def test_load_population():
reset_project()
initializer_ = SmellInitialization(
config.UDB_PATH,
population_size=config.POPULATION_SIZE,
lower_band=config.LOWER_BAND,
upper_band=config.UPPER_BAND
)
initializer_.load_population(config.PROJECT_PATH + '_population.json')
def test_generate_population():
reset_project()
initializer_ = SmellInitialization(
config.UDB_PATH,
population_size=config.POPULATION_SIZE,
lower_band=config.LOWER_BAND,
upper_band=config.UPPER_BAND
)
population_ = initializer_.generate_population()
print(population_)
def generate_population(self):
"""
Generate a biased initial population consists of first-time validated refactorings
Return:
list: list of refactoring sequences (list of refactoring operations)
"""
config.logger.debug(f'Generating a biased initial population ...')
for _ in range(0, self.population_size):
individual = []
individual_size = random.randint(self.lower_band, self.upper_band)
for j in range(individual_size):
main, params, name = self.select_random()
logger.debug(f'Refactoring name: {name}')
logger.debug(f'Refactoring params: {params}')
is_correct_refactoring = main(**params)
while is_correct_refactoring is False:
reset_project()
main, params, name = self.select_random()
logger.debug(f'Refactoring name: {name}')
logger.debug(f'Refactoring params: {params}')
is_correct_refactoring = main(**params)
####
# update_understand_database(self.udb_path)
# quit()
####
individual.append((main, params, name))
logger.debug(f'Append a refactoring "{name}" to "{j}th" gene of the individual {_}.')
reset_project()
logger.debug('-' * 100)
self.population.append(individual)
logger.debug(f'Append individual {_} to population, s')
logger.debug('=' * 100)
initial_pop_path = f'{config.PROJECT_LOG_DIR}initial_population_{config.global_execution_start_time}.json'
self.dump_population(path=initial_pop_path)
config.logger.debug(f'Generating a biased initial population was finished.')
return self.population
def execute_from_json_log(self, input_file_path=None, reset=True):
if input_file_path is None:
input_file_path = glob.glob(
os.path.join(self.log_directory,
'best_refactoring_sequences*.json'))[0]
# log_project_info(reset_=True, )
population = []
with open(input_file_path, 'r', encoding='utf-8') as fp:
population_trimmed = json.load(fp)
for chromosome in population_trimmed:
chromosome_new = []
for gene_ in chromosome:
refactoring_params = gene_[1]
if refactoring_params.get('udb_path'):
refactoring_params['udb_path'] = config.UDB_PATH
if refactoring_params.get('project_dir'):
refactoring_params['project_dir'] = config.PROJECT_PATH
if refactoring_params.get('file_path'):
refactoring_params['file_path'] = refactoring_params[
'file_path'].replace(self.file_path_base_dir_old,
config.PROJECT_ROOT_DIR)
if refactoring_params.get('class_path'):
refactoring_params['class_path'] = refactoring_params[
'class_path'].replace(self.file_path_base_dir_old,
config.PROJECT_ROOT_DIR)
chromosome_new.append((REFACTORING_MAIN_MAP[gene_[0]],
refactoring_params, gene_[0]))
population.append(chromosome_new)
# print(population)
# quit()
applicability_map = {
'Project': [],
'Sequence': [],
'Applied refactorings': [],
'Rejected refactorings': []
}
for k, refactoring_sequence in enumerate(population):
true_refactorings_count = 0
false_refactorings_count = 0
reset_project()
# Apply sequence X to system S
for refactoring_operation in refactoring_sequence:
res = refactoring_operation[0](**refactoring_operation[1])
update_understand_database(config.UDB_PATH)
if res:
true_refactorings_count += 1
else:
false_refactorings_count += 1
config.logger.info(
f"Executed {refactoring_operation[2]} with status {res}")
applicability_map['Project'].append(config.PROJECT_NAME)
applicability_map['Sequence'].append(k)
applicability_map['Applied refactorings'].append(
true_refactorings_count)
applicability_map['Rejected refactorings'].append(
false_refactorings_count)
continue
# Dump refactored project
dump_path = os.path.join(
config.PROJECT_ROOT_DIR,
f'{config.PROJECT_NAME}_refactored_with_algorithm{config.PROBLEM}_rand',
f'dump{k}')
if not os.path.exists(config.PROJECT_ROOT_DIR):
os.mkdir(dump_path)
shutil.copytree(config.PROJECT_PATH, dump_path)
# Compute quality metrics
log_project_info(
