Configuration discrimination fails if Pydantic algorithm models are passed

[jdeschamps]

Describe the bug

Since a recent PR, passing a configuration as dictionary, albeit with a an algorithm Pydantic model (rather than a dict itself ), will cause an error:

value = {'algorithm_config': N2VAlgorithm(algorithm='n2v', loss='n2v', model=UNetModel(architecture='UNet', conv_dims=2, num_c...e=True, stopping_threshold=None, divergence_threshold=None, check_on_train_epoch_end=False, log_rank_zero_only=False))}

    def _algorithm_config_discriminator(value: Union[dict, Configuration]) -> str:
        """Discriminate algorithm-specific configurations based on the algorithm.
    
        Parameters
        ----------
        value : Any
            Value to discriminate.
    
        Returns
        -------
        str
            Discriminator value.
        """
        if isinstance(value, dict):
>           return value["algorithm_config"]["algorithm"]
E           TypeError: 'N2VAlgorithm' object is not subscriptable

Indeed, we only check for the overall object being a dict, not the algorithm_config. However, it is perfectly legal to pass it as N2VAlgorithmConfig.

This will be superseded and solved by #367.