"""Handles genomes (individuals in the population)."""
import copy
import sys
from itertools import count
from random import choice, random, shuffle
from neat.activations import ActivationFunctionSet
from neat.aggregations import AggregationFunctionSet
from neat.config import ConfigParameter, write_pretty_params
from neat.genes import DefaultConnectionGene, DefaultNodeGene
from neat.graphs import creates_cycle
from neat.graphs import required_for_output
[docs]
class DefaultGenomeConfig:
"""Sets up and holds configuration information for the DefaultGenome class."""
allowed_connectivity = ['unconnected', 'fs_neat_nohidden', 'fs_neat', 'fs_neat_hidden',
'full_nodirect', 'full', 'full_direct',
'partial_nodirect', 'partial', 'partial_direct']
def __init__(self, params, section_name='DefaultGenome'):
# Create full set of available activation functions.
self.activation_defs = ActivationFunctionSet()
# ditto for aggregation functions - name difference for backward compatibility
self.aggregation_function_defs = AggregationFunctionSet()
self.aggregation_defs = self.aggregation_function_defs
self._params = [ConfigParameter('num_inputs', int),
ConfigParameter('num_outputs', int),
ConfigParameter('num_hidden', int),
ConfigParameter('feed_forward', bool),
ConfigParameter('compatibility_disjoint_coefficient', float),
ConfigParameter('compatibility_weight_coefficient', float),
ConfigParameter('conn_add_prob', float),
ConfigParameter('conn_delete_prob', float),
ConfigParameter('node_add_prob', float),
ConfigParameter('node_delete_prob', float),
ConfigParameter('single_structural_mutation', bool, 'false'),
ConfigParameter('structural_mutation_surer', str, 'default'),
ConfigParameter('initial_connection', str, 'unconnected')]
# Gather configuration data from the gene classes.
self.node_gene_type = params['node_gene_type']
self._params += self.node_gene_type.get_config_params()
self.connection_gene_type = params['connection_gene_type']
self._params += self.connection_gene_type.get_config_params()
# Use the configuration data to interpret the supplied parameters.
for p in self._params:
setattr(self, p.name, p.interpret(params, section_name))
self.node_gene_type.validate_attributes(self)
self.connection_gene_type.validate_attributes(self)
# By convention, input pins have negative keys, and the output
# pins have keys 0,1,...
self.input_keys = [-i - 1 for i in range(self.num_inputs)]
self.output_keys = [i for i in range(self.num_outputs)]
self.connection_fraction = None
# Verify that initial connection type is valid.
# pylint: disable=access-member-before-definition
if 'partial' in self.initial_connection:
c, p = self.initial_connection.split()
self.initial_connection = c
self.connection_fraction = float(p)
if not (0 <= self.connection_fraction <= 1):
raise RuntimeError(
"'partial' connection value must be between 0.0 and 1.0, inclusive.")
assert self.initial_connection in self.allowed_connectivity
# Verify structural_mutation_surer is valid.
# pylint: disable=access-member-before-definition
if self.structural_mutation_surer.lower() in ['1', 'yes', 'true', 'on']:
self.structural_mutation_surer = 'true'
elif self.structural_mutation_surer.lower() in ['0', 'no', 'false', 'off']:
self.structural_mutation_surer = 'false'
elif self.structural_mutation_surer.lower() == 'default':
self.structural_mutation_surer = 'default'
else:
error_string = f"Invalid structural_mutation_surer {self.structural_mutation_surer!r}"
raise RuntimeError(error_string)
self.node_indexer = None
# Innovation tracker will be set by Population/Reproduction
# This enables same-generation deduplication per NEAT paper (Stanley & Miikkulainen, 2002)
self.innovation_tracker = None
[docs]
def add_activation(self, name, func):
self.activation_defs.add(name, func)
[docs]
def add_aggregation(self, name, func):
self.aggregation_function_defs.add(name, func)
[docs]
def save(self, f):
if 'partial' in self.initial_connection:
if not (0 <= self.connection_fraction <= 1):
raise RuntimeError(
"'partial' connection value must be between 0.0 and 1.0, inclusive.")
