Detector / scripts /verify_dataset.py
Benxelua's picture
Add reproducible six-run detector-domain training package
4409fdb verified
Raw History Blame Contribute Delete
1.72 kB
#!/usr/bin/env python3
"""Fail-fast verification for the immutable paired bright/dark LOD protocol."""
from __future__ import annotations
import argparse, json
from pathlib import Path
import sys
sys.path.insert(0, str(Path(__file__).resolve().parent))
from detector_lib import EXPECTED, LockedLODDataset
def main():
p=argparse.ArgumentParser(); p.add_argument("--dataset-root",type=Path,required=True); p.add_argument("--labels-root",type=Path,default=Path(__file__).resolve().parents[1]/"labels"); p.add_argument("--package-root",type=Path,default=Path(__file__).resolve().parents[1]); a=p.parse_args()
report={"dataset_root":str(a.dataset_root),"splits":{},"status":"passed"}
ids={}
for domain in ("bright","dark"):
for split, expected in EXPECTED.items():
manifest=a.package_root/"manifests"/f"{domain}_{split}.json"; ds=LockedLODDataset(manifest,a.dataset_root,a.labels_root)
if len(ds)!=expected: raise RuntimeError((domain,split,len(ds),expected))
# Resolve one exemplar and validate labels/class IDs without copying data.
_, target=ds[0]
report["splits"][f"{domain}_{split}"]={"count":len(ds),"first_labels":target["labels"].tolist()}
ids.setdefault(split,{})[domain]=[x["pair_id"] for x in ds.items]
for split, pair in ids.items():
if pair["bright"] != pair["dark"]: raise RuntimeError(f"bright/dark pairing mismatch for {split}")
train=set(ids["train"]["bright"]); val=set(ids["val"]["bright"]); test=set(ids["test"]["bright"])
if train & val or train & test or val & test: raise RuntimeError("locked split overlap")
print(json.dumps(report,indent=2))
if __name__=="__main__": main()