def init(context, end_date, C1, C2):
'''
Stores initial values for each relevant variable
'''
symbol = C1 + C2
context.database = MagicDB(context.timestep, context.time, end_date, bDB())
first_100 = []
first_100V = []
timestep = HF.get_timestep_as_timedelta('1m')
for i in range(200):
price = context.get_historical_open_rate(symbol, context.time)
volume = context.get_historical_volume(symbol, context.time)
context.step(timestep)
first_100.append(price)
first_100V.append(volume)
for num in ["10", "20", "30", "40", "50", "70", "90", "100", "150", "200"]:
val = int(num)
context.variables["P{}".format(num)] = first_100[-val:]
if val <= 50:
context.variables["V{}".format(num)] = first_100V[-val:]
context.buys[symbol] = []
context.sells[symbol] = []
context.variables[symbol] = price
context.variables["Buy?"] = True
context.variables["NN"] = load_NN('Algorithms/NeuralNets/NET6.pth')
context.graphs = [(symbol, )]
def strat(context, C1, C2):
symbol = C1 + C2
price = context.get_historical_open_rate(symbol, context.time)
volume = context.get_historical_volume(symbol, context.time)
if isinstance(price, float):
vec = [price]
mavgs = [sum(context.variables["P{}".format(num)])/int(num)
for num in ["10", "20", "30", "40", "50", "60", "70", "80", "90", "100"]]
vec.extend(mavgs)
vec = normalize(vec)
vavgs = normalize([sum(context.variables["V{}".format(num)])/int(num)
for num in ["10", "20", "30", "40", "50"]])
vec.extend(vavgs)
NN = context.variables["NN"]
xi = torch.from_numpy(np.array(vec)).float()
pred = NN(xi)
selection = pred.max(0)[1].item()
if selection == 0 and pred[0].item() > .9997 and context.positions[C1] >0:
context.sell(C1,C2, .3*context.positions[C1])
elif selection == 1 and pred[1].item() > .9996 and context.positions[C2] >0:
context.buy(C1,C2, .3*context.positions[C2])
for num in ["10", "20", "30", "40", "50", "60", "70", "80", "90", "100"]:
last = context.variables['P{}'.format(num)][1:]
last.append(price)
context.variables['P{}'.format(num)] = last
if int(num) <= 50:
last = context.variables['V{}'.format(num)][1:]
last.append(volume)
context.variables['V{}'.format(num)] = last
context.variables[symbol] = price
def strat(context, C1, C2):
symbol = C1 + C2
price = context.get_historical_open_rate(symbol, context.time)
volume = context.get_historical_volume(symbol, context.time)
if isinstance(price, float):
vec = [price]
mavgs = [
sum(context.variables["P{}".format(num)]) / int(num) for num in
["10", "20", "30", "40", "50", "70", "90", "100", "150", "200"]
]
vec.extend(mavgs)
vec = normalize(vec)
vavgs = [volume]
vavgs.extend([
sum(context.variables["V{}".format(num)]) / int(num)
for num in ["10", "20", "30", "40", "50"]
])
vavgs = normalize(vavgs)
vec.extend(vavgs)
NN = context.variables["NN"]
xi = torch.from_numpy(np.array(vec)).float()
pred = NN(xi)
selection = pred.max(0)[1].item()
# print(vec)
buy_ = context.variables["Buy?"]
