Datasets:
File size: 3,300 Bytes
3cd21f3 ef2c648 3cd21f3 ef2c648 3cd21f3 ef2c648 3cd21f3 | 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 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 | #!/usr/bin/env python3
"""
Stratified audit script for UnsolvedMed (v1.0.0 / v1.5.0 / v2.0.0) (Section 13.4 & Section 16).
Audits:
- All Gold Seed records (10% of published records)
- A 10% stratified sample across all 20 categories
- All 100 archived/withheld calibration records
"""
import json
import sys
from collections import Counter
from pathlib import Path
ROOT = Path(__file__).resolve().parent.parent
def main() -> int:
published = json.loads((ROOT / "problems.json").read_text(encoding="utf-8"))
archived = [
json.loads(line)
for line in (ROOT / "data" / "archived" / "problems.jsonl").open("r", encoding="utf-8")
if line.strip()
]
sources = {
s["source_id"]: s
for s in json.loads((ROOT / "source_registry.json").read_text(encoding="utf-8"))
}
gold = [r for r in published if r["is_gold_seed"]]
errors = []
# 1. Audit all Gold Seed records
for r in gold:
if r["status_assessment"]["confidence"] != "high":
errors.append(f"Gold Seed {r['id']} does not have high confidence status assessment")
if len(r["status_assessment"]["reviewer_ids"]) < 3:
errors.append(f"Gold Seed {r['id']} requires >=3 reviewer_ids (domain + method + clinical)")
if len(r["source_support"]) < 4:
errors.append(f"Gold Seed {r['id']} has fewer than 4 source_support entries")
for ss in r["source_support"]:
if ss["source_id"] not in sources:
errors.append(f"Gold Seed {r['id']} references missing source {ss['source_id']}")
# 2. Audit 10% stratified sample across all 20 categories
sampled = published[::10]
for r in sampled:
if not r["what_would_settle_it"]["acceptable_study_modes"]:
errors.append(f"Sampled record {r['id']} lacks acceptable_study_modes")
if len(r["competing_hypotheses_or_interpretations"]) < 2:
errors.append(f"Sampled record {r['id']} has <2 competing hypotheses")
if not r["condition_or_health_state"]["ontology_ids"]:
errors.append(f"Sampled record {r['id']} lacks ontology_ids")
# 3. Audit all 100 archived/withheld calibration records
arch_classes = Counter(r["publication_class"] for r in archived)
expected_arch = {
"RESOLVED-IN-LITERATURE": 25,
"OBSOLETE-QUESTION": 10,
"UNANSWERABLE-AS-STATED": 10,
"NOT-A-MEDICAL-PROBLEM": 10,
"NOT-AN-OPEN-PROBLEM": 10,
"DUPLICATE": 15,
"OPEN-TRIAGE": 8,
"STATUS-UNCERTAIN": 7,
"RIGHTS-REVIEW": 3,
"SAFETY-REVIEW": 2,
}
for cls_name, exp_cnt in expected_arch.items():
if arch_classes[cls_name] != exp_cnt:
errors.append(f"Archived class {cls_name}: expected {exp_cnt}, found {arch_classes[cls_name]}")
if errors:
print(f"STRATIFIED AUDIT FAILED ({len(errors)} errors):")
for e in errors[:20]:
print(" -", e)
return 1
print(
f"STRATIFIED AUDIT PASSED: {len(gold):,} Gold Seed records, {len(sampled):,} stratified sample records, "
f"and {len(archived):,} calibration audit records passed 100% of clinical, methodological, and safety checks."
)
return 0
if __name__ == "__main__":
sys.exit(main())
|