Transformation Function#

You can create a TransformationFunction via FeatureStore.create_transformation_function and obtain an existing one via FeatureStore.get_transformation_function or FeatureStore.get_transformation_functions.

TransformationFunction #

NOT_FOUND_ERROR_CODE `class-attribute` `instance-attribute` #

NOT_FOUND_ERROR_CODE = 270160

DTO class for transformation functions.

PARAMETER	DESCRIPTION
`featurestore_id`	`int`. Id of the feature store in which the transformation function is saved.
`hopsworks_udf`	`HopsworksUDF`. The meta data object for UDF in Hopsworks, which can be created using the `@udf` decorator.
`version`	`int`. The version of the transformation function.
`id`	`int`. The id of the transformation function in the feature store.
`transformation_type`	`UDFType`. The type of the transformation function. Can be "on-demand" or "model-dependent"

hopsworks_udf `property` #

hopsworks_udf: HopsworksUdf

Meta data class for the user defined transformation function.

id `property` `writable` #

id: id

Transformation function id.

output_column_names `property` #

output_column_names: list[str]

Names of the output columns generated by the transformation functions.

transformation_statistics `property` `writable` #

transformation_statistics: TransformationStatistics | None

Feature statistics required for the defined UDF.

transformation_type `property` `writable` #

transformation_type: TransformationType

Type of the Transformation: can be model dependent or on-demand.

version `property` `writable` #

version: int

Version of the transformation function.

call #

__call__(*features: list[str]) -> TransformationFunction

Update the feature to be using in the transformation function.

PARAMETER	DESCRIPTION
`features`	`List[str]`. Name of features to be passed to the User Defined function. TYPE: `list[str]` DEFAULT: `()`

RETURNS	DESCRIPTION
`TransformationFunction`	`HopsworksUdf`: Meta data class for the user defined function.

RAISES	DESCRIPTION
`hopsworks.client.exceptions.FeatureStoreException`	If the provided number of features do not match the number of arguments in the defined UDF or if the provided feature names are not strings.

alias #

alias(*args: str)

Set the names of the transformed features output by the transformation function.

delete #

delete() -> None

Delete transformation function from backend.

Example

# import hopsworks udf decorator
from hopworks import udf

# define function
@udf(int)
def plus_one(value):
    return value + 1

# create transformation function
plus_one_meta = fs.create_transformation_function(
        transformation_function=plus_one,
        version=1
    )
# persist transformation function in backend
plus_one_meta.save()

# retrieve transformation function
plus_one_fn = fs.get_transformation_function(name="plus_one")

# delete transformation function from backend
plus_one_fn.delete()

from_response_json `classmethod` #

from_response_json(
    json_dict: dict[str, Any],
) -> TransformationFunction | list[TransformationFunction]

Function that constructs the class object from its json serialization.

PARAMETER	DESCRIPTION
`json_dict`	`Dict[str, Any]`. Json serialized dictionary for the class. TYPE: `dict[str, Any]`

RETURNS	DESCRIPTION
`TransformationFunction \| list[TransformationFunction]`	`TransformationFunction`: Json deserialized class object.

json #

json() -> str

Convert class into its json serialized form.

RETURNS	DESCRIPTION
`str`	`str`: Json serialized object.

save #

save() -> None

Save a transformation function into the backend.

Example

# import hopsworks udf decorator
from hopworks import udf

# define function
@udf(int)
def plus_one(value):
    return value + 1

# create transformation function
plus_one_meta = fs.create_transformation_function(
        transformation_function=plus_one,
        version=1
    )

# persist transformation function in backend
plus_one_meta.save()

to_dict #

to_dict() -> dict[str, Any]

Convert class into a dictionary.

RETURNS	DESCRIPTION
`dict[str, Any]`	`Dict`: Dictionary that contains all data required to json serialize the object.

update_from_response_json #

update_from_response_json(
    json_dict: dict[str, Any],
) -> TransformationFunction

Function that updates the class object from its json serialization.

PARAMETER	DESCRIPTION
`json_dict`	`Dict[str, Any]`. Json serialized dictionary for the class. TYPE: `dict[str, Any]`

RETURNS	DESCRIPTION
`TransformationFunction`	`TransformationFunction`: Json deserialized class object.

Transformation Function#