def test_random_diff_duplicates_and_size(stop, n, start):
ga = StandardGA(fitness_test_sin_func)
nums = ga._random_diff(stop, n, start=start)
assert len(nums) == len(set(nums))
assert len(nums) == n
def test_random_diff_duplicates_and_size(stop, n, start):
ga = StandardGA(fitness_test_sin_func)
nums = ga._random_diff(stop, n, start=start)
assert len(nums) == len(set(nums))
assert len(nums) == n
def test_sort_population(optim):
ga = StandardGA(fitness_test_sin_func, optim=optim)
ga.population = list(unsorted_population)
ga._sort_population()
required_population = sort_population(optim, list(unsorted_population))
assert ga.population == required_population
def test_sort_population(optim):
ga = StandardGA(fitness_test_sin_func, optim=optim)
ga.population = list(unsorted_population)
ga._sort_population()
required_population = sort_population(optim, list(unsorted_population))
assert ga.population == required_population
def test_update_solution(chrom, fitness, optim, result):
ga = StandardGA(fitness_test_sin_func, optim=optim)
ga.best_chromosome = 1
ga.best_fitness = 100
ga._update_solution(chrom, fitness)
assert ga.best_solution == result
def test_select_parents_tournament():
population = list(unsorted_population)
population = sort_population('min', population)
ga = StandardGA(fitness_test_sin_func, optim='min', selection='tournament', tournament_size=len(population))
parent1, parent2 = ga._select_parents(population)
assert (parent1.fitness_val, parent2.fitness_val) == (1, 2)
def test_select_parents_tournament():
population = list(unsorted_population)
population = sort_population('min', population)
ga = StandardGA(fitness_test_sin_func,
optim='min',
selection='tournament',
tournament_size=len(population))
parent1, parent2 = ga._select_parents(population)
assert (parent1.fitness_val, parent2.fitness_val) == (1, 2)
def test_conduct_tournament_population_part(optim):
ga = StandardGA(fitness_test_sin_func, optim=optim)
population = list(unsorted_population)
population = sort_population(optim, population)
size = 2
best1, best2 = ga._conduct_tournament(population, size)
if optim == 'max':
assert population[best1].fitness_val >= population[best2].fitness_val
else:
assert population[best1].fitness_val <= population[best2].fitness_val
def test_conduct_tournament_population_part(optim):
ga = StandardGA(fitness_test_sin_func, optim=optim)
population = list(unsorted_population)
population = sort_population(optim, population)
size = 2
best1, best2 = ga._conduct_tournament(population, size)
if optim == 'max':
assert population[best1].fitness_val >= population[best2].fitness_val
else:
assert population[best1].fitness_val <= population[best2].fitness_val
def test_extend_population():
ga = StandardGA(fitness_test_sin_func, optim='min')
ga.population = [IndividualGA(1, 100)]
new_elems = [IndividualGA(2, 50), IndividualGA(3, 150)]
ga.extend_population(new_elems)
assert ga.best_solution == (2, 50)
result = []
for i, individ in zip(range(len(ga.population)), ga.population):
result.append((individ.chromosome, individ.fitness_val))
assert result == [(3, 150), (1, 100), (2, 50)]
def test_extend_population():
ga = StandardGA(fitness_test_sin_func, optim='min')
ga.population = [IndividualGA(1, 100)]
new_elems = [IndividualGA(2, 50), IndividualGA(3, 150)]
ga.extend_population(new_elems)
assert ga.best_solution == (2, 50)
result = []
for i, individ in zip(range(len(ga.population)), ga.population):
result.append((individ.chromosome, individ.fitness_val))
assert result == [(3, 150), (1, 100), (2, 50)]
def test_conduct_tournament_whole_population(optim):
ga = StandardGA(fitness_test_sin_func, optim=optim)
population = list(unsorted_population)
population = sort_population(optim, population)
size = len(population)
if optim == 'max':
correct_out = (7, 6)
else:
correct_out = (1, 2)
best1, best2 = ga._conduct_tournament(population, size)
assert (population[best1].fitness_val, population[best2].fitness_val) == correct_out
def test_init_valid_cross_type(cross_type):
"""
This function tests only common valid values of crossover type.
DO NOT FORGET THAT SUBCLASSES (of StandardGA) HAVE ITS OWN RESTRICTIONS
THAT MUST BE THOROUGHLY TESTED.
"""
StandardGA(fitness_test_sin_func, cross_type=cross_type)
def test_conduct_tournament_whole_population(optim):
ga = StandardGA(fitness_test_sin_func, optim=optim)
population = list(unsorted_population)
population = sort_population(optim, population)
size = len(population)
if optim == 'max':
correct_out = (7, 6)
else:
correct_out = (1, 2)
best1, best2 = ga._conduct_tournament(population, size)
assert (population[best1].fitness_val,
population[best2].fitness_val) == correct_out
def test_init_invalid_mut_type(mut_type):
"""
This function tests only common invalid values of mutation type.
