from dataclasses import dataclass
import yaml
@dataclass
class EnvConfig:
base: float = 5.0
amplitude: float = 2.0
period_s: float = 7200.0
door_drop_C: float = 5.0
door_start_s: float = 1200.0
door_duration_s: float = 120.0
@dataclass
class SensorConfig:
sigma: float = 0.2
bias: float = 0.3
dropout_prob: float = 0.02
@dataclass
class ControllerConfig:
type: str = 'predictive_onoff' # 'onoff' or 'predictive_onoff'
setpoint: float = 21.0
deadband: float = 1.0
tau: float = 30.0
safety_high: float = 26.0
@dataclass
class ModelConfig:
R: float = 0.5
C: float = 10000.0
P: float = 200.0
process_sigma: float = 0.0
@dataclass
class SimConfig:
dt: float = 1.0
duration_s: float = 3600.0
seed: int = 42
init_T: float = 18.0
@dataclass
class Config:
env: EnvConfig
sensor: SensorConfig
controller: ControllerConfig
model: ModelConfig
sim: SimConfig
def load_config(path: str) -> Config:
with open(path, "r") as f:
raw = yaml.safe_load(f)
env = EnvConfig(**raw.get("env", {}))
sensor = SensorConfig(**raw.get("sensor", {}))
controller = ControllerConfig(**raw.get("controller", {}))
model = ModelConfig(**raw.get("model", {}))
sim = SimConfig(**raw.get("sim", {}))
return Config(env=env, sensor=sensor, controller=controller, model=model, sim=sim)
