Elemeno MLOps Python Client’s documentation

Modules

MLOps Client API

Datasource API

Datasource Type

class mlops_client.datasource.datasource_type.DatasourceType(value)

This is a class that defines the type of datasource that is being used.

The DatasourceType class is an enumeration class that defines the type of datasource that is being used.

Attributes
REDSHIFTDatasourceType

The REDSHIFT datasource type

BIGQUERYDatasourceType

The BIGQUERY datasource type

CSVDatasourceType

The CSV datasource type

exception mlops_client.datasource.datasource_type.InvalidTypeError(dstype: str)

GCP Authentication Type

Redshift Authentication Type

Feature Store API

Feature Key Type

class mlops_client.feature_store.feature_key.FeatureKey

This class is used to build a feature key.

Object building functions:

  • with_key_name (str): The name of the feature key.

  • with_key_value_type (FeatureValueType): The type of the feature key.

  • build: Returns a complete instance of the object.

Returns:

FeatureKey: A feature key.

Feature Value Type

class mlops_client.feature_store.feature_value_type.FeatureValueType(value)

FeatureValueType is an enumeration of the possible types of values that a feature can have.

STRING: A string value. FLOAT: A floating point value. INTEGER: An integer value. ARRAY: An array of values.

exception mlops_client.feature_store.feature_value_type.InvalidFeatureValueTypeError(value_type: str)

Feature Key Type

Inference Server API

class mlops_client.inference_server.inferenceserver_client.InferenceServer(headers: Dict[str, str], host: str, client: Optional[ClientSession] = None)
async create_rest(model_path: str, num_instances: int, sources: List[FeatureSource]) → Any

Creates a REST inference server for a given model.

Args:

  • model_path: Path to the model file.

  • num_instances: Number of instances to create.

  • sources: A list of FeatureSource objects.

Returns:

  • A list of InferenceServer objects.

async list(offset: Optional[str] = None, limit: Optional[int] = None) → Any

List all inference servers.

Parameters:

  • offset: An optional string that represents the starting item, should be the value of ‘next’ field from the previous response.

  • limit: An optional integer to limit the number of returned items.

Returns:

  • A list of InferenceServer objects.

Feature Source Type

class mlops_client.inference_server.feature_source_type.FeatureSourceType(value)

FeatureSourceType is an enumeration of the possible sources of features for a feature set.

FEATURE_TABLE: The feature set is based on a feature table. REQUEST_BODY: The feature set is based on a request body. REQUEST_BODY_KEY: The feature set is based on a request body key.

Feature Source

class mlops_client.inference_server.feature_source.FeatureSource
Object building functions:

  • with_source_type: The type of the feature source.

  • with_feature_table_id: The id of the feature table.

  • with_feature_name: The name of the feature.

  • with_body_json_path: The path to the feature in the request body.

  • build: Returns a complete instance of the object.

Raises:

  • MissingFieldError: If a required field is not provided.

  • InvalidFeatureValueTypeError: If the provided source type is invalid.

Returns:

  • A feature source.

Indices and tables

Indices and tables