DO NOT FORGET THAT SUBCLASSES (of StandardGA) HAVE ITS OWN RESTRICTIONS
THAT MUST BE THOROUGHLY TESTED.
"""
with pytest.raises(ValueError):
StandardGA(fitness_test_sin_func, mut_type=mut_type)
def test_update_solution(chrom, fitness, optim, result):
ga = StandardGA(fitness_test_sin_func, optim=optim)
ga.best_chromosome = 1
ga.best_fitness = 100
ga._update_solution(chrom, fitness)
assert ga.best_solution == result
def test_invalid_random_diff():
ga = StandardGA(fitness_test_sin_func)
with pytest.raises(ValueError):
ga._random_diff(2, 10, start=0)
def test_select_parents_unknown_type():
ga = StandardGA(fitness_test_sin_func)
ga.selection = 'unknown'
with pytest.raises(ValueError):
ga._select_parents([])
def test_invalid_conduct_tournament(population, size):
ga = StandardGA(fitness_test_sin_func)
with pytest.raises(ValueError):
ga._conduct_tournament(population, size)
def test_invalid_init_population(chromosomes):
ga = StandardGA(fitness_test_sin_func)
with pytest.raises(ValueError):
ga.init_population(chromosomes)
def test_init_invalid_optim():
with pytest.raises(ValueError):
StandardGA(fitness_test_sin_func, optim='LIE!')
def test_init_valid_mutation_prob(prob):
StandardGA(fitness_test_sin_func, mut_prob=prob)
def test_init_valid_crossover_prob(prob):
StandardGA(fitness_test_sin_func, cross_prob=prob)
def test_init_valid_selection_1(selection):
StandardGA(fitness_test_sin_func, selection=selection)
def test_init_valid_selection_tournament():
StandardGA(fitness_test_sin_func,
selection='tournament',
tournament_size=2)
def test_init_valid_elitism(elitism):
StandardGA(fitness_test_sin_func, elitism=elitism)
def test_compute_rank_wheel_sum(size, result):
ga = StandardGA(fitness_test_sin_func)
assert ga._compute_rank_wheel_sum(size) == result
def test_random_diff_whole_interval():
ga = StandardGA(fitness_test_sin_func)
nums = ga._random_diff(5, 5, start=0)
assert nums == list(range(5))
def test_invalid_random_diff():
ga = StandardGA(fitness_test_sin_func)
with pytest.raises(ValueError):
ga._random_diff(2, 10, start=0)
def test_best_solution():
ga = StandardGA(fitness_test_sin_func)
ga.best_chromosome = [1, 2]
ga.best_fitness = 155
assert ga.best_solution == ([1, 2], 155)
def test_init_invalid_elitism(elitism):
with pytest.raises(ValueError):
StandardGA(fitness_test_sin_func, elitism=elitism)
def test_init_fitness_func():
with pytest.raises(ValueError):
StandardGA()
def test_random_diff_whole_interval():
ga = StandardGA(fitness_test_sin_func)
nums = ga._random_diff(5, 5, start=0)
assert nums == list(range(5))
def test_init_valid_optim(optim):
StandardGA(fitness_test_sin_func, optim=optim)
def test_invalid_init_population(chromosomes):
ga = StandardGA(fitness_test_sin_func)
with pytest.raises(ValueError):
ga.init_population(chromosomes)
def test_select_parents_unknown_type():
ga = StandardGA(fitness_test_sin_func)
ga.selection = 'unknown'
with pytest.raises(ValueError):
ga._select_parents([])
def test_init_invalid_selection_type():
with pytest.raises(ValueError):
StandardGA(fitness_test_sin_func, selection='unknown')
def test_select_parents_wheel_sum(selection, wheel_sum):
ga = StandardGA(fitness_test_sin_func, selection=selection)
with pytest.raises(ValueError):
ga._select_parents([], wheel_sum)
def test_init_invalid_mutation_prob(prob):
with pytest.raises(ValueError):
StandardGA(fitness_test_sin_func, mut_prob=prob)
def test_select_parents_wheel_sum(selection, wheel_sum):
ga = StandardGA(fitness_test_sin_func, selection=selection)
with pytest.raises(ValueError):
ga._select_parents([], wheel_sum)
def test_init_invalid_crossover_prob(prob):
with pytest.raises(ValueError):
StandardGA(fitness_test_sin_func, cross_prob=prob)
def test_compute_rank_wheel_sum(size, result):
ga = StandardGA(fitness_test_sin_func)
assert ga._compute_rank_wheel_sum(size) == result
def test_invalid_conduct_tournament(population, size):
ga = StandardGA(fitness_test_sin_func)
with pytest.raises(ValueError):
ga._conduct_tournament(population, size)
def test_best_solution():
ga = StandardGA(fitness_test_sin_func)
ga.best_chromosome = [1, 2]
ga.best_fitness = 155
assert ga.best_solution == ([1,2], 155)
