def evaluate_feature_model():
"""
:return: evaluate the feature entropy and accuracy metrics.
"""
prefix = "C:\\Users\\yukimula\\git\\jcsa\\JCMuta\\results\\data"
postfix = "C:\\Users\\yukimula\\git\\jcsa\\PyMuta\\output\\evaluate"
print("Start testing here.")
evaluator = FeatureEvaluation()
for filename in os.listdir(prefix):
directory = os.path.join(prefix, filename)
program = cprog.CProgram(directory)
mutant_space = mut.MutantSpace(program)
evaluator.train(mutant_space.get_mutants())
print("\tTraining from", filename, "for", len(mutant_space.get_mutants()), "mutants.")
evaluator.evaluate(os.path.join(postfix, "feature.error.txt"))
return
def encoding_feature_vectors():
"""
:return: generate .corpus, .path and .csv
"""
prefix = "C:\\Users\\yukimula\\git\\jcsa\\JCMuta\\results\\data"
postfix = "C:\\Users\\yukimula\\git\\jcsa\\PyMuta\\output\\encoding"
print("Start testing here.")
max_distance = 3
encoder = FeatureEncoding()
csv_files = list()
for filename in os.listdir(prefix):
directory = os.path.join(prefix, filename)
program = cprog.CProgram(directory)
mutant_space = mut.MutantSpace(program)
encoder.reset()
print("Testing for", filename, "starts.")
print("\t\t1. Parsing program and", len(mutant_space.get_mutants()), "mutants from code.")
encoder.encode_paths(mutant_space.get_mutants(), max_distance)
encoder.print_paths(mutant_space.get_mutants(), os.path.join(postfix, filename + ".path.txt"), max_distance)
print("\t\t2. Encode the documents in mutation with", max_distance, "length.")
encoder.generate_doc2vec(os.path.join(postfix, filename + ".encoding.model"),
os.path.join(postfix, filename + "paths.corpus"))
print("\t\t3. Generate .corpus and .model file in", postfix)
csv_files.append(os.path.join(postfix, filename + ".csv"))
encoder.generate_data_file(mutant_space.get_mutants(), max_distance, os.path.join(postfix, filename + ".csv"))
print("\t\t4. Generate .csv file in", postfix)
print("\tTesting for", program.get_file_name(), "finished.")
csv_total_file = os.path.join(postfix, "training_data.csv")
with open(csv_total_file, 'w') as writer:
first = True
for csv_file in csv_files:
first2 = True
with open(csv_file, 'r') as reader:
for line in reader:
line = line.strip()
if first2:
first2 = False
if first:
first = False
writer.write(line.strip() + "\n")
elif len(line) > 0:
writer.write(line + "\n")
print("Testing end for all.")
return
def __output_state_error_main__():
prefix = "C:\\Users\\yukimula\\git\\jcsa\\JCMuta\\results\\data"
distance, sym_depth, vector_size, optimize, epochs = 6, 3, 128, True, 4
for filename in os.listdir(prefix):
directory = os.path.join(prefix, filename)
program = cpro.CProgram(directory)
mutant_space = mut.MutantSpace(program)
common_corpus.reset() # reset the encoding feature space
print("Get mutation information for", program.get_file_name())
output_file = os.path.join(
"C:\\Users\\yukimula\\git\\jcsa\\PyMuta\\output\\feature_paths",
filename + ".path")
with open(output_file, 'w') as writer:
for mutant in mutant_space.get_mutants():
writer.write(str(mutant.get_id()) + "\t")
writer.write(str(mutant.mutation.get_mutation_class()) + "\t")
writer.write(
str(mutant.mutation.get_mutation_operator()) + "\t")
location = mutant.mutation.get_location()
location: astree.AstNode
writer.write(str(location.get_beg_line()) + "\t")
writer.write("\"" + location.get_code(True) + "\"\t")
if mutant.mutation.has_parameter():
writer.write(str(mutant.mutation.get_parameter()))
else:
writer.write("")
writer.write("\t")
labels = mutant.labels
labels: mut.MutantLabels
writer.write(str(labels.get_category()) + "\t")
writer.write("\n")
state_infection = mutant.get_features()
state_infection: mut.StateInfection
writer.write("coverage at:\t" +
str(state_infection.get_faulty_execution()) +
"\n")
for state_error, constraints in state_infection.error_infections.items(
):
constraints: mut.StateConstraints
__output_state_error__(program, state_error, constraints,
writer, distance, sym_depth)
writer.write("\n")
print("Testing end for all.")
