black

Author: HenrikMettler

cgp/local_search/random_search.py

@@ -4,11 +4,11 @@
 
 class RandomLocalSearch:
     def __init__(
-            self,
-            objective: Callable,
-            seed: int,
-            n_steps: int = 10,  # todo: should this be dependent on the number of parameters?
-            permissible_values: np.ndarray = np.logspace(start=0, stop=7, num=8, base=2, dtype=int),
+        self,
+        objective: Callable,
+        seed: int,
+        n_steps: int = 10,  # todo: should this be dependent on the number of parameters?
+        permissible_values: np.ndarray = np.logspace(start=0, stop=7, num=8, base=2, dtype=int),
     ) -> None:
         self.objective = objective
         self.seed = seed
@@ -23,15 +23,21 @@ def __call__(self, ind) -> None:
         if len(params_values) > 0:
             for _ in range(self.n_steps):
                 # sample a new set of parameter values randomly
-                params_sampled = [rng.choice(self.permissible_values) for param_value in params_values]
+                params_sampled = [
+                    rng.choice(self.permissible_values) for param_value in params_values
+                ]
                 # write the parameters into a clone of individual
                 new_ind = ind.clone()
-                new_ind.update_parameters_from_numpy_array(params=params_sampled, params_names=params_names)
+                new_ind.update_parameters_from_numpy_array(
+                    params=params_sampled, params_names=params_names
+                )
                 # evaluate fitness
                 self.objective(new_ind)
                 # if fitness improved: replace parameter values and fitness
                 if new_ind.fitness >= ind.fitness:  # todo: should this be >= or > ??
-                    ind.update_parameters_from_numpy_array(params=params_sampled, params_names=params_names)
+                    ind.update_parameters_from_numpy_array(
+                        params=params_sampled, params_names=params_names
+                    )
                     ind.fitness = new_ind.fitness
 
 
@@ -49,8 +55,12 @@ def objective(ind):
     ind = cgp.IndividualSingleGenome(genome=genome)
 
     objective(ind)
-    print(f'Node parameters before local search {ind.parameters_to_numpy_array(only_active_nodes=True)} \n')
+    print(
+        f"Node parameters before local search {ind.parameters_to_numpy_array(only_active_nodes=True)} \n"
+    )
 
     rls = RandomLocalSearch(objective=objective, seed=seed, n_steps=1000)
     rls(ind)
-    print(f'Node parameters after local search {ind.parameters_to_numpy_array(only_active_nodes=True)} \n')
+    print(
+        f"Node parameters after local search {ind.parameters_to_numpy_array(only_active_nodes=True)} \n"
+    )

cgp/node.py

@@ -178,7 +178,7 @@ def __init_subclass__(cls: Type["OperatorNode"]) -> None:
         node_validation.check_to_func(cls)
         node_validation.check_to_numpy(cls)
         node_validation.check_to_torch(cls)
-        #node_validation.check_to_sympy(cls)
+        node_validation.check_to_sympy(cls)
 
     @classmethod
     def _extract_input_names_from_def_output(self, cls: Type["OperatorNode"]) -> None: