def run_optimizer(length, minutes, minutes_until_sale):
if length == '1m':
data_length_multiplier = 360 # 6 hours # 3x as long, but 12 sections
if length == '5m':
data_length_multiplier = 60*24*2/5 # 2 days
if length == '15m':
data_length_multiplier = 60*24*7/15 # 1 wk
if length == '30m':
data_length_multiplier = 60*24*14/30 # 2 wks
if length == '1h':
data_length_multiplier = 60*24*14/60 # 2 wks
if length == '2h':
data_length_multiplier = 60*24*30/120 # 1 month
if length == '6h':
data_length_multiplier = 60*24*45/360 # 1.5 months
if length == '12h':
data_length_multiplier = 60*24*90/720 # 3 months
if length == '1d':
data_length_multiplier = 60*24*120/1440 # 4 months
while True:
print('starting..')
################################################################################ RUN OPTIMIZER
symbols = ut.pickle_read('/home/ec2-user/environment/botfarming/Development/3_binance_btc_symbols.pklz')
symbols_trimmed = {}
total_btc_coins = 0
for s in symbols:
symbol = symbols[s]
if float(symbol['24hourVolume']) > 300:
total_btc_coins += 1
symbols_trimmed[s] = symbol
print('total symbols', total_btc_coins)
symbols_started = 0
for s in symbols_trimmed:
symbols_started += 1
print('symbols_started', symbols_started)
#get data in right format for save function (could fix this)
symbol = symbols_trimmed[s]
symbols_trimmed_one = {}
for s_one in symbols:
symbol_one = symbols[s_one]
if symbol_one['symbol'] == symbol['symbol']:
symbols_trimmed_one[s] = symbol_one
# 777
optimizing_array= [1,3,5,7,9,11,13,15]
end_time = int(time.time())
results = get_optimization_factors(optimizing_array, data_length_multiplier, end_time, symbol, symbols_trimmed_one, length, minutes, minutes_until_sale)
for i, result in enumerate(results):
optimization_factors_path = '/home/ec2-user/environment/botfarming/Development/optimization_factors/1_' + length + '_optimal_for_' + symbol['symbol'] + '_' + str(result['look_back']) + '.pklz'
if result['lowest_buy_factor'] != 1:
ut.pickle_write(optimization_factors_path, result)
print('###################################################################')
print('###################################################################')
print('######### LOOK_BACK', result['look_back'], '###########', symbol['symbol'])
print('lowest_buy_factor', result['lowest_buy_factor'])
print('highest_sale_factor', result['highest_sale_factor'])
else:
ut.pickle_write(optimization_factors_path, False)
# end optimization cycle, sleep
if length == '1d':
time.sleep(4*60*60)
if length == '12h':
time.sleep(3*60*60)
if length == '6h':
time.sleep(2*60*60)
if length == '2h':
time.sleep(1*60*60)
if length == '1h':
time.sleep(30*60)
if length == '30m':
time.sleep(20*60)
def run_optimizer_multi(lengths):
while True:
for length in lengths:
symbols = ut.pickle_read('/home/ec2-user/environment/botfarming/Development/3_binance_btc_symbols.pklz')
symbols_trimmed = {}
total_btc_coins = 0
for s in symbols:
symbol = symbols[s]
if float(symbol['24hourVolume']) > 300:
total_btc_coins += 1
symbols_trimmed[s] = symbol
print('total symbols', total_btc_coins)
if length == '1m':
data_length_multiplier = 360 # 6 hours # 3x as long, but 12 sections
minutes = 1
minutes_until_sale = 12
if length == '5m':
data_length_multiplier = 60*24*2/5 # 2 days
minutes = 5
minutes_until_sale = 12
if length == '15m':
data_length_multiplier = 60*24*7/15 # 1 wk
minutes = 15
minutes_until_sale = 12
if length == '30m':
data_length_multiplier = 60*24*14/30 # 2 wks
