Generator Model¶

Encapsulates Hazy generator model file archives.

class hazy_client2.generator_model.GeneratorModel(model_file: str)¶

Encapsulates Hazy generator model file archives.

file¶

Path to the Hazy model file

Returns a list of tables in a multitable model.

scores() → Optional[dict]¶

Returns all model scores.

column_scores()¶

Returns all table column scores.

score(score_type, score_name, table_name=None)¶

Returns a specific score.

Parameters:

score_type (str) – Type of score (i.e. similarity, utility, privacy)
score_name (str) – Name of specific score
table_name (str) – Name of table (for multi table models only). Defaults to None.

Returns:

Score within the range 0 - 100. A higher score indicates better performance

Return type:

float

sample_data()¶

Returns sample data that was generated and stored in the model file during the training stage.

property image¶

Returns the name of the Docker image used to train the model.

property commit¶

Returns the commit sha of the Docker image used to train the model.

property release_version¶

Returns the release version of the Docker image used to train the model in the form x.y.z.

property generation_schema_version: Optional[SchemaVersion]¶

Return the generation schema version locked in by the model file once trained.

Returns:: If the schema version is None. It means the model was trained before the introduction of saving schema versions to the model file which makes it either SchemaVersion.fs(“0.0.0”) or SchemaVersion.fs(“0.1.0”). If you are using `limit_strings and batch_magnitude on and this function returns None
Return type:: Optional[SchemaVersion]

custom_metadata()¶

Returns the custom metadata for this model.

Synthesiser