(self, in_memory, tmp_dir, *datasets)
| 196 | return dset |
| 197 | |
| 198 | def _to(self, in_memory, tmp_dir, *datasets): |
| 199 | if in_memory: |
| 200 | datasets = [dataset.map(keep_in_memory=True) for dataset in datasets] |
| 201 | else: |
| 202 | start = 0 |
| 203 | while os.path.isfile(os.path.join(tmp_dir, f"dataset{start}.arrow")): |
| 204 | start += 1 |
| 205 | datasets = [ |
| 206 | dataset.map(cache_file_name=os.path.join(tmp_dir, f"dataset{start + i}.arrow")) |
| 207 | for i, dataset in enumerate(datasets) |
| 208 | ] |
| 209 | return datasets if len(datasets) > 1 else datasets[0] |
| 210 | |
| 211 | def test_dummy_dataset(self, in_memory): |
| 212 | with tempfile.TemporaryDirectory() as tmp_dir: |
no test coverage detected