def __init__(self, dataset, rows=1, grid=True):
# We need to handle a strategy being passed
if isinstance(dataset, strategy.Strategy):
self.strategy = dataset
self.data_frame = self.strategy.dataset.data_frame
self.realtime_data_frame = dataset.realtime_data_frame
else: # Assume a dataset was passed
self.strategy = None
self.data_frame = dataset.data_frame
self.realtime_data_frame = None
self.data_frame.reset_index(inplace=True)
date_conversion = lambda date: date2num(date.to_pydatetime())
self.data_frame['DATE'] = self.data_frame['Date'].apply(date_conversion)
self.rows = rows
if grid:
plt.rc('axes', grid=True)
plt.rc('grid', color='0.75', linestyle='-', linewidth='0.2')
self.current_figure = None
self.figure_first_ax = None
self.figure_rows = 1
self.legend = []
self.legend_labels = []
self.add_figure(self.rows)
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info('dataset: %s rows: %s grid: %s' %(dataset, rows, grid))
def __init__(self, strategy, trading_profile):
self.strategy = strategy
self.trading_profile = trading_profile
self.trade_history = {}
self.available_money = self.trading_profile.capital
self.capital = self.trading_profile.capital
self.available_money_history = pd.Series(
index=self.strategy.dataset.data_frame.index)
self.available_capital_history = pd.Series(
index=self.strategy.dataset.data_frame.index)
self.ongoing_trades = {}
self.trades = 0
self.average_gain = 0.0
self.sharpe_ratio = 0.0
self.average_trading_amount = 0.0
self.average_bars = 0.0
self.total_fees = 0.0
self.average_fees = 0.0
self.total_slippage = 0.0
self.average_slippage = 0.0
self.winning_trades = 0
self.losing_trades = 0
self.average_winning_gain = 0.0
self.average_losing_gain = 0.0
self.percent_profitable = 0.0
self.gross_profit = 0.0
self.gross_loss = 0.0
self.net_profit = 0.0
self.lacking_capital = 0
self._require_finalize_calculations = False
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info('strategy: %s trading_profile: %s' %
(strategy, trading_profile))
def __init__(self,
train_data,
prediction_data,
network_type=FEED_FORWARD_NETWORK,
network_dataset_type=SUPERVISED_DATASET,
trainer_type=BACKPROP_TRAINER):
self.train_data = train_data
self.prediction_data = prediction_data
self.network_type = network_type
self.network_dataset_type = network_dataset_type
self.trainer_type = trainer_type
self.network = None
self.network_dataset = None
self.dataset = None
self.trainer = None
self.trained_iterations = 0
self.momentum = None
self.learning_rate = None
self.hidden_layers = None
self.prediction_window = None
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info('train_data: %s prediction_data: %s, network_type: %s, \
network_dataset_type: %s, trainer_type: %s'
%(train_data, prediction_data, network_type, \
network_dataset_type, trainer_type))
def __init__(self,
train_data,
prediction_data,
ensemble_type=RANDOM_FOREST_REGRESSOR):
self.train_data = train_data
self.prediction_data = prediction_data
self.ensemble_type = ensemble_type
self.ensemble = None
self.prediction_window = None
self.look_back_window = None
self.training_set = []
self.target_set = []
self.normalize = True
self.number_of_estimators = 150
self.max_depth = None
self.random_state = 0
self.min_samples_split = 2
self.number_of_jobs = 2
self.learning_rate = 1.0 # For Gradient Boosting
self.rating = None
self.feature_importances = None
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info(
'train_data: %s prediction_data: %s ensemble_type: %s' %
(train_data, prediction_data, ensemble_type))
def __init__(self, symbol_list, data_connection, start_datetime=None, \
end_datetime=None, periods=None, granularity=None):
self.symbol_list = symbol_list
# Either specify a start and end date or a number of periods since now
assert periods != None or start_datetime != None
self.start_datetime = start_datetime
self.end_datetime = end_datetime
self.periods = periods
self.granularity = granularity
self.data_connection = data_connection
self.data_frame = pd.DataFrame()
self.technical_indicators = []
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info('symbol_list: %s \
data_connection: %s \
start_datetime: %s \
end_datetime: %s \
periods: %s \
granularity: %s'
%(symbol_list, \
data_connection, \
start_datetime, \
end_datetime, \
periods, \
granularity))
def __init__(self, capital, trading_amount, trading_fee, slippage=0.0):
self.capital = capital
self.trading_amount = trading_amount
self.trading_fee = trading_fee
self.slippage = slippage
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info('Initialized - %s' %self)
def __init__(self, dataset, criteria_groups, trading_profile):
self.dataset = dataset
self.criteria_groups = criteria_groups
self.trading_profile = trading_profile
self.name = 'Strategy'
self.report = report.Report(self, self.trading_profile)
self.realtime_data_frame = pd.DataFrame() # Used for backtesting
self.first_pass = True # Flag to execute certain actions on first bar of backtest
self.upcoming_actions = {}
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info('Initialized - %s' %self)
def __init__(self, datetime, action, symbol, price, shares, money, fee,
slippage):
self.datetime = datetime
self.action = action
self.symbol = symbol
self.price = price
self.shares = shares
self.money = money
self.fee = fee
self.slippage = slippage
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info('Initialized - %s' % self)
def __init__(self,
username,
password,
recipient,
server='smtp.gmail.com',
port=587):
self.username = username
self.password = password
self.recipient = recipient
self.server = server
self.port = port
self.session = self._get_session()
self.headers = None
self.body = None
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info(
'username: %s password: * recipient: %s server: %s port: %s' %
(username, recipient, server, port))
def __init__(self, criteria_list, action, symbol):
self.criteria_list = criteria_list
self.action = action
self._action = 0
self.symbol = str(symbol)
# Add Actions Criteria Automatically
if self.action == Long():
self.criteria_list.append(Not(InMarket(self.symbol)))
self._action = LONG
elif self.action == LongExit():
self.criteria_list.append(IsLong(self.symbol))
self._action = LONG_EXIT
elif self.action == Short():
self.criteria_list.append(Not(InMarket(self.symbol)))
self._action = SHORT
elif self.action == ShortExit():
self.criteria_list.append(IsShort(self.symbol))
self._action = SHORT_EXIT
else:
raise InvalidAction()
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info('New Criteria Group - %s for %s (%s)' \
%(str(action).upper(), symbol, criteria_list))
def __init__(self):
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info('Initialized')
def __init__(self, name):
self.name = name.replace(' ', '').replace('_', '').replace('-',
'').lower()
self.logger = logger.Logger(self.__class__.__name__)
def __init__(self, symbols):
self.symbols = [Symbol(symbol) for symbol in symbols]
self.logger = logger.Logger(self.__class__.__name__)
self.logger.info(symbols)
def __init__(self):
self.logger = logger.Logger(self.__class__.__name__)
def __init__(self):
self.logger = logger.Logger(self.__class__.__name__)
# Some TIs don't require the entire data_frame history.
self.num_bars_required = None # Requires all history by default
