def verify_serialization(explanation, extra_path=None, exist_ok=False):
path = 'brand/new/path'
if extra_path is not None:
path = os.path.join(path, extra_path)
save_explanation(explanation, path, exist_ok=exist_ok)
loaded_explanation = load_explanation(path)
_assert_explanation_equivalence(explanation, loaded_explanation)
def _load(path, model_analysis):
"""Load the ExplainerManager from the given path.
:param path: The directory path to load the ExplainerManager from.
:type path: str
:param model_analysis: The loaded parent ModelAnalysis.
:type model_analysis: ModelAnalysis
"""
# create the ExplainerManager without any properties using the __new__
# function, similar to pickle
inst = ExplainerManager.__new__(ExplainerManager)
top_dir = Path(path)
explanation_path = top_dir / ManagerNames.EXPLAINER
if explanation_path.exists():
explanation = load_explanation(explanation_path)
inst.__dict__[EXPLANATION] = explanation
inst.__dict__['_' + MODEL] = model_analysis.model
with open(top_dir / META_JSON, 'r') as meta_file:
meta = meta_file.read()
meta = json.loads(meta)
inst.__dict__['_' + IS_RUN] = meta[IS_RUN]
inst.__dict__['_' + CLASSES] = model_analysis._classes
target_column = model_analysis.target_column
train = model_analysis.train.drop(columns=[target_column])
test = model_analysis.test.drop(columns=[target_column])
inst.__dict__[U_INITIALIZATION_EXAMPLES] = train
inst.__dict__[U_EVALUATION_EXAMPLES] = test
inst.__dict__['_' + FEATURES] = list(train.columns)
inst.__dict__['_' + IS_ADDED] = False
# reset self._surrogate_model
if meta[IS_ADDED]:
inst.add()
return inst
def _load(path, rai_insights):
"""Load the ExplainerManager from the given path.
:param path: The directory path to load the ExplainerManager from.
:type path: str
:param rai_insights: The loaded parent RAIInsights.
:type rai_insights: RAIInsights
:return: The ExplainerManager manager after loading.
:rtype: ExplainerManager
"""
# create the ExplainerManager without any properties using the __new__
# function, similar to pickle
inst = ExplainerManager.__new__(ExplainerManager)
all_cf_dirs = DirectoryManager.list_sub_directories(path)
if len(all_cf_dirs) != 0:
directory_manager = DirectoryManager(
parent_directory_path=path,
sub_directory_name=all_cf_dirs[0])
data_directory = directory_manager.get_data_directory()
with open(data_directory / META_JSON, 'r') as meta_file:
meta = meta_file.read()
meta = json.loads(meta)
inst.__dict__['_' + IS_RUN] = meta[IS_RUN]
inst.__dict__['_' + IS_ADDED] = meta[IS_ADDED]
inst.__dict__[EXPLANATION] = None
explanation_path = data_directory / ManagerNames.EXPLAINER
if explanation_path.exists():
explanation = load_explanation(explanation_path)
inst.__dict__[EXPLANATION] = explanation
else:
inst.__dict__['_' + IS_RUN] = False
inst.__dict__['_' + IS_ADDED] = False
inst.__dict__[EXPLANATION] = None
inst.__dict__['_' + MODEL] = rai_insights.model
inst.__dict__['_' + CLASSES] = rai_insights._classes
inst.__dict__['_' + CATEGORICAL_FEATURES] = \
rai_insights.categorical_features
target_column = rai_insights.target_column
train = rai_insights.train.drop(columns=[target_column])
test = rai_insights.test.drop(columns=[target_column])
inst.__dict__[U_INITIALIZATION_EXAMPLES] = train
inst.__dict__[U_EVALUATION_EXAMPLES] = test
inst.__dict__['_' + FEATURES] = list(train.columns)
# reset the surrogate model
inst._initialize_surrogate_model()
return inst
def verify_serialization(explanation):
paramkeys = ['MODEL_TYPE', 'MODEL_TASK', 'METHOD', 'FEATURES', 'CLASSES']
log_items = dict()
for paramkey in paramkeys:
param = getattr(ExplainParams, paramkey)
value = getattr(explanation, param, None)
if value is not None:
if isinstance(value, np.ndarray):
log_items[param] = value.tolist()
else:
log_items[param] = value
comment = json.dumps(log_items)
test_logger.setLevel(logging.INFO)
test_logger.info("validating serialization of explanation:\n%s", comment)
expljson = save_explanation(explanation)
deserialized_explanation = load_explanation(expljson)
_assert_explanation_equivalence(deserialized_explanation, explanation)
def test_basic_upload(self, iris, tabular_explainer):
x_train = iris[DatasetConstants.X_TRAIN]
x_test = iris[DatasetConstants.X_TEST]
y_train = iris[DatasetConstants.Y_TRAIN]
model = create_sklearn_random_forest_classifier(x_train, y_train)
explainer = tabular_explainer(model, x_train)
global_explanation = explainer.explain_global(x_test)
mlflow.set_experiment(TEST_EXPERIMENT)
client = mlflow.tracking.MlflowClient()
with mlflow.start_run() as run:
_log_explanation(TEST_EXPLANATION, global_explanation)
os.mkdir(TEST_EXPLANATION)
download_path = client.download_artifacts(
run.info.run_uuid, '', dst_path=TEST_EXPLANATION)
downloaded_explanation = load_explanation(download_path)
_assert_explanation_equivalence(global_explanation,
downloaded_explanation)
