output_dir = Path(args.output_directory)
output_dir.mkdir(exist_ok=True)
output_path = output_dir / args.output_file_name
for file_path in json_files:
log.info(f"Loading {file_path}")
with file_path.open("r") as fp:
lines = fp.readlines()
log.info(f"Found {len(lines)} lines")
for line in lines:
parsed_line = json.loads(line)
source_value = select_and_merge_jsonpaths(
parsed_line, args.source_jsonpaths)
target_value = select_and_merge_jsonpaths(
parsed_line, [args.target_jsonpath])
output_object = {
args.source_key: source_value,
args.target_key: target_value,
}
output_line = json.dumps(output_object)
with output_path.open("a") as out_fp:
out_fp.write(output_line)
out_fp.write(os.linesep)
if __name__ == "__main__":
parser = ArgumentParser(Args)
args = parser.parse_args()
logging.basicConfig(level="INFO")
main(args)
def start(config_params: Config):
start_time = time.time()
log = ImportData(config_params.event_log_file)
logging.info(log.trace_list)
logging.info(log.unique_events)
logging.info(log.event_map)
population = InitialPopulation(
log.unique_events, config_params.initial_population_size
)
best_tree = utility.run(
population.trees, log.unique_events, log.trace_list, config_params
)
logging.info(f"Execution time: {time.time() - start_time}")
logging.info(
f"Tree: {best_tree} Replay fitness: {best_tree.replay_fitness} Precision: {best_tree.precision} Simplicity: {best_tree.simplicity} Generalization: {best_tree.generalization} Fitness: {best_tree.fitness}"
)
for k, v in config_params.__dict__.items():
logging.info(f"{k}: {v}")
logging.info("Tree class values")
for k, v in best_tree.__dict__.items():
logging.info(f"{k}: {v}")
if __name__ == "__main__":
parser = ArgumentParser(Config)
config = parser.parse_args()
start(config)
class ParseOptions:
M1_dim: Tuple[int, int] = field(metadata=dict(args=["-M1_dim"]),
default=(2, 2))
M2_dim: Tuple[int, int] = field(metadata=dict(args=["-M2_dim"]),
default=(2, 2))
hidden_layers: str = field(metadata=dict(args=["-hiddens"]), default=None)
log_dir: str = field(metadata=dict(args=["-log-dir"]), default=None)
learning_rate: float = field(metadata=dict(args=["-learning_rate"]),
default=1e-3)
buffer_size: int = field(metadata=dict(args=["-buf_size"]), default=1000)
batch_size: int = field(metadata=dict(args=["-batch_size"]), default=32)
loss: str = field(metadata=dict(args=["-loss"]), default="mse")
optimizer: str = field(metadata=dict(args=["-optimizer"]), default="adam")
activation: str = field(metadata=dict(args=["-activation"]),
default="ReLU")
layer: str = field(metadata=dict(args=["-layer"]), default="affine")
if __name__ == "__main__":
parser = ArgumentParser(ParseOptions)
print(parser.parse_args())
args = parser.parse_args()
hiddens = args.hidden_layers[1:len(args.hidden_layers) - 1]
hiddens = hiddens.split(',')
hiddens = [int(i) for i in hiddens]
args.hidden_layers = hiddens
subprocess.Popen(["tensorboard", "--logdir", args.log_dir])
webbrowser.open("127.0.0.1:6006")
trainer = Trainer(**vars(args))
trainer.train()