Source code for grid2op.Observation.observationSpace

# Copyright (c) 2019-2020, RTE (https://www.rte-france.com)
# See AUTHORS.txt
# This Source Code Form is subject to the terms of the Mozilla Public License, version 2.0.
# If a copy of the Mozilla Public License, version 2.0 was not distributed with this file,
# you can obtain one at http://mozilla.org/MPL/2.0/.
# SPDX-License-Identifier: MPL-2.0
# This file is part of Grid2Op, Grid2Op a testbed platform to model sequential decision making in power systems.

import sys
import copy
import logging
import os
from grid2op.Exceptions.envExceptions import EnvError

from grid2op.Observation.serializableObservationSpace import (
    SerializableObservationSpace,
)
from grid2op.Reward import RewardHelper
from grid2op.Observation.completeObservation import CompleteObservation


[docs]class ObservationSpace(SerializableObservationSpace): """ Helper that provides useful functions to manipulate :class:`BaseObservation`. BaseObservation should only be built using this Helper. It is absolutely not recommended to make an observation directly form its constructor. This class represents the same concept as the "BaseObservation Space" in the OpenAI gym framework. Attributes ---------- with_forecast: ``bool`` If ``True`` the :func:`BaseObservation.simulate` will be available. If ``False`` it will deactivate this possibility. If `simulate` function is not used, setting it to ``False`` can lead to non neglectible speed-ups. observationClass: ``type`` Class used to build the observations. It defaults to :class:`CompleteObservation` _simulate_parameters: :class:`grid2op.Parameters.Parameters` Type of Parameters used to compute powerflow for the forecast. rewardClass: Union[type, BaseReward] Class used by the :class:`grid2op.Environment.Environment` to send information about its state to the :class:`grid2op.Agent.BaseAgent`. You can change this class to differentiate between the reward of output of :func:`BaseObservation.simulate` and the reward used to train the BaseAgent. action_helper_env: :class:`grid2op.Action.ActionSpace` BaseAction space used to create action during the :func:`BaseObservation.simulate` reward_helper: :class:`grid2op.Reward.RewardHelper` BaseReward function used by the the :func:`BaseObservation.simulate` function. obs_env: :class:`grid2op.Environment._Obsenv._ObsEnv` Instance of the environment used by the BaseObservation Helper to provide forcecast of the grid state. _empty_obs: :class:`BaseObservation` An instance of the observation with appropriate dimensions. It is updated and will be sent to he BaseAgent. """
[docs] def __init__( self, gridobj, env, rewardClass=None, observationClass=CompleteObservation, actionClass=None, with_forecast=True, kwargs_observation=None, observation_bk_class=None, observation_bk_kwargs=None, logger=None, _with_obs_env=True, # pass _local_dir_cls=None, ): """ INTERNAL .. warning:: /!\\\\ Internal, do not use unless you know what you are doing /!\\\\ Env: requires :attr:`grid2op.Environment.BaseEnv.parameters` and :attr:`grid2op.Environment.BaseEnv.backend` to be valid """ if actionClass is None: from grid2op.Action import CompleteAction actionClass = CompleteAction self._init_observationClass = observationClass SerializableObservationSpace.__init__( self, gridobj, observationClass=observationClass, _local_dir_cls=_local_dir_cls, logger=logger, ) self.with_forecast = with_forecast self._simulate_parameters = copy.deepcopy(env.parameters) self._legal_action = env._game_rules.legal_action self._env_param = copy.deepcopy(env.parameters) if rewardClass is None: self._reward_func = env._reward_helper.template_reward else: self._reward_func = rewardClass # helpers self.action_helper_env = env._helper_action_env self.reward_helper = RewardHelper(reward_func=self._reward_func, logger=self.logger) self.