Target

Store information about the target model and data.

class sacroml.attacks.target.Target(model: Any = None, model_path: str = '', model_module_path: str = '', model_name: str = '', model_params: dict | None = None, train_module_path: str = '', train_params: dict | None = None, dataset_name: str = '', dataset_module_path: str = '', features: dict | None = None, X_train: ndarray | None = None, y_train: ndarray | None = None, X_test: ndarray | None = None, y_test: ndarray | None = None, X_orig: ndarray | None = None, y_orig: ndarray | None = None, X_train_orig: ndarray | None = None, y_train_orig: ndarray | None = None, X_test_orig: ndarray | None = None, y_test_orig: ndarray | None = None, proba_train: ndarray | None = None, proba_test: ndarray | None = None)[source]

Store information about the target model and data.

Methods

`add_feature`(name, indices, encoding)	Add a feature description to the data dictionary.
`add_processed_data`(X_train, y_train, X_test, ...)	Add a processed and split dataset.
`add_raw_data`(X_orig, y_orig, X_train_orig, ...)	Add original unprocessed dataset.
`add_safemodel_results`(data)	Add the results of safemodel disclosure checking.
`has_data`()	Return whether the target has all processed data.
`has_model`()	Return whether the target has a loaded model.
`has_probas`()	Return whether the target has all probability data.
`has_raw_data`()	Return whether the target has all raw data.
`load`([path])	Load the target class from persistent storage.
`load_array`(arr_path, name)	Load a data array variable from file.
`save`([path, ext])	Save the target class to persistent storage.

__init__(model: Any = None, model_path: str = '', model_module_path: str = '', model_name: str = '', model_params: dict | None = None, train_module_path: str = '', train_params: dict | None = None, dataset_name: str = '', dataset_module_path: str = '', features: dict | None = None, X_train: ndarray | None = None, y_train: ndarray | None = None, X_test: ndarray | None = None, y_test: ndarray | None = None, X_orig: ndarray | None = None, y_orig: ndarray | None = None, X_train_orig: ndarray | None = None, y_train_orig: ndarray | None = None, X_test_orig: ndarray | None = None, y_test_orig: ndarray | None = None, proba_train: ndarray | None = None, proba_test: ndarray | None = None) → None[source]

Store information about a target model and associated data.

Parameters:

modelAny: Trained target model.
model_pathstr: Path to a saved model.
model_module_pathstr: Path to module containing model class.
model_namestr: Class name of model.
model_paramsdict | None: Hyperparameters for instantiating the model.
train_module_pathstr: Path to module containing training function.
train_paramsdict | None: Hyperparameters for training the model.
dataset_namestr: The name of the dataset.
dataset_module_pathstr: Path to module containing dataset loading function.
featuresdict: Dictionary describing the dataset features.
X_trainnp.ndarray | None: The (processed) training inputs.
y_trainnp.ndarray | None: The (processed) training outputs.
X_testnp.ndarray | None: The (processed) testing inputs.
y_testnp.ndarray | None: The (processed) testing outputs.
X_orignp.ndarray | None: The original (unprocessed) dataset inputs.
y_orignp.ndarray | None: The original (unprocessed) dataset outputs.
X_train_orignp.ndarray | None: The original (unprocessed) training inputs.
y_train_orignp.ndarray | None: The original (unprocessed) training outputs.
X_test_orignp.ndarray | None: The original (unprocessed) testing inputs.
y_test_orignp.ndarray | None: The original (unprocessed) testing outputs.
proba_trainnp.ndarray | None: The model predicted training probabilities.
proba_testnp.ndarray | None: The model predicted testing probabilities.

add_feature(name: str, indices: list[int], encoding: str) → None[source]: Add a feature description to the data dictionary.

add_processed_data(X_train: ndarray, y_train: ndarray, X_test: ndarray, y_test: ndarray) → None[source]: Add a processed and split dataset.

add_raw_data(X_orig: ndarray, y_orig: ndarray, X_train_orig: ndarray, y_train_orig: ndarray, X_test_orig: ndarray, y_test_orig: ndarray) → None[source]: Add original unprocessed dataset.

add_safemodel_results(data: list) → None[source]

Add the results of safemodel disclosure checking.

Parameters:

datalist: The results of safemodel disclosure checking.

has_data() → bool[source]: Return whether the target has all processed data.

has_model() → bool[source]: Return whether the target has a loaded model.

has_probas() → bool[source]: Return whether the target has all probability data.

has_raw_data() → bool[source]: Return whether the target has all raw data.

load(path: str = 'target') → None[source]

Load the target class from persistent storage.

Parameters:

pathstr: Name of the output folder containing a target yaml file.

load_array(arr_path: str, name: str) → None[source]

Load a data array variable from file.

Handles both .pkl and .csv files.

Parameters:

arr_pathstr: Filename of a data array.
namestr: Name of the data array to load.

save(path: str = 'target', ext: str = 'pkl') → None[source]

Save the target class to persistent storage.

Parameters:

pathstr: Name of the output folder to save target information.
extstr: File extension defining the model saved format, e.g., “pkl” or “sav”.

sacroml.attacks.target.get_array_csv(path: str, name: str)[source]: Load a data array from csv.

sacroml.attacks.target.get_array_pkl(path: str, name: str)[source]: Load a data array from pickle.