Put some error message when the year_list exceeds the size of the file we load so that it is clear what is happening

[georgemilosh]

Command:

python hyperparameter_optimization.py study=test n_trials=6 year_list="range(200)"

Error message:

n on folder = './11--ResNet__True--batch_normalizations__True--batch_size__38--conv_channel_first__32--conv_dropouts__0.32--conv_l2coef__47235...' failed due to AttributeError("'NoneType' object has no attribute 'shape'")
Traceback (most recent call last):
  File "/ClimateDynamics/MediumSpace/ClimateLearningFR/gmiloshe/CESM/models/hypop_France_CNN_deep/t3-test/Learn2_new.py", line 3040, in run
    score, info = k_fold_cross_val(folder, self.X, self.Y, **k_fold_cross_val_kwargs)
  File "/ClimateDynamics/MediumSpace/ClimateLearningFR/gmiloshe/CESM/models/hypop_France_CNN_deep/t3-test/ERA/utilities.py", line 234, in wrapper
    r = func(*args, **kwargs)
  File "/ClimateDynamics/MediumSpace/ClimateLearningFR/gmiloshe/CESM/models/hypop_France_CNN_deep/t3-test/ERA/utilities.py", line 196, in wrapper_inner
    r = func(*args, **kwargs)
  File "/ClimateDynamics/MediumSpace/ClimateLearningFR/gmiloshe/CESM/models/hypop_France_CNN_deep/t3-test/Learn2_new.py", line 2252, in k_fold_cross_val
    X_tr, Y_tr, X_va, Y_va = k_fold_cross_val_split(i, X, Y, nfolds=nfolds, val_folds=val_folds)
  File "/ClimateDynamics/MediumSpace/ClimateLearningFR/gmiloshe/CESM/models/hypop_France_CNN_deep/t3-test/ERA/utilities.py", line 234, in wrapper
    r = func(*args, **kwargs)
  File "/ClimateDynamics/MediumSpace/ClimateLearningFR/gmiloshe/CESM/models/hypop_France_CNN_deep/t3-test/ERA/utilities.py", line 196, in wrapper_inner
    r = func(*args, **kwargs)
  File "/ClimateDynamics/MediumSpace/ClimateLearningFR/gmiloshe/CESM/models/hypop_France_CNN_deep/t3-test/Learn2_new.py", line 1963, in k_fold_cross_val_split
    fold_len = arr[0].shape[0]//nfolds
AttributeError: 'NoneType' object has no attribute 'shape'

[I 2023-02-22 18:50:25,285] Trial 5 pruned. Run failed: pruning.
Completed 6 runs, 6 of which failed due to invalid network architecture

[AlessandroLovo]

This should not have happened due to year list problems:
The Plasim_Field.select_years function checks for invalid years

@ut.exec_time(logger)
    def select_years(self, year_list=None):
        '''
        Select a subset of years

        Parameters
        ----------
        year_list : array-like, optional
            list of the years to keep, by default None
        '''
        # check if the given year list is within the range of the data
        if year_list is not None:
            invalid_years = set(year_list) - set(self.field.time.dt.year.data)
            if invalid_years:
                raise IndexError(f'Data year range is {self.year_range} which does not include {pretty_set_of_int(invalid_years)}')
            self.field = self.field.sel(time=self.field.time.dt.year.isin(year_list))
            self.years = len(year_list)

So I think the error was somewhere else