MinMaxScaler

class MinMaxScaler(Scaler):
    def __init__(self):
        super().__init__()

    def apply(self, dataset: Dataset) -> Dataset:
        data = dataset.data.copy()
        numeric_data = data.select_dtypes(include="number")
        if not numeric_data.empty:
            self.min_values = numeric_data.min()
            self.max_values = numeric_data.max()
            data[numeric_data.columns] = (numeric_data - self.min_values) / (
                self.max_values - self.min_values
            )

        array_data = data.select_dtypes(include=[list, np.ndarray])
        skip_col = False
        for col in array_data.columns:
            scaled_column = []
            for x in data[col].tolist():
                if (not self.is_numeric_array(x) and col != "data") or col in ["target", "title"]: 
                    skip_col = True
                    break
                x_array = np.array(x)

                if x_array.ndim == 1:
                    min_val = x_array.min()
                    max_val = x_array.max()
                    scaled_array = (
                        (x_array - min_val) / (max_val - min_val)
                        if max_val != min_val
                        else x_array
                    )

                elif x_array.ndim == 2:
                    min_vals = x_array.min(axis=0)
                    max_vals = x_array.max(axis=0)
                    scaled_array = (
                        (x_array - min_vals) / (max_vals - min_vals)
                        if not np.all(max_vals == min_vals)
                        else x_array
                    )

                else:
                    raise ValueError(
                        f"Unsupported array shape {x_array.shape} in column {col}"
                    )

                scaled_column.append(scaled_array)

            if skip_col: continue
            data[col] = scaled_column

        dataset.data = data
        return dataset
    
    def is_numeric_array(self, data):
        """Checks if the data is a numeric type or a list/array of numeric types.

        Args:
            data: The data to check.

        Returns:
            bool: True if the data is numeric, False otherwise.
        """
        if isinstance(data, (int, float)):
            return True
        elif isinstance(data, (list, np.ndarray)):
            return all(isinstance(elem, (int, float)) for elem in data)
        else:
            return False