-
Notifications
You must be signed in to change notification settings - Fork 6
/
main.py
202 lines (159 loc) · 7.77 KB
/
main.py
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
import csv
from datetime import date, datetime, timedelta
from multiprocessing import Pool
from decimal import Decimal
import multiprocessing
import pickle
import pathlib
import os
from dataTypes import *
from simulation import *
import hypothesis
# ********* Saba *********
import pandas as pd
import numpy as np
# ********* Saba *********
class HypothesisTesterStupid:
def __init__(self, startingDate, shortTermWindow, endingDate, shortTermData, longTermData, startingCash):
self.startingDate = startingDate
self.shortTermWindow = shortTermWindow
self.endingDate = endingDate
self.shortTermData = shortTermData
self.longTermData = longTermData
self.startingCash = startingCash
def testHypothesis(self, hypothesis):
assert callable(hypothesis)
return simulation(self.startingDate, self.shortTermWindow, self.endingDate, self.shortTermData, self.longTermData, hypothesis, self.startingCash)["success"]
def DataFrame(data, filename):
assert isinstance(filename, str)
assert ".xlsx" in filename
short = pd.DataFrame([val.__dict__ for val in data['short']]).dropna()
long = pd.DataFrame([val.__dict__ for val in data['long']]).dropna()
#plt.plot_date(short.date,short.safeMeanPrice, linestyle='solid', marker='')
#check for linearity with scatter plots
#plt.scatter(short.safeMeanDeltaVolumePerTransaction, short.safeMeanPrice)
#create additional data
# def signMomentum(sign):
# momentum = []
# mom = 0
# LastSignPositive = True
# for index, sign in enumerate(sign):
# if sign == 0: #concern, we should probably count very small differences as 0 instead of increasing momentum
# mom = 0
# if sign == 1:
# if LastSignPositive == True:
# mom += 1
# else:
# mom = 1
# LastSignPositive = True
# if sign == -1:
# if LastSignPositive == True:
# mom = -1
# else:
# mom -= 1
# LastSignPositive = False
# if np.isnan(sign):
# momentum.append(np.nan)
# else:
# momentum.append(mom)
# return momentum
# short['deltaPrice1Row'] = short['safeMeanPrice'].diff()
# short['deltaPrice5Row'] = short['safeMeanPrice'].diff(periods=5)
# short['deltaPrice10Row'] = short['safeMeanPrice'].diff(periods=10)
# short['deltaPrice25Row'] = short['safeMeanPrice'].diff(periods=25)
# short['deltaPrice50Row'] = short['safeMeanPrice'].diff(periods=50)
# short['deltaPrice100Row'] = short['safeMeanPrice'].diff(periods=100)
# short['deltaPrice200Row'] = short['safeMeanPrice'].diff(periods=200)
# short['deltaPrice500Row'] = short['safeMeanPrice'].diff(periods=500)
# #^^ if you graph all of these with the above scatter, you will find linearity starting to increase at >100 rows ^^ which makes some sense
# short['deltaSign1Row'] = np.sign(short['deltaPrice1Row'])
# short['signMomentum1Row'] = signMomentum(short['deltaSign1Row'])
# short['deltaSign500Row'] = np.sign(short['deltaPrice500Row'])
# short['signMomentum500Row'] = signMomentum(short['deltaSign500Row'])
# short['std5Row'] = short['safeMeanPrice'].rolling(5).std()
# short['std100Row'] = short['safeMeanPrice'].rolling(100).std()
# short['volume5Row'] = short['volume'].rolling(5).sum()
# short['volume100Row'] = short['volume'].rolling(100).sum()
# short['volume500Row'] = short['volume'].rolling(500).sum()
# short['movingAverage5'] = short['safeMeanPrice'].rolling(5).sum()/5
# short['movingAverage50'] = short['safeMeanPrice'].rolling(50).sum()/50
# short['movingAverage500'] = short['safeMeanPrice'].rolling(500).sum()/500
# plt.scatter(short.deltaPrice500Row, short.safeMeanPrice)
# plt.scatter(short.signMomentum1Row, short.safeMeanPrice)
#linear-ish combos: std5row, safemeanprice
#signmomentum500row, deltaprice500row -- also on log scale
#signmomentum500row * volume500Row, deltaprice500row
with pd.ExcelWriter(filename) as writer:
short.to_excel(writer, sheet_name='short')
long.to_excel(writer, sheet_name='long')
return short,long
def convertDataListMode(convertMe):
assert isinstance(convertMe, list)
startingDate = datetime(year=2018, month=1, day=2, hour=0, minute=0, second=0)
endingDate = datetime(year=2020, month=12, day=31)
shortTermWindow = timedelta(hours=1)
longTermWindow = timedelta(hours=24)
for filename in convertMe:
try:
file = open(filename) # Check to see if exists and openable
file.close()
except:
raise Exception(f"{convertMe} file not found!")
