# Copyright 2019 The TensorTrade Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License
import numpy as np
import pandas as pd
from abc import abstractmethod
from gym.spaces import Space, Box
from typing import List, Dict
from tensortrade.trades import Trade, TradeType
from tensortrade.exchanges import InstrumentExchange
from tensortrade.slippage import RandomUniformSlippageModel
from tensortrade.features import FeaturePipeline
[docs]class SimulatedExchange(InstrumentExchange):
"""An instrument exchange, in which the price history is based off the supplied data frame and
trade execution is largely decided by the designated slippage model.
If the `data_frame` parameter is not supplied upon initialization, it must be set before
the exchange can be used within a trading environment.
"""
def __init__(self, data_frame: pd.DataFrame = None, **kwargs):
super().__init__(base_instrument=kwargs.get('base_instrument', 'USD'),
dtype=kwargs.get('dtype', np.float16),
feature_pipeline=kwargs.get('feature_pipeline', None))
self._commission_percent = kwargs.get('commission_percent', 0.3)
self._base_precision = kwargs.get('base_precision', 2)
self._instrument_precision = kwargs.get('instrument_precision', 8)
self._min_trade_price = kwargs.get('min_trade_price', 1E-6)
self._max_trade_price = kwargs.get('max_trade_price', 1E6)
self._min_trade_amount = kwargs.get('min_trade_amount', 1E-3)
self._max_trade_amount = kwargs.get('max_trade_amount', 1E6)
self._min_order_amount = kwargs.get('min_order_amount', 1E-3)
self._initial_balance = kwargs.get('initial_balance', 1E4)
self._observation_columns = kwargs.get(
'observation_columns', ['open', 'high', 'low', 'close', 'volume'])
self._window_size = kwargs.get('window_size', 1)
self._pretransform = kwargs.get('pretransform', True)
self.data_frame = data_frame
max_allowed_slippage_percent = kwargs.get('max_allowed_slippage_percent', 1.0)
SlippageModelClass = kwargs.get('slippage_model', RandomUniformSlippageModel)
self._slippage_model = SlippageModelClass(max_allowed_slippage_percent)
@property
def data_frame(self) -> pd.DataFrame:
"""The underlying data model backing the price and volume simulation."""
return getattr(self, '_data_frame', None)
@data_frame.setter
def data_frame(self, data_frame: pd.DataFrame):
if not isinstance(data_frame, pd.DataFrame):
self._data_frame = data_frame
return
self._data_frame = data_frame[self._observation_columns]
if self._pretransform:
self.transform_data_frame()
@property
def feature_pipeline(self) -> FeaturePipeline:
return self._feature_pipeline
@feature_pipeline.setter
def feature_pipeline(self, feature_pipeline=FeaturePipeline):
self._feature_pipeline = feature_pipeline
if isinstance(self.data_frame, pd.DataFrame) and self._pretransform:
self.transform_data_frame()
return self._feature_pipeline
@property
def initial_balance(self) -> float:
return self._initial_balance
@property
def balance(self) -> float:
return self._balance
@property
def portfolio(self) -> Dict[str, float]:
return self._portfolio
@property
def trades(self) -> pd.DataFrame:
return self._trades
@property
def performance(self) -> pd.DataFrame:
return self._performance
@property
def generated_space(self) -> Space:
low = np.array([self._min_trade_price, ] * 4 + [self._min_trade_amount, ])
high = np.array([self._max_trade_price, ] * 4 + [self._max_trade_amount, ])
return Box(low=low, high=high, dtype='float')
@property
def generated_columns(self) -> List[str]:
return list(self._observation_columns)
@property
def has_next_observation(self) -> bool:
return self._current_step < len(self._data_frame) - 1
def _next_observation(self) -> pd.DataFrame:
lower_range = max((self._current_step - self._window_size, 0))
upper_range = max(min(self._current_step, len(self._data_frame)), 1)
obs = self._data_frame.iloc[lower_range:upper_range]
if len(obs) < self._window_size:
padding = np.zeros((len(self.generated_columns), self._window_size - len(obs)))
obs = pd.concat([pd.DataFrame(padding), obs], ignore_index=True)
if not self._pretransform and self._feature_pipeline is not None:
obs = self._feature_pipeline.transform(obs, self.generated_space)
self._current_step += 1
return obs
[docs] def current_price(self, symbol: str) -> float:
if self.data_frame is not None:
frame = self._data_frame.iloc[self._current_step]
if frame.empty is False:
return frame['close']
return 0
def _is_valid_trade(self, trade: Trade) -> bool:
if trade.trade_type is TradeType.MARKET_BUY or trade.trade_type is TradeType.LIMIT_BUY:
return trade.amount >= self._min_order_amount and self._balance >= trade.amount * trade.price
elif trade.trade_type is TradeType.MARKET_SELL or trade.trade_type is TradeType.LIMIT_SELL:
return trade.amount >= self._min_order_amount and self._portfolio.get(trade.symbol, 0) >= trade.amount
return True
def _update_account(self, trade: Trade):
if self._is_valid_trade(trade) and not trade.is_hold:
self._trades = self._trades.append({
'step': self._current_step,
'symbol': trade.symbol,
'type': trade.trade_type,
'amount': trade.amount,
'price': trade.price
}, ignore_index=True)
if trade.is_buy:
self._balance -= trade.amount * trade.price
self._portfolio[trade.symbol] = self._portfolio.get(trade.symbol, 0) + trade.amount
elif trade.is_sell:
self._balance += trade.amount * trade.price
self._portfolio[trade.symbol] = self._portfolio.get(trade.symbol, 0) - trade.amount
self._portfolio[self._base_instrument] = self._balance
self._performance = self._performance.append({
'balance': self.balance,
'net_worth': self.net_worth,
}, ignore_index=True)
[docs] def execute_trade(self, trade: Trade) -> Trade:
current_price = self.current_price(symbol=trade.symbol)
commission = self._commission_percent / 100
filled_trade = trade.copy()
if filled_trade.is_hold or not self._is_valid_trade(filled_trade):
filled_trade.amount = 0
return filled_trade
if filled_trade.is_buy:
price_adjustment = price_adjustment = (1 + commission)
filled_trade.price = max(round(current_price * price_adjustment,
self._base_precision), self.base_precision)
filled_trade.amount = round(
(filled_trade.price * filled_trade.amount) / filled_trade.price, self._instrument_precision)
elif filled_trade.is_sell:
price_adjustment = (1 - commission)
filled_trade.price = round(current_price * price_adjustment, self._base_precision)
filled_trade.amount = round(filled_trade.amount, self._instrument_precision)
filled_trade = self._slippage_model.fill_order(filled_trade, current_price)
self._update_account(filled_trade)
return filled_trade
[docs] def reset(self):
super().reset()
self._current_step = 0
self._balance = self.initial_balance
self._portfolio = {self.base_instrument: self.balance}
self._trades = pd.DataFrame([], columns=['step', 'symbol', 'type', 'amount', 'price'])
self._performance = pd.DataFrame([], columns=['balance', 'net_worth'])