scorer.py

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

GRADIENTS = {"flat", "shallow", "moderate", "moderate_to_steep", "steep", "flat_to_positive"}

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

def score(sample: Dict[str, Any], prediction: str) -> ScoreResult:
    p = (prediction or "").lower()
    words_ok = len(p.split()) <= 800

    has_traj = "trajectory" in p or "stressed" in p or "decoupling" in p or "collapse" in p
    has_gradient = any(g in p for g in GRADIENTS) or "gradient" in p
    has_levers = ">" in p or "rank" in p or "leverage" in p
    has_gain = "coherence" in p and ("gain" in p or "medium" in p or "high" in p or "low" in p)
    has_tradeoffs = "tradeoff" in p or "cost" in p or "risk" in p
    has_monitor = "monitor" in p or "indicator" in p or "metric" in p

    raw = (
        0.15 * int(words_ok) +
        0.20 * int(has_traj) +
        0.20 * int(has_gradient) +
        0.20 * int(has_levers) +
        0.10 * int(has_gain) +
        0.10 * int(has_tradeoffs) +
        0.05 * int(has_monitor)
    )
    return ScoreResult(score=min(1.0, raw), details={"id": sample.get("id")})

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)}