f.write(f'initial_connection = {self.initial_connection} {self.connection_fraction}\n')
else:
f.write(f'initial_connection = {self.initial_connection}\n')
assert self.initial_connection in self.allowed_connectivity
write_pretty_params(f, self, [p for p in self._params
if 'initial_connection' not in p.name])
[docs]
def get_new_node_key(self, node_dict):
if self.node_indexer is None:
if node_dict:
self.node_indexer = count(max(list(node_dict)) + 1)
else:
# No existing nodes; start from num_outputs.
self.node_indexer = count(self.num_outputs)
new_id = next(self.node_indexer)
assert new_id not in node_dict
return new_id
[docs]
def check_structural_mutation_surer(self):
if self.structural_mutation_surer == 'true':
return True
elif self.structural_mutation_surer == 'false':
return False
elif self.structural_mutation_surer == 'default':
return self.single_structural_mutation
else:
error_string = f"Invalid structural_mutation_surer {self.structural_mutation_surer!r}"
raise RuntimeError(error_string)
def __getstate__(self):
"""Prepare config for pickling by converting node_indexer to a picklable form."""
state = self.__dict__.copy()
# Convert the itertools.count object to an integer representing the next value
# We peek at the value by calling next() and storing it, then the counter advances
if self.node_indexer is not None:
# We need to save the next value that would be returned
# Since calling next() consumes the value, we save it and will recreate
# the counter starting from that value when unpickling
state['_node_indexer_next_value'] = next(self.node_indexer)
state['node_indexer'] = None
else:
state['_node_indexer_next_value'] = None
return state
def __setstate__(self, state):
"""Restore config from pickled state, recreating node_indexer."""
_node_indexer_next_value = state.pop('_node_indexer_next_value', None)
self.__dict__.update(state)
# Recreate the count object starting from the saved next value
if _node_indexer_next_value is not None:
# Recreate counter starting from the value we saved
self.node_indexer = count(_node_indexer_next_value)
else:
self.node_indexer = None
[docs]
class DefaultGenome:
"""
A genome for generalized neural networks.
Terminology
pin: Point at which the network is conceptually connected to the external world;
pins are either input or output.
node: Analog of a physical neuron.
connection: Connection between a pin/node output and a node's input, or between a node's
output and a pin/node input.
key: Identifier for an object, unique within the set of similar objects.
Design assumptions and conventions.
1. Each output pin is connected only to the output of its own unique
neuron by an implicit connection with weight one. This connection
is permanently enabled.
2. The output pin's key is always the same as the key for its
associated neuron.
3. Output neurons can be modified but not deleted.
4. The input values are applied to the input pins unmodified.
"""
[docs]
@classmethod
def parse_config(cls, param_dict):
param_dict['node_gene_type'] = DefaultNodeGene
param_dict['connection_gene_type'] = DefaultConnectionGene
return DefaultGenomeConfig(param_dict, cls.__name__)
[docs]
@classmethod
def write_config(cls, f, config):
config.save(f)
def __init__(self, key):
# Unique identifier for a genome instance.
self.key = key
# (gene_key, gene) pairs for gene sets.
self.connections = {}
self.nodes = {}
# Fitness results.
self.fitness = None
[docs]
def mutate(self, config):
""" Mutates this genome. """
if config.single_structural_mutation:
div = max(1, (config.node_add_prob + config.node_delete_prob +
config.conn_add_prob + config.conn_delete_prob))
r = random()
if r < (config.node_add_prob / div):
self.mutate_add_node(config)
elif r < ((config.node_add_prob + config.node_delete_prob) / div):
self.mutate_delete_node(config)
elif r < ((config.node_add_prob + config.node_delete_prob +
config.conn_add_prob) / div):
self.mutate_add_connection(config)
elif r < ((config.node_add_prob + config.node_delete_prob +
config.conn_add_prob + config.conn_delete_prob) / div):
self.mutate_delete_connection()
else:
if random() < config.node_add_prob:
self.mutate_add_node(config)
if random() < config.node_delete_prob:
self.mutate_delete_node(config)
if random() < config.conn_add_prob:
self.mutate_add_connection(config)
if random() < config.conn_delete_prob:
self.mutate_delete_connection()
# Mutate connection genes.
for cg in self.connections.values():
cg.mutate(config)
# Mutate node genes (bias, response, etc.).
for ng in self.nodes.values():
ng.mutate(config)
[docs]
def mutate_add_node(self, config):
"""
Add a new node by splitting an existing connection.
