Cache for GARCH variance. Speed increase x10.

backtest.py

@@ -26,27 +26,27 @@
 start_date = "2020-01-02"
 end_date = "2020-01-05" # "2023-05-31"
 date_range = pd.date_range(start=start_date, end=end_date, freq='B')
-solutions = {}
+solutions = {"2020-01-01": inital_weights}
 times = {}
 
 for date in date_range:
     t1 = time.time()
     universe.date_today = date
     simulator = Simulator(universe)
     simulator.calibrate()
-    simulator.run_simulation(time_steps=10, number_of_simulations=1000)
-    optimizer = MPCMaximumReturn(universe, portfolio, simulator.return_tensor, gamma=1, l1_penalty=0, l2_penalty=0.05,
+    simulator.run_simulation(time_steps=10, number_of_simulations=1_000)
+    optimizer = MPCMaximumReturn(universe, portfolio, simulator.return_tensor, gamma=2, l1_penalty=0, l2_penalty=0.15,
                                  covariance_matrix=simulator.covariance_matrix)
     optimizer.solve()
-
     solution = optimizer.solution
     times[date] = time.time() - t1
     portfolio.weights = solution
     solutions[date] = solution
 
+
 print('\n' * 10)
 for date, solution in solutions.items():
-    print(f"Took {times[date]} seconds.")
+    print(f"Took {times.get(date, -1)} seconds.")
     for a, w in solution.items():
         print(a, w)
     print()

src/abacus/models/garch.py

@@ -1,4 +1,8 @@
 # -*- coding: utf-8 -*-
+import time
+
+from cytoolz import memoize
+
 import torch
 import numpy as np
 
@@ -59,7 +63,6 @@ def transform_to_true(self, uniform_sample: torch.Tensor) -> torch.Tensor:
         parameters = torch.tensor(self._solution.x)
         mu_corr, mu_ewma = self._intermediary_parameters(parameters=parameters)
 
-
         variance = self._compute_variance(parameters=torch.tensor(self._solution.x))[-1]
         squared_return = self._squared_returns[-1]
 
@@ -125,6 +128,7 @@ def _compute_inital_variance(self) -> torch.Tensor:
             return torch.square(torch.std(self._data[:INITIAL_VARIANCE_GARCH_OBSERVATIONS]))
         return self._initial_squared_returns
 
+    @memoize
     def _compute_variance(self, parameters: torch.Tensor) -> torch.Tensor:
         initial_variance = self._initial_variance
         variances = torch.zeros(self._number_of_observations)

src/abacus/simulator/simulator.py

@@ -1,4 +1,6 @@
 # -*- coding: utf-8 -*-
+import time
+
 import torch
 import numpy as np
 import pyvinecopulib as pv