import requests
import json
from datetime import datetime, timedelta
from decimal import Decimal
import math
class HyperliquidAnalyzer:
def __init__(self, address):
self.address = address
self.base_url = "https://api.hyperliquid.xyz/info"
def post_request(self, data):
"""发送 POST 请求到 Hyperliquid API"""
headers = {'Content-Type': 'application/json'}
response = requests.post(self.base_url, json=data, headers=headers)
return response.json()
def get_user_state(self):
"""获取用户当前状态（包括持仓和账户信息）"""
data = {
"type": "clearinghouseState",
"user": self.address
}
return self.post_request(data)
def get_user_fills(self):
"""获取用户历史成交记录"""
data = {
"type": "userFills",
"user": self.address
}
return self.post_request(data)
def get_user_funding_history(self):
"""获取用户资金费率历史"""
data = {
"type": "userFunding",
"user": self.address
}
return self.post_request(data)
def get_snapshot_and_history(self):
"""获取账户快照和历史数据"""
data = {
"type": "userSnapshot",
"user": self.address
}
return self.post_request(data)
def calculate_profit_factor(self, fills_data, user_state):
"""
计算 Profit Factor
Profit Factor = 总盈利 / 总亏损
"""
total_profit = Decimal('0')
total_loss = Decimal('0')
# 1. 统计已实现盈亏（从历史成交中计算）
for fill in fills_data:
if 'closedPnl' in fill and fill['closedPnl'] != '0':
closed_pnl = Decimal(str(fill['closedPnl']))
if closed_pnl > 0:
total_profit += closed_pnl
else:
total_loss += abs(closed_pnl)
# 2. 统计未实现盈亏（从当前持仓中获取）
if user_state and 'assetPositions' in user_state:
for position in user_state['assetPositions']:
if 'position' in position and 'unrealizedPnl' in position['position']:
unrealized_pnl = Decimal(str(position['position']['unrealizedPnl']))
if unrealized_pnl > 0:
total_profit += unrealized_pnl
else:
total_loss += abs(unrealized_pnl)
# 3. 计算 Profit Factor
if total_loss != 0:
profit_factor = float(total_profit / total_loss)
return round(profit_factor, 2)
else:
return "1000+" if total_profit > 0 else 0
def calculate_sharpe_ratio(self, account_history, period='all'):
"""
计算 Sharpe 比率
Sharpe = 平均日收益率 / 日收益率标准差
"""
if not account_history or len(account_history) < 2:
return 0
# 1. 计算平均账户价值
total_value = sum(Decimal(str(item['accountValue']))
for item in account_history)
avg_value = total_value / len(account_history)
if avg_value == 0:
return 0
# 2. 计算每日收益率（百分比）
daily_returns = []
for i in range(1, len(account_history)):
prev_pnl = Decimal(str(account_history[i-1].get('pnl', 0)))
curr_pnl = Decimal(str(account_history[i].get('pnl', 0)))
daily_pnl = curr_pnl - prev_pnl
# 日收益率 = (日PNL / 平均账户价值) * 100
daily_return = float((daily_pnl / avg_value) * 100)
daily_returns.append(daily_return)
if len(daily_returns) < 2:
return 0
# 3. 计算平均日收益率
mean_return = sum(daily_returns) / len(daily_returns)
# 4. 计算标准差
variance = sum((r - mean_return) ** 2 for r in daily_returns) / (len(daily_returns) - 1)
std_dev = math.sqrt(variance)
# 5. 计算 Sharpe 比率
if std_dev == 0:
return 0
sharpe = mean_return / std_dev
return round(sharpe, 2)
def get_account_value_history(self):
"""
获取账户价值历史
注意：Hyperliquid 可能需要通过多个 API 调用来构建完整的历史数据
"""
# 尝试获取快照数据
snapshot_data = self.get_snapshot_and_history()
