Trade
2025-06-18 10:25:05 +08:00
symbol = order.symbol
volume = order.volume
2025-06-23 22:21:59 +08:00
2025-06-18 10:25:05 +08:00
# 尝试计算成交价格
fill_price = self._calculate_fill_price(order, current_bar)
2025-06-23 22:21:59 +08:00
if fill_price <= 0: # 未成交或不满足限价条件
return None
2025-06-18 10:25:05 +08:00
# --- 以下是订单成功成交的逻辑 ---
trade_value = volume * fill_price
commission = trade_value * self.commission_rate
current_position = self.positions.get(symbol, 0)
current_average_cost = self.average_costs.get(symbol, 0.0)
2025-06-23 22:21:59 +08:00
realized_pnl = 0.0
# 根据 direction 判断开平仓意图
# 如果 direction 是 CLOSE_LONG 或 CLOSE_SELL (平多), CLOSE_SHORT (平空) 则是平仓交易
is_close_trade = order.direction in ["CLOSE_LONG", "CLOSE_SELL", "CLOSE_SHORT"]
# 如果 direction 是 BUY 或 SELL 且不是平仓意图,则是开仓交易
is_open_trade = (order.direction in ["BUY", "SELL"]) and (not is_close_trade)
# 区分实际的买卖方向
actual_execution_direction = ""
if order.direction == "BUY" or order.direction == "CLOSE_SHORT":
actual_execution_direction = "BUY"
elif order.direction == "SELL" or order.direction == "CLOSE_LONG" or order.direction == "CLOSE_SELL":
actual_execution_direction = "SELL"
else:
print(f"[{current_bar.datetime}] 模拟器: 收到未知订单方向 {order.direction} for Order ID: {order.id}. 订单未处理。")
if order.id in self.pending_orders: del self.pending_orders[order.id]
return None
2025-06-23 22:21:59 +08:00
if actual_execution_direction == "BUY": # 处理实际的买入 (开多 / 平空)
2025-06-18 10:25:05 +08:00
if current_position >= 0: # 当前持有多仓或无仓位 (开多)
new_total_cost = (current_average_cost * current_position) + (fill_price * volume)
new_total_volume = current_position + volume
self.average_costs[symbol] = new_total_cost / new_total_volume if new_total_volume > 0 else 0.0
self.positions[symbol] = new_total_volume
else: # 当前持有空仓 (平空)
2025-06-23 22:21:59 +08:00
pnl_per_share = current_average_cost - fill_price # 空头平仓盈亏
2025-06-18 10:25:05 +08:00
realized_pnl = pnl_per_share * volume
self.positions[symbol] += volume
2025-06-23 22:21:59 +08:00
if self.positions[symbol] == 0:
2025-06-18 10:25:05 +08:00
del self.positions[symbol]
if symbol in self.average_costs: del self.average_costs[symbol]
2025-06-23 22:21:59 +08:00
elif self.positions[symbol] > 0 and current_position < 0: # 空转多
self.average_costs[symbol] = fill_price # 新多头仓位成本以成交价为准
2025-06-18 10:25:05 +08:00
2025-06-23 22:21:59 +08:00
if self.cash < trade_value + commission:
print(f"[{current_bar.datetime}] 模拟器: 资金不足,无法执行买入 {volume} {symbol} @ {fill_price:.2f}")
if order.id in self.pending_orders: del self.pending_orders[order.id]
2025-06-18 10:25:05 +08:00
return None
2025-06-23 22:21:59 +08:00
self.cash -= (trade_value + commission)
2025-06-18 10:25:05 +08:00
2025-06-23 22:21:59 +08:00
elif actual_execution_direction == "SELL": # 处理实际的卖出 (开空 / 平多)
2025-06-18 10:25:05 +08:00
if current_position <= 0: # 当前持有空仓或无仓位 (开空)
new_total_value = (current_average_cost * abs(current_position)) + (fill_price * volume)
new_total_volume = abs(current_position) + volume
