def __init__(self, dispatcher_config: DispatcherConfig):
"""
:param dispatcher_config:
:raise
"""
super().__init__()
print(f"version: {__version__.__version__}")
# Validate python version
Validation.python_version(FileObserver.PY_VERSION_MIN, f"Python version required >= {FileObserver.PY_VERSION_MIN}")
self.__dispatcher_config = dispatcher_config
# Load loggers
self.__init_loggers(dispatcher_config.log_filename)
# Validate rules
self.__validate_rules()
# Check permissions
self.__check_permissions()
self.__dir_obs_dict = self.__fill_dict(dispatcher_config.dispatcher_sources)
self.__event_handler = FileEventHandler(
RuleBuilder.build_list(
dispatcher_config.dispatcher_rules,
dispatcher_config.dispatcher_formats,
dispatcher_config.dispatcher_sources,
dispatcher_config.dispatcher_destinations
),
dispatcher_config.general_threads
)
def __init__(self, conf: Conf):
super().__init__()
# validate python version
Validation.python_version(
Evaluator.REQUIRED_PYTHON, f"Unsupported Python version.\n"
f"Required Python {Evaluator.REQUIRED_PYTHON[0]}.{Evaluator.REQUIRED_PYTHON[1]} or higher."
)
Validation.can_read(
conf.dataset_train,
f"Training set file *must* exists and be readable. "
f"Current file: '{conf.dataset_train}'.\n"
f"Training set path (fully qualified) can be specified in conf.ini file or using Conf object."
)
Validation.can_read(
conf.dataset_test, f"Test set file *must* exists and be readable. "
f"Current file: '{conf.dataset_test}'.\n"
f"Test set path (fully qualified) can be specified in conf.ini file or using Conf object."
)
self.__LOG = LogManager.get_instance().logger(LogManager.Logger.EVAL)
self.__conf = conf
# using full dataset as training set
self.__training = Set(pd.read_csv(self.conf.dataset_train))
# load test set if it has same format as training_set.csv provided
# as example file see ./res/dataset/test_set_no_index.csv
self.__test = Set(pd.read_csv(self.conf.dataset_test))
# load test set if it has header (F1-20 and CLASS row) and index, so a test test saved using
# command pd.to_csv('/path', index=True)
# as example file see ./res/dataset/test_set_index.csv
# self.__test = Set(pd.read_csv(self.conf.dataset_test, index_col=0))
# load test set if it does not have header row (does not have F1-20 and CLASS row) and
# it is was not saved using command pd.to_csv('/path', index=True), so it has not index
# as example file see ./res/dataset/test_set_no_index_features.csv
# self.__test = Set(pd.read_csv(self.conf.dataset_test, header=None,
# names=[f"F{i}" for i in range(1, 21)] + ["CLASS"]))
# current classifiers used
self.__classifiers = {
Evaluator._MULTILAYER_PERCEPTRON:
None,
Evaluator._SUPPORT_VECTOR_MACHINE:
None,
# Evaluator._DECISION_TREE: None,
Evaluator._RANDOM_FOREST:
None,
# Evaluator._KNEAREST_NEIGHBORS: None,
# Evaluator._STOCHASTIC_GRADIENT_DESCENT: None,
Evaluator._ADA_BOOST:
None,
# Evaluator._NAIVE_BAYES: None,
# Evaluator._KMEANS: None
}
