Skip to content

~95% on SimpleQA (e.g. Qwen3.6-27B on a 3090). Supports all local and cloud LLMs (llama.cpp, Ollama, Google, ...). 10+ search engines - arXiv, PubMed, your private documents. Everything Local & Encrypted.

34.4 KB
Raw Blame
"""
Tests for benchmarks/optimization/optuna_optimizer.py

Tests cover:
- OptunaOptimizer initialization
- Default parameter space
- Weight normalization
- Convenience optimization functions
"""

from unittest.mock import Mock, patch
import pytest


class TestOptunaOptimizerInit:
    """Tests for OptunaOptimizer initialization."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_init_with_defaults(self, mock_evaluator):
        """Test initialization with default values."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(
            base_query="test query",
            output_dir="/tmp/test_output",
        )

        assert optimizer.base_query == "test query"
        assert optimizer.output_dir == "/tmp/test_output"
        assert optimizer.n_trials == 30
        assert optimizer.n_jobs == 1
        assert optimizer.temperature == 0.7

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_init_with_custom_values(self, mock_evaluator):
        """Test initialization with custom values."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(
            base_query="test query",
            output_dir="/tmp/test",
            model_name="custom-model",
            provider="openai",
            search_tool="google",
            temperature=0.5,
            n_trials=50,
            n_jobs=4,
        )

        assert optimizer.model_name == "custom-model"
        assert optimizer.provider == "openai"
        assert optimizer.search_tool == "google"
        assert optimizer.temperature == 0.5
        assert optimizer.n_trials == 50
        assert optimizer.n_jobs == 4

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_init_normalizes_weights(self, mock_evaluator):
        """Test that metric weights are normalized."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(
            base_query="test",
            metric_weights={"quality": 2, "speed": 2},
        )

        # Weights should be normalized to sum to 1
        total = sum(optimizer.metric_weights.values())
        assert total == pytest.approx(1.0)
        assert optimizer.metric_weights["quality"] == pytest.approx(0.5)
        assert optimizer.metric_weights["speed"] == pytest.approx(0.5)

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_init_default_benchmark_weights(self, mock_evaluator):
        """Test default benchmark weights."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        assert "simpleqa" in optimizer.benchmark_weights
        assert optimizer.benchmark_weights["simpleqa"] == 1.0

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_init_custom_benchmark_weights(self, mock_evaluator):
        """Test custom benchmark weights."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(
            base_query="test",
            benchmark_weights={"simpleqa": 0.6, "browsecomp": 0.4},
        )

        assert optimizer.benchmark_weights["simpleqa"] == 0.6
        assert optimizer.benchmark_weights["browsecomp"] == 0.4

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_init_generates_study_name(self, mock_evaluator):
        """Test that study name is generated if not provided."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        assert optimizer.study_name.startswith("ldr_opt_")

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_init_uses_custom_study_name(self, mock_evaluator):
        """Test that custom study name is used."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(
            base_query="test",
            study_name="my_custom_study",
        )

        assert optimizer.study_name == "my_custom_study"


class TestDefaultParamSpace:
    """Tests for default parameter space."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_get_default_param_space(self, mock_evaluator):
        """Test getting default parameter space."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")
        param_space = optimizer._get_default_param_space()

        assert "iterations" in param_space
        assert "questions_per_iteration" in param_space
        assert "search_strategy" in param_space
        assert "max_results" in param_space

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_iterations_param_space(self, mock_evaluator):
        """Test iterations parameter space."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")
        param_space = optimizer._get_default_param_space()

        iterations = param_space["iterations"]
        assert iterations["type"] == "int"
        assert iterations["low"] == 1
        assert iterations["high"] == 5

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_search_strategy_param_space(self, mock_evaluator):
        """Test search strategy parameter space."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")
        param_space = optimizer._get_default_param_space()

        strategy = param_space["search_strategy"]
        assert strategy["type"] == "categorical"
        assert "choices" in strategy
        assert "source-based" in strategy["choices"]


class TestConvenienceFunctions:
    """Tests for convenience optimization functions."""