Uses innovation tracking per NEAT paper (Stanley & Miikkulainen, 2002):
If multiple genomes split the same connection in one generation, the resulting
connections receive matching innovation numbers.
"""
if not self.connections:
if config.check_structural_mutation_surer():
self.mutate_add_connection(config)
return
assert config.innovation_tracker is not None, (
"Innovation tracker must be set before genome mutations. "
"This should be set by the reproduction module."
)
# Choose a random connection to split
conn_to_split = choice(list(self.connections.values()))
new_node_id = config.get_new_node_key(self.nodes)
ng = self.create_node(config, new_node_id)
# Make the new node as neutral as possible with respect to the
# existing connection: start with zero bias regardless of the
# global bias initialization distribution.
if hasattr(ng, "bias"):
ng.bias = 0.0
self.nodes[new_node_id] = ng
# Disable this connection and create two new connections joining its nodes via
# the given node. The new node+connections have roughly the same behavior as
# the original connection (depending on the activation function of the new node).
conn_to_split.enabled = False
i, o = conn_to_split.key
# Get innovation numbers for the two new connections
# These are keyed by the connection being split, so multiple genomes splitting
# the same connection get matching innovation numbers
in_innovation = config.innovation_tracker.get_innovation_number(
i, new_node_id, 'add_node_in'
)
out_innovation = config.innovation_tracker.get_innovation_number(
new_node_id, o, 'add_node_out'
)
# Add the two new connections with their innovation numbers
self.add_connection(config, i, new_node_id, 1.0, True, innovation=in_innovation)
self.add_connection(config, new_node_id, o, conn_to_split.weight, True, innovation=out_innovation)
[docs]
def add_connection(self, config, input_key, output_key, weight, enabled, innovation=None):
"""Add a connection to this genome. If innovation is None, gets a new one from tracker."""
# TODO: Add further validation of this connection addition?
assert isinstance(input_key, int)
assert isinstance(output_key, int)
assert output_key >= 0
assert isinstance(enabled, bool)
key = (input_key, output_key)
# Get innovation number if not provided
if innovation is None:
assert config.innovation_tracker is not None, "Innovation tracker must be set"
innovation = config.innovation_tracker.get_innovation_number(
input_key, output_key, 'add_connection'
)
connection = config.connection_gene_type(key, innovation=innovation)
connection.init_attributes(config)
connection.weight = weight
connection.enabled = enabled
self.connections[key] = connection
[docs]
def mutate_add_connection(self, config):
"""
Attempt to add a new connection, the only restriction being that the output
node cannot be one of the network input pins.
Uses innovation tracking per NEAT paper (Stanley & Miikkulainen, 2002):
If multiple genomes in the same generation add the same connection,
they receive the same innovation number.
"""
assert config.innovation_tracker is not None, (
"Innovation tracker must be set before genome mutations. "
"This should be set by the reproduction module."
)
possible_outputs = list(self.nodes)
out_node = choice(possible_outputs)
possible_inputs = list((set(self.nodes)- set(config.output_keys)) | set(config.input_keys) )
in_node = choice(possible_inputs)
# Don't duplicate connections.
key = (in_node, out_node)
if key in self.connections:
# TODO: Should this be using mutation to/from rates? Hairy to configure...
if config.check_structural_mutation_surer():
self.connections[key].enabled = True
return
# Don't allow connections between two output nodes
if in_node in config.output_keys and out_node in config.output_keys:
