def opt_cut_aggr(self, cuts, grades, seed, plot=True):
heatmap = np.zeros([len(cuts), len(grades)])
for i, cut in enumerate(cuts):
for j, grade in enumerate(grades):
# set cut of strateg
self.strat.cut = cut
self.strat.grade = grade
# set initial investment of Backtester
Backtester = backtester.Backtester(self.DataHelper, self.model,
self.strat, 1)
Backtester.initialInvest = seed
# append return
_, r, _ = Backtester.backtest_OOT(returnAll=False)
heatmap[i, j] = r
# change it to df
heatmap_df = pd.DataFrame(heatmap, index=cuts, columns=grades)
if plot:
plt.figure(figsize=(10, 8))
ax = sns.heatmap(heatmap_df, cmap="YlGnBu")
ax.set_title(
"aggressive strategy return with init invest {}".format(seed))
plt.show()
return heatmap_df
def opt_cut_conserv(self, cuts, seed, plot=True):
profits = []
for cut in cuts:
# set cut of strateg
self.strat.cut = cut
# set initial investment of Backtester
Backtester = backtester.Backtester(self.DataHelper, self.model,
self.strat, 1)
Backtester.initialInvest = seed
# run backtest
_, r, _ = Backtester.backtest_OOT(returnAll=False)
# append return
profits.append(r)
if plot:
plt.figure(figsize=(10, 8))
plt.plot(cuts, profits)
plt.show()
return profits
def opt_amountfunc_aggr(self, slopes, maxs, cut, grade, seed, plot=False):
# set cut of strateg
self.strat.cut = cut
self.strat.grade = grade
heatmap = np.zeros([len(slopes), len(maxs)])
for i, s in enumerate(slopes):
for j, m in enumerate(maxs):
# set cut of strateg
self.strat.amount_slope = s
self.strat.amount_max = m
# set initial investment of Backtester
Backtester = backtester.Backtester(self.DataHelper, self.model,
self.strat, 1)
Backtester.initialInvest = seed
# append return
_, r, _ = Backtester.backtest_OOT(returnAll=False)
heatmap[i, j] = r
# change it to df
heatmap_df = pd.DataFrame(heatmap, index=slopes, columns=maxs)
if plot:
plt.figure(figsize=(10, 8))
ax = sns.heatmap(heatmap_df, cmap="YlGnBu")
ax.set_title(
"aggressive strategy return with init invest {}".format(seed))
ax.set_ylabel('p1 - slope')
ax.set_xlabel('p2 - max')
plt.show()
return heatmap_df
def statisticalArb(cash, stock1, stock2, start, end, days_moving_avg,
hedge_method):
df = getStocks(stock1, stock2, start, end, days_moving_avg)
# first graph the stocks
df[stock1].plot(label=stock1, color='b')
df[stock2].plot(label=stock2, color='g')
plt.legend(loc='upper left')
plt.title('Stock Prices')
plt.show()
# calculate hedge ratio
df = calculateHedgeRatio(df, stock1, stock2, days_moving_avg, hedge_method)
# print "Hedge Ratio Used: " + str(hedge_ratio)
df = calculateResidual(df, stock1, stock2, days_moving_avg)
getStatistics(df, hedge_method)
graph(df)
Backtester = backtester.Backtester(capital_limit, df, stock1, stock2,
days_moving_avg)
Backtester.run()
# print df['HedgeRatio'].tolist()
# print df['2016-09-01':'2016-11-01']
# staticHedge(df[:'2016-09-19 00:00:00'], days_moving_avg) # DEBUGGING
# plt.plot(df.index, df['HedgeRatio'], label='Hedge Ratio')
# plt.title('Static Hedge Ratio Over Time (DEBUGGING)')
plt.show()
import backtester, math
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report, confusion_matrix
import pickle
from sklearn.preprocessing import MinMaxScaler
from keras.models import Sequential
from keras.layers import Dense, LSTM
initial = pd.read_csv('https://www.alphavantage.co/query?function=TIME_SERIES_INTRADAY&interval=1min&symbol=msft&apikey=OUMVBY0VK0HS8I9E&datatype=csv&outputsize=full')
initial = initial[::-1].reset_index(drop=True)
#initial = initial[:500]
scaler = MinMaxScaler(feature_range=(0,1))
model = 1
buy_price = 0.0
initialInvestment = 1000
b = backtester.Backtester(initialInvestment)
predictions = []
with open('lstm.pkl', 'rb') as file:
model = pickle.load(file)
scaler.fit(initial.filter(['close']))
for i in range(len(initial)-60):
new_df = initial.filter(['close'])
last_60_days = new_df[i:60+i].values
last_60_days_scaled = scaler.transform(last_60_days)
X_test = []
X_test.append(last_60_days_scaled)
X_test = np.array(X_test)
X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 1))
pred_price = model.predict(X_test)
content = json.loads(buy.content)
return float(content['candles'][-1]['mid']['c']) #close first, close second
except Exception as e:
return 1.257
# r = requests.get('https://finance.yahoo.com/quote/GBPUSD=X/')
# soup = bs4.BeautifulSoup(r.text, 'lxml')
# f = soup.find('div',{'class': 'My(6px) Pos(r) smartphone_Mt(6px)'}).find('span').text
# return float(f)
initial = pd.read_csv('https://www.alphavantage.co/query?function=FX_INTRADAY&from_symbol=aud&to_symbol=USD&interval=1min&apikey=OUMVBY0VK0HS8I9E&outputsize=full&datatype=csv')
initial = initial[::-1]
initial.drop(['timestamp', 'open', 'high', 'low'], axis=1, inplace=True)
b = backtester.Backtester(1000)
model = 1
buy_price = 0.0
predictions = [1.0]
with open('forex_lstm.pkl', 'rb') as file:
model = pickle.load(file)
scaler = MinMaxScaler(feature_range=(0,1))
scaler.fit(initial)
for i in range(10000):
now = datetime.now()
# dd/mm/YY H:M:S
new_dict = {}
close = get_close()
new_dict['close'] = float(close)
create_order('MSFT', 5, 'sell', 'market', 'day')
predictions.append(pred)
except Exception as e:
print(e)
socket = 'wss://data.alpaca.markets/stream'
API_KEY = 'PKOT8ZWPLGJ94Q705388'
SECRET_KEY = 'GgjCthiFXfl8b05cMsdZ4uhZ2aekaMNgX2ZzPB3T'
APCA_API_BASE_URL = 'https://paper-api.alpaca.markets'
api = tradeapi.REST(API_KEY, SECRET_KEY, APCA_API_BASE_URL, api_version='v2')
account = api.get_account()
indicators = pd.DataFrame()
i = 0
b = backtester.Backtester(12000)
X_test = pd.DataFrame()
initial = pd.DataFrame()
from sklearn.ensemble import RandomForestRegressor
rfc = RandomForestRegressor(n_estimators=100)
predictions = [209.8]
buy_price = 0.0
with open('stock_model.pkl', 'rb') as file:
rfc = pickle.load(file)
ws = websocket.WebSocketApp(socket, on_open=on_open, on_message=on_message)
ws.run_forever()