Re-upload CSV splits from splits_v2 (train/val/test/leaky_validation)

Replace old *_data.csv with new splits (train/validation/leaky/test)

New split based on week

update datafiles and readme

upload dataset and dataset card