return
# No need to check for connections between input nodes:
# they cannot be the output end of a connection (see above).
# For feed-forward networks, avoid creating cycles.
if config.feed_forward and creates_cycle(list(self.connections), key):
return
# Get innovation number for this connection
# Same connection added by multiple genomes in same generation gets same number
innovation = config.innovation_tracker.get_innovation_number(
in_node, out_node, 'add_connection'
)
cg = self.create_connection(config, in_node, out_node, innovation)
self.connections[cg.key] = cg
[docs]
def mutate_delete_node(self, config):
# Do nothing if there are no non-output nodes.
available_nodes = [k for k in self.nodes if k not in config.output_keys]
if not available_nodes:
return -1
del_key = choice(available_nodes)
connections_to_delete = set()
for k, v in self.connections.items():
if del_key in v.key:
connections_to_delete.add(v.key)
for key in connections_to_delete:
del self.connections[key]
del self.nodes[del_key]
return del_key
[docs]
def mutate_delete_connection(self):
if self.connections:
key = choice(list(self.connections.keys()))
del self.connections[key]
[docs]
def distance(self, other, config):
"""
Returns the genetic distance between this genome and the other. This distance value
is used to compute genome compatibility for speciation.
"""
# Compute node gene distance component.
node_distance = 0.0
if self.nodes or other.nodes:
disjoint_nodes = 0
for k2 in other.nodes:
if k2 not in self.nodes:
disjoint_nodes += 1
for k1, n1 in self.nodes.items():
n2 = other.nodes.get(k1)
if n2 is None:
disjoint_nodes += 1
else:
# Homologous genes compute their own distance value.
node_distance += n1.distance(n2, config)
max_nodes = max(len(self.nodes), len(other.nodes))
node_distance = (node_distance +
(config.compatibility_disjoint_coefficient *
disjoint_nodes)) / max_nodes
# Compute connection gene differences.
connection_distance = 0.0
if self.connections or other.connections:
disjoint_connections = 0
for k2 in other.connections:
if k2 not in self.connections:
disjoint_connections += 1
for k1, c1 in self.connections.items():
c2 = other.connections.get(k1)
if c2 is None:
disjoint_connections += 1
else:
# Homologous genes compute their own distance value.
connection_distance += c1.distance(c2, config)
max_conn = max(len(self.connections), len(other.connections))
connection_distance = (connection_distance +
(config.compatibility_disjoint_coefficient *
disjoint_connections)) / max_conn
distance = node_distance + connection_distance
return distance
[docs]
def size(self):
"""
Returns genome 'complexity', taken to be
(number of nodes, number of enabled connections)
"""
num_enabled_connections = sum([1 for cg in self.connections.values() if cg.enabled])
return len(self.nodes), num_enabled_connections
[docs]
def __str__(self):
s = f"Key: {self.key}\nFitness: {self.fitness}\nNodes:"
for k, ng in self.nodes.items():
s += f"\n\t{k} {ng!s}"
s += "\nConnections:"
connections = list(self.connections.values())
connections.sort()
for c in connections:
s += "\n\t" + str(c)
return s
[docs]
@staticmethod
def create_node(config, node_id):
node = config.node_gene_type(node_id)
node.init_attributes(config)
return node
[docs]
@staticmethod
def create_connection(config, input_id, output_id, innovation):
"""Create a new connection gene with the given innovation number."""
connection = config.connection_gene_type((input_id, output_id), innovation=innovation)
connection.init_attributes(config)
return connection
[docs]
def connect_fs_neat_nohidden(self, config):
"""
Randomly connect one input to all output nodes
(FS-NEAT without connections to hidden, if any).
Originally connect_fs_neat.
"""
assert config.innovation_tracker is not None, "Innovation tracker must be set"
input_id = choice(config.input_keys)
for output_id in config.output_keys:
innovation = config.innovation_tracker.get_innovation_number(
input_id, output_id, 'initial_connection'
)
connection = self.create_connection(config, input_id, output_id, innovation)
self.connections[connection.key] = connection
[docs]
def connect_fs_neat_hidden(self, config):
"""
Randomly connect one input to all hidden and output nodes
(FS-NEAT with connections to hidden, if any).
"""
assert config.innovation_tracker is not None, "Innovation tracker must be set"
input_id = choice(config.input_keys)
others = [i for i in self.nodes if i not in config.input_keys]
for output_id in others:
innovation = config.innovation_tracker.get_innovation_number(
input_id, output_id, 'initial_connection'
)
connection = self.create_connection(config, input_id, output_id, innovation)
self.connections[connection.key] = connection
[docs]
def compute_full_connections(self, config, direct):
"""
Compute connections for a fully-connected feed-forward genome--each
input connected to all hidden nodes
(and output nodes if ``direct`` is set or there are no hidden nodes),
each hidden node connected to all output nodes.
