mlrl.testbed.runnables module

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

Provides base classes for programs that can be configured via command line arguments.

class mlrl.testbed.runnables.Runnable

Bases: Recipe, ABC

An abstract base class for all programs that can be configured via the command line API. The programs functionality is implemented by individual extensions that are applied to the runnable.

class PredictionDatasetExtension(*dependencies: Extension)

Bases: Extension

An extension that configures the functionality to predict for different datasets.

configure_experiment(args: Namespace, experiment_builder: Builder, _: ExperimentMode)

See mlrl.testbed.extensions.extension.Extension.configure_experiment()

get_supported_modes() → set[ExperimentMode]

See mlrl.testbed.extensions.extension.Extension.get_supported_modes()

configure_arguments(cli: CommandLineInterface, mode: Mode) → tuple[set[Argument], set[Argument]]

Configures the command line interface according to the extensions applied to the runnable.

Parameters:

  • cli – The command line interface to be configured

  • mode – The mode of operation

Returns:

A set that contains the arguments that should be added to the command line interface according to the registered extensions, as well as a set that contains algorithmic arguments

configure_batch_mode(cli: CommandLineInterface) → BatchMode

Configures the batch mode according to the extensions applied to the runnable.

Parameters:

cli – The command line interface to be configured

configure_read_mode(cli: CommandLineInterface) → ReadMode

Configures the read mode according to the extensions applied to the runnable.

Parameters:

cli – The command line interface to be configured

configure_run_mode(cli: CommandLineInterface) → RunMode

Configures the run mode according to the extensions applied to the runnable.

Parameters:

cli – The command line interface to be configured

abstractmethod create_batch_config_file_factory() → Callable[[str], ConfigFile]

Must be implemented by subclasses in order to create the factory that allows to create the configuration file that configures the batch of experiments to be run in batch mode.

Returns:

The factory that has been created

get_algorithmic_arguments(mode: ExperimentMode, known_args: Namespace) → set[Argument]

May be overridden by subclasses in order to return the arguments for configuring algorithmic parameters that should be added to the command line API.

Parameters:

  • mode – The mode of operation

  • known_args – The command line arguments specified by the user

Returns:

A set that contains the arguments that should be added to the command line API

get_extensions() → list[Extension]

May be overridden by subclasses in order to return the extensions that should be applied to the runnable.

Returns:

A set that contains the extensions to be applied to the runnable

get_program_info() → ProgramInfo | None

May be overridden by subclasses in order to provide information about the program to be printed via the command line argument ‘-v’ or ‘–version’.

Returns:

A ProgramInfo or None, if no information is provided

get_supported_extensions(mode: ExperimentMode) → list[Extension]

Returns the extensions that should be applied to the runnable and support a given mode of operation.

Parameters:

mode – The mode to be supported

Returns:

A list that contains all extensions that should be applied to the runnable and support the given mode

run(mode: Mode, control_arguments: set[Argument], algorithmic_arguments: set[Argument], args: Namespace)

Executes the runnable.

Parameters:

  • mode – The mode of operation

  • control_arguments – The arguments that should be added to the command line interface for controlling mlrl-testbed’s behavior

  • algorithmic_arguments – The arguments that should be added to the command line interface for configuring the algorithm’s hyperparameters

  • args – The command line arguments specified by the user