palaestrai.entrypoints package

Submodules

palaestrai.entrypoints.palaestrai_runner module

palaestrai.entrypoints.palaestrai_runner.execute(experiment_run_definition: ExperimentRun | TextIO | str | Path | List[ExperimentRun | TextIO | str | Path], runtime_config: str | TextIO | dict | None = None, parallel_runs: int = 1) → Tuple[List[str], ExecutorState][source]

Provides a single-line command to start an experiment

This function is a high-level wrapper that takes a number of experiment run defintions and optionally a runtime config, executes the runs, and returns the results.

Note

This is a sync method. If you already have an event loop running, please use this function’s sibling, run()

Parameters:

  • experiment_run_definition –

    1. Any text stream

      1. The path to a file

    The configuration from which the experiment is loaded.

  • runtime_config –

    1. dict

      1. None

    The Runtime configuration applicable for the run. Note that even when no additional source is provided, runtime will load a minimal configuration from build-in defaults.

  • parallel_runs (int, default: 1) – Number of experiment runs that can be executed in parallel

Returns:

A tuple containing: 1. The list of all experiment run IDs 2. The final state of the executor

Return type:

Tuple[Sequence[str], ExecutorState]

async palaestrai.entrypoints.palaestrai_runner.run(experiment_run_definition: ExperimentRun | TextIO | str | Path | List[ExperimentRun | TextIO | str | Path], runtime_config: str | TextIO | dict | None = None, parallel_runs: int = 1) → Tuple[List[str], ExecutorState][source]

Provides a single-line command to start an experiment

This is the asynchronous variant of execute(). Use it if you have already an event loop running (e.g., in a Jupyter Notebook).

Module contents