    def test_optimize_parameters_exists(self):
        """Test that optimize_parameters function exists."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            optimize_parameters,
        )

        assert callable(optimize_parameters)

    def test_optimize_for_speed_exists(self):
        """Test that optimize_for_speed function exists."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            optimize_for_speed,
        )

        assert callable(optimize_for_speed)

    def test_optimize_for_quality_exists(self):
        """Test that optimize_for_quality function exists."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            optimize_for_quality,
        )

        assert callable(optimize_for_quality)

    def test_optimize_for_efficiency_exists(self):
        """Test that optimize_for_efficiency function exists."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            optimize_for_efficiency,
        )

        assert callable(optimize_for_efficiency)


class TestOptimizeFunctionSignatures:
    """Tests for optimization function signatures."""

    def test_optimize_for_speed_default_weights(self):
        """Test optimize_for_speed uses speed-focused weights."""
        import inspect
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            optimize_for_speed,
        )

        # Check function source for speed weights
        source = inspect.getsource(optimize_for_speed)
        assert "speed" in source.lower()

    def test_optimize_for_quality_default_weights(self):
        """Test optimize_for_quality uses quality-focused weights."""
        import inspect
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            optimize_for_quality,
        )

        source = inspect.getsource(optimize_for_quality)
        assert "quality" in source.lower()

    def test_optimize_for_efficiency_default_weights(self):
        """Test optimize_for_efficiency uses balanced weights."""
        import inspect
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            optimize_for_efficiency,
        )

        source = inspect.getsource(optimize_for_efficiency)
        assert "resource" in source.lower()


class TestOptimizerState:
    """Tests for optimizer state management."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_initial_state(self, mock_evaluator):
        """Test initial optimizer state."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        assert optimizer.best_params is None
        assert optimizer.study is None
        assert optimizer.trials_history == []

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_optimizer_stores_progress_callback(self, mock_evaluator):
        """Test that progress callback is stored."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        callback = Mock()
        optimizer = OptunaOptimizer(
            base_query="test",
            progress_callback=callback,
        )

        assert optimizer.progress_callback is callback


class TestPlottingAvailability:
    """Tests for plotting availability handling."""

    def test_plotting_available_flag_exists(self):
        """Test that PLOTTING_AVAILABLE flag exists."""
        from local_deep_research.benchmarks.optimization import optuna_optimizer

        assert hasattr(optuna_optimizer, "PLOTTING_AVAILABLE")
        assert isinstance(optuna_optimizer.PLOTTING_AVAILABLE, bool)


class TestObjectiveFunction:
    """Tests for the objective function."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_objective_method_exists(self, mock_evaluator):
        """Test that _objective method exists."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        assert hasattr(optimizer, "_objective")
        assert callable(optimizer._objective)

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_run_experiment_method_exists(self, mock_evaluator):
        """Test that _run_experiment method exists."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        assert hasattr(optimizer, "_run_experiment")
        assert callable(optimizer._run_experiment)


class TestVisualizationMethods:
    """Tests for visualization methods."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_create_visualizations_method_exists(self, mock_evaluator):
        """Test that _create_visualizations method exists."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        assert hasattr(optimizer, "_create_visualizations")

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_create_quick_visualizations_method_exists(self, mock_evaluator):
        """Test that _create_quick_visualizations method exists."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        assert hasattr(optimizer, "_create_quick_visualizations")

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_save_results_method_exists(self, mock_evaluator):
        """Test that _save_results method exists."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        assert hasattr(optimizer, "_save_results")


class TestOptimizeMethod:
    """Tests for the optimize method."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.optuna"
    )
    def test_optimize_creates_study(self, mock_optuna, mock_evaluator):
        """Test that optimize creates an Optuna study."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()
        mock_study = Mock()
        mock_study.best_params = {"iterations": 2}
        mock_study.best_value = 0.8
        mock_study.best_trial = Mock()
        mock_study.best_trial.user_attrs = {}
        mock_study.trials = []
        mock_optuna.create_study.return_value = mock_study

        optimizer = OptunaOptimizer(
            base_query="test query",
            n_trials=1,
        )

        # Mock _save_results to avoid file operations
        with patch.object(optimizer, "_save_results"):
            with patch.object(optimizer, "_create_visualizations"):
                optimizer.optimize()

        mock_optuna.create_study.assert_called_once()
        assert optimizer.study == mock_study

