def test_id_probs(tmp_path):
frames = dp.parse_csv(
SAMPLE_PATH /
'csv/2006 Ford Fusion/Test Data/2006 Ford Fusion Test.csv')
probs_file = str(tmp_path / 'idprobs.json')
dp.write_id_probs(frames, probs_file)
idprobs = dp.load_id_probs(probs_file)
idprobs_check = {
1072: 0.07521674607513865,
336: 0.15019917206904632,
352: 0.15019917206904632,
560: 0.15019917206904632,
71: 0.015621338748730767,
357: 0.0937280324923846,
368: 0.0937280324923846,
512: 0.0937280324923846,
513: 0.0937280324923846,
1056: 0.015621338748730767,
1058: 0.015621338748730767,
65: 0.014996485198781535,
832: 0.014996485198781535,
848: 0.014996485198781535,
613: 0.0014840271811294228,
1279: 0.0014840271811294228,
1108: 0.0014840271811294228,
1059: 0.0014840271811294228,
1107: 0.0014840271811294228
}
assert idprobs == idprobs_check
def test_canlist(tmp_path):
frames = dp.parse_csv(
str(SAMPLE_PATH /
'csv/2006 Ford Fusion/Test Data/2006 Ford Fusion Test.csv'))
dp.write_canlist(frames, tmp_path / 'canlist_validate.json')
frames_check = dp.load_canlist(tmp_path / 'canlist_validate.json')
assert frames == frames_check
assert dp.validate_can_data(frames)
def test_inject_malicious_packets():
frames = dp.parse_csv(
SAMPLE_PATH /
'csv/2006 Ford Fusion/Test Data/2006 Ford Fusion Test.csv')
prev_len = len(frames)
mg = MaliciousGenerator()
frames, labels = dp.inject_malicious_packets(frames, mg)
assert len(frames) == len(labels)
assert not all(x is None for x in labels)
assert len(frames) > prev_len
assert dp.validate_can_data(frames)
def test_feature_lists(tmp_path):
frames = dp.parse_csv(
SAMPLE_PATH /
'csv/2006 Ford Fusion/Test Data/2006 Ford Fusion Test.csv')
idprobs = dp.write_id_probs(frames, tmp_path / 'idprobs')
feature_lists = dp.generate_feature_lists(frames, idprobs)
feat_file = tmp_path / 'feature_lists'
dp.write_feature_lists(feature_lists, [], feat_file)
feature_lists, _ = dp.load_feature_lists(feat_file)
with open(SAMPLE_PATH / 'test_preprocessor/feature_lists_validate.txt') \
as validation_file:
feature_lists_check = literal_eval(validation_file.read())
assert feature_lists == feature_lists_check
def test_parse_csv():
# Check parse_csv
frames = dp.parse_csv(
str(SAMPLE_PATH /
'csv/2006 Ford Fusion/Test Data/2006 Ford Fusion Test.csv'))
with open(SAMPLE_PATH / 'test_preprocessor/csv_validate.txt') \
as verification_file:
frames_check = literal_eval(verification_file.read())
assert frames == frames_check
with pytest.raises(FileNotFoundError):
dp.parse_csv('asdf')
with pytest.raises(FileNotFoundError):
dp.parse_csv('.')
with pytest.raises(ValueError):
dp.parse_csv(str(SAMPLE_PATH.parent / 'test_preprocessor.py'))
def test_validate_can_data():
old_stdout = sys.stdout
sys.stdout = buffer = io.StringIO(
) # Write anything being printed to stdout to buffer instead
frames = dp.parse_csv(
str(SAMPLE_PATH /
'csv/2006 Ford Fusion/Test Data/2006 Ford Fusion Test.csv'))
assert dp.validate_can_data(frames)
frames[163].pop('id')
frames[124]['id'] = 3999
frames[777]['timestamp'] = 'asdf'
frames[32]['data'] = b'aaaaaaaaaaaa'
frames[32]['timestamp'] = -1
frames[777]['data'] = 3.1
assert not dp.validate_can_data(frames)
# Check that what was being printed is the correct output
sys.stdout = old_stdout
assert buffer.getvalue() == """This dataset is valid!
def start_simulation(self, can_file_url, idprobs_name):
if platform.system() == 'Windows':
file_path = can_file_url.toString()[8:]
else:
file_path = can_file_url.toString()[7:]
if file_path.endswith('.traffic'):
canlist = dp.parse_traffic(file_path)
elif file_path.endswith('.csv'):
canlist = dp.parse_csv(file_path)
elif file_path.endswith('.json'):
canlist = dp.load_canlist(file_path)
else:
raise ValueError(
f'Unknown type of CAN frame file provided: {file_path}.')
self._current_canlist, self._current_labels = dp.inject_malicious_packets(
canlist, self._malgen)
self._ids_manager._ids.change_ids_parameters(
'idprobs_path', idprobs_dir + '/' + idprobs_name + '.json')
self._ids_manager.start_simulation()
self.sim_thread = SimulationThread(self)
self.sim_thread.start()
def create_dataset(self, can_file_url, idprobs_name, malgen_probs, dataset_name):
# Convert the JS object to a dictionary.
malgen_probs = malgen_probs.toVariant()
if platform.system() == 'Windows':
file_path = can_file_url.toString()[8:]
else:
file_path = can_file_url.toString()[7:]
if file_path.endswith('.traffic'):
canlist = dp.parse_traffic(file_path)
elif file_path.endswith('.csv'):
canlist = dp.parse_csv(file_path)
elif file_path.endswith('.json'):
canlist = dp.load_canlist(file_path)
else:
raise ValueError(f'Unknown type of CAN frame file provided: {file_path}.')
idprobs = dp.load_id_probs(idprobs_dir + '/' + idprobs_name + '.json')
# Create a folder under the 'datasets' folder with the name being the
# name of the dataset. This is where all the relevant files for the
# dataset go.
dataset_folder = datasets_dir + '/' + dataset_name
os.mkdir(dataset_folder)
# Write the good CAN frame file.
dp.write_canlist(canlist, dataset_folder + '/good_canlist.json')
# Create the CAN frame file with malicious frames injected.
malgen = MaliciousGenerator()
malgen.adjust(malgen_probs)
bad_canlist, labels = dp.inject_malicious_packets(canlist, malgen)
dp.write_canlist(bad_canlist, dataset_folder + '/bad_canlist.json')
# Create the feature lists/labels file.
features = dp.generate_feature_lists(bad_canlist, idprobs)
dp.write_feature_lists(features, labels, dataset_folder + '/features_labels.json')
# Update the available datasets.
self._available_datasets.append(dataset_name)
self.get_availableDatasets.emit()
def create_idprobs_file(self, list_of_can_files, idprobs_name):
canlist = []
# Here, each entry in `list_of_can_files` is a QUrl.
for file_url in list_of_can_files:
# toString() returns 'file://<actual file path>', so strip the
# beginning and open the rest.
if platform.system() == 'Windows':
file_path = file_url.toString()[8:]
else:
file_path = file_url.toString()[7:]
if file_path.endswith('.traffic'):
canlist += dp.parse_traffic(file_path)
elif file_path.endswith('.csv'):
canlist += dp.parse_csv(file_path)
elif file_path.endswith('.json'):
canlist += dp.load_canlist(file_path)
else:
raise ValueError(f'Unknown type of CAN frame file provided: {file_path}.')
dp.write_id_probs(canlist, idprobs_dir + '/' + idprobs_name + '.json')
# After writing the file, update the available ID probs files by
# appending to the list and then emitting the "notify" signal.
self._available_idprobs.append(idprobs_name)
self.get_availableIdprobs.emit()