def setUpClass(cls) -> None:
cls.window = Window(window_start=date(2008, 1, 3),
trading_win_len=timedelta(days=90),
repository=DataRepository())
cls.empty_portfolio: Portfolio = Portfolio(100_000, cls.window)
cls.portfolio_to_be_augmented = Portfolio(100_000, cls.window)
cls.decision_maker = SignalGenerator(cls.empty_portfolio,
entry_z=0.5,
exit_z=0.1,
emergency_z=0.8)
cls.pair1 = CointegratedPair(pair=(SnpTickers.AAPL, SnpTickers.MSFT),
mu_x_ann=0.0,
sigma_x_ann=0.0,
scaled_beta=1.2,
hl=0.0,
ou_mean=0.0,
ou_std=0.0,
ou_diffusion_v=0.0,
recent_dev=0.0,
recent_dev_scaled=0.1)
cls.pair2 = CointegratedPair(pair=(SnpTickers.AAPL, SnpTickers.TDG),
mu_x_ann=0.0,
sigma_x_ann=0.0,
scaled_beta=0.5,
hl=0.0,
ou_mean=0.0,
ou_std=0.0,
ou_diffusion_v=0.0,
recent_dev=0.0,
recent_dev_scaled=0.2)
class Application:
def __init__(self):
self.running = True
self.window = Window(self, 900, 600, 'Boxes')
self.game_state = GameState(self)
def run(self):
"""
Main function of the game
:return: None
"""
self.window.init()
self.game_state.preupdate()
while self.running:
self.game_state.update()
self.window.display()
def goToPlay():
witdthofacell = 5
numberOfLine = 100
numberOfColumn = 100
mondamier = Damier(witdthofacell, numberOfLine, numberOfColumn)
regles = Rule('classic')
myWindow = Window(mondamier, regles)
def __init__(self,
logger: logging.Logger,
backtest_start: date = date(2008, 1, 2),
backtest_end: Optional[date] = None,
target_number_of_coint_pairs: int = 100,
max_active_pairs: int = 10,
clust_and_coint_frequency_per_window=8,
window_length: timedelta = timedelta(days=120),
trading_freq: timedelta = timedelta(days=1),
adf_confidence_level: AdfPrecisions = AdfPrecisions.ONE_PCT,
max_mean_rev_time: int = 30,
hurst_exp_threshold: float = 0.15,
entry_z_lower_bound: float = 2,
entry_z_upper_bound: float = 3,
emergency_delta_z: float = 2,
exit_delta_z: float = 1):
# If end_date is None, run for the entirety of the dataset
# Window is the lookback period (from t=window_length to t=0 (today) over which we analyse data
# to inform us on trades to make on t=0 (today).
self.logger: logging.Logger = logger
self.backtest_start: date = backtest_start
self.max_active_pairs: int = max_active_pairs
self.clust_and_coint_frequency_per_window: int = clust_and_coint_frequency_per_window
self.window_length: timedelta = window_length
self.trading_freq: timedelta = trading_freq
self.adf_confidence_level: AdfPrecisions = adf_confidence_level # e.g. "5%" or "1%"
self.max_mean_rev_time: int = max_mean_rev_time
self.hurst_exp_threshold: float = hurst_exp_threshold
self.entry_z_lower_bound: float = entry_z_lower_bound
self.entry_z_upper_bound: float = entry_z_upper_bound
self.emergency_delta_z: float = emergency_delta_z
self.exit_delta_z: float = exit_delta_z
# Last SNP date, hard coded for now...
self.backtest_end = date(year=2019, month=12, day=31) if backtest_end is None else backtest_end
self.repository = DataRepository(self.window_length)
self.current_window: Window = Window(window_start=backtest_start,
win_len=window_length,
repository=self.repository)
self.today = self.current_window.window_business_days[-1]
self.day_count: int = 0
self.last_traded_date: Optional[date] = None
self.clusterer = Clusterer()
self.cointegrator = Cointegrator(self.repository,
target_number_of_coint_pairs,
self.adf_confidence_level,
self.max_mean_rev_time,
previous_cointegrated_pairs=[])
self.filters = Filters()
self.portfolio: Portfolio = Portfolio(100_000, self.current_window)
self.signalgenerator = SignalGenerator(self.portfolio, entry_z_lower_bound,
entry_z_upper_bound, exit_delta_z,
emergency_delta_z, max_active_pairs)
def __init__(self,
backtest_start: date = date(2008, 1, 2),
trading_window_length: timedelta = timedelta(days=90),
backtest_end: Optional[date] = None,
adf_confidence_level: AdfPrecisions = AdfPrecisions.FIVE_PCT,
max_mean_rev_time: int = 15,
hurst_exp_threshold: float = 0.35,
entry_z: float = 1.5,
emergency_z: float = 3,
exit_z: float = 0.5):
