Objective wrapper in utils. Check if individual has fitness in wrappe…

…r. Print progres depending on generation or evaluation

cgp/hl_api.py

@@ -53,8 +53,11 @@ def evolve(
     None
     """
     if max_generations == np.inf and max_evaluations == np.inf:
-        raise ('Faulty configuration of evolve parameters: '
-               'Either max_generations or max_evaluation must be finite, to avoid non terminal evolutionary process')
+        raise (
+            "Faulty configuration of evolve parameters: "
+            "Either max_generations or max_evaluation must be finite, "
+            "to avoid non terminal evolutionary process"
+        )
 
     ea.initialize_fitness_parents(pop, objective)
     if callback is not None:
@@ -75,11 +78,20 @@ def evolve(
             max_fitness = pop.champion.fitness
 
         if print_progress:
-            print(
-                f"\r[{pop.generation + 1}/{max_generations}] max fitness: {max_fitness}\033[K",
-                end="",
-                flush=True,
-            )
+            if not max_generations == np.inf:
+                print(
+                    f"\r[{pop.generation + 1}/{max_generations}] max fitness: {max_fitness}\033[K",
+                    end="",
+                    flush=True,
+                )
+            elif not max_evaluations == np.inf:
+                print(
+                    f"\r[{objective.calls}/{max_evaluations}] max fitness: {max_fitness}\033[K",
+                    end="",
+                    flush=True,
+                )
+            else:
+                assert False  # should never be reached
 
         if callback is not None:
             callback(pop)

cgp/utils.py

@@ -6,7 +6,7 @@
 
 import numpy as np
 
-from .individual import IndividualMultiGenome, IndividualSingleGenome
+from .individual import IndividualBase, IndividualMultiGenome, IndividualSingleGenome
 from .node import Node, primitives_dict
 
 
@@ -201,3 +201,26 @@ def primitives_from_class_names(primitives_str: Tuple[str, ...]) -> Tuple[Type[N
         primitives.append(primitives_dict[s])
 
     return tuple(primitives)
+
+
+def objective(f: Callable[[IndividualBase], IndividualBase]):
+    """ Wrapper for the objective function.
+
+    Wraps the objective function and counts the calls of the function.
+
+    Parameters
+    ----------
+    f: Callable
+        The objective function.
+    """
+
+    @functools.wraps(f)
+    def wrapped(individual: IndividualBase):
+        if individual.fitness is None:
+            wrapped.calls += 1
+            return f(individual)
+        else:
+            return individual
+
+    wrapped.calls = 0
+    return wrapped

examples/example_minimal.py

@@ -20,7 +20,6 @@
 import matplotlib.pyplot as plt
 import numpy as np
 import scipy.constants
-import functools
 from docopt import docopt
 
 import cgp
@@ -41,20 +40,9 @@ def f_target(x):
 # represented by a given individual and the target function evaluated
 # on a set of random points.
 
-def counted(f):
-    @functools.wraps(f)
-    def wrapped(*args, **kwargs):
-        wrapped.calls += 1
-        return f(*args, **kwargs)
-    wrapped.calls = 0
-    return wrapped
 
-
-@counted
+@cgp.utils.objective
 def objective(individual):
-    if individual.fitness is not None:
-        objective.calls -= 1
-        return individual
 
     n_function_evaluations = 1000
 
@@ -92,7 +80,8 @@ def objective(individual):
 
 ea_params = {"n_offsprings": 4, "tournament_size": 2, "n_processes": 1}
 
-evolve_params = {"max_generations": int(args["--max-generations"]), "min_fitness": 0.0}
+evolve_params = {"max_generations": np.inf, "min_fitness": 0.0, "max_evaluations": 1000}
+# Todo: change max_generations: int(args["--max-generations"])
 
 # %%
 # We create a population that will be evolved