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.optuna"
    )
    def test_optimize_returns_best_params(self, mock_optuna, mock_evaluator):
        """Test that optimize returns best parameters."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()
        mock_study = Mock()
        mock_study.best_params = {"iterations": 3, "questions_per_iteration": 4}
        mock_study.best_value = 0.85
        mock_study.best_trial = Mock()
        mock_study.best_trial.user_attrs = {}
        mock_study.trials = []
        mock_optuna.create_study.return_value = mock_study

        optimizer = OptunaOptimizer(
            base_query="test",
            n_trials=1,
        )

        with patch.object(optimizer, "_save_results"):
            with patch.object(optimizer, "_create_visualizations"):
                best_params, best_value = optimizer.optimize()

        assert isinstance(best_params, dict)
        assert best_params["iterations"] == 3
        assert best_value == 0.85

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.optuna"
    )
    def test_optimize_stores_trials_history(self, mock_optuna, mock_evaluator):
        """Test that optimize stores trials history."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        # Create mock trials
        mock_trial1 = Mock()
        mock_trial1.params = {"iterations": 2}
        mock_trial1.value = 0.7
        mock_trial1.user_attrs = {}

        mock_trial2 = Mock()
        mock_trial2.params = {"iterations": 3}
        mock_trial2.value = 0.8
        mock_trial2.user_attrs = {}

        mock_study = Mock()
        mock_study.best_params = {"iterations": 3}
        mock_study.best_value = 0.8
        mock_study.best_trial = mock_trial2
        mock_study.trials = [mock_trial1, mock_trial2]
        mock_optuna.create_study.return_value = mock_study

        optimizer = OptunaOptimizer(
            base_query="test",
            n_trials=2,
        )

        with patch.object(optimizer, "_save_results"):
            with patch.object(optimizer, "_create_visualizations"):
                optimizer.optimize()

        # Trials history should be populated from the study callback
        assert optimizer.study is not None


class TestObjectiveFunctionExecution:
    """Tests for objective function execution."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_objective_suggests_parameters(self, mock_evaluator):
        """Test that objective function suggests parameters from trial."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        # Create a mock trial
        mock_trial = Mock()
        mock_trial.suggest_int.return_value = 2
        mock_trial.suggest_float.return_value = 0.7
        mock_trial.suggest_categorical.return_value = "iterdrag"
        mock_trial.set_user_attr = Mock()

        # Mock _run_experiment to return a score
        with patch.object(optimizer, "_run_experiment") as mock_run:
            mock_run.return_value = {
                "score": 0.75,
                "quality_score": 0.8,
                "speed_score": 0.7,
            }

            param_space = optimizer._get_default_param_space()
            score = optimizer._objective(mock_trial, param_space=param_space)

            assert score == 0.75
            mock_trial.suggest_int.assert_called()
            mock_trial.suggest_categorical.assert_called()

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_objective_handles_experiment_error(self, mock_evaluator):
        """Test that objective handles experiment errors gracefully."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        mock_trial = Mock()
        mock_trial.suggest_int.return_value = 2
        mock_trial.suggest_float.return_value = 0.7
        mock_trial.suggest_categorical.return_value = "iterdrag"
        mock_trial.set_user_attr = Mock()

        # Mock _run_experiment to raise an exception
        with patch.object(optimizer, "_run_experiment") as mock_run:
            mock_run.side_effect = Exception("Experiment failed")

            param_space = optimizer._get_default_param_space()
            score = optimizer._objective(mock_trial, param_space=param_space)

            # Should return -inf on error (worst possible score for maximization)
            assert score == float("-inf")


class TestRunExperiment:
    """Tests for run experiment functionality."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.SpeedProfiler"
    )
    def test_run_experiment_calculates_score(
        self, mock_profiler, mock_evaluator
    ):
        """Test that run_experiment calculates weighted score."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        # Setup mock evaluator - evaluate() returns dict with "quality_score"
        mock_eval_instance = Mock()
        mock_eval_instance.evaluate.return_value = {
            "quality_score": 0.8,
            "benchmark_results": {},
        }
        mock_evaluator.return_value = mock_eval_instance