# If end_date is None, run for the entirety of the dataset
# Window is the lookback period (from t=window_length to t=0 (today) over which we analyse data
# to inform us on trades to make on t=0 (today).
# We assume an expanding window for now.
self.backtest_start: date = backtest_start
self.window_length: timedelta = trading_window_length
self.adf_confidence_level: AdfPrecisions = adf_confidence_level # e.g. "5%" or "1%"
self.max_mean_rev_time: int = max_mean_rev_time
self.entry_z: float = entry_z
self.exit_z: float = exit_z
self.hurst_exp_threshold: float = hurst_exp_threshold
# if the pair crosses this boundary, we don't believe their cointegrated anymore
# - close the position at a loss
# require: emergency_z > entry_z > exit_z
self.emergency_z: float = emergency_z
# Last SNP date, hard coded for now...
self.backtest_end = date(
year=2020, month=12,
day=31) if backtest_end is None else backtest_end
self.repository = DataRepository()
initial_window = Window(window_start=backtest_start,
trading_win_len=trading_window_length,
repository=self.repository)
self.current_window: Window = initial_window
self.history: List[Window] = [initial_window]
self.trading_days = initial_window.window_trading_days
self.today = self.trading_days[-1]
self.days_alive: int = (self.today - self.backtest_start).days
self.clusterer = Clusterer()
self.cointegrator = Cointegrator(self.repository,
self.adf_confidence_level,
self.max_mean_rev_time, self.entry_z,
self.exit_z, initial_window,
self.history[-1])
self.risk_manager = RiskManager(self.entry_z, self.exit_z)
self.filters = Filters()
self.portfolio: Portfolio = Portfolio(100_000, self.current_window)
self.dm = SignalGenerator(self.portfolio, entry_z, emergency_z, exit_z)
def run_one_window(win_info):
window = Window(win_info)
window.run_window_call_variant()
return window
def run_one_window(win_info):
window = Window(win_info, tech=get_value("paras")["tech"])
window.run_window_benchmark()
return window
def test_window_initialised_with_non_trading_day(self):
# 1st of Jan isn't a trading day; the production code should raise a KeyError
with self.assertRaises(KeyError):
Window(date(2008, 1, 1), timedelta(days=10), DataRepository())
def setUpClass(cls) -> None:
cls.initial = Window(date(2008, 1, 2), timedelta(days=10),
DataRepository())
def run_GUI_demo_2():
app = QtWidgets.QApplication(sys.argv)
wTitle: str = progName + "run_GUI_demo_2 (version : " + version + ")"
GUI = Window()
GUI.show()
sys.exit(app.exec_())
def __init__(self):
self.running = True
self.window = Window(self, 900, 600, 'Boxes')
self.game_state = GameState(self)
class UI:
def __init__(self, initial, final, level, pivot):
self.simulator = Simulator(initial, final, level, pivot)
self.background = Window(*(0, 0), *os.get_terminal_size(),
*os.get_terminal_size(), '01 ')
def render_background_with_form(self, form):
x = self.background.x + (self.background.width - form.width) // 2
y = self.background.y + (self.background.height - form.height) // 2
form.set_x_y(x, y)
for i in range(100000):
self.background.render_window(form)
time.sleep(0.3)
self.background.push_new_line_on_window()
def one_game(self, form):
tick = 0.3
form.set_string('Enter size:', *(1, 1))
time.sleep(tick)
size = form.get_and_set_string_from_input(2, 1)
data = self.simulator.get_data(int(size))
initial = self.simulator.get_initial(data)
form.set_string(initial, *(3, 1))
time.sleep(tick)
form.set_string('Enter string:', *(4 + 2, 1))
time.sleep(tick)
user_input = form.get_and_set_string_from_input(5 + 2, 1)
form.set_string('Result:', *(6 + 2, 1))
time.sleep(tick)
result = self.simulator.validate_input(user_input, data)
if result['status'] == 'SUCCESS':
i = 7
form.set_string('SUCCESS', *(7 + 2, 1))
time.sleep(tick)
else:
for i, message in enumerate(result['messages'], start=7 + 2):
result = ''
if message[0] == 'incompatible':
result = f'#{message[1]}, expected {message[3]} but was {message[2]}'
else:
result = f'Incorrect size'
form.set_string(result, *(i, 1))
time.sleep(tick)
form.set_string('pause', *(i + 1 + 2, 1))
time.sleep(tick)
user_input = form.get_and_set_string_from_input(i + 2 + 2, 1)
form.clean()
def run(self):
width, height = os.get_terminal_size()
form = Form(width // 2, height // 2)
thread = threading.Thread(target=self.render_background_with_form,
