mlrl.testbed.experiments.input.data module

Author Michael Rapp (michael.rapp.ml@gmail.com)

Provides classes for representing input data.

class mlrl.testbed.experiments.input.data.DatasetInputData(properties: Properties, context: Context = Context(include_dataset_type=True, include_prediction_scope=True, include_fold=True))

Bases: InputData

Input data that can be converted into a dataset.

class mlrl.testbed.experiments.input.data.InputData(properties: Properties, context: Context = Context(include_dataset_type=True, include_prediction_scope=True, include_fold=True))

Bases: object

Represents input data.

get_key(state: ExperimentState) → str

Returns the key that is used to add the input data to the extras of an ExperimentState.

Parameters:

state – The state to be updated

Returns:

The key

update_state(state: ExperimentState, input_data: Any)

Updates the state of an experiment based on given input data.

Parameters:

  • state – The state to be updated

  • input_data – The input data

class mlrl.testbed.experiments.input.data.StructuralInputData(properties: Properties, context: Context = Context(include_dataset_type=True, include_prediction_scope=True, include_fold=True))

Bases: InputData

An abstract base class for all classes that represent input data that can be converted into a structural representation, e.g., YAML or JSON.

update_state(state: ExperimentState, input_data: Any)

See mlrl.testbed.experiments.input.data.InputData.update_state()

class mlrl.testbed.experiments.input.data.TabularInputData(properties: TabularProperties, context: Context = Context(include_dataset_type=True, include_prediction_scope=True, include_fold=True))

Bases: InputData

Input data that can be converted into a tabular representation.

update_state(state: ExperimentState, input_data: Any)

Updates the state of an experiment based on given input data.

Parameters:

  • state – The state to be updated

  • input_data – The input data

class mlrl.testbed.experiments.input.data.TextualInputData(properties: Properties, context: Context = Context(include_dataset_type=True, include_prediction_scope=True, include_fold=True))

Bases: InputData

Input data that can be converted into a textual representation.

update_state(state: ExperimentState, input_data: Any)

See mlrl.testbed.experiments.input.data.InputData.update_state()