def __init__(self,
action_scheme: ActionScheme,
reward_scheme: RewardScheme,
observer: Observer,
stopper: Stopper,
informer: Informer,
renderer: Renderer,
min_periods: int = None,
**kwargs) -> None:
super().__init__()
self.clock = Clock()
self.action_scheme = action_scheme
self.reward_scheme = reward_scheme
self.observer = observer
self.stopper = stopper
self.informer = informer
self.renderer = renderer
self.min_periods = min_periods
for c in self.components.values():
c.clock = self.clock
self.action_space = action_scheme.action_space
self.observation_space = observer.observation_space
self._enable_logger = kwargs.get('enable_logger', False)
if self._enable_logger:
self.logger = logging.getLogger(kwargs.get('logger_name',
__name__))
self.logger.setLevel(kwargs.get('log_level', logging.DEBUG))
def test_basic_clock_init():
clock = Clock()
assert clock
assert clock.start == 0
assert clock.step == 0
class TradingEnv(gym.Env, TimeIndexed):
"""A trading environment made for use with Gym-compatible reinforcement
learning algorithms.
Parameters
----------
action_scheme : `ActionScheme`
A component for generating an action to perform at each step of the
environment.
reward_scheme : `RewardScheme`
A component for computing reward after each step of the environment.
observer : `Observer`
A component for generating observations after each step of the
environment.
informer : `Informer`
A component for providing information after each step of the
environment.
renderer : `Renderer`
A component for rendering the environment.
kwargs : keyword arguments
Additional keyword arguments needed to create the environment.
"""
agent_id: str = None
episode_id: str = None
def __init__(self,
action_scheme: ActionScheme,
reward_scheme: RewardScheme,
observer: Observer,
stopper: Stopper,
informer: Informer,
renderer: Renderer,
min_periods: int = None,
**kwargs) -> None:
super().__init__()
self.clock = Clock()
self.action_scheme = action_scheme
self.reward_scheme = reward_scheme
self.observer = observer
self.stopper = stopper
self.informer = informer
self.renderer = renderer
self.min_periods = min_periods
for c in self.components.values():
c.clock = self.clock
self.action_space = action_scheme.action_space
self.observation_space = observer.observation_space
self._enable_logger = kwargs.get('enable_logger', False)
if self._enable_logger:
self.logger = logging.getLogger(kwargs.get('logger_name',
__name__))
self.logger.setLevel(kwargs.get('log_level', logging.DEBUG))
@property
def components(self) -> 'Dict[str, Component]':
"""The components of the environment. (`Dict[str,Component]`, read-only)"""
return {
"action_scheme": self.action_scheme,
"reward_scheme": self.reward_scheme,
"observer": self.observer,
"stopper": self.stopper,
"informer": self.informer,
"renderer": self.renderer
}
def step(self, action: Any) -> 'Tuple[np.array, float, bool, dict]':
"""Makes on step through the environment.
Parameters
----------
action : Any
An action to perform on the environment.
Returns
-------
`np.array`
The observation of the environment after the action being
performed.
float
The computed reward for performing the action.
bool
Whether or not the episode is complete.
dict
The information gathered after completing the step.
"""
self.action_scheme.perform(self, action)
obs = self.observer.observe(self)
reward = self.reward_scheme.reward(self)
done = self.stopper.stop(self)
info = self.informer.info(self)
self.clock.increment()
return obs, reward, done, info
def reset(self) -> 'np.array':
"""Resets the environment.
Returns
-------
obs : `np.array`
The first observation of the environment.
"""
self.episode_id = str(uuid.uuid4())
self.clock.reset()
for c in self.components.values():
if hasattr(c, "reset"):
c.reset()
obs = self.observer.observe(self)
self.clock.increment()
return obs
def render(self, **kwargs) -> None:
"""Renders the environment."""
self.renderer.render(self, **kwargs)
def save(self) -> None:
"""Saves the rendered view of the environment."""
self.renderer.save()
def close(self) -> None:
"""Closes the environment."""
self.renderer.close()
def test_basic_clock_increment():
clock = Clock()
clock.increment()
assert clock.step == 1