args=([form]))
thread.start()
while True:
self.one_game(form)
def __init__(self, initial, final, level, pivot):
self.simulator = Simulator(initial, final, level, pivot)
self.background = Window(*(0, 0), *os.get_terminal_size(),
*os.get_terminal_size(), '01 ')
import pygame
from src.Game import Game
from src.Window import Window
from src.Menu import Menu
from src.State import State
if __name__ == '__main__':
pygame.init()
window = Window()
current_state = Menu(window)
while window.is_open():
window.fill_base()
for event in pygame.event.get():
if event.type == pygame.QUIT or State.is_key_pressed(
pygame.K_ESCAPE):
window.close()
pygame.quit()
print("UPDATE: Game window closed")
break
state_name = current_state.state
current_state.keys_pressed_reaction()
if state_name != current_state.state:
if current_state.state == State.GAME_STATE or current_state.state == State.RESTART_STATE:
# Create new game
current_state = Game(window)
elif current_state.state == State.AI_GAME_STATE or current_state.state == State.AI_RESTART_STATE:
current_state = Game(window, State.AI_GAME_STATE)
elif current_state.state == State.MENU_STATE:
current_state = Menu(window)
def __init__(self):
Window.Window()
return pair_expected_return
def __mean_variance_optimizer(self, cov_matrix, expected_return):
cov_matrix_inv = np.linalg.inv(cov_matrix.values)
weight = cov_matrix_inv.dot(expected_return)
scale = weight.sum()
weight = weight / scale
return weight[:,0]
from datetime import date, timedelta
from src.DataRepository import DataRepository
if __name__ == '__main__':
win = Window(window_start=date(2008, 1, 2),
trading_win_len=timedelta(days=90),
repository=DataRepository())
test_input = [
CointegratedPair(pair = (SnpTickers.CRM, SnpTickers.WU),
mu_x_ann = 0.1,
sigma_x_ann = 0.3,
scaled_beta = 1.2,
hl = 7.5,
ou_mean = -0.017,
ou_std = 0.043,
ou_diffusion_v = 0.70,
recent_dev= 0.071,
recent_dev_scaled = 2.05),
CointegratedPair(pair=(SnpTickers.ABC, SnpTickers.ABT),
def generate_pairs(self, clustering_results: Dict[int,
Tuple[Tuple[Tickers]]],
hurst_exp_threshold: float, current_window: Window):
# run cointegration_analysis on all poss combinations of pairs within the same cluster
current_cointegrated_pairs = []
n_cointegrated = 0
tickers_per_cluster = [i for i in clustering_results.values()]
for cluster in tickers_per_cluster:
for pair in itertools.combinations(list(cluster), 2):
t1 = current_window.get_data(tickers=[pair[0]],
features=[Features.CLOSE])
t2 = current_window.get_data(tickers=[pair[1]],
features=[Features.CLOSE])
try:
# sometimes there are no price data, in which case, skip
residuals, beta, reg_output = self.__logged_lin_reg(t1, t2)
except ValueError:
continue
adf_test_statistic, adf_critical_values = self.__adf(
residuals.flatten())
hl_test = self.__hl(residuals)
he_test = self.__hurst_exponent_test(residuals, current_window)
ou_mean, ou_std, ou_diffusion_v, \
recent_dev, recent_dev_scaled = self.__ou_params(residuals)
ols_stdzed_residuals = (residuals - ou_mean) / ou_std
is_cointegrated = self.__acceptance_rule(
adf_test_statistic,
adf_critical_values,
self.adf_confidence_level,
hl_test,
self.max_mean_rev_time,
he_test,
hurst_exp_threshold,
ols_stdzed_residuals,
at_least=int(current_window.window_length.days / 6))
if is_cointegrated:
#a = pd.concat([t1, t2], axis=1).iplot(asFigure=True)
#b = pd.concat([np.log(t1), np.log(t2)], axis=1).iplot(asFigure=True)
#a.show()
#b.show()
n_cointegrated += 1
t1_most_recent = float(t1.iloc[-1, :])
t2_most_recent = float(t2.iloc[-1, :])
hedge_ratio = beta * t1_most_recent / t2_most_recent
scaled_beta = hedge_ratio / (hedge_ratio - 1)
recent_dev_scaled_hist = [recent_dev_scaled]
cointegration_rank = self.__score_coint(
adf_test_statistic, self.adf_confidence_level,
adf_critical_values, he_test, hurst_exp_threshold, 10)
#a = pd.DataFrame(ols_stdzed_residuals).iplot(asFigure=True)
#a.show()
position = Position(pair[0], pair[1])
current_cointegrated_pairs.append(
CointegratedPair(pair, reg_output, scaled_beta,
hl_test, ou_mean, ou_std,
ou_diffusion_v, recent_dev,
recent_dev_scaled,
recent_dev_scaled_hist,
cointegration_rank,
ols_stdzed_residuals, position))
if n_cointegrated == self.target_number_of_coint_pairs:
current_cointegrated_pairs = sorted(
current_cointegrated_pairs,
key=lambda coint_pair: coint_pair.
