refactor: move optimize test files into tests/learning/

Move 12 optimize-related test files from tests/ root into tests/learning/
to match the module reorganization. All 305 tests pass from new locations.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

tests/evals/test_deepplanning.py

@@ -0,0 +1,87 @@
+"""Tests for DeepPlanning benchmark."""
+
+from unittest.mock import MagicMock
+
+from openjarvis.evals.core.types import EvalRecord
+from openjarvis.evals.datasets.deepplanning import DeepPlanningDataset
+from openjarvis.evals.scorers.deepplanning_scorer import DeepPlanningScorer
+
+
+def _mock_backend() -> MagicMock:
+    backend = MagicMock()
+    backend.generate.return_value = "CORRECT"
+    return backend
+
+
+class TestDeepPlanningDataset:
+    def test_instantiation(self) -> None:
+        ds = DeepPlanningDataset()
+        assert ds.dataset_id == "deepplanning"
+        assert ds.dataset_name == "DeepPlanning"
+
+    def test_has_required_methods(self) -> None:
+        ds = DeepPlanningDataset()
+        assert hasattr(ds, "load")
+        assert hasattr(ds, "iter_records")
+        assert hasattr(ds, "size")
+
+
+class TestDeepPlanningScorer:
+    def test_instantiation(self) -> None:
+        s = DeepPlanningScorer(_mock_backend(), "test-model")
+        assert s.scorer_id == "deepplanning"
+
+    def test_correct_plan(self) -> None:
+        s = DeepPlanningScorer(_mock_backend(), "test-model")
+        record = EvalRecord(
+            record_id="dp-1",
+            problem="## Task (travel)\nPlan a trip to Paris",
+            reference="Day 1: Flight to Paris...",
+            category="agentic",
+            subject="travel",
+            metadata={"task_type": "travel"},
+        )
+        is_correct, meta = s.score(record, "Day 1: Fly to Paris, visit Eiffel Tower")
+        assert is_correct is True
+        assert meta["match_type"] == "llm_judge"
+
+    def test_incorrect_plan(self) -> None:
+        backend = MagicMock()
+        backend.generate.return_value = "INCORRECT - Missing budget constraint"
+        s = DeepPlanningScorer(backend, "test-model")
+        record = EvalRecord(
+            record_id="dp-2",
+            problem="## Task (shopping)\nBuild a cart",
+            reference="Cart: item A, item B, total $50",
+            category="agentic",
+            subject="shopping",
+            metadata={"task_type": "shopping"},
+        )
+        is_correct, meta = s.score(record, "Cart: item C, total $100")
+        assert is_correct is False
+
+    def test_empty_response(self) -> None:
+        s = DeepPlanningScorer(_mock_backend(), "test-model")
+        record = EvalRecord(
+            record_id="dp-3", problem="task",
+            reference="answer", category="agentic",
+        )
+        is_correct, meta = s.score(record, "")
+        assert is_correct is False
+        assert meta["reason"] == "empty_response"
+
+
+class TestDeepPlanningCLI:
+    def test_in_benchmarks(self) -> None:
+        from openjarvis.evals.cli import BENCHMARKS
+        assert "deepplanning" in BENCHMARKS
+
+    def test_build_dataset(self) -> None:
+        from openjarvis.evals.cli import _build_dataset
+        ds = _build_dataset("deepplanning")
+        assert ds.dataset_id == "deepplanning"
+
+    def test_build_scorer(self) -> None:
+        from openjarvis.evals.cli import _build_scorer
+        s = _build_scorer("deepplanning", _mock_backend(), "test-model")
+        assert s.scorer_id == "deepplanning"

