class LongTestSGcsSimulator(LongTestGcsSimulator):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.algorithm_variant = CykServiceVariationManager(is_stochastic=True)
self.configuration = self.algorithm_variant.create_default_configuration()
self.sut = GcsSimulator(self.randomizer, self.algorithm_variant)
self.set_parameters()
def set_parameters(self):
self.configuration.should_run_evolution = True
self.configuration.evolution.selectors = [
EvolutionRouletteSelectorConfiguration.create(),
EvolutionRouletteSelectorConfiguration.create()
]
self.configuration.induction.grammar_correction.should_run = False
self.configuration.induction.coverage.operators.terminal.chance = 1
self.configuration.induction.coverage.operators.universal.chance = 0
self.configuration.induction.coverage.operators.starting.chance = 1
self.configuration.induction.coverage.operators.starting.adding_hint = \
AddingRuleStrategyHint.expand_population
self.configuration.induction.coverage.operators.full.chance = 1
self.configuration.induction.coverage.operators.full.adding_hint = \
AddingRuleStrategyHint.expand_population
self.configuration.max_algorithm_steps = 5000
self.configuration.rule.max_non_terminal_symbols = 19
self.configuration.rule.random_starting_population_size = 30
self.configuration.max_execution_time = 8000
self.configuration.satisfying_fitness = 1
self.configuration.evolution.operators.crossover.chance = 0.2
self.configuration.evolution.operators.mutation.chance = 0.8
self.configuration.evolution.operators.inversion.chance = 0.8
self.configuration.induction.coverage.operators.aggressive.chance = 0
self.configuration.induction.coverage.operators.aggressive.adding_hint = \
AddingRuleStrategyHint.expand_population
self.configuration.rule.adding.crowding.factor = 18
self.configuration.rule.adding.crowding.size = 3
self.configuration.rule.adding.elitism.is_used = True
self.configuration.rule.adding.elitism.size = 3
self.configuration.rule.adding.max_non_terminal_rules = 40
class LongTestGcsSimulator(unittest.TestCase):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
logging.basicConfig(level=logging.INFO,
filename=r"C:\Users\Michał\PycharmProjects\mgr\sgcs\log.log",
format='%(asctime)s %(message)s')
self.algorithm_variant = CykServiceVariationManager(is_stochastic=False)
self.configuration = self.algorithm_variant.create_default_configuration()
self.randomizer = Randomizer(Random())
self.sut = GcsSimulator(self.randomizer, self.algorithm_variant)
self.set_parameters()
def mk_path(self, relative):
return os.path.join(r"C:\Users\Michał\PycharmProjects\mgr\sgcs\sgcs\data\example gramatics",
relative)
def set_parameters(self):
self.configuration.should_run_evolution = True
self.configuration.evolution.selectors = [
EvolutionRouletteSelectorConfiguration.create(),
EvolutionRouletteSelectorConfiguration.create()
]
self.configuration.induction.grammar_correction.should_run = False
self.configuration.induction.coverage.operators.terminal.chance = 1
self.configuration.induction.coverage.operators.universal.chance = 0
self.configuration.induction.coverage.operators.starting.chance = 1
self.configuration.induction.coverage.operators.starting.adding_hint = \
AddingRuleStrategyHint.expand_population
self.configuration.induction.coverage.operators.full.chance = 1
self.configuration.induction.coverage.operators.full.adding_hint = \
AddingRuleStrategyHint.expand_population
self.configuration.max_algorithm_steps = 5000
self.configuration.rule.max_non_terminal_symbols = 19
self.configuration.rule.random_starting_population_size = 30
self.configuration.max_execution_time = 8000
self.configuration.satisfying_fitness = 1
self.configuration.evolution.operators.crossover.chance = 0.2
self.configuration.evolution.operators.mutation.chance = 0.8
self.configuration.evolution.operators.inversion.chance = 0.8
self.configuration.induction.coverage.operators.aggressive.chance = 0
self.configuration.induction.coverage.operators.aggressive.adding_hint = \
AddingRuleStrategyHint.expand_population
self.configuration.rule.adding.crowding.factor = 18
self.configuration.rule.adding.crowding.size = 3
self.configuration.rule.adding.elitism.is_used = True
self.configuration.rule.adding.elitism.size = 3
self.configuration.rule.adding.max_non_terminal_rules = 40
self.configuration.statistics.positive_weight = 1
self.configuration.statistics.negative_weight = 2