# 如果 API 不直接提供历史数据，我们需要从成交记录重建
fills = self.get_user_fills()
# 构建账户历史
account_history = []
# 按时间排序成交记录
if fills:
sorted_fills = sorted(fills, key=lambda x: x.get('time', 0))
# 模拟账户价值变化
running_pnl = Decimal('0')
daily_data = {}
for fill in sorted_fills:
timestamp = fill.get('time', 0)
date = datetime.fromtimestamp(timestamp / 1000).date()
closed_pnl = Decimal(str(fill.get('closedPnl', 0)))
running_pnl += closed_pnl
if date not in daily_data:
daily_data[date] = {
'accountValue': 0,
'pnl': float(running_pnl),
'time': timestamp
}
else:
daily_data[date]['pnl'] = float(running_pnl)
# 获取当前账户价值
user_state = self.get_user_state()
if user_state and 'marginSummary' in user_state:
current_value = Decimal(str(user_state['marginSummary'].get('accountValue', 0)))
# 填充账户价值
for date in sorted(daily_data.keys()):
daily_data[date]['accountValue'] = float(current_value)
account_history.append(daily_data[date])
return account_history
def analyze(self):
"""执行完整分析"""
print(f"分析地址: {self.address}\n")
print("=" * 60)
# 1. 获取数据
print("正在获取数据...")
user_state = self.get_user_state()
fills = self.get_user_fills()
account_history = self.get_account_value_history()
# 2. 显示基本信息
if user_state and 'marginSummary' in user_state:
margin_summary = user_state['marginSummary']
print(f"\n账户基本信息:")
print(f" 账户价值: ${float(margin_summary.get('accountValue', 0)):,.2f}")
print(f" 已用保证金: ${float(margin_summary.get('totalMarginUsed', 0)):,.2f}")
print(f" 未实现盈亏: ${float(margin_summary.get('totalNtlPos', 0)):,.2f}")
# 3. 计算 Profit Factor
print(f"\n计算 Profit Factor...")
profit_factor = self.calculate_profit_factor(fills, user_state)
print(f" Profit Factor: {profit_factor}")
# 4. 计算 Sharpe 比率
print(f"\n计算 Sharpe 比率...")
sharpe_ratio = self.calculate_sharpe_ratio(account_history)
print(f" Sharpe 比率: {sharpe_ratio}")
# 5. 显示持仓信息
if user_state and 'assetPositions' in user_state:
positions = user_state['assetPositions']
if positions:
print(f"\n当前持仓 ({len(positions)} 个):")
for pos in positions:
coin = pos['position']['coin']
szi = float(pos['position']['szi'])
unrealized_pnl = float(pos['position']['unrealizedPnl'])
print(f" {coin}: 数量={szi:.4f}, 未实现盈亏=${unrealized_pnl:.2f}")
# 6. 显示交易统计
if fills:
print(f"\n交易统计:")
print(f" 总成交数: {len(fills)}")
profitable_trades = sum(1 for f in fills if float(f.get('closedPnl', 0)) > 0)
total_trades = len([f for f in fills if 'closedPnl' in f and f['closedPnl'] != '0'])
if total_trades > 0:
win_rate = (profitable_trades / total_trades) * 100
print(f" 盈利交易: {profitable_trades}/{total_trades}")
print(f" 胜率: {win_rate:.2f}%")
print("\n" + "=" * 60)
return {
'profit_factor': profit_factor,
'sharpe_ratio': sharpe_ratio,
'user_state': user_state,
'fills': fills
}
# 使用示例
if __name__ == "__main__":
address = "0x7717a7a245d9f950e586822b8c9b46863ed7bd7e"
analyzer = HyperliquidAnalyzer(address)
results = analyzer.analyze()
# 可以进一步处理结果
print(f"\n最终结果:")
print(f"Profit Factor: {results['profit_factor']}")
print(f"Sharpe Ratio: {results['sharpe_ratio']}")
drake
Posted on
从Hyperliquid获取交易数据，计算Sharpe 和 Profit Factor
For further actions, you may consider blocking this person and/or reporting abuse
Top comments (0)