Create scorer.py

scorer.py

@@ -0,0 +1,48 @@
+import re
+from dataclasses import dataclass
+from typing import Dict, Any, List
+
+BASINS = {
+    "basin_a_inflammatory_autonomic",
+    "basin_b_mito_metabolic_fatigue",
+    "basin_c_neuroimmune_cognitive",
+    "basin_d_mast_cell_histamine_like",
+    "basin_e_autoimmune_multisystem",
+}
+
+@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 _extract_ints(text: str) -> List[int]:
+    return [int(x) for x in re.findall(r"\b\d{1,3}\b", text) if 0 <= int(x) <= 100]
+
+def score(sample: Dict[str, Any], prediction: str) -> ScoreResult:
+    p = (prediction or "").lower().strip()
+    words_ok = len(p.split()) <= 360
+
+    basin_ok = any(b in p for b in BASINS)
+    nums = _extract_ints(p)
+    has_score = any(0 <= n <= 100 for n in nums)
+
+    features_ref = _has(p, [r"top5", r"features", r"pem", r"pots", r"il6", r"ana", r"tryptase", r"esr"])
+    evidence_ref = _has(p, [r"because", r"domin", r"pattern", r"defines", r"stable"])
+
+    raw = (
+        0.30 * int(basin_ok) +
+        0.20 * int(has_score) +
+        0.20 * int(features_ref) +
+        0.20 * int(evidence_ref) +
+        0.10 * int(words_ok)
+    )
+    return ScoreResult(score=min(1.0, raw), details={"basin_ok": basin_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)}