def testDnaSize(): dna = Dna(1, 3) assert len(dna.genes) == 12 assert len(dna.genes) == dna.length dna2 = Dna(2, 3) assert dna2.length == 22 dna3 = Dna(2, 4) assert dna3.length == 26
def testShapeExtraction(): dna = Dna(2, 3) dna.genes = [2, 3, 255, 0, 0, 255, 1, 2, 3, 4, 5, 6, 255, 0, 0, 255, 7, 8, 9, 10, 11, 12] assert sorted(dna.polyAtIndex(1)) == sorted([(255, 0, 0, 255), [(7, 8), (9, 10), (11, 12)]])
def testImageCreation(): dna = Dna(2, 3) dna.genes = [2, 3, 255, 0, 0, 100, 10, 10, 250, 250, 10, 250, 0, 255, 0, 50, 10, 10, 250, 10, 10, 250] image = ImageTools.imageFromDna(dna) image.show()
def initialize_population(self):
"""Initializes the population.
"""
for _ in range(self.params.get('pop_size')):
tmp_dna = Dna(self.initialize_bits(),
self.params.get('item_weights'),
self.params.get('item_values'),
Bag(self.params.get('bag_size')))
tmp_dna.eval_vals()
self.pop.append(tmp_dna)
def testMutationRanges(): dna = Dna(1, 3) for i in xrange(100): dna.mutate(1) for i in xrange(dna.headerSize, dna.length): gene = dna.genes[i] if dna.indexIsColour(i): assert gene > 0 and gene <= 255 else: assert gene > 0 and gene <= Dna.imgSize
def __init__(self, size, polyCount, vertexCount, targetImage):
self.size = size
self.targetImage = targetImage
self.artists = [Dna(polyCount, vertexCount) for i in xrange(size)]
[a.randomisePolys() for a in self.artists]
self.cycles = 0
self.improvements = 0
self.mutationLevel = 0
def __init__(self):
self.dna = Dna()
self.acc = [0, 0]
self.pos = [0, 0]
self.vel = [0, 0]
self.timer = 20
pass
def __init__(self, dna=None):
if dna:
self.dna = dna
else:
self.dna = Dna()
self.keys = [Keys.UP, Keys.DOWN, Keys.LEFT, Keys.RIGHT]
self.fitness = 0
self.count = -1
def crossover(self, parent_a: Creature, parent_b: Creature) -> Dna:
"""
Generates a new child with crossover.
"""
# Compare both parent's genes.
matching, disjoint, excess, max_number, a_connections, b_connections = self.compare_genomes(parent_a, parent_b)
fit_parent = parent_a if parent_a.fitness > parent_b.fitness else parent_b \
if parent_a.fitness < parent_b.fitness else None
non_matching = disjoint + excess
# Decide which genes the child will inherit.
# For each gene, add the parent it will be inherited from. After everything is decided add the genes.
child_gene_sources = dict()
for number in matching:
# Inherit weight from a random parent.
if random() < 0.5:
child_gene_sources[number] = parent_a
else:
child_gene_sources[number] = parent_b
# If there is a fit parent, inherit disjoint and matching genes from it.
if fit_parent:
for number in non_matching:
if number in fit_parent.dna.connections:
child_gene_sources[number] = fit_parent
# If both parents are equally fit, inherit disjoint and excess genes randomly.
else:
for number in non_matching:
if random() < 0.5 and number in a_connections:
child_gene_sources[number] = parent_a
elif number in b_connections:
child_gene_sources[number] = parent_b
# Add all genes the child should inherit.
child_connections = dict()
child_nodes = dict()
for number, parent in child_gene_sources.items():
connection = parent.dna.connections[number]
child_connections[number] = connection
src_number, dst_number = connection.src_number, connection.dst_number
child_nodes[src_number] = parent.dna.nodes[src_number]
child_nodes[dst_number] = parent.dna.nodes[dst_number]
# Generate child.
child_dna = Dna(nodes=child_nodes, connections=child_connections)
return child_dna
def crossover(self, parent1, parent2, crossover_point):
"""Applies one point crossover operation to between two Dna objects.
Returns the results of the crossover by two Dna object.
Keyword Arguments:
parent1 -- Dna object that represents firs parent of crossover.
parent2 -- Dna object that represents second parent of crossover.
crossover_point -- Integer index of the crossover point.
"""
crossover_point += 1
print('Applying Crossover:\t at', crossover_point)
print('Parents: ', (parent1, parent2))
child1_bits = parent1.bits[:crossover_point] + parent2.bits[
crossover_point:]
child2_bits = parent2.bits[:crossover_point] + parent1.bits[
crossover_point:]
child1 = Dna(child1_bits, self.params.get('item_weights'),
self.params.get('item_values'),
Bag(self.params.get('bag_size')))
child2 = Dna(child2_bits, self.params.get('item_weights'),
self.params.get('item_values'),
Bag(self.params.get('bag_size')))
print('Children: ', (child1_bits, child2_bits))
return (child1, child2)
def base_dna() -> Tuple[Dna, List[ConnectionMutation]]:
"""
Generate base Dna to start creatures from, connect nodes with connection mutations.
