Adapt evaluate_metrics and create_tabular for datasets that have a ground truth

In evaluate_metrics:

• Include the computation of special metrics

In create_tabular:

• Allow for the read out of partially existent metrics
• Add an option to display either all datasets (with empty spaces) or only those were all metrics exist
• Add a flag for the latter

changed title from compare_ground_truth.py for datasets that have a ground truth to Adapt evaluate_metrics and create_tabular for datasets that have a ground truth

changed the description

added priority label

closed

Fixed in e9b0d362. The flag idea was omitted