__can_never_use_simulate = False _with_obs_env = _with_obs_env and self._create_backend_obs(env, observation_bk_class, observation_bk_kwargs, _local_dir_cls) self._ObsEnv_class = None if _with_obs_env: self._create_obs_env(env, observationClass) self.reward_helper.initialize(self.obs_env) for k, v in self.obs_env.other_rewards.items(): v.reset(self.obs_env) else: self.with_forecast = False self.obs_env = None self._backend_obs = None self.__can_never_use_simulate = True self._empty_obs = self._template_obj self._update_env_time = 0.0 self.__nb_simulate_called_this_step = 0 self.__nb_simulate_called_this_episode = 0 self._highres_sim_counter = env.highres_sim_counter # extra argument to build the observation if kwargs_observation is None: kwargs_observation = {} self._ptr_kwargs_observation = kwargs_observation self._real_env_kwargs = {} self._observation_bk_class = observation_bk_class self._observation_bk_kwargs = observation_bk_kwargs
def set_real_env_kwargs(self, env): if not self.with_forecast: return # I don't need the backend nor the chronics_handler from grid2op.Environment import Environment self._real_env_kwargs = Environment.get_kwargs(env, False, False) # remove the parameters anyways (the 'forecast parameters will be used # when building the forecasted_env) del self._real_env_kwargs["parameters"] # i also "remove" the opponent from grid2op.Action import DontAct from grid2op.Opponent import BaseOpponent, NeverAttackBudget self._real_env_kwargs["opponent_action_class"] = DontAct self._real_env_kwargs["opponent_class"] = BaseOpponent self._real_env_kwargs["opponent_init_budget"] = 0. self._real_env_kwargs["opponent_budget_per_ts"] = 0. self._real_env_kwargs["opponent_budget_class"] = NeverAttackBudget self._real_env_kwargs["opponent_attack_duration"] = 0 self._real_env_kwargs["opponent_attack_cooldown"] = 999999 # and finally I remove the extra bk_class and bk_kwargs if "observation_bk_class" in self._real_env_kwargs: del self._real_env_kwargs["observation_bk_class"] if "observation_bk_kwargs" in self._real_env_kwargs: del self._real_env_kwargs["observation_bk_kwargs"] def _create_obs_env(self, env, observationClass): if self._ObsEnv_class is None: # lazy import to prevent circular references (Env -> Observation -> Obs Space -> _ObsEnv -> Env) from grid2op.Environment._obsEnv import _ObsEnv # self._ObsEnv_class = _ObsEnv.init_grid( # type(env.backend), force_module=_ObsEnv.__module__, force=_local_dir_cls is not None # ) # self._ObsEnv_class._INIT_GRID_CLS = _ObsEnv # otherwise it's lost self._ObsEnv_class = _ObsEnv.init_grid( type(env.backend), _local_dir_cls=env._local_dir_cls ) self._ObsEnv_class._INIT_GRID_CLS = _ObsEnv # otherwise it's lost other_rewards = {k: v.rewardClass for k, v in env.other_rewards.items()} self.obs_env = self._ObsEnv_class( init_env_path=None, # don't leak the path of the real grid to the observation space init_grid_path=None, # don't leak the path of the real grid to the observation space backend_instanciated=self._backend_obs, obsClass=CompleteObservation, # do not put self.observationClass otherwise it's initialized twice parameters=self._simulate_parameters, reward_helper=self.reward_helper, action_helper=self.action_helper_env, thermal_limit_a=env.get_thermal_limit(), legalActClass=copy.deepcopy(env._legalActClass), other_rewards=other_rewards, helper_action_class=env._helper_action_class, helper_action_env=env._helper_action_env, epsilon_poly=env._epsilon_poly, tol_poly=env._tol_poly, has_attention_budget=env._has_attention_budget, attention_budget_cls=env._attention_budget_cls, kwargs_attention_budget=env._kwargs_attention_budget, max_episode_duration=env.max_episode_duration(), delta_time_seconds=env.delta_time_seconds, logger=self.logger, highres_sim_counter=env.highres_sim_counter, _complete_action_cls=env._complete_action_cls, _ptr_orig_obs_space=self, _local_dir_cls=env._local_dir_cls, _read_from_local_dir=env._read_from_local_dir, ) for