        # Setup mock profiler - source calls start(), stop(), get_summary()
        mock_profiler_instance = Mock()
        mock_profiler_instance.get_summary.return_value = {
            "total_duration": 10.0,
        }
        mock_profiler.return_value = mock_profiler_instance

        optimizer = OptunaOptimizer(
            base_query="test",
            metric_weights={"quality": 0.7, "speed": 0.3},
        )

        params = {
            "iterations": 2,
            "questions_per_iteration": 3,
            "search_strategy": "iterdrag",
            "max_results": 50,
        }

        result = optimizer._run_experiment(params)

        assert "score" in result
        assert "quality_score" in result
        assert "speed_score" in result
        assert result["success"] is True


class TestSaveResults:
    """Tests for save results functionality."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.joblib"
    )
    @patch(
        "local_deep_research.security.file_write_verifier.write_json_verified"
    )
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_save_results_creates_json(
        self, mock_evaluator, mock_write_json, mock_joblib
    ):
        """Test that _save_results creates JSON output."""
        import tempfile
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        with tempfile.TemporaryDirectory() as tmpdir:
            optimizer = OptunaOptimizer(
                base_query="test",
                output_dir=tmpdir,
            )

            # Setup mock study
            mock_study = Mock()
            mock_study.best_params = {"iterations": 2}
            mock_study.best_value = 0.8
            mock_study.best_trial = Mock()
            mock_study.best_trial.user_attrs = {}
            mock_study.trials = [Mock()]
            optimizer.study = mock_study
            optimizer.best_params = {"iterations": 2}
            optimizer.trials_history = [
                {"params": {"iterations": 2}, "score": 0.8}
            ]

            optimizer._save_results()

            # write_json_verified should have been called for history and best params
            assert mock_write_json.call_count >= 1
            # joblib.dump should have been called for the study
            assert mock_joblib.dump.called

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.joblib"
    )
    @patch(
        "local_deep_research.security.file_write_verifier.write_json_verified"
    )
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_save_results_handles_numpy_types(
        self, mock_evaluator, mock_write_json, mock_joblib
    ):
        """Test that _save_results handles numpy types properly."""
        import tempfile
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        with tempfile.TemporaryDirectory() as tmpdir:
            optimizer = OptunaOptimizer(
                base_query="test",
                output_dir=tmpdir,
            )

            mock_study = Mock()
            mock_study.best_params = {"iterations": 2}
            mock_study.best_value = 0.8
            mock_study.best_trial = Mock()
            mock_study.best_trial.user_attrs = {}
            mock_study.trials = []
            optimizer.study = mock_study
            optimizer.best_params = {"iterations": 2}
            optimizer.trials_history = []

            # Should not raise even with potential numpy types
            optimizer._save_results()


class TestVisualizationCreation:
    """Tests for visualization creation."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_create_visualizations_handles_no_plotting(self, mock_evaluator):
        """Test that visualization creation handles missing matplotlib gracefully."""
        import tempfile
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        with tempfile.TemporaryDirectory() as tmpdir:
            optimizer = OptunaOptimizer(
                base_query="test",
                output_dir=tmpdir,
            )

            mock_study = Mock()
            mock_study.trials = []
            optimizer.study = mock_study
            optimizer.trials_history = []

            # Should not raise even if plotting is unavailable
            optimizer._create_visualizations()

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.PLOTTING_AVAILABLE",
        True,
    )
    @patch("local_deep_research.benchmarks.optimization.optuna_optimizer.plt")
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.plot_optimization_history"
    )
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.plot_param_importances"
    )
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.plot_slice"
    )
    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.plot_contour"
    )
    def test_create_visualizations_generates_plots(
        self,
        mock_contour,
        mock_slice,
        mock_importances,
        mock_history,
        mock_plt,
        mock_evaluator,
    ):
        """Test that visualizations are generated when matplotlib is available."""
        import tempfile
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        with tempfile.TemporaryDirectory() as tmpdir:
            optimizer = OptunaOptimizer(
                base_query="test",
                output_dir=tmpdir,
            )

            # Need at least 2 trials for visualizations to proceed
            mock_study = Mock()
            mock_study.trials = [Mock(), Mock()]
            mock_study.best_params = {"iterations": 2}
            optimizer.study = mock_study
            optimizer.trials_history = [
                {
                    "params": {"iterations": 2},
                    "score": 0.8,
                    "result": {
                        "success": True,
                        "quality_score": 0.85,
                        "speed_score": 0.75,
                    },
                },
                {
                    "params": {"iterations": 3},
                    "score": 0.7,
                    "result": {
                        "success": True,
                        "quality_score": 0.75,
                        "speed_score": 0.65,
                    },
                },
            ]

            optimizer._create_visualizations()

