Source code for causal_world.intervention_actors.visual_actor

from causal_world.intervention_actors.base_actor import \
    BaseInterventionActorPolicy
import numpy as np


[docs]class VisualInterventionActorPolicy(BaseInterventionActorPolicy):
[docs] def __init__(self, **kwargs): """ This intervention actor intervenes on all visual components of the robot, (i.e: colors). :param kwargs: """ super(VisualInterventionActorPolicy, self).__init__() self.task_intervention_space = None
[docs] def initialize(self, env): """ This functions allows the intervention actor to query things from the env, such as intervention spaces or to have access to sampling funcs for goals..etc :param env: (causal_world.env.CausalWorld) the environment used for the intervention actor to query different methods from it. :return: """ self.task_intervention_space = env.get_variable_space_used() return
def _act(self, variables_dict): """ :param variables_dict: :return: """ interventions_dict = dict() for variable in self.task_intervention_space: if isinstance(self.task_intervention_space[variable], dict): if 'color' in self.task_intervention_space[variable]: interventions_dict[variable] = dict() interventions_dict[variable]['color'] = np.random.uniform( self.task_intervention_space[variable]['color'][0], self.task_intervention_space[variable]['color'][1]) elif 'color' in variable: interventions_dict[variable] = np.random.uniform( self.task_intervention_space[variable][0], self.task_intervention_space[variable][1]) return interventions_dict
[docs] def get_params(self): """ returns parameters that could be used in recreating this intervention actor. :return: (dict) specifying paramters to create this intervention actor again. """ return {'visual_actor': dict()}