k, v in self.obs_env.other_rewards.items(): v.initialize(self.obs_env) def _aux_create_backend(self, env, observation_bk_class, observation_bk_kwargs, path_grid_for, _local_dir_cls): if observation_bk_kwargs is None: observation_bk_kwargs = env.backend._my_kwargs observation_bk_class_used = observation_bk_class.init_grid(type(env.backend), _local_dir_cls=_local_dir_cls) self._backend_obs = observation_bk_class_used(**observation_bk_kwargs) self._backend_obs.set_env_name(env.name) self._backend_obs.load_grid(path_grid_for) self._backend_obs.assert_grid_correct() self._backend_obs.runpf() self._backend_obs.assert_grid_correct_after_powerflow() self._backend_obs.set_thermal_limit(env.get_thermal_limit()) def _create_backend_obs(self, env, observation_bk_class, observation_bk_kwargs, _local_dir_cls): _with_obs_env = True path_sim_bk = os.path.join(env.get_path_env(), "grid_forecast.json") if observation_bk_class is not None or observation_bk_kwargs is not None: # backend used for simulate is of a different class (or build with different arguments) if observation_bk_class is not None: self.logger.warn("Using a backend for the 'forecast' of a different class. Make sure the " "elements of the grid are in the same order and have the same name ! " "Do not hesitate to use a 'BackendConverter' if that is not the case.") else: observation_bk_class = env._raw_backend_class if os.path.exists(path_sim_bk) and os.path.isfile(path_sim_bk): path_grid_for = path_sim_bk else: path_grid_for = os.path.join(env.get_path_env(), "grid.json") self._aux_create_backend(env, observation_bk_class, observation_bk_kwargs, path_grid_for, _local_dir_cls) elif os.path.exists(path_sim_bk) and os.path.isfile(path_sim_bk): # backend used for simulate will use the same class with same args as the env # backend, but with a different grid observation_bk_class = env._raw_backend_class self._aux_create_backend(env, observation_bk_class, observation_bk_kwargs, path_sim_bk, _local_dir_cls) elif env.backend._can_be_copied: # case where I can copy the backend for the 'simulate' and I don't need to build # it (uses same class and same grid) try: self._backend_obs = env.backend.copy() except Exception as exc_: self._backend_obs = None self.logger.warn(f"Backend cannot be copied, simulate feature will " f"be unsusable. Error was: {exc_}") self._deactivate_simulate(env) _with_obs_env = False self.__can_never_use_simulate = True else: # no 'simulate' can be made unfortunately self._backend_obs = None self._deactivate_simulate(env) _with_obs_env = False self.__can_never_use_simulate = True return _with_obs_env def _deactivate_simulate(self, env): if self._backend_obs is not None: self._backend_obs.close() self._backend_obs = None self.with_forecast = False if env is not None: env.deactivate_forecast() env.backend._can_be_copied = False self.logger.warning("Forecasts have been deactivated because " "the backend cannot be copied.") def reactivate_forecast(self, env): if self.__can_never_use_simulate: raise EnvError("You cannot use `simulate` for this environment, either because the " "backend you used cannot be copied, or because this observation space " "does not support this feature.") if self.obs_env is None or self._backend_obs is None: # force create of everything in this case if self._backend_obs is not None: self._backend_obs.close() self._backend_obs = None self._create_backend_obs(env, self._observation_bk_class, self._observation_bk_kwargs, env._local_dir_cls) if self.obs_env is not None: self.obs_env.close() self.obs_env = None self._create_obs_env(env, self._init_observationClass) self.set_real_env_kwargs(env) self.with_forecast = True
