tinymind-native-8b-remote-handoff/bundle/data/distill/scripts/audit_apexdistill.py

#!/usr/bin/env python
from __future__ import annotations

import json
from collections import Counter
from pathlib import Path


ROOT = Path(r"D:\ad\tinymind\data\distill")
JSONL = ROOT / "jsonl" / "apexdistill_gold.jsonl"
OUT = ROOT / "manifests" / "distill_dimension_audit.json"


def main() -> int:
    counts = {
        "records": 0,
        "domains": Counter(),
        "axes": Counter(),
        "lenses": Counter(),
        "pressures": Counter(),
        "rubric_fields": Counter(),
        "avg_quality": 0.0,
    }
    total_quality = 0.0
    with JSONL.open("r", encoding="utf-8") as f:
        for line in f:
            if not line.strip():
                continue
            item = json.loads(line)
            counts["records"] += 1
            counts["domains"][item["domain"]] += 1
            inputs = item.get("inputs", {})
            counts["axes"][inputs.get("distill_axis", "base")] += 1
            counts["lenses"][inputs.get("lens", "base")] += 1
            counts["pressures"][inputs.get("pressure", "base")] += 1
            rubric = item.get("quality", {}).get("rubric", {})
            for key in rubric:
                counts["rubric_fields"][key] += 1
            total_quality += item.get("quality", {}).get("score", 0.0)

    result = {
        "records": counts["records"],
        "domains": dict(counts["domains"]),
        "axes": dict(counts["axes"]),
        "lenses": dict(counts["lenses"]),
        "pressures": dict(counts["pressures"]),
        "rubric_fields": dict(counts["rubric_fields"]),
        "avg_quality": round(total_quality / max(counts["records"], 1), 4),
        "jsonl_mb": round(JSONL.stat().st_size / 1024 / 1024, 3),
    }
    OUT.write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding="utf-8")
    print(json.dumps(result, indent=2, ensure_ascii=False))
    return 0


if __name__ == "__main__":
    raise SystemExit(main())