(Recurrent genomes will also include node self-connections.)
"""
hidden = [i for i in self.nodes if i not in config.output_keys]
output = [i for i in self.nodes if i in config.output_keys]
connections = []
if hidden:
for input_id in config.input_keys:
for h in hidden:
connections.append((input_id, h))
for h in hidden:
for output_id in output:
connections.append((h, output_id))
if direct or (not hidden):
for input_id in config.input_keys:
for output_id in output:
connections.append((input_id, output_id))
# For recurrent genomes, include node self-connections.
if not config.feed_forward:
for i in self.nodes:
connections.append((i, i))
return connections
[docs]
def connect_full_nodirect(self, config):
"""
Create a fully-connected genome
(except without direct input-output unless no hidden nodes).
"""
assert config.innovation_tracker is not None, "Innovation tracker must be set"
for input_id, output_id in self.compute_full_connections(config, False):
innovation = config.innovation_tracker.get_innovation_number(
input_id, output_id, 'initial_connection'
)
connection = self.create_connection(config, input_id, output_id, innovation)
self.connections[connection.key] = connection
[docs]
def connect_full_direct(self, config):
""" Create a fully-connected genome, including direct input-output connections. """
assert config.innovation_tracker is not None, "Innovation tracker must be set"
for input_id, output_id in self.compute_full_connections(config, True):
innovation = config.innovation_tracker.get_innovation_number(
input_id, output_id, 'initial_connection'
)
connection = self.create_connection(config, input_id, output_id, innovation)
self.connections[connection.key] = connection
[docs]
def connect_partial_nodirect(self, config):
"""
Create a partially-connected genome,
with (unless no hidden nodes) no direct input-output connections.
"""
assert 0 <= config.connection_fraction <= 1
assert config.innovation_tracker is not None, "Innovation tracker must be set"
all_connections = self.compute_full_connections(config, False)
shuffle(all_connections)
num_to_add = int(round(len(all_connections) * config.connection_fraction))
for input_id, output_id in all_connections[:num_to_add]:
innovation = config.innovation_tracker.get_innovation_number(
input_id, output_id, 'initial_connection'
)
connection = self.create_connection(config, input_id, output_id, innovation)
self.connections[connection.key] = connection
[docs]
def connect_partial_direct(self, config):
"""
Create a partially-connected genome,
including (possibly) direct input-output connections.
"""
assert 0 <= config.connection_fraction <= 1
assert config.innovation_tracker is not None, "Innovation tracker must be set"
all_connections = self.compute_full_connections(config, True)
shuffle(all_connections)
num_to_add = int(round(len(all_connections) * config.connection_fraction))
for input_id, output_id in all_connections[:num_to_add]:
innovation = config.innovation_tracker.get_innovation_number(
input_id, output_id, 'initial_connection'
)
connection = self.create_connection(config, input_id, output_id, innovation)
self.connections[connection.key] = connection
def get_pruned_copy(self, genome_config):
used_node_genes, used_connection_genes = get_pruned_genes(self.nodes, self.connections,
genome_config.input_keys, genome_config.output_keys)
new_genome = DefaultGenome(None)
new_genome.nodes = used_node_genes
new_genome.connections = used_connection_genes
return new_genome
def get_pruned_genes(node_genes, connection_genes, input_keys, output_keys):
used_nodes = required_for_output(input_keys, output_keys, connection_genes)
used_pins = used_nodes.union(input_keys)
# Copy used nodes into a new genome.
used_node_genes = {}
for n in used_nodes:
used_node_genes[n] = copy.deepcopy(node_genes[n])
# Copy enabled and used connections into the new genome.
used_connection_genes = {}
for key, cg in connection_genes.items():
in_node_id, out_node_id = key
if cg.enabled and in_node_id in used_pins and out_node_id in used_pins:
used_connection_genes[key] = copy.deepcopy(cg)
return used_node_genes, used_connection_genes