self.average_costs[symbol] = new_total_value / new_total_volume if new_total_volume > 0 else 0.0
2025-06-23 22:21:59 +08:00
self.positions[symbol] -= volume
2025-06-18 10:25:05 +08:00
else: # 当前持有多仓 (平多)
2025-06-23 22:21:59 +08:00
pnl_per_share = fill_price - current_average_cost # 多头平仓盈亏
2025-06-18 10:25:05 +08:00
realized_pnl = pnl_per_share * volume
self.positions[symbol] -= volume
2025-06-23 22:21:59 +08:00
if self.positions[symbol] == 0:
2025-06-18 10:25:05 +08:00
del self.positions[symbol]
if symbol in self.average_costs: del self.average_costs[symbol]
2025-06-23 22:21:59 +08:00
elif self.positions[symbol] < 0 and current_position > 0: # 多转空
self.average_costs[symbol] = fill_price # 新空头仓位成本以成交价为准
if self.cash < commission: # 卖出交易,佣金先扣
print(f"[{current_bar.datetime}] 模拟器: 资金不足(佣金),无法执行卖出 {volume} {symbol} @ {fill_price:.2f}")
if order.id in self.pending_orders: del self.pending_orders[order.id]
2025-06-18 10:25:05 +08:00
return None
2025-06-23 22:21:59 +08:00
self.cash -= commission
self.cash += trade_value
# 创建 Trade 对象时direction 使用原始订单的 direction
2025-06-18 10:25:05 +08:00
executed_trade = Trade(
order_id=order.id, fill_time=current_bar.datetime, symbol=symbol,
2025-06-23 22:21:59 +08:00
direction=order.direction, # 使用原始订单的 direction
2025-06-18 10:25:05 +08:00
volume=volume, price=fill_price, commission=commission,
cash_after_trade=self.cash, positions_after_trade=self.positions.copy(),
2025-06-23 22:21:59 +08:00
realized_pnl=realized_pnl,
2025-06-18 10:25:05 +08:00
is_open_trade=is_open_trade,
is_close_trade=is_close_trade
)
self.trade_log.append(executed_trade)
2025-06-23 22:21:59 +08:00
# 订单成交,从待处理订单中移除
2025-06-18 10:25:05 +08:00
if order.id in self.pending_orders:
del self.pending_orders[order.id]
return executed_trade
def cancel_order(self, order_id: str) -> bool:
"""
尝试取消一个待处理订单
"""
if order_id in self.pending_orders:
del self.pending_orders[order_id]
return True
return False
2025-06-23 22:21:59 +08:00
# --- 新增:强制平仓指定合约的所有持仓 ---
def force_close_all_positions_for_symbol(self, symbol_to_close: str, closing_bar: Bar) -> List[Trade]:
"""
强制平仓指定合约的所有持仓
Args:
symbol_to_close (str): 需要平仓的合约代码
closing_bar (Bar): 用于获取平仓价格的当前K线数据通常是旧合约的最后一根K线
Returns:
List[Trade]: 因强制平仓而产生的交易记录
"""
closed_trades: List[Trade] = []
# 仅处理指定symbol的持仓
if symbol_to_close in self.positions and self.positions[symbol_to_close] != 0:
volume_to_close = self.positions[symbol_to_close]
# 根据持仓方向决定平仓订单的方向
direction = "SELL" if volume_to_close > 0 else "BUY" # 多头平仓是卖出,空头平仓是买入
# 构造一个市价平仓订单
rollover_order = Order(
id=f"FORCE_CLOSE_{symbol_to_close}_{closing_bar.datetime.strftime('%Y%m%d%H%M%S%f')}",
symbol=symbol_to_close,
direction=direction,
volume=abs(volume_to_close),
price_type="MARKET",
limit_price=None,
submitted_time=closing_bar.datetime,
)
# 使用内部的执行逻辑进行撮合
trade = self._execute_single_order(rollover_order, closing_bar)
if trade:
closed_trades.append(trade)
else:
print(f"[{closing_bar.datetime}] 警告: 强制平仓 {symbol_to_close} 失败!")
