def __init__(self):
self.login_df=
self.un = self.login_df['Username'].iloc[0]
self.pw = self.login_df['Password'].iloc[0]
self.my_portfolio = Portfolio()
self.mt = ModelTools()
self.L = Log()
class ModelTools:
def __init__(self):
self.model_loc = r'/home/jeb/TraderBotLogs/testmodel.dat'
self.my_model = pickle.load(open(self.model_loc, 'rb'))
self.FE = FeatureEngineerer()
self.L = Log()
def get_model_inputs(self, symbol, _date):
""" Current features --> ['Volume', 'Dividends', 'Stock Splits', 'Pct_Change', 'Three_Day_Movement', 'Five_Day_Movement']"""
if _date.weekday() == 0:
_date = _date - timedelta(days=2)
tckr = yf.Ticker(symbol)
result = pd.DataFrame(
tckr.history(start=_date - timedelta(days=1), end=_date))
if result.empty:
print('No info for this day.')
return pd.DataFrame()
result = result.iloc[-1]
running_result = pd.DataFrame(
tckr.history(start=_date - timedelta(days=5), end=_date))
running_result['Date'] = running_result.index
running_result['Date'] = running_result['Date'].dt.date
model_inputs = pd.DataFrame()
model_inputs['Date'] = [_date]
try:
model_inputs['Volume'] = [int(result['Volume'])]
except:
print('NaNs in inputs')
return pd.DataFrame()
model_inputs['Dividends'] = [int(result['Dividends'])]
model_inputs['Stock Splits'] = [int(result['Stock Splits'])]
model_inputs['Pct_Change'] = [self.FE.daily_pct_change(result)]
model_inputs['Three_Day_Movement'] = [
self.FE.prior_trend(running_result, _date, 3)
]
model_inputs['Five_Day_Movement'] = [
self.FE.prior_trend(running_result, _date, 5)
]
return model_inputs
def make_prediction(self, symbol, _date):
"""Returns True for whether or not a stock should be bought."""
inputs = self.get_model_inputs(symbol, _date)
if not inputs.empty:
inputs.drop('Date', inplace=True, axis=1)
results = self.my_model.predict(inputs)
if results == 1:
return True
else:
return False
self.L.add_line(f'No valid model inputs found for {symbol}', symbol)
return False
def __init__(self):
self.model_loc = r'/home/jeb/TraderBotLogs/testmodel.dat'
self.my_model = pickle.load(open(self.model_loc, 'rb'))
self.FE = FeatureEngineerer()
self.L = Log()
class ATRaC:
def __init__(self):
self.login_df=
self.un = self.login_df['Username'].iloc[0]
self.pw = self.login_df['Password'].iloc[0]
self.my_portfolio = Portfolio()
self.mt = ModelTools()
self.L = Log()
def _login(self):
chirp.login(self.un, self.pw)
def amount_to_buy(self):
"""
Current Rules:
1. Never buy more than 10% of total accound value
2. If there isn't 10% of total available, spend roughly what is left.
3. If less than a dollar is left of buying power, return 0
"""
buying_power = float(chirp.load_account_profile()['portfolio_cash'])
if buying_power<1:
return 0
total_equity = float(chirp.load_portfolio_profile()['equity'])
ten_percent_of_portfolio = (buying_power+total_equity)*.1
if ten_percent_of_portfolio<buying_power:
return ten_percent_of_portfolio
else:
return buying_power*.9
def buy_stock(self, symbol, value):
value = self.amount_to_buy()
if value==0:
self.L.add_line(f'Not enough money to buy {symbol}, purchase not made.')
else:
chirp.order_buy_fractional_by_price(symbol, value)
self.L.add_line('', symbol, 'BOUGHT', value)
self.my_portfolio.add_to_portfolio(symbol)
def sell_stock(self, symbol):
""""Sell all of a stock """
amount_to_sell = self.get_equity(symbol)
chirp.order_sell_fractional_by_price(symbol, amount_to_sell)
self.L.add_line('', symbol, 'SOLD', amount_to_sell)
def get_equity(self, symbol):
self.update_portfolio()
return float(self.my_portfolio[symbol]['equity'])
def update_portfolio(self):
self.my_portfolio = chirp.build_holdings()
def stocks_to_sell():
port = self.my_portfolio.portfolio_contents()
today = date.today()
sell_me = []
if today.weekday() not in [5, 6]:
for i in list(port['Symbol'].unique()):
bought_date = port[port['Symbol'] == i]['Date'].iloc[0]
if bought_date+timedelta(days=2)<today:
sell_me.append(i)
return sell_me