[docs] def simulate_called(self): """ INTERNAL .. warning:: /!\\\\ Internal, do not use unless you know what you are doing /!\\\\ Tells this class that the "obs.simulate" function has been called. """ self.__nb_simulate_called_this_step += 1 self.__nb_simulate_called_this_episode += 1
@property def nb_simulate_called_this_episode(self): return self.__nb_simulate_called_this_episode @property def nb_simulate_called_this_step(self): return self.__nb_simulate_called_this_step @property def total_simulate_simulator_calls(self): return self._highres_sim_counter.total_simulate_simulator_calls
[docs] def can_use_simulate(self) -> bool: """ This checks on the rules if the agent has not made too many calls to "obs.simulate" this step """ return self._legal_action.can_use_simulate( self.__nb_simulate_called_this_step, self.__nb_simulate_called_this_episode, self._env_param, )
def _change_parameters(self, new_param): """ INTERNAL .. warning:: /!\\\\ Internal, do not use unless you know what you are doing /!\\\\ change the parameter of the "simulate" environment """ if self.obs_env is not None: self.obs_env.change_parameters(new_param) self._simulate_parameters = new_param
[docs] def change_other_rewards(self, dict_reward): """ this function is used to change the "other rewards" used when you perform simulate. This can be used, for example, when you want to do faster call to "simulate". In this case you can remove all the "other_rewards" that will be used by the simulate function. Parameters ---------- dict_reward: ``dict`` see description of :attr:`grid2op.Environment.BaseEnv.other_rewards` Examples --------- If you want to deactivate the reward in the simulate function, you can do as following: .. code-block:: python import grid2op from grid2op.Reward import CloseToOverflowReward, L2RPNReward, RedispReward env_name = "l2rpn_case14_sandbox" other_rewards = {"close_overflow": CloseToOverflowReward, "l2rpn": L2RPNReward, "redisp": RedispReward} env = grid2op.make(env_name, other_rewards=other_rewards) env.observation_space.change_other_rewards({}) """ from grid2op.Reward import BaseReward from grid2op.Exceptions import Grid2OpException if self.obs_env is not None: self.obs_env.other_rewards = {} for k, v in dict_reward.items(): if not issubclass(v, BaseReward): raise Grid2OpException( 'All values of "rewards" key word argument should be classes that inherit ' 'from "grid2op.BaseReward"' ) if not isinstance(k, str): raise Grid2OpException( 'All keys of "rewards" should be of string type.' ) self.obs_env.other_rewards[k] = RewardHelper(v) for k, v in self.obs_env.other_rewards.items(): v.initialize(self.obs_env)
def change_reward(self, reward_func): if self.obs_env is not None: if self.obs_env.is_valid(): self.obs_env._reward_helper.change_reward(reward_func) else: raise EnvError("Impossible to change the reward of the simulate " "function when you cannot simulate (because the " "backend could not be copied)") def set_thermal_limit(self, thermal_limit_a): if self.obs_env is not None: self.obs_env.set_thermal_limit(thermal_limit_a) if self._backend_obs is not None: self._backend_obs.set_thermal_limit(thermal_limit_a) def reset_space(self): if self.with_forecast: if self.obs_env.is_valid(): self.obs_env.reset_space() else: raise EnvError("Impossible to reset_space " "function when you cannot simulate (because the " "backend could not be copied)") self.action_helper_env.actionClass.reset_space()
[docs] def __call__(self, env, _update_state=True): obs_env_obs = None if self.with_forecast: self.obs_env.update_grid(env) obs_env_obs = self.obs_env if self.obs_env.is_valid() else None res = self.observationClass( obs_env=obs_env_obs, action_helper=self.action_helper_env, random_prng=self.space_prng, kwargs_env=self._real_env_kwargs, **self._ptr_kwargs_observation ) self.__nb_simulate_called_this_step = 0 if _update_state: # TODO how to make sure that whatever the number of time i call "simulate" i still get the same observations # TODO use self.obs_prng when updating actions res.update(env=env, with_forecast=self.with_forecast) return res