2025-06-18 10:25:05 +08:00
2025-06-23 22:21:59 +08:00
return closed_trades
# --- 新增:取消指定合约的所有挂单 ---
def cancel_all_pending_orders_for_symbol(self, symbol_to_cancel: str) -> int:
2025-06-18 10:25:05 +08:00
"""
2025-06-23 22:21:59 +08:00
取消指定合约的所有待处理订单
2025-06-18 10:25:05 +08:00
"""
2025-06-23 22:21:59 +08:00
cancelled_count = 0
order_ids_to_cancel = [
order_id for order_id, order in self.pending_orders.items()
if order.symbol == symbol_to_cancel
]
for order_id in order_ids_to_cancel:
if self.cancel_order(order_id): # 调用现有的 cancel_order 方法
cancelled_count += 1
return cancelled_count
def get_pending_orders(self) -> Dict[str, Order]:
2025-06-18 10:25:05 +08:00
return self.pending_orders.copy()
def get_portfolio_value(self, current_bar: Bar) -> float:
"""
计算当前的投资组合总价值包括现金和持仓市值
2025-06-23 22:21:59 +08:00
此方法需要兼容多合约持仓的场景
2025-06-18 10:25:05 +08:00
Args:
2025-06-23 22:21:59 +08:00
current_bar (Bar): 当前的Bar数据用于计算**当前活跃合约**的持仓市值
注意如果 simulator 中持有多个合约这里需要更复杂的逻辑
目前假设主力合约回测时simulator.positions 主要只包含当前主力合约
2025-06-18 10:25:05 +08:00
Returns:
float: 当前的投资组合总价值
"""
total_value = self.cash
2025-06-19 15:28:26 +08:00
2025-06-23 22:21:59 +08:00
# 遍历所有持仓,计算市值。
# 注意:这里假设 current_bar 提供了当前活跃主力合约的价格。
# 如果 self.positions 中包含其他非 current_bar.symbol 的旧合约,
# 它们的市值将无法用 current_bar.open 来准确计算。
# 在换月模式下,旧合约会被强制平仓,因此 simulator.positions 通常只包含一个合约。
for symbol, quantity in self.positions.items():
# 这里简单处理:如果持仓合约与 current_bar.symbol 相同,则使用 current_bar.open 计算。
# 如果是其他合约,则需要外部提供其最新价格,但这超出了本函数当前的能力范围。
# 考虑到换月模式,旧合约会被平仓,所以大部分时候这不会是问题。
if symbol == current_bar.symbol:
total_value += quantity * current_bar.open
else:
# 警告:如果这里出现,说明有未平仓的旧合约持仓,且没有其最新价格来计算市值。
# 在严谨的主力连续回测中,这不应该发生,因为换月会强制平仓。
print(f"[{current_bar.datetime}] 警告持仓中存在非当前K线合约 {symbol},无法准确计算其市值。")
# 可以选择将这部分持仓价值计为0或者使用上一个已知价格需要额外数据结构
# 这里我们假设它不影响总价值计算,因为换月时会处理掉
pass
2025-06-19 15:28:26 +08:00
2025-06-18 10:25:05 +08:00
return total_value
def get_current_positions(self) -> Dict[str, int]:
return self.positions.copy()
def get_trade_history(self) -> List[Trade]:
return self.trade_log.copy()
2025-06-19 15:28:26 +08:00
def reset(self, new_initial_capital: float = None, new_initial_positions: Dict[str, int] = None) -> None:
"""
重置模拟器状态到新的初始条件
2025-06-23 22:21:59 +08:00
此方法不用于换月时的平仓它用于整个回测开始前的初始化
2025-06-19 15:28:26 +08:00
"""
print("ExecutionSimulator: 重置状态。")
self.cash = new_initial_capital if new_initial_capital is not None else self.initial_capital
self.positions = new_initial_positions.copy() if new_initial_positions is not None else {}
2025-06-23 22:21:59 +08:00
self.average_costs = {}
for symbol, qty in self.positions.items(): # 重置平均成本
self.average_costs[symbol] = 0.0
self.trade_log = []
self.pending_orders = {} # 清空挂单
self._current_time = None
2025-06-19 15:28:26 +08:00
2025-06-23 22:21:59 +08:00
# Removed clear_trade_history as trade_log is cleared in reset
def get_average_position_price(self, symbol: str) -> Optional[float]:
if symbol in self.positions and self.positions[symbol] != 0:
return self.average_costs.get(symbol)
2025-06-23 22:21:59 +08:00
return None