206 lines
6.9 KiB
Python
206 lines
6.9 KiB
Python
"""
|
||
山寨币策略回测脚本
|
||
对 DB 中所有有完整入场方案（stop_loss/tp1/tp2）的推荐做模拟跟踪。
|
||
"""
|
||
import json
|
||
import os
|
||
import sqlite3
|
||
import sys
|
||
from datetime import datetime
|
||
from pathlib import Path
|
||
|
||
import ccxt
|
||
import pandas as pd
|
||
|
||
REPO_ROOT = Path(__file__).resolve().parents[1]
|
||
sys.path.insert(0, str(REPO_ROOT))
|
||
|
||
exchange = ccxt.binance({'enableRateLimit': True})
|
||
DB = os.getenv("ALPHAX_DB_PATH", str(REPO_ROOT / "data" / "altcoin_monitor.db"))
|
||
OUTPUT_PATH = REPO_ROOT / "reports" / "backtest_result.json"
|
||
|
||
|
||
def fetch_klines_since(symbol, timeframe, since_ms, limit=500):
|
||
"""Fetch K-lines from a specific timestamp."""
|
||
try:
|
||
ohlcv = exchange.fetch_ohlcv(symbol, timeframe, since=since_ms, limit=limit)
|
||
if not ohlcv:
|
||
return None
|
||
df = pd.DataFrame(ohlcv, columns=['timestamp', 'open', 'high', 'low', 'close', 'volume'])
|
||
df['timestamp'] = pd.to_datetime(df['timestamp'], unit='ms')
|
||
return df
|
||
except Exception as e:
|
||
print(f" fetch_klines error for {symbol}: {e}")
|
||
return None
|
||
|
||
|
||
def simulate_trade(rec, klines_df):
|
||
"""Simulate one trade: walk through K-lines, check TP/stop."""
|
||
entry_price = float(rec['entry_price'])
|
||
stop_loss = float(rec['stop_loss'] or 0)
|
||
tp1 = float(rec['tp1'] or 0)
|
||
tp2 = float(rec['tp2'] or 0)
|
||
|
||
if stop_loss <= 0 or tp1 <= 0:
|
||
return {'result': 'no_entry_plan', 'exit_price': 0, 'exit_time': '', 'pnl_pct': 0, 'hours': 0}
|
||
|
||
result = 'expired'
|
||
exit_price = entry_price
|
||
exit_time = ''
|
||
max_profit_pct = 0
|
||
max_loss_pct = 0
|
||
|
||
for _, row in klines_df.iterrows():
|
||
high = float(row['high'])
|
||
low = float(row['low'])
|
||
close = float(row['close'])
|
||
ts = row['timestamp']
|
||
|
||
max_profit_pct = max(max_profit_pct, (high / entry_price - 1) * 100)
|
||
max_loss_pct = min(max_loss_pct, (low / entry_price - 1) * 100)
|
||
|
||
# Check TP2 first (higher target)
|
||
if tp2 > 0 and high >= tp2:
|
||
result = 'hit_tp2'
|
||
exit_price = tp2
|
||
exit_time = str(ts)
|
||
break
|
||
|
||
# Check TP1
|
||
if tp1 > 0 and high >= tp1:
|
||
result = 'hit_tp1'
|
||
exit_price = tp1
|
||
exit_time = str(ts)
|
||
break
|
||
|
||
# Check stop loss
|
||
if stop_loss > 0 and low <= stop_loss:
|
||
result = 'stopped_out'
|
||
exit_price = stop_loss
|
||
exit_time = str(ts)
|
||
break
|
||
|
||
pnl_pct = round((exit_price / entry_price - 1) * 100, 2)
|
||
|
||
# Calculate holding hours
|
||
if exit_time:
|
||
try:
|
||
et = datetime.fromisoformat(exit_time)
|
||
rt = datetime.fromisoformat(rec['rec_time'])
|
||
hours = round((et - rt).total_seconds() / 3600, 1)
|
||
except:
|
||
hours = 0
|
||
else:
|
||
hours = 0
|
||
|
||
return {
|
||
'result': result,
|
||
'exit_price': round(exit_price, 6),
|
||
'exit_time': exit_time,
|
||
'pnl_pct': pnl_pct,
|
||
'max_profit_pct': round(max_profit_pct, 2),
|
||
'max_loss_pct': round(max_loss_pct, 2),
|
||
'hours': hours,
|
||
}
|
||
|
||
|
||
def main():
|
||
conn = sqlite3.connect(DB)
|
||
conn.row_factory = sqlite3.Row
|
||
|
||
# Only backtest "爆发" with full entry plans
|
||
rows = conn.execute("""
|
||
SELECT id, symbol, rec_time, rec_state, rec_score, entry_price,
|
||
