import pandas as pd from utils import initialize, shutdown, get_rates, close_all, send_order from backtest import BacktestEngine from logger import logger from config import INITIAL_CAPITAL, SYMBOL, TIMEFRAME, BACKTEST_COUNT, BACKTEST_START_DATE, BACKTEST_END_DATE, USE_DATE_RANGE from risk_management import RiskController from dynamic_weights import DynamicWeightManager from trade_logger import trade_logger from logger import logger # 导入优化器 from optimizer import run_optimizer def run_realtime(): if not initialize(): logger.error("MT5初始化失败") return # 初始化风险管理和动态权重 risk_controller = RiskController() weight_manager = DynamicWeightManager(risk_controller) # 获取动态策略配置 strategies_with_weights = weight_manager.get_current_strategies_and_weights() weight_info = weight_manager.get_weight_info() logger.info(f"市场状态: {weight_info['market_state']}, 置信度: {weight_info['confidence']:.2f}") # 执行策略信号生成 signals = [] weights = [] for strat, weight in strategies_with_weights: try: logger.info(f"执行策略：{strat.__class__.__module__}, 权重: {weight:.2f}") # 特殊处理风险管理策略 if strat.__class__.__module__ == 'risk_management': signal = strat.generate_signal_with_sync(risk_controller) else: signal = strat.generate_signal() signals.append(signal) weights.append(weight) except Exception as e: logger.exception(f"运行策略 {strat.__class__.__module__} 时出错：{e}") # 计算加权信号 weighted_signal_sum = sum(s * w for s, w in zip(signals, weights)) logger.info(f"加权信号总和: {weighted_signal_sum:.2f}") # 获取当前价格用于风险管理 current_price = get_current_price(SYMBOL) current_time = pd.Timestamp.now() # 检查风险管理条件 if current_price: risk_action, risk_reason = risk_controller.check_risk_management(current_price) if risk_action != "none": logger.info(f"触发风险管理: {risk_action}, 原因: {risk_reason}") # 记录平仓交易 if trade_logger.current_position: trade_logger.close_position(SYMBOL, current_price, current_time, f"risk_management_{risk_action}") risk_controller.execute_risk_action(risk_action, risk_reason) shutdown() return # 获取当前持仓状态，取消交易间隔限制 current_position = trade_logger.current_position min_trade_interval_minutes = 0 # 取消最小交易间隔限制 # 取消交易间隔限制检查 can_trade = True # 调试信息：打印当前状态 logger.info(f"当前持仓状态: {current_position is not None}") if current_position: time_since_open = (current_time - current_position['open_time']).total_seconds() / 60 logger.info(f"持仓信息: {current_position['direction']} @ {current_position['open_price']:.2f}, 持仓时间: {time_since_open:.1f}分钟") # 执行交易决策 logger.info(f"加权信号总和: {weighted_signal_sum:.2f}, 买入阈值: 0, 卖出阈值: 0") logger.info(f"风险管理允许买入: {risk_controller.should_allow_trade('buy')}, 允许卖出: {risk_controller.should_allow_trade('sell')}") logger.info(f"允许交易: {can_trade}") if weighted_signal_sum > 0: logger.info(f"检测到买入信号: {weighted_signal_sum:.2f} > 0") if risk_controller.should_allow_trade("buy") and can_trade: # 检查是否已经有相同方向的持仓 if current_position and current_position['direction'] == 'buy': logger.info(f"已持有多头仓位，信号强度: {weighted_signal_sum:.2f}，不重复开仓") else: logger.info(f"=== 准备执行买入交易 ===") logger.info(f"买入信号: 加权总和({weighted_signal_sum:.2f}) > 0") logger.info(f"当前价格: {current_price:.2f}") logger.info(f"风险管理允许: {risk_controller.should_allow_trade('buy')}") logger.info(f"交易间隔检查通过: {can_trade}") # 先平掉现有仓位（如果有） if current_position: close_all(SYMBOL) trade_logger.close_position(SYMBOL, current_price, current_time, "close_before_buy") # 开新仓 send_order(SYMBOL, 'buy') trade_logger.open_position(SYMBOL, 'buy', current_price, current_time, f"weighted_signal_{weighted_signal_sum:.2f}") if current_price: risk_controller.update_position_entry(current_price, "long") else: if not can_trade: logger.info(f"买入信号被阻止: 交易间隔限制") else: logger.info(f"买入信号被风险管理阻止: