def __init__(self, symbol, dates, start_val, verbose=False):
# Initialize Genetic Algorithm
self.POP_SIZE = 100
self.DNA_SIZE = 8
self.MAX_GENERATIONS = 500
# Mutation probability = 1/mutation_chance
self.mutation_chance = 100
self.mutation_factor = 0.1
self.population_replacement = 0.7
self.best_population_replacement = 0.5
self.verbose = verbose
self.start_val = start_val
self.symbol = symbol
self.dates = dates
self.fitness_dict = {}
# Get data
self.prices_df = util.get_data([symbol],
pd.date_range(dates[0], dates[1]),
addSPY=False).dropna()
# Get Candlestick data
self.df_aux = util.get_all_data(symbol, dates)
self.ohlc = ind.get_ohlc(self.df_aux)
self.patterns_signal, self.trends, self.points_signal, self.sma_signal, self.rsi_signal, self.macd_signal, self.bb_signal = self.get_signals(
self.prices_df, self.ohlc)
def testPolicy(self,
symbol,
dates=['2011-1-1', '2011-12-31'],
start_val=100000,
ga_train=False):
#Test a trading policy for a stock wthin a date range and output a trades dataframe.
# Get data
prices_df = util.get_data([symbol],
pd.date_range(dates[0], dates[1]),
addSPY=False).dropna()
# Get Candlestick data
df_aux = util.get_all_data(symbol, dates)
ohlc = ind.get_ohlc(df_aux)
#Generate order signals
order_signals = self.trade_strategy(prices_df, ohlc, ga_train)
# Remove 0 signals
order_signals = order_signals[order_signals != 0.0]
# Create trades dataframe
trades = []
# Double-Down and Double-Up is Active
position = 0
for date in order_signals.index:
if order_signals.loc[date].values == 1:
# Buy Order
if position == 0:
trades.append((date, symbol, "BUY", 1000))
position = 1
elif position == -1:
# Double-Up
trades.append((date, symbol, "BUY", 2000))
position = 1
elif order_signals.loc[date].values == -1:
# Sell Order
if position == 0:
trades.append((date, symbol, "SELL", 1000))
position = -1
elif position == 1:
# Double-Down
trades.append((date, symbol, "SELL", 2000))
position = -1
if date == order_signals.index[-1]:
# Last day, close open positions
last_trade = trades[-1]
trades[-1] = (last_trade[0], last_trade[1], last_trade[2],
last_trade[3] - 1000)
position = 0
self.df_trades = pd.DataFrame(
trades, columns=['Date', 'Symbol', 'Order', 'Shares'])
self.df_trades.set_index('Date', inplace=True)
#pdb.set_trace()
return self.df_trades
# Search for Morning Star patterns
found, signal = is_morning_star_doji(ohlc[i-1], row, ohlc[i+1])
if found:
found_signals.append([mdates.num2date(ohlc[i+1][date]), signal])
if "Abandoned Baby" in patterns:
# Search for Abandoned baby patterns
found, signal = is_abandoned_baby(ohlc[i-1], row, ohlc[i+1])
if found:
found_signals.append([mdates.num2date(ohlc[i+1][date]), signal])
# Search for 2 candle patterns
if "Harami Cross" in patterns:
# Search for Harumi Cross patterns
found, signal = is_harami_cross(row, ohlc[i+1])
if found:
#pdb.set_trace()
found_signals.append([mdates.num2date(ohlc[i+1][date]), signal])
return found_signals
if __name__ == "__main__":
symbols = ['SPY','AAPL', 'GOOG', 'IBM', 'XOM']
obs_dates = ['2011-03-01','2011-04-01']
symbol = "GOOG"
df_aapl = util.get_all_data(symbol, obs_dates)
ohlc = ind.get_ohlc(df_aapl)
for i, row in enumerate(ohlc[:-1]):
if is_harami_cross(row, ohlc[i+1])[0]:
print ("Found Harami Cross!")
print (mdates.num2date(row[0]), row[1:])
print (mdates.num2date(ohlc[i+1][0]), ohlc[i+1][1:])
#pdb.set_trace()
split_date = df.index[pos]
# Use GeneticAlgorithm to optimize signal weights
params = False
gen_alg = ga.GeneticAlgorithm(symbol=symbol,
dates=[dates[0], split_date],
start_val=start_val,
verbose=True)
params, sharpe_ratio = gen_alg.start_ga()
#pdb.set_trace()
ft = FakeTicker(symbol, df, testing_pctg)
#print (ft.get_training_data().shape)
prices_df = ft.get_training_data()['Adj Close']
ohlc = ind.get_ohlc(ft.get_training_data())
trades = []
ctr = 0
while not ft.is_end_ticker():
start = time.time()
# Get tick data
ticker_data = ft.get_ticker_data()
# Add new ticker to training data
new_index = prices_df.index.insert(len(prices_df), ticker_data.name)
prices_df = np.append(prices_df.values.flatten(),
ticker_data['Adj Close'])
prices_df = pd.DataFrame(prices_df, index=new_index, columns=[symbol])
# Add new OHLC to ohlc array