def main():
#########################################
# Parse the files.
##########################################
pcapPar = pcapParser(
tcpdump,
tcpdump_list,
timezone_offset=timezone_offset,
max_packets=None,
start_time=None,
stop_time=None) #start and stop time are in unix time.
tensors = pcapPar.tensors()
labeledTensors = pcapPar.labeledTensors()
##########################################
#Set up the SOM and train
##########################################
som = SOM(n=10, m=10, dim=len(tensors[0]), n_iterations=2)
som.train(tensors)
som.label(labeledTensors)
##########################################
# Get and print result
##########################################
percentages = som.get_percentages()
#Get accuracy matrix.
# Note: Once the SOM is working, labeledTensors should belong to the test data opposed to the training data.
matrix = som.get_accuracy_matrix(labeledTensors)
print("attacks identified as attacks:" + str(matrix[0][0]))
print("attacks identified as normal:" + str(matrix[0][1]))
print("attacks identifed as other:" + str(matrix[0][2]))
print("normal identified as attack:" + str(matrix[1][0]))
print("normal identified as normal:" + str(matrix[1][1]))
print("normal identified as other:" + str(matrix[1][2]))
###########################################
# Plot
##########################################
# Heat Map
som.plot_heat_map()
# Attack lines.
som.plot_line_graph(labeledTensors, pcapPar.labler)
plt.show()