risk_controller.should_allow_trade('buy') = {risk_controller.should_allow_trade('buy')}") elif weighted_signal_sum < 0: logger.info(f"检测到卖出信号: {weighted_signal_sum:.2f} < 0") if risk_controller.should_allow_trade("sell") and can_trade: # 检查是否已经有相同方向的持仓 if current_position and current_position['direction'] == 'sell': logger.info(f"已持有空头仓位，信号强度: {weighted_signal_sum:.2f}，不重复开仓") else: logger.info(f"=== 准备执行卖出交易 ===") logger.info(f"卖出信号: 加权总和({weighted_signal_sum:.2f}) < 0") logger.info(f"当前价格: {current_price:.2f}") logger.info(f"风险管理允许: {risk_controller.should_allow_trade('sell')}") logger.info(f"交易间隔检查通过: {can_trade}") # 先平掉现有仓位（如果有） if current_position: close_all(SYMBOL) trade_logger.close_position(SYMBOL, current_price, current_time, "close_before_sell") # 开新仓 send_order(SYMBOL, 'sell') trade_logger.open_position(SYMBOL, 'sell', current_price, current_time, f"weighted_signal_{weighted_signal_sum:.2f}") if current_price: risk_controller.update_position_entry(current_price, "short") else: logger.info(f"卖出信号被阻止") else: # 信号在中间区域，使用RiskController决定是否平仓 if current_position: # 使用RiskController检查风险管理条件 risk_action, risk_reason = risk_controller.check_risk_management(current_price) if risk_action != "none": close_all(SYMBOL) trade_logger.close_position(SYMBOL, current_price, current_time, f"signal_neutral_{risk_action}") risk_controller.execute_risk_action(risk_action, risk_reason) shutdown() def get_current_price(symbol): """获取当前价格""" try: import MetaTrader5 as mt5 tick = mt5.symbol_info_tick(symbol) return tick.bid if tick else None except Exception as e: logger.error(f"获取当前价格失败: {e}") return None def run_backtest(): if not initialize(): logger.error("MT5初始化失败") return # 重置交易日志记录器 trade_logger.__init__() # 根据配置选择获取数据的方式 if USE_DATE_RANGE: rates = get_rates(SYMBOL, TIMEFRAME, BACKTEST_COUNT, BACKTEST_START_DATE, BACKTEST_END_DATE) else: rates = get_rates(SYMBOL, TIMEFRAME, BACKTEST_COUNT) if rates is None: logger.error("获取历史数据失败") shutdown() return logger.info(f"初始资金: {INITIAL_CAPITAL}") df = pd.DataFrame(rates) engine = BacktestEngine(df) # 使用动态权重管理器（需要在MT5连接状态下获取市场状态） risk_controller = RiskController() weight_manager = DynamicWeightManager(risk_controller) # 获取当前策略配置（回测时使用固定权重或模拟动态权重） strategies_with_weights = weight_manager.get_current_strategies_and_weights() weight_info = weight_manager.get_weight_info() logger.info(f"回测使用市场状态: {weight_info['market_state']}") # 获取市场状态后关闭MT5连接 shutdown() signals_list = [] weights = [] strategy_names = [] for strat, weight in strategies_with_weights: try: logger.info(f"回测策略：{strat.__class__.__module__}, 权重: {weight:.2f}") signals = engine.run_strategy(strat) signals_list.append(signals) weights.append(weight) strategy_names.append(strat.__class__.__module__) except Exception as e: logger.exception(f"回测策略 {strat.__class__.__module__} 时出错：{e}") combined_signal = engine.combine_signals(signals_list, weights) # 使用带交易记录的收益率计算 cum_ret = engine.calc_returns_with_trades(combined_signal, SYMBOL) final_capital = INITIAL_CAPITAL * (1 + cum_ret.iloc[-1]) logger.info("策略组合回测完成") logger.info(f"最终资金: {final_capital:.2f}") logger.info(f"总收益率: {(final_capital/INITIAL_CAPITAL - 1):.2%}") logger.info("使用的策略权重:") for name, weight in zip(strategy_names, weights): logger.info(f" {name}: {weight:.2f}") logger.info("最近收益率:") logger.info(cum_ret.tail()) # 打印详细的交易记录 logger.info("\n" + "="*50) logger.info("详细交易记录:") trade_logger.print_all_trades() # 保存交易记录到文件 trade_logger.save_to_csv("backtest_trades.csv") trade_logger.save_to_json("backtest_trades.json") if __name__ == "__main__": # --- 选择运行模式 --- # 1. 运行一次回测 (使用config.py中的默认权重) run_backtest() # 2. 运行实盘交易 (使用config.py中的默认权重) # run_realtime() # 3. 运行遗传算法优化，寻找最佳权重 # run_optimizer()