def setUp(self):
super(ModelFactoryTest, self).setUp()
self.addCleanup(mock.patch.stopall)
mock.patch.object(utils,
'fetch_explanation_metadata',
return_value=explain_metadata.ExplainMetadata(
inputs=[], framework='xgboost')).start()
def get_metadata(self):
"""Returns the current metadata as a dictionary."""
current_md = explain_metadata.ExplainMetadata(
inputs=list(self._inputs.values()),
outputs=list(self._outputs.values()),
framework=explain_metadata.Framework.TENSORFLOW2,
tags=[constants.METADATA_TAG])
return current_md.to_dict()
def get_metadata(self):
"""Returns the current metadata."""
current_md = explain_metadata.ExplainMetadata(
inputs=list(self._inputs.values()),
outputs=list(self._outputs.values()),
framework='Tensorflow',
tags=[constants.METADATA_TAG])
return current_md.to_dict()
def _create_metadata_dict(self, features_dict, crossed_columns,
desired_columns, output_dict):
"""Creates metadata from given tensor information.
Args:
features_dict: Dictionary from feature name to FeatureTensors class.
crossed_columns: A set of crossed column names.
desired_columns: A list of feature column names. Only the columns in
this list will be added to input metadata.
output_dict: Dictionary from tf.feature_columns to list of dense tensors.
Returns:
A dictionary that abides to explanation metadata.
"""
return explain_metadata.ExplainMetadata(
inputs=self._create_input_metadata(features_dict, crossed_columns,
desired_columns),
outputs=self._create_output_metadata(output_dict),
framework='Tensorflow',
tags=[constants.METADATA_TAG]).to_dict()
def _create_metadata_dict(
self,
features_dict: Dict[Text, List[monkey_patch_utils.FeatureTensors]],
crossed_columns: Set[Text],
desired_columns: List[Text],
output_dict: Dict[Text, tf.Tensor],
drop_duplicate_features: bool = False,
group_duplicate_features: bool = False) -> Dict[Text, Any]:
"""Creates metadata from given tensor information.
Args:
features_dict: Dictionary from feature name to FeatureTensors class.
crossed_columns: A set of crossed column names.
desired_columns: A list of feature column names. Only the columns in
this list will be added to input metadata.
output_dict: Dictionary from tf.feature_columns to list of dense tensors.
drop_duplicate_features: If there are multiple inputs for the same
feature column, then we will drop all but one if drop_duplicate_features
is True. If False, we will include them all with unique suffix added
to the input names to disambiguate.
group_duplicate_features: If there are multiple inputs for the same
feature column, then we will group them all as one feature group
if this parameter is set to True.
Returns:
A dictionary that abides to explanation metadata.
"""
return explain_metadata.ExplainMetadata(
inputs=self._create_input_metadata(
features_dict,
crossed_columns,
desired_columns,
drop_duplicate_features=drop_duplicate_features,
group_duplicate_features=group_duplicate_features),
outputs=self._create_output_metadata(output_dict),
framework='Tensorflow',
tags=[constants.METADATA_TAG]).to_dict()
