"""Phase 1: Raw Data Audit — source integrity, schema, statistics. Critical check: every source entry has all required fields. """ import json, os, time from pathlib import Path from collections import Counter, defaultdict OUT = Path("scripts/audit_reports") OUT.mkdir(parents=True, exist_ok=True) DATASET = "dataset/entries_final_v3.json" AUDIT_DIR = Path.cwd() if Path.cwd().name == "Scandium-Dataset" else Path("/home/shamique/Scandium Labs SSB/Scandium-Dataset") REQUIRED_FIELDS = [ "formula", "source", "source_id", "formation_energy_per_atom", "elements", "nsites", "structure_json", ] REQUIRED_STRUCTURE_KEYS = ["lattice", "sites", "@module", "@class"] severity_counts = {"CRITICAL": 0, "HIGH": 0, "MEDIUM": 0, "LOW": 0, "PASS": 0} findings = [] def finding(severity, phase, check, status, detail): severity_counts[severity] += 1 findings.append({ "severity": severity, "phase": phase, "check": check, "status": status, "detail": detail, "timestamp": time.strftime("%Y-%m-%d %H:%M:%S") }) s = "🔴" if severity == "CRITICAL" else "🟠" if severity == "HIGH" else "🟡" if severity == "MEDIUM" else "🔵" if severity == "LOW" else "✅" print(f" {s} [{severity:8s}] {check}: {detail[:120]}") def main(): print("=" * 60) print(" PHASE 1: RAW DATA AUDIT") print("=" * 60) with open(AUDIT_DIR / DATASET) as f: entries = json.load(f) N = len(entries) print(f"\n Loaded {N:,} entries") # 1. Source integrity print("\n--- Source Integrity ---") sources = Counter(e.get("source", "unknown") for e in entries) for src, cnt in sources.most_common(): finding("PASS", "source_integrity", f"source_{src}", f"{cnt:,} entries", "") # Check for unknown sources unknown = [e for e in entries if e.get("source") not in ("mp", "oqmd", "jarvis")] if unknown: finding("CRITICAL", "source_integrity", "unknown_sources", f"{len(unknown)} entries", f"sources: {set(e.get('source') for e in unknown)}") else: finding("PASS", "source_integrity", "all_sources_known", "3 valid sources", "") # 2. Schema consistency print("\n--- Schema Consistency ---") missing_fields = defaultdict(set) wrong_types = [] corrupted_json = [] duplicate_ids = defaultdict(list) for i, e in enumerate(entries): for field in REQUIRED_FIELDS: if e.get(field) is None: src = e.get("source", "?") missing_fields[field].add(src) # Check structure_json parseable sj = e.get("structure_json") if sj: if isinstance(sj, str): try: sd = json.loads(sj) if not all(k in sd for k in REQUIRED_STRUCTURE_KEYS): corrupted_json.append((i, "missing_keys")) except json.JSONDecodeError: corrupted_json.append((i, "parse_error")) elif isinstance(sj, dict): if not all(k in sj for k in REQUIRED_STRUCTURE_KEYS): corrupted_json.append((i, "missing_keys_dict")) # Check duplicate IDs per source sid = e.get("source_id") src = e.get("source") if sid and src: duplicate_ids[(src, sid)].append(i) # Report missing fields for field, srcs in missing_fields.items(): finding("HIGH" if field in ("formula", "source", "structure_json") else "MEDIUM", "schema", f"missing_field_{field}", f"missing in {', '.join(sorted(srcs))}", f"{field} should never be None") # Report corrupted JSON if corrupted_json: finding("CRITICAL", "schema", "corrupted_structure_json", f"{len(corrupted_json)} entries", "") else: finding("PASS", "schema", "structure_json_valid", "all parseable", "") # Report duplicate IDs dup_ids_found = {k: v for k, v in duplicate_ids.items() if len(v) > 1} if dup_ids_found: finding("HIGH", "schema", "duplicate_source_ids", f"{len(dup_ids_found)} groups", "same (source, source_id) pairs exist — potential dedup gap") else: finding("PASS", "schema", "no_duplicate_ids", "all source IDs unique", "") # 3. Source statistics print("\n--- Source Statistics ---") for src in ["mp", "oqmd", "jarvis"]: subset = [e for e in entries if e.get("source") == src] print(f"\n {src.upper()}: {len(subset):,} entries") # Missing labels for prop in ["formation_energy_per_atom", "energy_above_hull", "band_gap", "space_group", "volume", "density"]: missing = sum(1 for e in subset if e.get(prop) is None) if missing > 0: pct = 100 * missing / len(subset) sev = "CRITICAL" if pct > 50 else "HIGH" if pct > 10 else "MEDIUM" if pct > 0 else "PASS" finding(sev, "source_stats", f"{src}_{prop}_missing", f"{missing:,} / {len(subset):,} ({pct:.1f}%)", "") else: finding("PASS", "source_stats", f"{src}_{prop}_present", f"0 missing (100% coverage)", "") # 4. License compatibility check print("\n--- License Compatibility ---") oqmd = [e for e in entries if e.get("source") == "oqmd"] finding("PASS", "license", "oqmd_license", f"{len(oqmd):,} entries: non-commercial use OK", "OQMD allows non-commercial use with attribution") finding("PASS", "license", "mp_license", f"{sources.get('mp', 0):,} entries: CC BY 4.0", "MP requires attribution") finding("PASS", "license", "jarvis_license", f"{sources.get('jarvis', 0):,} entries: CC0", "No restrictions") # Summary print(f"\n{'=' * 60}") print(f" PHASE 1 SUMMARY") print(f" CRITICAL: {severity_counts['CRITICAL']}") print(f" HIGH: {severity_counts['HIGH']}") print(f" MEDIUM: {severity_counts['MEDIUM']}") print(f" LOW: {severity_counts['LOW']}") print(f" PASS: {severity_counts['PASS']}") print(f"{'=' * 60}") report = { "phase": "Phase 1: Raw Data Audit", "timestamp": time.strftime("%Y-%m-%d %H:%M:%S"), "total_entries": N, "sources": dict(sources.most_common()), "findings": findings, "summary": dict(severity_counts), } with open(OUT / "phase1_raw_data_audit.json", "w") as f: json.dump(report, f, indent=2) print(f"\n Report: {OUT / 'phase1_raw_data_audit.json'}") if __name__ == "__main__": main()