def eval_pcap(pcap, labels, time_const, label=None, rnn_size=100, model_path='/models/OneLayerModel.pkl', model_type='RandomForest'):
logger = logging.getLogger(__name__)
try:
if 'LOG_LEVEL' in os.environ and os.environ['LOG_LEVEL'] != '':
logger.setLevel(os.environ['LOG_LEVEL'])
except Exception as e:
logger.error(
'Unable to set logging level because: {0} defaulting to INFO.'.format(str(e)))
data = create_dataset(pcap, time_const, label=label,
model_path=model_path, model_type=model_type)
# Create an iterator
iterator = BatchIterator(
data,
labels,
perturb_types=['random data']
)
logger.debug('Created iterator')
rnnmodel = SoSModel(rnn_size=rnn_size, label_size=len(labels))
logger.debug('Created model')
rnnmodel.load(os.path.join(working_set.find(Requirement.parse(
'poseidonml')).location, 'poseidonml/models/SoSmodel'))
logger.debug('Loaded model')
X_list = iterator.X
L_list = iterator.L
sessions = iterator.sessions
num_total = 0
max_score = 0
scores = {}
for i, X in enumerate(X_list):
L = L_list[i]
out = rnnmodel.get_output(
np.expand_dims(X, axis=0),
np.expand_dims(L, axis=0),
)
for j, o in enumerate(out):
for k, s in enumerate(o):
num_total += 1
session = sessions[i][k]['session info']
p = session['protocol']
if p == '06':
p = 'TCP'
if p == '17':
p = 'UDP'
if p == '01':
p = 'ICMP'
flowlike = p + ' '
if session['initiated by source']:
flowlike += session['source']+' to '+session['destination']
else:
flowlike += session['destination']+' to '+session['source']
scores[num_total] = str(s)
if s > max_score:
max_score = s
logger.info(max_score)
return max_score
labels,
perturb_types=['random data', 'port swap', 'direction_swap'])
logger.info('Created iterator')
rnnmodel = SoSModel(rnn_size=100, label_size=len(labels))
logger.info('Created model')
try:
rnnmodel.load('/models/SoSmodel')
logger.info('Loaded model')
except Exception as e:
rnnmodel.initialize()
logger.info('Initialized model')
X_v, L_v, Y_v = iterator.gen_batch(split='validation', batch_size=64)
cost = rnnmodel.get_cost(X_v, L_v, Y_v)
out = rnnmodel.get_output(X_v, L_v)
logger.info('Initial validation cost: %s', np.mean(cost))
min_cost = cost
last_save = 0
for i in range(100000):
tick = time.clock()
X, L, Y = iterator.gen_batch(split='train', batch_size=64)
tock = time.clock()
_ = rnnmodel.train_on_batch(X, L, Y)
if (i + 1) % 100 == 0:
cost = rnnmodel.get_cost(X_v, L_v, Y_v)
logger.info('Validation cost after %s batches: %s', i, cost)
if cost < min_cost:
min_cost = cost
rnnmodel.save('/models/SoSmodel')