for filename in convertMe:
data = getData(filename, startingDate, endingDate, shortTermWindow, longTermWindow)
excelName = filename.replace(".csv", ".xlsx")
DataFrame(data, excelName)
def main():
with open('DAILY.csv', newline='') as f:
reader = csv.reader(f)
dlist = list(reader)
format_d = r'%Y-%m-%d %H:%M:%S'
for x in range(1, len(dlist)):
dlist[x][0] = datetime.strptime(dlist[x][0], format_d)
dlist[x][1] = Decimal(dlist[x][1])
with open('hourlyfixed.csv', newline='') as f:
reader = csv.reader(f)
hlist = list(reader)
format_h = r'%Y-%m-%d %H:%M:%S'
for x in range(1, len(hlist)):
hlist[x][0] = datetime.strptime(hlist[x][0], format_h)
hlist[x][1] = Decimal(hlist[x][1])
THREAD_COUNT = os.cpu_count()
print("I have {} cores".format(THREAD_COUNT))
FILENAME = "XMRUSD.csv"
# print(hypothesisTester(FILENAME, hypothesis.equationMethod))
startingDate = datetime(year=2018, month=1, day=2, hour=0, minute=0, second=0)
endingDate = datetime(year=2020, month=12, day=31)
shortTermWindow = timedelta(hours=1)
longTermWindow = timedelta(hours=24)
# import time
# start = time.time()
data = getData(FILENAME, startingDate, endingDate, shortTermWindow, longTermWindow)
# shortdf,longdf = DataFrame(data)
# print("Took {} seconds".format(time.time() - start))
# for x in longTerm:
# print("L RANGE:", x.date, " - ", x.endDate)
# # for x in shortTerm:
# # print("S RANGE:", x.date, " - ", x.endDate)
#add hlist, dlist at end
result = simulation(startingDate, shortTermWindow, endingDate, data["short"], data["long"], hypothesis.testing, Decimal(1_000), hlist,dlist)
def ParameterPrint():
return "{:.3f}% success\n{:.3f}% market risk\n{} Buys\n{} Sells"\
.format(result["success"],result["marketRisk"],result["numberOfBuys"],result["numberOfSells"])
print(ParameterPrint())
simulationPlotter(data["long"], result)
# print(hypothesisTester(FILENAME, hypothesis.hold))
# hypothesisTester = HypothesisTester(startingDate, shortTermWindow, endingDate, data["short"], data["long"], Decimal(1_000)).testHypothesis
# inputList = np.arange(.05, 3, .05)
# hypothesisList = [hypothesis.HypothesisVariation(hypothesis.bollingerBandsSafe, bollinger_number_of_stdev=i).hypothesis for i in inputList]
# pool = multiprocessing.Pool(THREAD_COUNT)
# results = pool.map(hypothesisTester, hypothesisList)
# associatedDict = {}
# for i in range(len(results)):
# associatedDict[inputList[i]] = results[i]
# print("Stdev {} : {}% profit".format(inputList[i], results[i]))
# bestResult = max(results)
# bestResultIndex = results.index(bestResult)
# print("Best: {} : {}% profit".format(inputList[bestResultIndex], results[bestResultIndex]))
if __name__ == "__main__":
main()