minutes = 30
minutes_until_sale = 12
if length == '1h':
data_length_multiplier = 60*24*14/60 # 2 wks
minutes = 60
minutes_until_sale = 12
if length == '2h':
data_length_multiplier = 60*24*30/120 # 1 month
minutes = 60 * 2
minutes_until_sale = 8
if length == '6h':
data_length_multiplier = 60*24*45/360 # 1.5 months
minutes = 60 * 6
minutes_until_sale = 4
if length == '12h':
data_length_multiplier = 60*24*90/720 # 3 months
minutes = 60 * 12
minutes_until_sale = 3
if length == '1d':
data_length_multiplier = 60*24*120/1440 # 4 months
minutes = 60 * 24
minutes_until_sale = 2
hours_per_period = round(float(data_length_multiplier*minutes/60.0), 3)
symbols_started = 0
for s in symbols_trimmed:
symbols_started += 1
print('########## symbols_started', symbols_started, 'for', length, ut.get_time())
#get data in right format for save function (could fix this)
symbol = symbols_trimmed[s]
symbols_trimmed_one = {}
for s_one in symbols:
symbol_one = symbols[s_one]
if symbol_one['symbol'] == symbol['symbol']:
symbols_trimmed_one[s] = symbol_one
# 777
optimizing_array = [1,3,5,7,9,11,13,15]
results_by_look_back = {}
for lb in optimizing_array:
results_by_look_back[str(lb)] = []
step_back_periods = 12
for step_back in range(0, step_back_periods):
end_time = int(time.time()) - data_length_multiplier * step_back * minutes * 60
results = get_optimization_factors(optimizing_array, data_length_multiplier, end_time, symbol, symbols_trimmed_one, length, minutes, minutes_until_sale)
for r in results:
results_by_look_back[str(r['look_back'])].append(r)
# results = [r, r, r...]
# r['lowest_buy_factor'] = lowest_buy_factor
# r['highest_sale_factor'] = highest_sale_factor
# r['look_back'] = optimizing
# r['optimal_buy_factor'] = lowest_buy_factor
# r['optimal_sell_factor'] = highest_sale_factor
for lb in results_by_look_back:
lowest_buy_factors = []
highest_sale_factors = []
results = results_by_look_back[lb]
for r in results:
lowest_buy_factors.append(r['optimal_buy_factor'])
highest_sale_factors.append(r['optimal_sell_factor'])
lowest_buy_factor_stats = fn.std_dev(lowest_buy_factors)
highest_sale_factor_stats = fn.std_dev(highest_sale_factors)
# print('#####################################################')
# print('LOWEST BUY FACTORS ######## LOOK_BACK', lb, s, step_back_periods, '*', hours_per_period, 'hrs')
# pprint(lowest_buy_factor_stats)
# print(lowest_buy_factors)
# print('HIGHEST SALE FACTORS ######## LOOK_BACK', lb, s, step_back_periods, '*', hours_per_period, 'hrs')
# pprint(highest_sale_factor_stats)
# print(highest_sale_factors)
# print('#####################################################')
result = {}
result['lowest_buy_factor_std_dev'] = lowest_buy_factor_stats['std_dev']
result['lowest_buy_factor_min'] = lowest_buy_factor_stats['min']
result['highest_sale_factor_std_dev'] = highest_sale_factor_stats['std_dev']
result['highest_sale_factor_min'] = highest_sale_factor_stats['min']
result['look_back'] = int(lb)
result['optimal_buy_factor'] = lowest_buy_factor_stats['mean']
result['optimal_sell_factor'] = highest_sale_factor_stats['mean']
optimization_factors_path = '/home/ec2-user/environment/botfarming/Development/optimization_factors/2_' + length + '_optimal_for_' + symbol['symbol'] + '_' + str(result['look_back']) + '.pklz'
ut.pickle_write(optimization_factors_path, result)
length = '1m'
minutes = 1
max_price_to_buy_factor = .985
buy_sell_starting_gap = .01
while True:
print('starting..')