tests/evals/test_paperarena.py

@@ -0,0 +1,115 @@
+"""Tests for PaperArena benchmark."""
+
+from unittest.mock import MagicMock
+
+from openjarvis.evals.core.types import EvalRecord
+from openjarvis.evals.datasets.paperarena import PaperArenaDataset
+from openjarvis.evals.scorers.paperarena_judge import PaperArenaScorer
+
+
+def _mock_backend() -> MagicMock:
+    backend = MagicMock()
+    backend.generate.return_value = "CORRECT"
+    return backend
+
+
+class TestPaperArenaDataset:
+    def test_instantiation(self) -> None:
+        ds = PaperArenaDataset()
+        assert ds.dataset_id == "paperarena"
+        assert ds.dataset_name == "PaperArena"
+
+    def test_has_required_methods(self) -> None:
+        ds = PaperArenaDataset()
+        assert hasattr(ds, "load")
+        assert hasattr(ds, "iter_records")
+        assert hasattr(ds, "size")
+
+
+class TestPaperArenaScorer:
+    def test_instantiation(self) -> None:
+        s = PaperArenaScorer(_mock_backend(), "test-model")
+        assert s.scorer_id == "paperarena"
+
+    def test_mc_correct(self) -> None:
+        s = PaperArenaScorer(_mock_backend(), "test-model")
+        record = EvalRecord(
+            record_id="pa-1",
+            problem="## Question\nWhat is X?\nOptions:\n  A) foo\n  B) bar",
+            reference="A",
+            category="agentic",
+            subject="easy_mc",
+            metadata={"question_type": "MC"},
+        )
+        is_correct, meta = s.score(record, "The answer is A")
+        assert is_correct is True
+        assert meta["match_type"] == "exact_letter"
+
+    def test_mc_wrong(self) -> None:
+        s = PaperArenaScorer(_mock_backend(), "test-model")
+        record = EvalRecord(
+            record_id="pa-2",
+            problem="Question",
+            reference="B",
+            category="agentic",
+            subject="medium_mc",
+            metadata={"question_type": "MC"},
+        )
+        is_correct, meta = s.score(record, "The answer is C")
+        assert is_correct is False
+        assert meta["candidate_letter"] == "C"
+
+    def test_open_answer_judge(self) -> None:
+        s = PaperArenaScorer(_mock_backend(), "test-model")
+        record = EvalRecord(
+            record_id="pa-3",
+            problem="## Question\nExplain X",
+            reference="X is a method for...",
+            category="agentic",
+            subject="hard_oa",
+            metadata={"question_type": "OA"},
+        )
+        is_correct, meta = s.score(record, "X is a technique used to...")
+        assert is_correct is True
+        assert meta["match_type"] == "llm_judge"
+
+    def test_closed_answer_judge(self) -> None:
+        backend = MagicMock()
+        backend.generate.return_value = "INCORRECT"
+        s = PaperArenaScorer(backend, "test-model")
+        record = EvalRecord(
+            record_id="pa-4",
+            problem="## Question\nWhat is the value?",
+            reference="42.5",
+            category="agentic",
+            subject="medium_ca",
+            metadata={"question_type": "CA"},
+        )
+        is_correct, meta = s.score(record, "The value is 100")
+        assert is_correct is False
+
+    def test_empty_response(self) -> None:
+        s = PaperArenaScorer(_mock_backend(), "test-model")
+        record = EvalRecord(
+            record_id="pa-5", problem="q",
+            reference="a", category="agentic",
+        )
+        is_correct, meta = s.score(record, "")
+        assert is_correct is False
+        assert meta["reason"] == "empty_response"
+
+
+class TestPaperArenaCLI:
+    def test_in_benchmarks(self) -> None:
+        from openjarvis.evals.cli import BENCHMARKS
+        assert "paperarena" in BENCHMARKS
+
+    def test_build_dataset(self) -> None:
+        from openjarvis.evals.cli import _build_dataset
+        ds = _build_dataset("paperarena")
+        assert ds.dataset_id == "paperarena"
+
+    def test_build_scorer(self) -> None:
+        from openjarvis.evals.cli import _build_scorer
+        s = _build_scorer("paperarena", _mock_backend(), "test-model")
+        assert s.scorer_id == "paperarena"