Lab Space

Experiments

This script is intended to run single or multithreaded experiments.

class lab_space.experiment.Experiment(trial_config=None, expt_config: Optional[dict] = None, log_level: str = 'WARNING')

Bases: object

Perform trials for a given set of experiments.

Users should provide two parameter files.

Parameters:

  • trial_config – (list(dict)/generator) Configurations for each trial, default: None

  • expt_config – (dict) Experiment configuration file containing the following keys: - default: None - “experiment”: (func) Reference to function under test - “n_trials”: (int) Number of trials to run for each set of parameters, default: 1 - “n_processes”: (int) Number of processes to use, default: 1 - “data_file”: (str) file to data data, if none does not data, default: None - “clear_data”: (bool) clears data from pickle before running experiment, default: False

  • log_level – (str) Logging level, default: “WARNING”

reset(trial_config=None, expt_config: Optional[dict] = None)

Reset experiment with new configurations

Parameters:

  • trial_config – (list(dict)/generator) Configurations for each trial, default: None

  • expt_config – (dict) Experiment configuration, default: None

run(trial_config=None, expt_config: Optional[dict] = None)

Run experiment with new configurations

Parameters:

  • trial_config – (list(dict)/generator) Configurations for each trial, default: None

  • expt_config – (dict) Experiment configuration, default: None

Returns:

(pandas.DataFrame) data

Labspace

This script is handle command line argmuents for starting experiments and analyzing data.

lab_space.labspace.deregister_experiment(experiment: str)

Deregisters experiment to be run

Parameters:

experiment – (str) Experiment to deregister

lab_space.labspace.get_registered_experiment(experiment: str)

Gets all registered experiment

Parameters:

experiment – (module) Experiment module

Returns:

(function) Experiment function reference

lab_space.labspace.register_experiment(module_name: str, func_name: Optional[str] = None, func_key_name: Optional[str] = None, module_path: Optional[str] = None)

Registers experiment to be run

Parameters:

  • module_name – (str) Name of experiment file

  • func_name – (str) Function to call register

  • module_path – (str) Path to module

  • func_key_name – (str) Key to identify function