cointegration_rank,
reverse=True)
self.previous_cointegrated_pairs = current_cointegrated_pairs
return current_cointegrated_pairs
self.previous_cointegrated_pairs = current_cointegrated_pairs
return current_cointegrated_pairs
def generate_pairs(self, clustering_results: Dict[int,
Tuple[Tuple[Tickers]]],
hurst_exp_threshold: float, current_window: Window):
# run cointegration_analysis on all poss combinations of pairs
current_cointegrated_pairs = []
n_cointegrated = 0
list_of_lists = [i for i in clustering_results.values()]
flattened = [pair for x in list_of_lists for pair in x]
sorted_cluster_results = sorted(flattened, key=lambda x: x[0].value)
for pair in sorted_cluster_results:
t1 = current_window.get_data(universe=Universes.SNP,
tickers=[pair[0]],
features=[Features.CLOSE])
t2 = current_window.get_data(universe=Universes.SNP,
tickers=[pair[1]],
features=[Features.CLOSE])
try:
# sometimes there are no price data
residuals, beta, reg_output = self.__logged_lin_reg(t1, t2)
except:
continue
# for some reason residuals is a (60,1) array not (60,) array when i run the code so have changed input to residuals.flatten
adf_test_statistic, adf_critical_values = self.__adf(
residuals.flatten())
hl_test = self.__hl(residuals)
he_test = self.__hurst_exponent_test(residuals, current_window)
is_cointegrated = self.__acceptance_rule(
adf_test_statistic, adf_critical_values,
self.adf_confidence_level, hl_test, self.max_mean_rev_time,
he_test, hurst_exp_threshold)
target_coint = 300
if is_cointegrated:
n_cointegrated += 1
print(n_cointegrated, " / ", target_coint, "cointegrated")
r_x = self.__log_returner(t1)
mu_x_ann = float(250 * np.mean(r_x))
sigma_x_ann = float(250**0.5 * np.std(r_x))
ou_mean, ou_std, ou_diffusion_v, recent_dev, recent_dev_scaled = self.__ou_params(
residuals)
scaled_beta = beta / (beta - 1)
recent_dev_scaled_hist = [recent_dev_scaled]
cointegration_rank = self.__score_coint(
adf_test_statistic, self.adf_confidence_level,
adf_critical_values)
current_cointegrated_pairs.append(
CointegratedPair(pair, mu_x_ann, sigma_x_ann, reg_output,
scaled_beta, hl_test, ou_mean, ou_std,
ou_diffusion_v, recent_dev,
recent_dev_scaled, recent_dev_scaled_hist,
cointegration_rank))
if n_cointegrated == target_coint:
# logic to be fixed and made more efficient by: 1) having proper
# clustering algorithm; 2) not running clustering and cointegration
# everyday 3) taking best 10 pairs according to some score
current_cointegrated_pairs = sorted(
current_cointegrated_pairs,
key=lambda coint_pair: coint_pair.cointegration_rank,
reverse=True)
self.previous_cointegrated_pairs = current_cointegrated_pairs
return current_cointegrated_pairs
self.previous_cointegrated_pairs = current_cointegrated_pairs
return current_cointegrated_pairs