if pred[2].item() > .9999999995 and buy_ == True:
context.buy(C1, C2, context.positions[C2])
context.variables["Buy?"] = price
elif isinstance(buy_, float):
# if pred[0].item() > .99995:
# context.sell(C1,C2, context.positions[C1])
# context.variables["Buy?"] = True
if price >= buy_ * 1.01:
context.sell(C1, C2, context.positions[C1])
context.variables["Buy?"] = True
elif price <= buy_ * .99:
context.sell(C1, C2, context.positions[C1])
context.variables["Buy?"] = True
for num in [
"10", "20", "30", "40", "50", "70", "90", "100", "150", "200"
]:
last = context.variables['P{}'.format(num)][1:]
last.append(price)
context.variables['P{}'.format(num)] = last
if int(num) <= 50:
last = context.variables['V{}'.format(num)][1:]
last.append(volume)
context.variables['V{}'.format(num)] = last
context.variables[symbol] = price
def init(context, end_date, C1, C2):
'''
Stores initial values for each relevant variable
'''
symbol = C1 + C2
context.database = MagicDB(context.timestep, context.time, end_date, bDB())
first_50 = []
first_50V = []
timestep = HF.get_timestep_as_timedelta('1m')
for i in range(50):
price = context.get_historical_open_rate(symbol, context.time)
volume = context.get_historical_volume(symbol, context.time)
first_50.append(price)
first_50V.append(volume)
context.step(timestep)
V10 = first_50V[40:]
V20 = first_50V[30:]
V30 = first_50V[20:]
V40 = first_50V[10:]
last_10 = first_50[40:]
last_20 = first_50[30:]
last_30 = first_50[20:]
last_40 = first_50[10:]
context.buys[symbol] = []
context.sells[symbol] = []
context.variables['P10'] = last_10
context.variables['P20'] = last_20
context.variables['P30'] = last_30
context.variables['P40'] = last_40
context.variables['P50'] = first_50
context.variables['PMA10'] = sum(last_10)/10
context.variables['PMA20'] = sum(last_20)/20
context.variables['PMA30'] = sum(last_30)/30
context.variables['PMA40'] = sum(last_40)/40
context.variables['PMA50'] = sum(first_50)/50
context.variables['V10'] = V10
context.variables['V20'] = V20
context.variables['V30'] = V30
context.variables['V40'] = V40
context.variables['V50'] = first_50V
context.variables['VMA10'] = sum(V10)/10
context.variables['VMA20'] = sum(V20)/20
context.variables['VMA30'] = sum(V30)/30
context.variables['VMA40'] = sum(V40)/40
context.variables['VMA50'] = sum(first_50V)/50
context.variables[symbol] = price
context.variables["NN"] = load_NN('Algorithms/NeuralNets/NET2.pth')
context.graphs = [(symbol,'PMA10', 'PMA50')]
def strat(context, C1, C2):
symbol = C1 + C2
price = context.get_historical_open_rate(symbol, context.time)
volume = context.get_historical_volume(symbol, context.time)
if isinstance(price, float):
for num in ['10', '20', '30', '40', '50']:
context.variables['PMA{}'.format(num)] = sum(context.variables['P{}'.format(num)])/int(num)
context.variables['VMA{}'.format(num)] = sum(context.variables['V{}'.format(num)])/int(num)
context.variables[symbol] = price
[V10,V20,V30,V40,V50] = [context.variables['VMA{}'.format(num)] for num in ["10", "20", "30", "40", "50"]]
[P10,P20,P30,P40,P50] = [context.variables['PMA{}'.format(num)] for num in ["10", "20", "30", "40", "50"]]
NN = context.variables["NN"]
vec = []
pavgs = [price, P10,P20,P30,P40,P50]
pavgs = normalize(pavgs)
for p in pavgs: vec.append(p)
vavgs = [V10,V20,V30,V40,V50]
vavgs = normalize(vavgs)
for v in vavgs: vec.append(v)
xi = torch.from_numpy(np.array(vec)).float()
pred = NN(xi)
selection = pred.max(0)[1].item()
if selection == 0 and pred[0].item() > .999999:
context.sell(C1,C2, .3*context.positions[C1])
elif selection == 2 and pred[2].item() > .9988:
context.buy(C1,C2, .3*context.positions[C2])
for num in ['10', '20', '30', '40', '50']:
last = context.variables['P{}'.format(num)][1:]
last.append(price)
context.variables['P{}'.format(num)] = last
last = context.variables['V{}'.format(num)][1:]
last.append(volume)
context.variables['V{}'.format(num)] = last