Configuration Dataclasses

You these to run automl(...) and predict(...).

Account Configs

SnowFlake Account Config

class SnowflakeAccountConfig:
    account: str
    username: str
    password: str

AWS Account Config

class AWSAccountConfig:
    access_key_id: str
    secret_access_key: str

Pipelline Configs

SnowFlake Table Config

class SnowflakeTableConfig:
    warehouse: str
    database: str
    schema: str
    table_name: str
    # We can specify the SQL query to get the data from Snowflake.
    # If not specified, then a sample of 1000 rows will be used.
    # Only set this on input table configs.
    sql_query: Optional[str] = None
    # Specifies whether we replace the table if it already exists.
    # Only set this on output, evals, & feature importance tables.
    force_replace_table: bool = False

Model Config

class ModelConfig:
    label: str
    model_name: Optional[str] = None
    model_description: str = ""
    train_frac: float = 0.8
    add_random_features: bool = True
    init_kwargs: Dict[str, Any] = field(default_factory=dict)
    fit_kwargs: Dict[str, Any] = field(default_factory=dict)
    feature_importance_kwargs: Dict[str, Any] = field(default_factory=dict)

Artifact Cloud Save Config

class ArtifactsCloudSaveConfig:
    local_save_path: str
    bucket_name: str
    prefix: str = ""
    auto_reload: bool = False
    force_reload: bool = False