Merge pull request #153 from jakobj/maint/simplify-individual

Simplify Individual: move mutate() to Genome

cgp/cartesian_graph.py

@@ -17,18 +17,20 @@
 except ModuleNotFoundError:
     torch_available = False
 
-from typing import Callable, DefaultDict, Dict, List, Optional, Set
+from typing import Callable, DefaultDict, Dict, List, Optional, Set, TYPE_CHECKING
 
 from .node import Node, InputNode, OutputNode, Parameter
-from .genome import Genome
+
+if TYPE_CHECKING:
+    from .genome import Genome
 
 
 class CartesianGraph:
     """Class representing a particular Cartesian graph defined by a
     Genome.
     """
 
-    def __init__(self, genome: Genome) -> None:
+    def __init__(self, genome: "Genome") -> None:
         """Init function.
 
         Parameters
@@ -95,7 +97,7 @@ def empty_node_str() -> str:
 
         return s
 
-    def parse_genome(self, genome: Genome) -> None:
+    def parse_genome(self, genome: "Genome") -> None:
         if genome.dna is None:
             raise RuntimeError("dna not initialized")
 

cgp/genome.py

@@ -10,6 +10,7 @@
 
 from typing import DefaultDict, Generator, List, Optional, Tuple, Type
 
+from .cartesian_graph import CartesianGraph
 from .node import Node
 from .primitives import Primitives
 
@@ -355,25 +356,26 @@ def _is_active_input_gene(self, gene_idx: int) -> bool:
         else:
             assert False  # should never be reached
 
-    def mutate(self, n_mutations: int, active_regions: List[int], rng: np.random.RandomState):
+    def mutate(self, mutation_rate: float, rng: np.random.RandomState):
         """Mutate the genome.
 
         Parameters
         ----------
-        n_mutations : int
-            Number of entries in the genome to be mutated.
-        active_regions: List[int]
-            Regions in the genome that are currently used in the
-            computational graph. Used to check whether mutations are
-            silent or require reevaluation of fitness.
+        mutation_rate : float
+            Proportion of genes to be mutated, between 0 and 1.
         rng : numpy.random.RandomState
             Random number generator instance to use for crossover.
 
         Returns
         ----------
-        True if only inactive regions of the genome were mutated, False otherwise.
+        bool
+            True if only inactive regions of the genome were mutated, False otherwise.
         """
-        assert isinstance(n_mutations, int) and 0 < n_mutations
+        n_mutations = int(mutation_rate * len(self.dna))
+        assert n_mutations > 0
+
+        graph = CartesianGraph(self)
+        active_regions = graph.determine_active_regions()
 
         successful_mutations = 0
         only_silent_mutations = True

cgp/individual.py

@@ -59,46 +59,10 @@ def update_parameters_from_torch_class(self, torch_cls):
 
     @staticmethod
     def _mutate_genome(genome: Genome, mutation_rate: float, rng: np.random.RandomState) -> bool:
-        """Mutate a given genome.
-
-        Parameters
-        ----------
-        genome : Genome
-            Genome to be mutated.
-        mutation_rate : float
-            Proportion of genes to be mutated, between 0 and 1.
-        rng : numpy.RandomState
-            Random number generator instance.
-
-        Returns
-        -------
-        bool
-            Whether all mutations were silent.
-        """
-        n_mutations = int(mutation_rate * len(genome.dna))
-        assert n_mutations > 0
-
-        graph = CartesianGraph(genome)
-        active_regions = graph.determine_active_regions()
-        only_silent_mutations = genome.mutate(n_mutations, active_regions, rng)
-
-        return only_silent_mutations
+        return genome.mutate(mutation_rate, rng)
 
     @staticmethod
     def _randomize_genome(genome: Genome, rng: np.random.RandomState) -> None:
-        """Randomize the individual's genome.
-
-        Parameters
-        ----------
-        genome : Genome
-            Genome to be randomized.
-        rng : numpy.RandomState
-            Random number generator instance to use for crossover.
-
-        Returns
-        -------
-        None
-        """
         genome.randomize(rng)
 
     @staticmethod