Create index mapping for dataset with large features

What is the mapping you'd like to have with those documents?