fix(benchmarks): accept reordered cubic residue coverage (#1422)

Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>

Author: kaoru0822-kitauji

benchmarks/datasets/mathematical-benchmarks-v1/cubic-image-classification/tests/Dockerfile

@@ -1,6 +1,6 @@
 FROM python:3.12-slim@sha256:57cd7c3a7a273101a6485ba99423ee568157882804b1124b4dd04266317710de
 RUN python -m pip install --no-cache-dir attrs==26.1.0 jsonschema==4.26.0 jsonschema-specifications==2025.9.1 referencing==0.37.0 rpds-py==2026.6.3 typing-extensions==4.16.0
-LABEL jacobian.checksum="842a6675b735e9db965a4673a5b8b642d171cc06580689905b2c9e7b8a711433"
+LABEL jacobian.checksum="64fae5a499672a82e654a58279f153e624a21180049647dc6742c18dddb97545"
 LABEL jacobian.task="jacobian/cubic-image-classification"
 COPY expected.json input.json test.sh verifier.py verifier_support.py public_contract.json /tests/
 COPY input.json /app/input.json

benchmarks/datasets/mathematical-benchmarks-v1/cubic-image-classification/tests/verifier.py

@@ -103,7 +103,11 @@ def _family(value: object) -> set[int] | None:
         return None
     residues = {(target[1] * (minimum + step) + target[0]) % 9 for step in range(9)}
     declared = value["covered_residues"]
-    if not _strict_int_list(declared) or declared != sorted(residues):
+    if (
+        not _strict_int_list(declared)
+        or len(declared) != len(residues)
+        or set(declared) != residues
+    ):
         return None
     return residues
 

benchmarks/validation/mathematical_benchmarks_v1/test_cubic_image_classification.py

@@ -0,0 +1,39 @@
+from __future__ import annotations
+
+import json
+from pathlib import Path
+
+from benchmarks.validation.mathematical_benchmarks_v1 import support
+
+TASK = "cubic-image-classification"
+
+
+def _case(tmp_path: Path):
+    return support._prepare_case(tmp_path, TASK, "computed")
+
+
+def _rewrite(app: Path, submission: dict) -> None:
+    support._bind_result_evidence(app, submission)
+    support._write_json(app / "submission.json", submission)
+
+
+def test_accepts_reordered_covered_residues(tmp_path: Path) -> None:
+    task, app, logs = _case(tmp_path)
+    submission = json.loads((app / "submission.json").read_text())
+    submission["result"]["families"][0]["covered_residues"].reverse()
+    _rewrite(app, submission)
+
+    accepted = support._run_verifier(task, app, logs)
+    assert accepted.details["correctness"] == 1.0
+    assert accepted.reward == 1.0
+
+
+def test_rejects_duplicate_covered_residue(tmp_path: Path) -> None:
+    task, app, logs = _case(tmp_path)
+    submission = json.loads((app / "submission.json").read_text())
+    submission["result"]["families"][0]["covered_residues"] = [1, 1, 4, 7]
+    _rewrite(app, submission)
+
+    rejected = support._run_verifier(task, app, logs)
+    assert rejected.details["correctness"] == 0.0
+    assert rejected.reward == 0.0