################################################################################ RUN OPTIMIZER
trailing_and_current_candles_array = {}
smart_trailing_candles_array = {}
symbols = ut.pickle_read(
'/home/ec2-user/environment/botfarming/Development/3_binance_btc_symbols.pklz'
)
total_btc_coins = 0
symbols_trimmed = {}
for s in symbols:
symbol = symbols[s]
if float(symbol['24hourVolume']) > 300:
total_btc_coins += 1
symbols_trimmed[s] = symbol
best_gain = 0
best_price_to_buy_factor = 0
best_bollingers_percentage_increase_factor = 0
best_lower_band_buy_factor = 0
def get_optimization_factors(optimizing_array, data_length_multiplier, end_time, symbol, symbols_trimmed_one, length, minutes, minutes_until_sale):
continuous_mode = True
continous_length = 1
datapoints_trailing = 7
min_volume = 0
end_of_period = end_time
one_period_ago = end_of_period - data_length_multiplier*minutes*60
readable_time_end_of_period = datetime.datetime.fromtimestamp(end_of_period-7*60*60).strftime('%Y-%m-%d %H:%M')
readable_time_start_of_period = datetime.datetime.fromtimestamp(one_period_ago-7*60*60).strftime('%Y-%m-%d %H:%M')
readable_time_end_of_period_folder = datetime.datetime.fromtimestamp(end_of_period-7*60*60).strftime('%Y%m%d_%H:%M')
readable_time_start_of_period_folder = datetime.datetime.fromtimestamp(one_period_ago-7*60*60).strftime('%Y%m%d_%H:%M')
period = readable_time_start_of_period_folder + '_to_' + readable_time_end_of_period_folder
#print('########## fetching sample period data', period, 'for', symbol['symbol'], length)
save_params = [
[period, readable_time_start_of_period, readable_time_end_of_period]
]
ut.save_data(save_params, datapoints_trailing, min_volume, minutes, symbols_trimmed_one)
symbol_data_path = '/home/ec2-user/environment/botfarming/Development/binance_training_data/'+ symbol['symbol'] +'_data_'+str(minutes)+'m.pklz'
results = []
for optimizing in optimizing_array:
lowest_buy_factor = 1
highest_sale_factor = 1
data = ut.pickle_read(symbol_data_path)
if data == False:
print('no data for symbol', symbol['symbol'])
continue
continous_length = len(data)
if not data:
continue
for index, candle in enumerate(data):
# ERROR: local variable 'total_of_prices' referenced before assignment
try:
if optimizing == 1:
total_of_prices = 6*float(data[index-1][4])+2*float(data[index-2][4])+float(data[index-3][4])
total_counts_of_prices = 9
else:
for x in range(1,optimizing+1):
total_of_prices += float(data[index-x][4])*x**2
total_counts_of_prices += x**2
for x in range(1,optimizing):
total_of_prices += float(data[index-2*optimizing+x][4])*x**2
total_counts_of_prices += x**2
except Exception as e:
print(e)
continue
compare_price = float(total_of_prices)/float(total_counts_of_prices)
low_buy_factor = float(candle[3])/compare_price
if low_buy_factor < lowest_buy_factor:
lowest_buy_factor = low_buy_factor
try:
max_price = -1
for i in range(1,minutes_until_sale):
if float(data[index + i][2]) > max_price:
max_price = float(data[index + i][2])
highest_sale_factor = max_price/compare_price
except IndexError:
break
result = {}
result['look_back'] = optimizing
result['optimal_buy_factor'] = lowest_buy_factor
result['optimal_sell_factor'] = highest_sale_factor
results.append(result)
if os.path.isfile(symbol_data_path):
os.remove(symbol_data_path)
return results
import time
from pprint import pprint
import numpy
import sys
import pickle
import gzip
import utility_3 as ut
import json
import math
from os import listdir
from os.path import isfile, join
try:
symbols = ut.pickle_read('/home/ec2-user/environment/botfarming/Development/3_binance_btc_symbols.pklz')