[docs] def size_obs(self): """ Size if the observation vector would be flatten :return: """ return self.n
[docs] def get_empty_observation(self): """ INTERNAL .. warning:: /!\\\\ Internal, do not use unless you know what you are doing /!\\\\ return an empty observation, for internal use only. """ return copy.deepcopy(self._empty_obs)
[docs] def reset(self, real_env): """reset the observation space with the new values of the environment""" self.__nb_simulate_called_this_step = 0 self.__nb_simulate_called_this_episode = 0 if self.with_forecast: self.obs_env._reward_helper.reset(self.obs_env) for k, v in self.obs_env.other_rewards.items(): v.reset(self.obs_env) self.obs_env.reset() self._env_param = copy.deepcopy(real_env.parameters)
def _custom_deepcopy_for_copy(self, new_obj, env=None): """implements a faster "res = copy.deepcopy(self)" to use in "self.copy" Do not use it anywhere else... """ # TODO clean that after it is working... (ie make this method per class...) # fill the super classes super()._custom_deepcopy_for_copy(new_obj) # now fill my class new_obj._init_observationClass = self._init_observationClass new_obj.with_forecast = self.with_forecast new_obj._simulate_parameters = copy.deepcopy(self._simulate_parameters) new_obj._reward_func = copy.deepcopy(self._reward_func) new_obj.action_helper_env = self.action_helper_env # const new_obj.reward_helper = copy.deepcopy(self.reward_helper) new_obj._backend_obs = self._backend_obs # ptr to a backend for simulate new_obj.obs_env = self.obs_env # it is None anyway ! new_obj._update_env_time = self._update_env_time new_obj.__can_never_use_simulate = self.__can_never_use_simulate new_obj.__nb_simulate_called_this_step = self.__nb_simulate_called_this_step new_obj.__nb_simulate_called_this_episode = ( self.__nb_simulate_called_this_episode ) # never copied (keep track of it) new_obj._highres_sim_counter = ( self._highres_sim_counter ) new_obj._env_param = copy.deepcopy(self._env_param) # as it's a "pointer" it's init from the env when needed here # this is why i don't deep copy it here ! new_obj._ptr_kwargs_observation = self._ptr_kwargs_observation # real env kwargs, these is a "pointer" anyway if env is not None: from grid2op.Environment import Environment new_obj._real_env_kwargs = Environment.get_kwargs(env, False, False) else: new_obj._real_env_kwargs = self._real_env_kwargs new_obj._observation_bk_class = self._observation_bk_class new_obj._observation_bk_kwargs = self._observation_bk_kwargs new_obj._ObsEnv_class = self._ObsEnv_class
[docs] def copy(self, copy_backend=False, env=None): """ INTERNAL .. warning:: /!\\\\ Internal, do not use unless you know what you are doing /!\\\\ Perform a deep copy of the Observation space. """ backend = self._backend_obs self._backend_obs = None obs_ = self._empty_obs self._empty_obs = None obs_env = self.obs_env self.obs_env = None # performs the copy # res = copy.deepcopy(self) # painfully slow... # create an empty "me" my_cls = type(self) res = my_cls.__new__(my_cls) self._custom_deepcopy_for_copy(res, env) if not copy_backend: res._backend_obs = backend res._empty_obs = obs_.copy() res.obs_env = obs_env else: # backend needs to be copied if obs_env is not None: # I also need to copy the obs env res.obs_env = obs_env.copy(env=env, new_obs_space=res) res._backend_obs = res.obs_env.backend res._empty_obs = obs_.copy() res._empty_obs._obs_env = res.obs_env else: # no obs env: I do nothing res.obs_env = None # assign back the results self._backend_obs = backend self._empty_obs = obs_ self.obs_env = obs_env return res
def close(self): if self.obs_env is not None: self.obs_env.close() del self.obs_env self.obs_env = None