reset_=False,
quality_attributes_path=os.path.join(
config.PROJECT_LOG_DIR,
'best_refactoring_sequences_objectives_extended.csv'),
generation='-1', # config.MAX_ITERATIONS,
testability_verbose=True,
testability_log_path=os.path.join(
config.PROJECT_LOG_DIR,
f'classes_testability2_for_problem_{config.PROBLEM}_best_sequence_{k}.csv'
))
# Log applied and rejected refactorings
df = pd.DataFrame(data=applicability_map)
df.to_csv(os.path.join(config.PROJECT_LOG_DIR,
'applicability_map.csv'),
index=False)
if reset:
reset_project()
res = REFACTORING_MAIN_MAP[refactoring_name](**params)
print(f"Executed {refactoring_name} with status {res}")
update_understand_database(config.UDB_PATH)
print('-' * 100)
def execute_refactoring_sequence():
directory_name_ = os.path.dirname(__file__)
refactoring_sequence_input_file = os.path.join(directory_name_,
r'utils/input.txt')
# reset_project()
# execute_from_txt_log(input_file_path=refactoring_sequence_input_file)
def measure_ad_hoc_project_quality():
# create_understand_database(project_dir=config.PROJECT_PATH, db_dir=config.UDB_ROOT_DIR)
log_project_info()
if __name__ == '__main__':
# measure_ad_hoc_project_quality()
# quit()
reset_project()
# quit()
# execute_refactoring_sequence()
eval_ = RefactoringSequenceEvaluation(log_directory=config.PROJECT_LOG_DIR)
eval_.execute_from_json_log(reset=False)
# eval_.evaluate_sequences()
# eval_.analysis_refactoring_sequences()
def log_project_info(reset_=True, design_metrics_path=None, quality_attributes_path=None,
generation=0, testability_verbose=True, testability_log_path=None):
"""
Logging project metrics and information
"""
if reset_:
reset_project()
if quality_attributes_path is None:
quality_attributes_path = os.path.join(config.PROJECT_LOG_DIR, 'quality_attrs_initial_values.csv')
if design_metrics_path is None:
design_metrics_path = os.path.join(config.PROJECT_LOG_DIR, 'design_metrics.csv')
design_quality_attribute = DesignQualityAttributes(config.UDB_PATH)
avg_, sum_ = design_quality_attribute.average_sum
predicted_testability = testability_main(
config.UDB_PATH,
initial_value=config.CURRENT_METRICS.get("TEST", 1.0),
verbose=testability_verbose,
log_path=testability_log_path
)
mdg_modularity = modularity_main(
config.UDB_PATH,
initial_value=config.CURRENT_METRICS.get("MODULE", 1.0)
)
design_metrics = {
"DSC": [design_quality_attribute.DSC],
"NOH": [design_quality_attribute.NOH],
"ANA": [design_quality_attribute.ANA],
"MOA": [design_quality_attribute.MOA],
"DAM": [design_quality_attribute.DAM],
"CAMC": [design_quality_attribute.CAMC],
"CIS": [design_quality_attribute.CIS],
"NOM": [design_quality_attribute.NOM],
"DCC": [design_quality_attribute.DCC],
"MFA": [design_quality_attribute.MFA],
"NOP": [design_quality_attribute.NOP]
}
quality_objectives = {
"generation": [generation],
"reusability": [design_quality_attribute.reusability],
"understandability": [design_quality_attribute.understandability],
"flexibility": [design_quality_attribute.flexibility],
"functionality": [design_quality_attribute.functionality],
"effectiveness": [design_quality_attribute.effectiveness],
"extendability": [design_quality_attribute.extendability],
"testability": [predicted_testability],
"modularity": [mdg_modularity],
}
logger.info('QMOOD design metrics (N):')
logger.info(design_metrics)
logger.info('Objectives:')
logger.info(quality_objectives)
logger.info('QMOOD quality attributes sum:')
logger.info(sum_)
logger.info('QMOOD quality attributes mean:')
logger.info(avg_)
df_quality_attributes = pd.DataFrame(data=quality_objectives)
if os.path.exists(quality_attributes_path):
df = pd.read_csv(quality_attributes_path, index_col=False)
df_result = pd.concat([df, df_quality_attributes], ignore_index=True)
df_result.to_csv(quality_attributes_path, index=False)
else:
df_quality_attributes.to_csv(quality_attributes_path, index=False)
df_design_metrics = pd.DataFrame(data=design_metrics)
if os.path.exists(design_metrics_path):
df = pd.read_csv(design_metrics_path, index_col=False)
df_results = pd.concat([df, df_design_metrics], ignore_index=True)
# df = df.append(df_design_metrics, ignore_index=True)
df_results.to_csv(design_metrics_path, index=False)
else:
df_design_metrics.to_csv(design_metrics_path, index=False)