def generate_predict_csv_files():
prefix = "C:\\Users\\yukimula\\git\\jcsa\\JCMuta\\results\\pdata"
distance, sym_depth, vector_size, optimize, epochs = 6, 3, 128, True, 4
encoder = PathFeatureEncoding(path_distance=distance,
sym_depth=sym_depth,
optimize=optimize,
vector_size=vector_size)
encoder.doc2vec = Doc2Vec.load(
os.path.join(
"C:\\Users\\yukimula\\git\\jcsa\\PyMuta\\output\\doc2vec.model"))
for filename in os.listdir(prefix):
directory = os.path.join(prefix, filename)
program = cpro.CProgram(directory)
mutant_space = mut.MutantSpace(program)
output_file = os.path.join(
"C:\\Users\\yukimula\\git\\jcsa\\PyMuta\\output\\predict_files",
filename + ".csv")
encoder.write_csv_file(mutant_space, output_file)
print("\tOutput csv file for", filename, "with",
len(mutant_space.mutants), "mutants.")
print("Testing end for all.")
def read_fault_detection_contexts():
"""
:return: Generate the k-degree of fault detection context for reachability, infection and propagation of each mutant
"""
prefix = "C:\\Users\\yukimula\\git\\jcsa\\JCMuta\\results\\data"
postfix = "C:\\Users\\yukimula\\git\\jcsa\\PyMuta\\output\\features"
max_distance = 3
print("Start testing here.")
for filename in os.listdir(prefix):
directory = os.path.join(prefix, filename)
program = cprog.CProgram(directory)
mutant_space = mut.MutantSpace(program)
output_file = os.path.join(postfix, filename)
mutant = mutant_space.get_mutant(random.randint(0, len(mutant_space.mutants) - 1))
print_reach_context(mutant, output_file, False, max_distance)
mutant = mutant_space.get_mutant(random.randint(0, len(mutant_space.mutants) - 1))
print_infect_context(mutant, output_file, False, max_distance)
mutant = mutant_space.get_mutant(random.randint(0, len(mutant_space.mutants) - 1))
print_propagate_context(mutant, output_file, False, max_distance)
mutant = mutant_space.get_mutant(random.randint(0, len(mutant_space.mutants) - 1))
print("Testing on", program.filename, "for mutant", mutant.id)
print("Testing end for all.")
return
def training_doc2vec_encoding():
prefix = "C:\\Users\\yukimula\\git\\jcsa\\JCMuta\\results\\data"
distance, sym_depth, vector_size, optimize, epochs = 6, 3, 128, True, 4
encoder = PathFeatureEncoding(path_distance=distance,
sym_depth=sym_depth,
optimize=optimize,
vector_size=vector_size)
print("Testing start for files.")
for filename in os.listdir(prefix):
directory = os.path.join(prefix, filename)
program = cpro.CProgram(directory)
mutant_space = mut.MutantSpace(program)
for mutant in mutant_space.get_mutants():
encoder.encoding_word_vectors(mutant)
print("\tComplete encoding", len(mutant_space.get_mutants()),
"mutants for", filename)
print("Start training doc2vec with", len(encoder.tagged_sentences),
"sentences.")
encoder.train_doc2vec(
4,
os.path.join(
"C:\\Users\\yukimula\\git\\jcsa\\PyMuta\\output\\doc2vec.model"))
print("Complete doc2vec training")
return
data_frame = MutantDataFrame()
prefix = "C:\\Users\\yukimula\\git\\jcsa\\JCMuta\\results\\data"
p_prefix = "C:\\Users\\yukimula\\git\\jcsa\\JCMuta\\results\\pdata"
postfix = "C:\\Users\\yukimula\\git\\jcsa\\PyMuta\\output\\training_files"
p_postfix = "C:\\Users\\yukimula\\git\\jcsa\\PyMuta\\output\\predict_files"
for filename in os.listdir(prefix):
csv_file = os.path.join(postfix, filename + ".csv")
data_frame.append(csv_file)
print("\tProcessing samples for", filename)
print("Load", len(data_frame), "mutant samples for training.")
classifier = data_frame.train_for(nnet.MLPClassifier())
for filename in os.listdir(p_prefix):
directory = os.path.join(p_prefix, filename)
program = cpro.CProgram(directory)
mutant_space = mut.MutantSpace(program)
new_data_frame = MutantDataFrame()
csv_file = os.path.join(p_postfix, filename + ".csv")
new_data_frame.append(csv_file)
stubborn_data_frame = new_data_frame.generate_predict_table(classifier)
print("\tIdentify", len(stubborn_data_frame.mutant_id_list), "stubborn mutants in",
len(new_data_frame.mutant_id_list), "for", filename)
output_directory = os.path.join(p_postfix, filename)
if not os.path.exists(output_directory):
os.mkdir(output_directory)
for mutant_key in stubborn_data_frame.mutant_id_list:
mutant_key: MutantKey
output_file = os.path.join(output_directory, filename + "." + str(mutant_key.get_mid()) + ".mc")
mutant_key.interpret(mutant_space, output_file)
print("Prediction ends for all.")