self.configuration.statistics.classical_fitness_weight = 1
self.configuration.statistics.fertility_weight = 0
self.configuration.statistics.base_fitness = 0.5
self.configuration.max_algorithm_runs = 50
def generic_simulation(self, learning_path, testing_path, name):
logging.info('starting %s', name)
with open(os.path.join(r'C:\Users\Michał\PycharmProjects\mgr\runs\auto',
name + '.txt'), 'w+') as file:
learning_set = SymbolTranslator.create(learning_path)
learning_set.negative_allowed = not self.algorithm_variant.is_stochastic
testing_set = SymbolTranslator.create(testing_path)
testing_set.negative_allowed = True
result, ngen, grammar_estimator, population, *_ = self.sut.perform_simulation(
learning_set, testing_set, self.configuration)
print(result)
print('NGen:', ngen)
file.write(str(result))
file.write(str(ngen))
def test_simulation_dummy_test(self):
learning_set = self.mk_path('tomita 1.txt')
testing_set = self.mk_path('tomita 1.txt')
self.generic_simulation(learning_set, testing_set, 'tomita 1')
def test_simulation_for_tomita_l1(self):
learning_set = self.mk_path('tomita 1.txt')
testing_set = self.mk_path('t1 opt 15 max')
self.generic_simulation(learning_set, testing_set, 'tomita 1')
def test_simulation_for_tomita_l2(self):
learning_set = self.mk_path('tomita 2.txt')
testing_set = self.mk_path('t2 opt 15 max')
self.generic_simulation(learning_set, testing_set, 'tomita 2')
def test_simulation_for_tomita_l3(self):
learning_set = self.mk_path('tomita 3.txt')
testing_set = self.mk_path('t3 opt 15 max')
self.generic_simulation(learning_set, testing_set, 'tomita 3')
def test_simulation_for_tomita_l4(self):
learning_set = self.mk_path('tomita 4.txt')
testing_set = self.mk_path('t4 opt 15 max')
self.generic_simulation(learning_set, testing_set, 'tomita 4')
def test_simulation_for_tomita_l5(self):
learning_set = self.mk_path('tomita 5.txt')
testing_set = self.mk_path('t5 opt 15 max')
self.generic_simulation(learning_set, testing_set, 'tomita 5')
def test_simulation_for_tomita_l6(self):
learning_set = self.mk_path('tomita 6.txt')
testing_set = self.mk_path('t6 opt 15 max')
self.generic_simulation(learning_set, testing_set, 'tomita 6')
def test_simulation_for_tomita_l7(self):
learning_set = self.mk_path('tomita 7.txt')
testing_set = self.mk_path('t7 opt 15 max')
self.generic_simulation(learning_set, testing_set, 'tomita 7')
def test_simulation_for_tomita_l5_l7(self):
print('tomita l5')
learning_set = self.mk_path('tomita 5.txt')
testing_set = self.mk_path('t5 opt 15 max')
self.generic_simulation(learning_set, testing_set, 'tomita 5')
print('tomita l6')
learning_set = self.mk_path('tomita 6.txt')
testing_set = self.mk_path('t6 opt 15 max')
self.generic_simulation(learning_set, testing_set, 'tomita 6')
print('tomita l7')
learning_set = self.mk_path('tomita 7.txt')
testing_set = self.mk_path('t7 opt 15 max')
self.generic_simulation(learning_set, testing_set, 'tomita 7')
def test_simulation_for_tomita_l1_l7(self):
self.generic_simulation(self.mk_path('tomita 1.txt'), self.mk_path('t1 opt 15 max'), 'tomita 1')
self.generic_simulation(self.mk_path('tomita 2.txt'), self.mk_path('t2 opt 15 max'), 'tomita 2')
self.generic_simulation(self.mk_path('tomita 3.txt'), self.mk_path('t3 opt 15 max'), 'tomita 3')
self.generic_simulation(self.mk_path('tomita 4.txt'), self.mk_path('t4 opt 15 max'), 'tomita 4')
self.generic_simulation(self.mk_path('tomita 5.txt'), self.mk_path('t5 opt 15 max'), 'tomita 5')
self.generic_simulation(self.mk_path('tomita 6.txt'), self.mk_path('t6 opt 15 max'), 'tomita 6')
self.generic_simulation(self.mk_path('tomita 7.txt'), self.mk_path('t7 opt 15 max'), 'tomita 7')
def test_simulation_for_other_gramatics(self):
self.configuration.max_algorithm_steps = 1000
self.configuration.max_algorithm_runs = 10
self.generic_simulation(self.mk_path('ab los 200 30'), self.mk_path('ab opt max 15'), 'ab')
self.generic_simulation(self.mk_path('anbn los 200 15'), self.mk_path('anbn opt 15 max'), 'anbn')
self.generic_simulation(self.mk_path('bra1 los2 200 30'), self.mk_path('bra1 opt 15 max'), 'bra1')
self.generic_simulation(self.mk_path('bra3 los 200 30'), self.mk_path('bra3 opt los 65534'), 'bra3')
self.generic_simulation(self.mk_path('toy los2 200 30'), self.mk_path('toy opt los 65534'), 'toy')
self.generic_simulation(self.mk_path('pal2 parz los 200'), self.mk_path('pal2 opt parzysty max 15'), 'pal2')