EpisodeBuffer

minari.data_collector.EpisodeBuffer

class minari.data_collector.EpisodeBuffer(id: int | None = None, seed: int | None = None, options: dict | None = None, observations: None | list | dict | tuple = None, actions: None | list | dict | tuple = None, rewards: list = <factory>, terminations: list = <factory>, truncations: list = <factory>, infos: dict | None = None)[source]

Contains the data of a single episode.

Attributes

EpisodeBuffer.id: int | None = None
EpisodeBuffer.seed: int | None = None
EpisodeBuffer.observations: None | list | dict | tuple = None
EpisodeBuffer.actions: None | list | dict | tuple = None
EpisodeBuffer.rewards: list
EpisodeBuffer.terminations: list
EpisodeBuffer.truncations: list
EpisodeBuffer.infos: dict | None = None

Methods

minari.data_collector.EpisodeBuffer.add_step_data(self, step_data: StepData) EpisodeBuffer

Add step data dictionary to episode buffer.

Parameters:

step_data (StepData) – dictionary with data for a single step

Returns:

EpisodeBuffer – episode buffer with appended data

minari.data_collector.EpisodeBuffer.__len__(self) int

Buffer length.