            # Optuna plot functions should have been called
            assert mock_history.called


class TestConvenienceFunctionImplementation:
    """Tests for convenience function implementation details."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.OptunaOptimizer"
    )
    def test_optimize_for_speed_uses_speed_weights(self, mock_optimizer_class):
        """Test that optimize_for_speed uses speed-focused weights."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            optimize_for_speed,
        )

        mock_optimizer = Mock()
        mock_optimizer.optimize.return_value = ({}, 0.0)
        mock_optimizer_class.return_value = mock_optimizer

        optimize_for_speed(query="test", n_trials=1)

        # Check that metric_weights have higher speed weight
        call_kwargs = mock_optimizer_class.call_args[1]
        assert (
            call_kwargs["metric_weights"]["speed"]
            > call_kwargs["metric_weights"]["quality"]
        )

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.OptunaOptimizer"
    )
    def test_optimize_for_quality_uses_quality_weights(
        self, mock_optimizer_class
    ):
        """Test that optimize_for_quality uses quality-focused weights."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            optimize_for_quality,
        )

        mock_optimizer = Mock()
        mock_optimizer.optimize.return_value = ({}, 0.0)
        mock_optimizer_class.return_value = mock_optimizer

        optimize_for_quality(query="test", n_trials=1)

        # Check that metric_weights have higher quality weight
        call_kwargs = mock_optimizer_class.call_args[1]
        assert (
            call_kwargs["metric_weights"]["quality"]
            > call_kwargs["metric_weights"]["speed"]
        )

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.OptunaOptimizer"
    )
    def test_optimize_for_efficiency_uses_balanced_weights(
        self, mock_optimizer_class
    ):
        """Test that optimize_for_efficiency uses balanced weights."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            optimize_for_efficiency,
        )

        mock_optimizer = Mock()
        mock_optimizer.optimize.return_value = ({}, 0.0)
        mock_optimizer_class.return_value = mock_optimizer

        optimize_for_efficiency(query="test", n_trials=1)

        # Check that metric_weights include resource
        call_kwargs = mock_optimizer_class.call_args[1]
        assert "resource" in call_kwargs["metric_weights"]


class TestProgressCallback:
    """Tests for progress callback functionality."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_progress_callback_invoked(self, mock_evaluator):
        """Test that progress callback is invoked during optimization."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        callback_calls = []

        def progress_callback(trial_num, n_trials, best_value, best_params):
            callback_calls.append(
                {
                    "trial_num": trial_num,
                    "n_trials": n_trials,
                    "best_value": best_value,
                }
            )

        optimizer = OptunaOptimizer(
            base_query="test",
            progress_callback=progress_callback,
        )

        # The callback should be stored
        assert optimizer.progress_callback is progress_callback

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_optimization_callback_method_exists(self, mock_evaluator):
        """Test that _optimization_callback method exists."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        assert hasattr(optimizer, "_optimization_callback")
        assert callable(optimizer._optimization_callback)


class TestCustomParameterSpace:
    """Tests for custom parameter space handling."""

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_custom_param_space_used(self, mock_evaluator):
        """Test that optimize() accepts a custom parameter space."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )
        import inspect

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        # Verify optimize() accepts param_space parameter
        sig = inspect.signature(optimizer.optimize)
        assert "param_space" in sig.parameters

        # Verify _get_default_param_space returns a dict with expected keys
        default_space = optimizer._get_default_param_space()
        assert isinstance(default_space, dict)
        assert "iterations" in default_space

    @patch(
        "local_deep_research.benchmarks.optimization.optuna_optimizer.CompositeBenchmarkEvaluator"
    )
    def test_default_param_space_used_when_none_provided(self, mock_evaluator):
        """Test that default parameter space is used when none provided."""
        from local_deep_research.benchmarks.optimization.optuna_optimizer import (
            OptunaOptimizer,
        )

        mock_evaluator.return_value = Mock()

        optimizer = OptunaOptimizer(base_query="test")

        # Should use default space
        default_space = optimizer._get_default_param_space()
        assert "iterations" in default_space
        assert "questions_per_iteration" in default_space