total_btc_coins = 0
symbols_trimmed = {}
min_symbol_volume = 300
global socket_list
socket_list = []
for s in symbols:
symbol = symbols[s]
lengths = ['1m', '5m', '15m', '30m', '1h', '2h', '6h', '12h', '1d']
for length in lengths:
try:
f = gzip.open('/home/ec2-user/environment/botfarming/Development/program_state/program_state_' + length + '_0_' + symbol['symbol'] + '.pklz','rb')
current_state = pickle.load(f)
f.close()
# min_gain = .001
# minutes = 2*60
# look_back_array = [1,3,5,7,9,11,13,15]
# optimization_length = 360
# version = '1_'
# length = '1h'
# min_gain = .001
# minutes = 1*60
# look_back_array = [1,5,9,15]
# optimization_length = 360
# version = '1_'
######
symbols = ut.pickle_read(
'/home/ec2-user/environment/botfarming/Development/3_binance_btc_symbols.pklz'
)
total_btc_coins = 0
symbols_trimmed = {}
for s in symbols:
symbol = symbols[s]
if float(symbol['24hourVolume']) > 300:
total_btc_coins += 1
symbols_trimmed[s] = symbol
total_gain = 0
total_symbols = 0
total_trades = 0
total_buy_sell_difference = 0
path = '/home/ec2-user/environment/botfarming/Development/binance_all_trades_history/'
onlyfiles = [f for f in listdir(path) if isfile(join(path, f))]
look_backs = ['1m', '5m', '15m', '30m', '1h', '2h', '6h', '12h', '1d']
for look_back in look_backs:
print('look_back', look_back)
bot_trades = []
for file in onlyfiles:
if file.startswith(look_back + "_0"):
#print('/home/ec2-user/environment/botfarming/Development/binance_all_trades_history/' + file)
trade = ut.pickle_read('/home/ec2-user/environment/botfarming/Development/binance_all_trades_history/' + file)
if len(trade) == 11:
trade.append(look_back)
trade.append(trade[0])
trade[0] = ut.convert_date_to_epoch(trade[0])
trade[10] = ut.convert_date_to_epoch(trade[10])
bot_trades.append(trade)
total_profit = 0
total_trades = 0
total_profit_a = 0
total_trades_a = 0
total_profit_b = 0
total_trades_b = 0
for i, bot_trade in enumerate(bot_trades):
## SCHEMA ##
# 'time start', bot_trade[0]
# 'symbol', bot_trade[1]
# 'asolute profit', bot_trade[2]
# 'percentage profit', bot_trade[3]
# 'bit coin invested', bot_trade[4]
# 'look_back', bot_trade[5]
# 'a_b', bot_trade[6]
# 'price_to_buy_factor', bot_trade[7]
# 'price_to_sell_factor', bot_trade[8]
# 'original_buy_time', bot_trade[9]
# 'sell time', bot_trade[10])
symbols_saved = ut.pickle_read(
'/home/ec2-user/environment/botfarming/Development/3_binance_btc_symbols.pklz'
)
path = '/home/ec2-user/environment/botfarming/Development/binance_all_trades_history/'
onlyfiles = [f for f in listdir(path) if isfile(join(path, f))]
look_backs = ['1m', '5m', '15m', '30m', '1h', '2h', '6h', '12h', '1d']
look_backs = ['1m']
for look_back in look_backs:
print('look_back', look_back)
bot_trades = []
for file in onlyfiles:
print 'python', sys.version
import requests
import time
from pprint import pprint
import numpy
import sys
import pickle
import gzip
import utility_3 as ut
import json
import math
from binance.client import Client
indicator_trades = ut.pickle_read(
'/home/ec2-user/environment/botfarming/Development/variables/indicator_trades_1m'
)
indicator_trades_lengths = ut.get_indicator_trade_lengths(
indicator_trades, 15, 1)
for a in range(0, 16):
print(indicator_trades_lengths[a])
# indicator_trades_path = '/home/ec2-user/environment/botfarming/Development/variables/indicator_trades_1d'
# ut.pickle_write(indicator_trades_path, [])
# indicator_trades_old = ut.pickle_read('/home/ec2-user/environment/botfarming/Development/variables/indicator_trades_1m')
# print(len(indicator_trades_old))