Tutorial

This guide can help you start with Lomas. For more detailed examples, see the Lomas repository on GitHub. The following codes are the basic Python examples to show how to use Lomas.

from lomas.preprocessor import Preprocessor
from lomas.generator import Generator

def main(config):
    # Preprocessing. define the file path and file name of the raw data
    data = Preprocessor(f_path=config['path'],
                        f_name=config['filenames'],
                        f_type=config['type'],
                        column_names=config['cols'])
    # Initialization. init Lomas model using the preprocessed data
    model = Generator(ip_id_dict=data.ip_id_dict,
                       ordered_ippair=data.ordered_ippair,
                       cdf_iat=data.cdf_iat,
                       cdf_size=data.cdf_size)
    model.initialize(data.trace_input)
    # Hyperparameters setting. set the hyperparameters of Lomas model, and train the model
    model.train(num_topics=25,
                chunksize=2000,
                passes=20,
                iterations=400)
    # Generation. generate synthetic trace by iteratively sampling from Lomas model
    model.generate(time_limit=data.trace_input['ts'].max(),
                   time_unit=config['ts_unit'])
# Running
if __name__ == "__main__":
    main(config)

Contents:

• Lomas Preprocessor
• Lomas Generator