BaseChallenge

class gaitmap_challenges.challenge_base.BaseChallenge

Base class for all challenges.

Note, that we decided to use datas-classes for challenges to reduce the amount of code that needs to be written. This means, that you need to wrap your challenge into a dataclass decorator as well.

Besides providing an expected interface, this class provides a _measure_time context manager that can be used within your implementation of the run method to measure the execution time of an algorithm. This will automatically set a number of results attributes about the start/end and runtime of the execution.

Other Parameters:

VERSION

(Class Constant) The version of the challenge. In case of breaking changes, this should be increased.

optimizer

The optimizer passed to the run method.

Attributes:

run_start_datetime_utc_timestamp_

The start time of the execution of the challenge in UTC timestamp format.

run_start_datetime_

The start time of the execution of the challenge in ISO format.

end_start_datetime_utc_timestamp_

The end time of the execution of the challenge in UTC timestamp format.

end_start_datetime_

The end time of the execution of the challenge in ISO format.

runtime_

The runtime of the execution of the challenge in seconds.

dataset_

The instance of the dataset class actually used in the challenge. This is usually a copy of the dataset class instance supplied to the challenge and has potentially some parameters overwritten to ensure that the challenge is executed as expected.

Methods

clone()	Create a new instance of the class with all parameters copied over.
get_core_results()	Get the main results of the challenge.
get_params([deep])	Get parameters for this algorithm.
load_core_results(folder_path)	Load the core results from a folder that have been stored using save_core_results.
run(optimizer)	Run the challenge.
save_core_results(folder_path)	Save the main results of the challenge to a folder.
set_params(**params)	Set the parameters of this Algorithm.

__init__() → None

property _measure_time: Callable[[], Generator[None, None, None]]

Context manager to measure the execution time of an algorithm.

Use this within your implementation of the run method to measure the execution time of an algorithm.

get_core_results() → Dict[str, Any]

Get the main results of the challenge.

classmethod load_core_results(folder_path: Path) → Dict[str, Any]

Load the core results from a folder that have been stored using save_core_results.

The assumption is, that the output is identical to calling get_core_results on the challenge instance directly.

When implementing this method, make sure that it remains compatible with results saved using older versions of the challenge_class.

run(optimizer: BaseOptimize)

Run the challenge.

save_core_results(folder_path: Path)

Save the main results of the challenge to a folder.