Create scorer.py

scorer.py

@@ -0,0 +1,53 @@
+from dataclasses import dataclass
+from typing import Dict, Any, List
+import re
+
+REQ = [
+    "selected_policy_id",
+    "policy_mode",
+    "predicted_coherence_trajectory",
+    "intervention_intensity",
+    "communication_strategy",
+    "policy_switch_trigger",
+]
+
+POLICY = ["p1", "p2", "p3"]
+INTENSITY = ["low", "medium", "high"]
+
+@dataclass
+class ScoreResult:
+    score: float
+    details: Dict[str, Any]
+
+def _traj_ok(p: str):
+    return "->" in p and re.search(r"\b0\.\d+\s*->\s*0\.\d+", p) is not None
+
+def score(sample: Dict[str, Any], prediction: str) -> ScoreResult:
+    p = (prediction or "").lower()
+    words_ok = len(p.split()) <= 900
+
+    hits = sum(1 for k in REQ if k in p)
+
+    policy_ok = int("selected_policy_id" in p and any(x in p for x in POLICY))
+    mode_ok = int("policy_mode" in p)
+    traj_ok = int("predicted_coherence_trajectory" in p and _traj_ok(p))
+    inten_ok = int("intervention_intensity" in p and any(i in p for i in INTENSITY))
+    comm_ok = int("communication_strategy" in p and len(p) > 80)
+    trig_ok = int("policy_switch_trigger" in p and len(p) > 100)
+
+    raw = (
+        0.15 * int(words_ok) +
+        0.35 * (hits / len(REQ)) +
+        0.15 * policy_ok +
+        0.10 * traj_ok +
+        0.10 * inten_ok +
+        0.075 * comm_ok +
+        0.075 * trig_ok
+    )
+
+    return ScoreResult(score=min(1.0, raw), details={"id": sample.get("id"), "hits": hits})
+
+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)}