| |
| """Validate JSONL provenance, counts, links and release hashes without networking.""" |
| import sys |
| sys.dont_write_bytecode=True |
| import hashlib,json,re |
| from pathlib import Path,PurePosixPath |
| from export_dataset import rows |
|
|
| ROOT=Path(__file__).resolve().parent.parent |
|
|
| def main(): |
| expected=rows();counts={} |
| for name,source_rows in expected.items(): |
| actual=[json.loads(line) for line in (ROOT/'data'/(name+'.jsonl')).read_text(encoding='utf-8').splitlines()] |
| assert actual==source_rows,name |
| assert len({r.get('case_id',r.get('claim_id',r.get('suite_id'))) for r in actual})==len(actual) |
| columns=set(actual[0]) |
| assert all(set(row)==columns for row in actual) |
| assert all({k:type(v).__name__ for k,v in row.items()}=={k:type(v).__name__ for k,v in actual[0].items()} for row in actual) |
| counts[name]=len(actual) |
| summary=json.loads((ROOT/'metadata'/'dataset_summary.json').read_text()) |
| assert all(summary[k]==v for k,v in counts.items()) |
| assert sum(x['assertions_count'] for x in expected['verification_suites'])==summary['assertions_total']==711899 |
| assert sum(x['placements_checked'] for x in expected['verification_suites'])==summary['separate_marker_placements']==192780 |
| for file in [ROOT/'README.md',ROOT/'docs'/'DATASET_GUIDE.md']: |
| for target in re.findall(r'\]\(([^)]+)\)',file.read_text(encoding='utf-8')): |
| if '://' not in target and not target.startswith('#'): |
| assert (file.parent/target.split('#')[0]).is_file(),target |
| paths=set() |
| for line in (ROOT/'SHA256SUMS.txt').read_text().splitlines(): |
| digest,name=line.split(' ',1);path=PurePosixPath(name) |
| assert not path.is_absolute() and '..' not in path.parts and name not in paths |
| assert (ROOT/name).is_file() and not (ROOT/name).is_symlink() |
| assert hashlib.sha256((ROOT/name).read_bytes()).hexdigest()==digest,name |
| paths.add(name) |
| actual_files={f.relative_to(ROOT).as_posix() for f in ROOT.rglob('*') if f.is_file()} |
| assert actual_files==paths|{'SHA256SUMS.txt'},actual_files-paths |
| print('JSONL provenance, uniform row schemas, counts, documentation links and checksums passed.') |
| print(json.dumps(summary,indent=2)) |
|
|
| if __name__=='__main__':main() |
|
|
