scorer.py

import re
from dataclasses import dataclass
from typing import Dict, Any, List

BASINS = {
    "improved_stable_recovery",
    "delayed_recovery",
    "complication_basin",
    "iatrogenic_worsening",
    "unstable_oscillation",
    "no_material_change",
}

@dataclass
class ScoreResult:
    score: float
    details: Dict[str, Any]

def score(sample: Dict[str, Any], prediction: str) -> ScoreResult:
    p = (prediction or "").lower().strip()
    words_ok = len(p.split()) <= 360

    seq_ok = any(k in p for k in ["t+","day","week"]) and any(ch in p for ch in [":", ";"])
    cross_ok = any(k in p for k in ["immune", "renal", "cardio", "neuro", "resp", "gi", "autonomic", "metabolic", "subjective"])
    div_ok = any(k in p for k in ["time_to_divergence", "divergence", "t+"]) and bool(re.search(r"\b\d+\s*(h|hour|day|week)s?\b", p))
    basin_ok = any(b in p for b in BASINS)

    raw = (
        0.20 * int(words_ok) +
        0.30 * int(seq_ok) +
        0.25 * int(cross_ok) +
        0.15 * int(div_ok) +
        0.10 * int(basin_ok)
    )
    return ScoreResult(score=min(1.0, raw), details={"id": sample.get("id"), "seq_ok": seq_ok, "basin_ok": basin_ok})

def aggregate(results: List[ScoreResult]) -> Dict[str, Any]:
    if not results:
        return {"mean": 0.0, "n": 0}
    return {"mean": sum(r.score for r in results) / len(results), "n": len(results)}