Datasets:
File size: 2,290 Bytes
69c079d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 | """Validate manifest coverage, labels, folds, HDF5 schema, and file checksums."""
import csv
import hashlib
from collections import Counter
from pathlib import Path
import h5py
def main():
root = Path(__file__).resolve().parents[1]
with (root / 'manifest.csv').open(newline='') as f:
rows = list(csv.DictReader(f))
paths = [r['path'] for r in rows]
assert len(rows) == 38180 and len(set(paths)) == len(rows)
assert len({r['sample_id'] for r in rows}) == len(rows)
assert set(paths) == {str(p.relative_to(root)) for p in (root / 'data').rglob('*.h5')}
counts = Counter()
for row in rows:
path = Path(row['path'])
assert not path.is_absolute() and '..' not in path.parts
assert row['label'] == {'negative': '0', 'positive': '1'}[row['label_name']]
assert row['label_name'] in path.name
counts[row['source'], row['split'], row['fold'], row['label_name']] += 1
filename = root / path
assert hashlib.sha256(filename.read_bytes()).hexdigest() == row['sha256'], path
with h5py.File(filename) as f:
d = f[row['h5_key']]
assert d.shape == tuple(int(row[k]) for k in ('height', 'width', 'bins')), path
assert str(d.dtype) == row['dtype'], path
assert Counter(r['split'] for r in rows) == {
'train': 26600, 'validation': 5700, 'test': 5700, 'unassigned': 180}
for i in range(1, 10):
for label in ('negative', 'positive'):
assert counts['experimental', 'unassigned', f'fold_{i}', label] == 10
for split in ('train', 'validation', 'test'):
with (root / f'metadata/synthetic_{split}.csv').open(newline='') as f:
assert list(csv.DictReader(f)) == [r for r in rows if r['split'] == split]
expected = [r for r in rows if r['source'] == 'experimental']
for name in ('metadata/experimental.csv', 'data/experiment_data/manifest.csv'):
with (root / name).open(newline='') as f:
assert list(csv.DictReader(f)) == expected
assert (root / 'SHA256SUMS').read_text() == ''.join(
f"{r['sha256']} {r['path']}\n" for r in rows)
print('PASS: 38,180 files, hashes, schemas, portable paths, split manifests, and nine balanced folds')
if __name__ == '__main__':
main()
|