forked from moonnejs/PythonLab
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathcalcFunction.py
More file actions
256 lines (227 loc) · 9.98 KB
/
calcFunction.py
File metadata and controls
256 lines (227 loc) · 9.98 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
# encoding: UTF-8
import numba as nb
import numpy as np
import pandas as pd
from ctaBase import *
from datetime import timedelta
from numpy import inf,nan,float32,float64
from dataFunction import loadStrategyData
#----------------------------------------------------------------------
def formatNumber(n):
"""格式化数字到字符串"""
rn = round(n, 2) # 保留两位小数
return format(rn, ',') # 加上千分符
#------------------------------------------------
def calc_sharpe_ratio(returns, periods=250):
"""计算夏普比率"""
return np.sqrt(periods) * (np.mean(returns)) / np.std(returns)
#------------------------------------------------
def calc_drawdowns(caps):
"""计算最大回测和最大回测周期"""
hwm = [0]
eq_idxs = caps.index.values
eq_idx = len(caps.index.values)
drawdown = pd.Series(index = range(1,eq_idx))
duration = pd.Series(index = range(0,eq_idx))
duration[0] = 0
for t in range(1,eq_idx):
cur_hwm = max(hwm[t-1], caps[eq_idxs[t]])
hwm.append(cur_hwm)
drawdown[t]= hwm[t] - caps[eq_idxs[t]]
duration[t]= 0 if drawdown[t] == 0 else duration[t-1] + 1
return drawdown.max(), int(duration.max())
#------------------------------------------------
def reshape_min(datas):
"""数据按分钟切片"""
pnl_idxs = datas.index
start = pnl_idxs[0].replace(second = 0,microsecond = 0)
end = pnl_idxs[-1].replace(second = 0,microsecond = 0)
minutes = int((end-start).total_seconds()/60)+2
min_idx = [start + timedelta(minutes=i) for i in xrange(minutes)]
index = 0
datas_min = pd.Series(index = min_idx)
for i in xrange(minutes):
datas_min[start+timedelta(minutes=i)] = 0
while index < len(pnl_idxs.values) and start+timedelta(minutes=i) > pnl_idxs[index]:
datas_min[start+timedelta(minutes=i)] += datas[pnl_idxs.values[index]]
index += 1
return datas_min
# 计算结算表现
#------------------------------------------------
def calcPerf(times,pnls,fees):
"""数据按分钟切片"""
pnlList = [] # 每笔盈亏序列
capital = 0 # 资金
maxCapital = 0 # 资金最高净值
drawdown = 0 # 回撤
totalResult = 0 # 总成交数量
totalCommission = 0 # 总手续费
timeList = [] # 时间序列
capitalList = [] # 盈亏汇总的时间序列
drawdownList = [] # 回撤的时间序列
winningResult = 0 # 盈利次数
losingResult = 0 # 亏损次数
totalWinning = 0 # 总盈利金额
totalLosing = 0 # 总亏损金额
for t,pnl,fee in zip(times,pnls,fees):
if pnl !=0 :
capital += pnl
maxCapital = max(capital, maxCapital)
drawdown = round(capital,2)
pnlList.append(pnl)
# 交易的时间戳使用平仓时间
timeList.append(t)
capitalList.append(capital)
drawdownList.append(drawdown)
totalResult += 1
totalCommission += fee
if pnl >= 0:
winningResult += 1
totalWinning += pnl
else:
losingResult += 1
totalLosing += pnl
# 计算盈亏相关数据
averageWinning = 0
averageLosing = 0
profitLossRatio = 0
winningRate = 0 if totalResult==0 else winningResult*1.0/totalResult*100
averageWinning = 0 if winningResult==0 else totalWinning/winningResult
averageLosing = 0 if losingResult==0 else totalLosing/losingResult
profitLossRatio = 0 if averageLosing==0 else -averageWinning/averageLosing
# 返回回测结果
d = {}
d['name'] = u'向量回测'
d['capital'] = round(capital,2)
d['maxCapital'] = maxCapital
d['drawdown'] = drawdown
d['totalResult'] = round(totalResult,2)
d['totalCommission'] = round(totalCommission,2)
d['timeList'] = timeList
d['pnlList'] = pnlList
d['capitalList'] = capitalList
d['drawdownList'] = drawdownList
d['winningRate'] = round(winningRate,2)
d['averageWinning'] = round(averageWinning,2)
d['averageLosing'] = round(averageLosing,2)
d['profitLossRatio'] = round(profitLossRatio,2)
d['datas'] = None
return d
@nb.autojit
#------------------------------------------------
def get_capital_np(markets,signals,size,commiRate,climit = 4, wlimit = 2, op=True):
"""使用numpy回测，标签的盈亏, op 表示是否延迟一个tick以后撮合"""
postions = np.zeros(len(signals))
actions = np.zeros(len(signals))
costs = np.zeros(len(signals))
pnls = np.zeros(len(signals))
fees = np.zeros(len(signals))
lastsignal = 0
lastpos = 0
lastcost = 0
num = 0
