#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
深入诊断 - 检查卖出逻辑和移动止损问题
"""

import pandas as pd
import numpy as np
import sys
sys.path.insert(0, '/root/.openclaw/workspace/quant')

# 加载数据
df = pd.read_csv('/root/.openclaw/workspace/quant/cyb50_baostock.csv')
df['date'] = pd.to_datetime(df['date'])
df = df.set_index('date').sort_index()

for col in ['open', 'high', 'low', 'close', 'volume']:
    df[col] = pd.to_numeric(df[col], errors='coerce')

# 计算指标
df['ma10'] = df['close'].rolling(window=10).mean()
df['ma30'] = df['close'].rolling(window=30).mean()
df['high_20'] = df['high'].rolling(window=20).max()
df['low_20'] = df['low'].rolling(window=20).min()
df['ret_10'] = df['close'].pct_change(periods=10)

# 买入条件
buy_cond = (
    (df['close'] > df['ma10']) & 
    (df['ma10'] > df['ma30']) & 
    (df['close'] >= df['high_20'] * 0.995) & 
    (df['ret_10'] > 0.02)
)

# 卖出条件1: 跌破MA30
sell_cond1 = df['close'] < df['ma30']
# 卖出条件2: 创20日新低
sell_cond2 = df['close'] <= df['low_20'] * 1.005

print("="*60)
print("🔍 卖出条件分析")
print("="*60)

print(f"\n卖出条件1 (跌破MA30): {sell_cond1.sum()} 天")
print(f"卖出条件2 (创20日新低): {sell_cond2.sum()} 天")
print(f"任一条件触发: {(sell_cond1 | sell_cond2).sum()} 天")

# 检查卖出条件1和2的重叠
both_sell = sell_cond1 & sell_cond2
print(f"两个条件同时触发: {both_sell.sum()} 天")

# 模拟交易，详细检查每次卖出原因
print("\n📈 详细交易分析:")

df['signal'] = 0
df.loc[buy_cond, 'signal'] = 1
df.loc[sell_cond1 | sell_cond2, 'signal'] = -1

position = 0
entry_price = 0
peak_price = 0
capital = 1000000
trades = []

for i in range(30, len(df)):
    date = df.index[i]
    price = df['close'].iloc[i]
    signal = df['signal'].iloc[i]
    
    # 检查各种卖出原因
    sell_ma30 = price < df['ma30'].iloc[i]
    sell_low20 = price <= df['low_20'].iloc[i] * 1.005
    
    # 移动止损检查
    stop_loss_triggered = False
    stop_loss_price = None
    if position > 0:
        if price > peak_price:
            peak_price = price
        stop_loss_price = peak_price * 0.90
        if price < stop_loss_price:
            signal = -1
            stop_loss_triggered = True
    
    # 买入
    if signal == 1 and position == 0:
        position = 1
        entry_price = price
        peak_price = price
        trades.append({
            'date': date,
            'action': 'BUY',
            'price': price,
            'capital': capital
        })
    
    # 卖出
    elif signal == -1 and position == 1:
        pnl = (price / entry_price - 1) * capital
        capital += pnl
        
        hold_days = (date - trades[-1]['date']).days
        
        exit_reasons = []
        if stop_loss_triggered:
            exit_reasons.append(f"移动止损({peak_price:.2f}→{stop_loss_price:.2f})")
        if sell_ma30:
            exit_reasons.append("跌破MA30")
        if sell_low20:
            exit_reasons.append("创20日新低")
        
        trades.append({
            'date': date,
            'action': 'SELL',
            'price': price,
            'capital': capital,
            'pnl': pnl,
            'return_pct': (price / entry_price - 1) * 100,
            'reasons': ' + '.join(exit_reasons),
            'hold_days': hold_days
        })
        
        position = 0

# 显示最近10次卖出的详细原因
print("\n最近10次卖出详情:")
sell_trades = [t for t in trades if t['action'] == 'SELL']
for t in sell_trades[-10:]:
    print(f"  {t['date'].strftime('%Y-%m-%d')}: {t['return_pct']:+.2f}% ({t['hold_days']}天)")
    print(f"    原因: {t['reasons']}")

# 统计各卖出原因的占比
print("\n📊 卖出原因统计:")
reasons_count = {}
for t in sell_trades:
    for reason in t['reasons'].split(' + '):
        reasons_count[reason] = reasons_count.get(reason, 0) + 1

for reason, count in sorted(reasons_count.items(), key=lambda x: x[1], reverse=True):
    print(f"  {reason}: {count}次 ({count/len(sell_trades)*100:.1f}%)")

print("\n" + "="*60)

# 检查移动止损的触发情况
print("\n📉 移动止损检查:")
stop_loss_trades = [t for t in sell_trades if '移动止损' in t['reasons']]
print(f"  移动止损触发次数: {len(stop_loss_trades)}")

if stop_loss_trades:
    avg_return = np.mean([t['return_pct'] for t in stop_loss_trades])
    print(f"  移动止损平均收益: {avg_return:.2f}%")
    
    print("\n  最近5次移动止损:")
    for t in stop_loss_trades[-5:]:
        print(f"    {t['date'].strftime('%Y-%m-%d')}: {t['return_pct']:+.2f}%")