"""
# Generate base Dna to start creatures from.
base_nodes = {}
for num in range(1, CREATURE_INPUTS + CREATURE_OUTPUTS + 1):
base_nodes[num] = InputNode(num) if num < CREATURE_INPUTS + 1 else OutputNode(num, BIAS_RANGE)
# Connect nodes with connection mutations.
base_dna = Dna(CREATURE_INPUTS, CREATURE_OUTPUTS, nodes=base_nodes)
mutations = []
# Connect nodes by NEAT PARAMETER settings.
if BASE_DNA == 'CONNECTED':
for src_number in range(CREATURE_INPUTS):
for dst_number in range(CREATURE_INPUTS, CREATURE_INPUTS + CREATURE_OUTPUTS):
# Generate connection mutation between each Input node to every Output node.
mutation = ConnectionMutation(len(mutations) + 1, src_number + 1, dst_number + 1)
mutations.append(mutation)
base_dna.update(mutations)
return base_dna, mutations
from dna import Dna
import numpy as np
from numpy import random
gene = Dna(population=3000, mutation=0.4)
pop = gene.randomic_population()
generation = 1
def epoch(population):
global generation
print("=======================================")
print("""------- Generation -------""" + str(generation))
generation = generation + 1
s = 0
for i in range(len(population)):
# Fit
population[i].fitness()
s = s + population[i].fit
# mating Pool
mean = s / gene.population
print("- mean error = {}".format(mean))
print("========================================")
matingPool = [gene]
new_population = []
from dna import Dna
from motif_search import MotifSearch
if __name__ == '__main__':
file_name = 'dna'
dna = Dna(file_name, generate=False)
dna_lines = dna.lines
print('10-Mer: ', dna.ten_mer)
print('Mutated 10-Mers: ', dna.mutated_list)
for k in range(9, 12):
MotifSearch('random', dna_lines, k, check_period=500).print_results()
MotifSearch('gibbs', dna_lines, k, check_period=500).print_results()
def random(cls):
"""
Generate a Candidate with 50 totally random Triangles.
"""
dna = Dna.random()
return cls(dna)
self.name = generate_name()
self.fitness = 0
self.health = CREATURE_HEALTH
self.reach = CREATURE_REACH
self.line_of_sight = CREATURE_LINE_OF_SIGHT
def __str__(self):
return CREATURE_STRING.format(self.name, self.dna.hidden)
def __repr__(self):
return str(self)
def think(self, inputs: CreatureNetworkInput) -> CreatureNetworkOutput:
"""
Gets the creature decision based on the inputs it was given.
:param inputs: Sensory input.
:return: Creature decision.
"""
return CreatureNetworkOutput(*self.network.get_output(list(inputs)))
def update(self, mutations: List[Union[MutationObject]]):
self.dna.update(mutations)
self.network = Network(self.dna.nodes, self.dna.node_connections)
if __name__ == '__main__':
main_dna = Dna(inputs=2, outputs=6)
c = Creature(dna=main_dna)
print(c.think(CreatureNetworkInput(0, 0, 0)))
print(c.think(CreatureNetworkInput(100, 100, 100)))
def testDnaPolySwap(): dna = Dna(2, 3) dna.genes = [2, 3, 255, 0, 0, 100, 1, 2, 3, 4, 5, 6, 0, 255, 0, 50, 7, 8, 9, 10, 11, 12] expectedResult = [2, 3, 0, 255, 0, 50, 7, 8, 9, 10, 11, 12, 255, 0, 0, 100, 1, 2, 3, 4, 5, 6] dna.swapPolys(0, 1) assert not cmp(dna.genes, expectedResult)
def testDnaSpliceValidLength(): mother = Dna(3, 3) father = Dna(3, 3) child = mother.splice(father) assert len(child.genes) == len(mother.genes)
#!flask/bin/python
from flask import Flask
from flask import request, url_for
from flask import jsonify, abort
from dna import Dna
from werkzeug.utils import secure_filename
from alphabot_exceptions import *
import os
d = Dna()
app = Flask(__name__)
@app.route('/', methods = ['GET', 'POST'])
def post_image():
if request.method == 'GET':
return "GET \n"
if request.method == 'POST':
file = request.files['file']
filename = secure_filename(file.filename)
file.save(filename)
dirr = os.getcwd()
osname = os.path.join(dirr, '')
dest_img = osname + filename
try:
results = d.find_distance_and_angle(dest_img) ### pairnei path
os.remove(dest_img)
return jsonify(results)
except BeaconNotFoundError:
os.remove(dest_img)
def testDnaInit(): dna = Dna(3, 4) assert dna.genes[0] == 3 assert dna.genes[1] == 4 for i in range(2, dna.length-1): assert dna.genes[i] == 0
def testIndexCategorisation(): dna = Dna(1, 3) for i in xrange(2, 5): assert dna.indexIsColour(i) == True for i in xrange(6, 12): assert dna.indexIsColour(i) == False