def render(self, savefig=False, filename='myfig'):
"""Matlplotlib rendering of each step.
Args:
savefig (bool): Whether to save the figure as an image or not.
filename (str): Name of the image file.
"""
if self._first_render:
self._f, self._ax = plt.subplots(
len(self._spread_coefficients) + int(len(self._spread_coefficients) > 1),
sharex=True
)
if len(self._spread_coefficients) == 1:
self._ax = [self._ax]
self._f.set_size_inches(12, 6)
self._first_render = False
self._f.canvas.mpl_connect('close_event', self._handle_close)
if len(self._spread_coefficients) > 1:
# TODO: To be checked
for prod_i in range(len(self._spread_coefficients)):
bid = self._prices_history[-1][2 * prod_i]
ask = self._prices_history[-1][2 * prod_i + 1]
self._ax[prod_i].plot([self._iteration, self._iteration + 1],
[bid, bid], color='white')
self._ax[prod_i].plot([self._iteration, self._iteration + 1],
[ask, ask], color='white')
self._ax[prod_i].set_title('Product {} (spread coef {})'.format(
prod_i, str(self._spread_coefficients[prod_i])))
# Spread price
prices = self._prices_history[-1]
bid, ask = calc_spread(prices, self._spread_coefficients)
self._ax[-1].plot([self._iteration, self._iteration + 1],
[bid, bid], color='white')
self._ax[-1].plot([self._iteration, self._iteration + 1],
[ask, ask], color='white')
ymin, ymax = self._ax[-1].get_ylim()
yrange = ymax - ymin
if (self._action == self._actions['sell']).all():
self._ax[-1].scatter(self._iteration + 0.5, bid + 0.03 *
yrange, color='orangered', marker='v')
elif (self._action == self._actions['buy']).all():
self._ax[-1].scatter(self._iteration + 0.5, ask - 0.03 *
yrange, color='lawngreen', marker='^')
plt.suptitle('Cumulated Reward: ' + "%.2f" % self._total_reward + ' ~ ' +
'Cumulated PnL: ' + "%.2f" % self._total_pnl + ' ~ ' +
'Position: ' + ['flat', 'long', 'short'][list(self._position).index(1)] + ' ~ ' +
'Entry Price: ' + "%.2f" % self._entry_price)
self._f.tight_layout()
plt.xticks(range(self._iteration)[::5])
plt.xlim([max(0, self._iteration - 80.5), self._iteration + 0.5])
plt.subplots_adjust(top=0.85)
plt.pause(0.01)
if savefig:
plt.savefig(filename)
def test_calc_spread():
spread_coefficients = [1, -0.1]
prices = np.array([1, 2, 10, 20])
spread_price = (-1, 1)
assert calc_spread(prices, spread_coefficients) == spread_price
def step(self, action):
"""Take an action (buy/sell/hold) and computes the immediate reward.
Args:
action (numpy.array): Action to be taken, one-hot encoded.
Returns:
tuple:
- observation (numpy.array): Agent's observation of the current environment.
- reward (float) : Amount of reward returned after previous action.
- done (bool): Whether the episode has ended, in which case further step() calls will return undefined results.
- info (dict): Contains auxiliary diagnostic information (helpful for debugging, and sometimes learning).
"""
assert any([(action == x).all() for x in self._actions.values()])
self._action = action
self._iteration += 1
done = False
instant_pnl = 0
info = {}
reward = -self._time_fee
if all(action == self._actions['buy']):
reward -= self._trading_fee
if all(self._position == self._positions['flat']):
self._position = self._positions['long']
self._entry_price = calc_spread(
self._prices_history[-1],
self._spread_coefficients)[1] # Ask
elif all(self._position == self._positions['short']):
self._exit_price = calc_spread(
self._prices_history[-1],
self._spread_coefficients)[1] # Ask
instant_pnl = self._entry_price - self._exit_price
self._position = self._positions['flat']
self._entry_price = 0
elif all(action == self._actions['sell']):
reward -= self._trading_fee
if all(self._position == self._positions['flat']):
self._position = self._positions['short']
self._entry_price = calc_spread(
self._prices_history[-1],
self._spread_coefficients)[0] # Bid
elif all(self._position == self._positions['long']):
self._exit_price = calc_spread(
self._prices_history[-1],
self._spread_coefficients)[0] # Bid
instant_pnl = self._exit_price - self._entry_price
self._position = self._positions['flat']
self._entry_price = 0
reward += instant_pnl
self._total_pnl += instant_pnl
self._total_reward += reward
# Game over logic
try:
self._prices_history.append(self._data_generator.next())
except StopIteration:
done = True
info['status'] = 'No more data.'
if self._iteration >= self._episode_length:
done = True
info['status'] = 'Time out.'
if self._closed_plot:
info['status'] = 'Closed plot'
observation = self._get_observation()
return observation, reward, done, info
