Datasets:
File size: 3,341 Bytes
84a242a | 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 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 | #!/usr/bin/env python3
"""Audit existing exports. This does not run acquisition or regenerate results."""
import argparse
import csv
import json
from release_support import add_path_arguments
from run_experiment import METHODS
def audit_incremental(run_root, checkpoints=100):
"""Detect repeated prior blocks, including the historical resume-export bug."""
seen = set()
next_order = 1
issues = []
for percentage in range(1, checkpoints + 1):
path = run_root / "queried_cells/newly_queried" / f"queried_{percentage:03d}pct.csv"
if not path.is_file():
issues.append(f"Missing incremental checkpoint {percentage}")
continue
with path.open(newline="") as handle:
for row in csv.DictReader(handle):
order, flat = int(row["query_order"]), int(row["flat_index"])
if order != next_order:
issues.append(f"Checkpoint {percentage}: expected query order {next_order}, found {order}")
if flat in seen:
issues.append(f"Checkpoint {percentage}: repeated flat index {flat}")
seen.add(flat)
next_order = order + 1
return {"unique_cells": len(seen), "issues": issues}
def main():
parser = argparse.ArgumentParser(description=__doc__)
add_path_arguments(parser, inputs=False)
parser.add_argument("--methods", nargs="+", choices=METHODS, default=list(METHODS))
parser.add_argument("--folds", nargs="+", type=int, choices=range(5), default=list(range(5)))
parser.add_argument("--seeds", nargs="+", type=int, default=[0, 1, 2])
parser.add_argument("--smoke-test", action="store_true")
parser.add_argument("--check-incremental", action="store_true",
help="Read incremental logs to check contiguous query orders and uniqueness across checkpoints.")
args = parser.parse_args()
records = []
for method in args.methods:
for fold in args.folds:
for seed in args.seeds:
root = args.output_dir / ("smoke" if args.smoke_test else "runs") / method / f"fold{fold}" / f"seed{seed}"
counts = {name: len(list((root / subpath).glob("*.csv"))) for name, subpath in {
"matrices": "strict_score_matrices", "cumulative_logs": "queried_cells/cumulative",
"incremental_logs": "queried_cells/newly_queried"}.items()}
complete = (root / "run_state/complete.json").is_file()
item = {"method": method, "fold": fold, "seed": seed, **counts, "complete_marker": complete,
"pass": complete and all(count == 100 for count in counts.values())}
if args.check_incremental:
item["incremental_audit"] = audit_incremental(root)
item["pass"] = item["pass"] and not item["incremental_audit"]["issues"]
records.append(item)
report = {"scope": "file counts and incremental order/uniqueness" if args.check_incremental else "file counts only",
"expected_runs": len(records), "passing_runs": sum(x["pass"] for x in records), "runs": records}
print(json.dumps(report, indent=2))
if not all(x["pass"] for x in records):
raise SystemExit(1)
if __name__ == "__main__":
main()
|