Stone-Soup
Stone-Soup copied to clipboard
Have data readers store data
Some data readers build a dictionary when being iterated through. For example, the CSVGroundTruthReader
builds a groundtruth_dict
so that it can keep track of path ids that it has come across, and avoid yielding them in separate GroundTruthPath
.
In some situations, it might be useful to keep hold of these dictionaries (eg. for visualisation, evaluation of the entire dataset), perhaps just stored as an attribute of the reader. For large datasets this probably wouldn't be good, but could be an option that the user can (de-)select?