rlberry-scool API

Manager

Main classe

rlberry.manager.ExperimentManager(agent_class)

Class to train, optimize hyperparameters, evaluate and gather statistics about an agent.

Evaluation and plot

rlberry.manager.evaluate_agents(...[, ...])	Evaluate and compare each of the agents in experiment_manager_list.
rlberry.manager.plot_writer_data(...[, ...])	Given a list of ExperimentManager or a folder, plot data (corresponding to info) obtained in each episode.
rlberry.manager.read_writer_data(data_source)	Given a list of ExperimentManager or a folder, read data (corresponding to info) obtained in each episode.
rlberry.manager.compare_agents(agent_source)	Compare several trained agents using the mean over n_simulations evaluations for each agent.

rlberry.manager.AdastopComparator([n, K, B, ...])

Compare sequentially agents, with possible early stopping.

Agents & Environments

Basic agents

rlberry_scool.agents.linear.LSVIUCBAgent(...)	A version of Least-Squares Value Iteration with UCB (LSVI-UCB), proposed by Jin et al. (2020).
rlberry_scool.agents.tabular_rl.QLAgent(env)	Q-Learning Agent.
rlberry_scool.agents.tabular_rl.SARSAAgent(env)	SARSA Agent.

Basic environments

rlberry_scool.envs.GridWorld([nrows, ncols, ...])	Simple GridWorld environment.
rlberry_scool.envs.Chain([L, fail_prob])	Simple chain environment.

Agent importation tools

rlberry.agents.stable_baselines.StableBaselinesAgent(env)

Wraps an StableBaselines3 Algorithm with a rlberry Agent.

Environment tools

rlberry.envs.gym_make(id[, wrap_spaces])

Same as gym.make, but wraps the environment to ensure unified seeding with rlberry.

Seeding

rlberry.seeding.safe_reseed(obj, seeder[, ...])	Calls obj.reseed(seed_seq) method if available; If a obj.seed() method is available, call obj.seed(seed_val), where seed_val is generated by the seeder.
rlberry.seeding.set_external_seed(seeder)	Set seeds of external libraries.

Environment Wrappers

rlberry.wrappers.discretize_state.DiscretizeStateWrapper(...)	Discretize an environment with continuous states and discrete actions.
rlberry.wrappers.RescaleRewardWrapper(env, ...)	Rescale the reward function to a bounded range.
rlberry.wrappers.WriterWrapper(env, writer)	Wrapper for environment to automatically record reward or action in writer.