Create scorer.py

scorer.py

@@ -0,0 +1,56 @@
+import re
+from dataclasses import dataclass
+from typing import Dict, Any, List
+
+ACTIONS = {
+    "stay_course",
+    "adjust_dose",
+    "switch_class",
+    "pause_observe",
+    "add_support_module",
+    "deescalate_exit_loop",
+    "urgent_escalation",
+}
+
+LABELS = {"coherent_navigation","partially_coherent","incoherent_navigation"}
+
+@dataclass
+class ScoreResult:
+    score: float
+    details: Dict[str, Any]
+
+def _has(t: str, pats: List[str]) -> bool:
+    t = (t or "").lower()
+    return any(re.search(p, t) for p in pats)
+
+def score(sample: Dict[str, Any], prediction: str) -> ScoreResult:
+    p = (prediction or "").lower().strip()
+    words_ok = len(p.split()) <= 380
+
+    action_ok = any(a in p for a in ACTIONS)
+    label_ok = any(l in p for l in LABELS)
+    signature_ref = _has(p, [r"slope", r"tolerance", r"tradeoff", r"coupling", r"volatility", r"plateau", r"discord"])
+    constraint_ref = _has(p, [r"avoid", r"contra", r"risk", r"cannot", r"must not"])
+
+    raw = (
+        0.20 * int(words_ok) +
+        0.35 * int(action_ok) +
+        0.35 * int(label_ok) +
+        0.05 * int(signature_ref) +
+        0.05 * int(constraint_ref)
+    )
+    final = max(0.0, min(1.0, raw))
+
+    return ScoreResult(
+        score=final,
+        details={
+            "action_ok": action_ok,
+            "label_ok": label_ok,
+            "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)}