for num in range(1,len(signals)):
postions[num] = lastpos
actions[num] = 0
costs[num] = lastcost
pnls[num] = 0
# 止盈止损
if lastpos > 0 and \
(markets[num,1]<=lastcost-climit or markets[num,1]>=lastcost+wlimit):
postions[num] = 0
actions[num] = -1
costs[num] = 0
fees[num] = (markets[num,1]+lastcost)*size*commiRate
pnls[num] = (markets[num,1]-lastcost)*size-fees[num]
elif lastpos < 0 and \
(markets[num,0]>=lastcost+climit or markets[num,0]<=lastcost-wlimit):
postions[num] = 0
actions[num] = 1
costs[num] = 0
fees[num] = (markets[num,0]+lastcost)*size*commiRate
pnls[num] = (lastcost-markets[num,0])*size-fees[num]
# 开仓
if op:
lastsignal = signals[num]
if lastsignal > 0 and lastpos == 0:
postions[num] = 1
actions[num] = 1
costs[num] = markets[num,0]
elif lastsignal < 0 and lastpos == 0:
postions[num] = -1
actions[num] = -1
costs[num] = markets[num,1]
lastpos = postions[num]
lastcost = costs[num]
lastsignal = signals[num]
return pnls,actions,fees
#------------------------------------------------
def get_perf(datas,signals,size,commiRate):
"""并行回测，标签的盈亏"""
# 计算交易信号，并整理数据
predatas = datas[datas.columns.drop(['askPrice1','bidPrice1'])].values
datas = datas[['askPrice1','bidPrice1']]
datas['signals'] = signals
# 计算仓位信息
datas = datas[datas['signals']!=0]
datas['position'] = datas['signals'].diff()
datas = datas.dropna(axis=0, how='any')
posInfo = datas[datas['position']!=0]
# 根据仓位信息计算开仓成本
posInfo['cost'] = posInfo.apply(lambda x:x['position']*x['askPrice1'] if x['position']>0 else -x['position']*x['bidPrice1'],axis=1)
# 计算每笔的手续费，第一笔被重复计算
posInfo['feeP'] = abs(posInfo['cost'])*size*commiRate
# 计算每笔盈亏，第一笔是错误的
posInfo['pnl'] = (posInfo['cost']-posInfo['cost'].shift(1))/2
posInfo['pnl'] = posInfo.apply(lambda x:-x['pnl'] if x['position']>0 else x['pnl'],axis=1) - (posInfo['feeP'] + posInfo['feeP'].shift(1))/2
# 修正错误信息
posInfo['position'].iloc[0] = posInfo['position'].iloc[0]/2
posInfo['position'].iloc[-1] = posInfo['position'].iloc[-1]/2
posInfo['pnl'][0] = 0
# 计算总手续费和总资金
posInfo['fee'] = posInfo.apply(np.cumsum)['feeP']
posInfo['cap'] = posInfo.apply(np.cumsum)['pnl']
pnl_min = reshape_min(posInfo['pnl'])
cap_min = reshape_min(posInfo['cap'])
period = posInfo['pnl'].count()
mdd, ddt = calc_drawdowns(posInfo['cap'])
return posInfo['cap'][-1],mdd
#------------------------------------------------
def get_daily_rtn(strategyNames,strategyBases,startDate='20100101',endDate='20181030'):
"""获取每日盈亏"""
fields = ['name','date','pnl']
rtns = pd.DataFrame()
caps = pd.DataFrame()
for name,base in zip(strategyNames,strategyBases):
datas = loadStrategyData(CAPITAL_DB_NAME,name,startDate,endDate,fields)
datas['pnl']=datas['pnl']/base
rtns = pd.concat([rtns,datas],axis = 0)
datas=datas.set_index('date')
datas['cap']=datas.apply(np.cumsum)['pnl']
#datas.plot(kind='line',title = name)
datas.reset_index(drop=False,inplace=True)
caps = pd.concat([caps,datas],axis = 0)
rtn_table = pd.crosstab(rtns['date'],rtns['name'], values = rtns['pnl'], aggfunc = sum) # 一维表变为二维表
rtn_table.fillna(0, inplace = True)
cap_table = pd.crosstab(caps['date'],caps['name'], values = caps['cap'], aggfunc = sum) # 一维表变为二维表
cap_table.fillna(method='pad', inplace = True) # 将NaN置换为0
cap_table.fillna(0, inplace = True) # 将NaN置换为0
return rtn_table,cap_table
#plt.show()
#cap_table.head(20)
#------------------------------------------------
def get_best_wei(rtn_table,risk_aversion):
"""获取指定风险厌恶系数下的最优策略配置组合"""
from cvxopt import matrix, solvers
cov_mat = rtn_table.cov() * 250 # 协方差矩阵(1年250个交易日)
exp_rtn = rtn_table.mean() * 250 # 标的预期收益(1年250个交易日)
P = risk_aversion * matrix(cov_mat.values)
q = -1 * matrix(exp_rtn.values)
G = matrix(np.vstack((np.diag(np.ones(len(exp_rtn))),np.diag(-np.ones(len(exp_rtn))))))
h = matrix(np.array([np.ones(len(exp_rtn)),np.zeros(len(exp_rtn))]).reshape(len(exp_rtn)*2,1))
A = matrix(np.ones(len(exp_rtn)),(1,len(exp_rtn)))
b = matrix([1.0])
solvers.options['show_progress'] = True
sol = solvers.qp(P, q, G, h, A, b)
weis=pd.DataFrame(index=exp_rtn.index,data = np.round(sol['x'],2), columns = ['weight']) # 权重精确到小数点后两位
return weis