import os
from dataclasses import dataclass
from functools import lru_cache

import yaml

@dataclass
class GeneralConfig:
    run_mode: str 
    
@dataclass
class DataSetConfig:
    input_dir: str
    num_slices: int
    train_split: float
    seed: int
    val_split: float
    output_dir: str
    
@dataclass
class TrainingConfig:
    batch_size: int
    num_epochs: int
    learning_rate: float
    checkpoint_path: str
    use_gpu: bool

@dataclass
class InferenceConfig:
    model_path: str
    num_classes: int
    output_dir: str

@dataclass
class AppConfig:
    build_dir: str
    
class AppSettings:
    """Application settings, to be initialized from app.yaml configuration file."""

    def __init__(self, config_file: str):
        # Load the YAML configuration file
        with open(config_file, "r") as f:
            config_data = yaml.safe_load(f)

        # Parse the loaded YAML into dataclass objects
        self.general = GeneralConfig(**config_data["general"])
        self.dataset = DataSetConfig(**config_data["dataset"])
        self.training = TrainingConfig(**config_data["training"])
        self.inference = InferenceConfig(**config_data["inference"])
        self.app = AppConfig(**config_data["app"])

@lru_cache
def get_app_settings() -> AppSettings:
    """Return application configuration settings."""
    module_path = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
    config_file = os.path.join(module_path, "conf", "app.yaml")
    return AppSettings(config_file=config_file)


if __name__ == "__main__":
    s = get_app_settings()