stop_loss, tp1, tp2, status, entry_plan_json, signals, sector
|
||
FROM recommendation
|
||
WHERE stop_loss > 0 AND tp1 > 0
|
||
ORDER BY id
|
||
""").fetchall()
|
||
|
||
conn.close()
|
||
|
||
print(f"回测样本: {len(rows)} 条 (有完整入场方案)\n")
|
||
|
||
results = []
|
||
wins = losses = expired_count = 0
|
||
total_pnl = 0
|
||
max_win = -999
|
||
max_loss = 999
|
||
|
||
for i, rec in enumerate(rows, 1):
|
||
symbol = rec['symbol']
|
||
rec_time = datetime.fromisoformat(rec['rec_time'])
|
||
since_ms = int(rec_time.timestamp() * 1000)
|
||
|
||
print(f"[{i}/{len(rows)}] {symbol} rec_time={rec['rec_time'][:19]} "
|
||
f"entry={rec['entry_price']} stop={rec['stop_loss']} tp1={rec['tp1']} tp2={rec['tp2']}", end='')
|
||
|
||
klines = fetch_klines_since(symbol, '15m', since_ms, limit=2000)
|
||
|
||
if klines is None or len(klines) < 2:
|
||
print(" → 数据不足，跳过")
|
||
continue
|
||
|
||
sim = simulate_trade(rec, klines)
|
||
results.append({**sim, 'symbol': symbol, 'rec_time': rec['rec_time'],
|
||
'entry_price': rec['entry_price'], 'rec_state': rec['rec_state']})
|
||
|
||
tag = {'hit_tp1': '🟢', 'hit_tp2': '🟢🟢', 'stopped_out': '🔴', 'expired': '⏰', 'no_entry_plan': '❓'}
|
||
print(f" → {tag.get(sim['result'], '?')} {sim['result']} pnl={sim['pnl_pct']}% "
|
||
f"max_profit={sim['max_profit_pct']}% max_loss={sim['max_loss_pct']}% {sim['hours']}h")
|
||
|
||
if sim['result'] in ('hit_tp1', 'hit_tp2'):
|
||
wins += 1
|
||
total_pnl += sim['pnl_pct']
|
||
max_win = max(max_win, sim['pnl_pct'])
|
||
elif sim['result'] == 'stopped_out':
|
||
losses += 1
|
||
total_pnl += sim['pnl_pct']
|
||
max_loss = min(max_loss, sim['pnl_pct'])
|
||
else:
|
||
expired_count += 1
|
||
|
||
print(f"\n{'='*60}")
|
||
print(f"回测汇总 (n={len(results)})")
|
||
print(f"{'='*60}")
|
||
print(f"止盈(TP): {wins} 笔")
|
||
print(f"止损: {losses} 笔")
|
||
print(f"未触达: {expired_count} 笔")
|
||
closed = wins + losses
|
||
if closed > 0:
|
||
print(f"胜率: {wins}/{closed} = {round(wins/closed*100,1)}%")
|
||
print(f"平均盈亏: {round(total_pnl/closed, 2)}%")
|
||
print(f"最大盈利: {max_win}%")
|
||
print(f"最大亏损: {max_loss}%")
|
||
print(f"盈亏比: {round(abs(max_win/max_loss) if max_loss != 0 else 99, 1)}")
|
||
print(f"{'='*60}")
|
||
|
||
# Detail table
|
||
print(f"\n{'symbol':<14} {'time':<17} {'result':<14} {'pnl':>7} {'max+':>7} {'max-':>7} {'h':>5}")
|
||
print("-" * 70)
|
||
for r in results:
|
||
print(f"{r['symbol']:<14} {r['rec_time'][:16]:<17} {r['result']:<14} "
|
||
f"{r['pnl_pct']:>6.1f}% {r['max_profit_pct']:>6.1f}% {r['max_loss_pct']:>6.1f}% {r['hours']:>5.1f}")
|
||
|
||
# Save to JSON for HTML report
|
||
OUTPUT_PATH.parent.mkdir(parents=True, exist_ok=True)
|
||
with open(OUTPUT_PATH, 'w', encoding='utf-8') as f:
|
||
json.dump({
|
||
'generated_at': datetime.now().isoformat(),
|
||
'total': len(results),
|
||
'wins': wins,
|
||
'losses': losses,
|
||
'expired': expired_count,
|
||
'win_rate': round(wins/closed*100, 1) if closed > 0 else 0,
|
||
'avg_pnl': round(total_pnl/closed, 2) if closed > 0 else 0,
|
||
'max_win': max_win,
|
||
'max_loss': max_loss,
|
||
'details': [{k: str(v) if isinstance(v, (datetime, pd.Timestamp)) else v
|
||
for k, v in r.items()} for r in results],
|
||
}, f, ensure_ascii=False, indent=2, default=str)
|
||
print(f"\n结果已保存: {OUTPUT_PATH}")
|
||
|
||
|
||
if __name__ == '__main__':
|
||
main()
|