def bet_size_probability(events,
prob,
num_classes,
pred=None,
step_size=0.0,
average_active=False,
num_threads=1):
"""
Calculates the bet size using the predicted probability. Note that if 'average_active' is True, the returned
pandas.Series will be twice the length of the original since the average is calculated at each bet's open and close.
:param events: (pandas.DataFrame) Contains at least the column 't1', the expiry datetime of the product, with
a datetime index, the datetime the position was taken.
:param prob: (pandas.Series) The predicted probability.
:param num_classes: (int) The number of predicted bet sides.
:param pred: (pd.Series) The predicted bet side. Default value is None which will return a relative bet size
(i.e. without multiplying by the side).
:param step_size: (float) The step size at which the bet size is discretized, default is 0.0 which imposes no
discretization.
:param average_active: (bool) Option to average the size of active bets, default value is False.
:param num_threads: (int) The number of processing threads to utilize for multiprocessing, default value is 1.
:return: (pandas.Series) The bet size, with the time index.
"""
signal_0 = get_signal(prob, num_classes, pred)
events_0 = signal_0.to_frame('signal').join(events['t1'], how='left')
if average_active:
signal_1 = avg_active_signals(events_0, num_threads)
else:
signal_1 = events_0.signal
if abs(step_size) > 0:
signal_1 = discrete_signal(signal0=signal_1, step_size=abs(step_size))
return signal_1
def test_avg_active_signals(self):
"""
Tests the avg_active_signals function. Also implicitly tests the
molecular multiprocessing function mp_avg_active_signals.
"""
test_avg_active = avg_active_signals(self.events_2)
self.assertEqual(self.avg_active.equals(test_avg_active), True)
def test_bet_size_probability_avg_active(self):
"""
Tests for successful execution of 'bet_size_probability' with 'average_active' set to True.
"""
# Setup the test DataFrame.
dates_test = np.array([
dt.datetime(2000, 1, 1) + i * dt.timedelta(days=1)
for i in range(5)
])
shift_dt = np.array([dt.timedelta(days=0.5 * i + 1) for i in range(5)])
dates_shifted_test = dates_test + shift_dt
events_test = pd.DataFrame(data=[[0.55, 1], [0.7, -1], [0.95, 1],
[0.65, -1], [0.85, 1]],
columns=['prob', 'side'],
index=dates_test)
events_test['t1'] = dates_shifted_test
# Calculate correct output.
signal_0 = get_signal(events_test['prob'], 2, events_test['side'])
df_signal_0 = signal_0.to_frame('signal').join(events_test['t1'],
how='left')
signal_1 = avg_active_signals(df_signal_0, 1)
# Evaluate test.
self.assertTrue(
signal_1.equals(
bet_size_probability(events=events_test,
prob=events_test['prob'],
num_classes=2,
pred=events_test['side'],
average_active=True)))