Datasets:
| from pathlib import Path | |
| import gzip,json | |
| SPLITS={'train','validation','id','long','counterfactual'} | |
| def iter_rows(split:str): | |
| if split not in SPLITS:raise ValueError('Unknown source split') | |
| with gzip.open(Path(__file__).resolve().parent/'data'/f'{split}.jsonl.gz','rt',encoding='utf-8') as f: | |
| for line in f:yield json.loads(line) | |