~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.
Python academiaanthropicarxivbravedeep-researchencryptionhome-automationhomeserverlocallocal-deep-researchlocal-llmmistralollamaopenaipubmedresearchresearch-toolretrieval-augmented-generationsearxngself-hosted
test: remove shadow test file test_iterative_reasoning_strategy.py (#4095)
LearningCircuitb69d92fb146f
diff --git a/tests/advanced_search_system/strategies/test_iterative_reasoning_strategy.py b/tests/advanced_search_system/strategies/test_iterative_reasoning_strategy.pydeleted file mode 100644index 5623123d9c..0000000000--- a/tests/advanced_search_system/strategies/test_iterative_reasoning_strategy.py+++ /dev/null@@ -1,661 +0,0 @@-"""-Tests for Iterative Reasoning Strategy--Phase 18: Advanced Search Strategies - Iterative Reasoning Tests-Tests reasoning iterations, knowledge building, and convergence.-"""--from unittest.mock import patch, MagicMock---class TestIterativeReasoning:- """Tests for iterative reasoning functionality"""-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_initial_hypothesis_generation(self, mock_strategy_cls):- """Test initial hypothesis is generated"""- mock_strategy = MagicMock()- mock_strategy._generate_initial_hypothesis.return_value = {- "hypothesis": "Initial answer hypothesis",- "confidence": 0.3,- }-- hypothesis = mock_strategy._generate_initial_hypothesis("What is X?")-- assert "hypothesis" in hypothesis- assert hypothesis["confidence"] < 0.5-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_hypothesis_refinement_iteration(self, mock_strategy_cls):- """Test hypothesis is refined through iteration"""- mock_strategy = MagicMock()- mock_strategy._refine_hypothesis.return_value = {- "hypothesis": "Refined answer",- "confidence": 0.7,- "iteration": 3,- }-- refined = mock_strategy._refine_hypothesis(- current_hypothesis={"hypothesis": "Initial", "confidence": 0.3},- new_evidence=[{"text": "Supporting evidence"}],- )-- assert refined["confidence"] > 0.3-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_evidence_integration_per_iteration(self, mock_strategy_cls):- """Test evidence is integrated each iteration"""- mock_strategy = MagicMock()- mock_strategy._integrate_evidence.return_value = {- "key_facts": ["Fact 1", "Fact 2"],- "uncertainties_resolved": 1,- }-- integration = mock_strategy._integrate_evidence(- knowledge_state={"key_facts": []},- new_evidence=[{"text": "New fact"}],- )-- assert len(integration["key_facts"]) > 0-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_convergence_detection(self, mock_strategy_cls):- """Test convergence is detected"""- mock_strategy = MagicMock()- mock_strategy._has_converged.return_value = True-- converged = mock_strategy._has_converged(- {- "confidence": 0.95,- "key_facts": ["fact1", "fact2", "fact3"],- "uncertainties": [],- }- )-- assert converged is True-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_divergence_handling(self, mock_strategy_cls):- """Test divergence is handled"""- mock_strategy = MagicMock()- mock_strategy._handle_divergence.return_value = {- "action": "broaden_search",- "new_constraints": [],- }-- handling = mock_strategy._handle_divergence(- {"confidence_history": [0.5, 0.4, 0.3], "trend": "decreasing"}- )-- assert "action" in handling-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_iteration_limit_enforcement(self, mock_strategy_cls):- """Test iteration limit is enforced"""- mock_strategy = MagicMock()- mock_strategy.max_iterations = 10- mock_strategy._should_stop.return_value = True-- should_stop = mock_strategy._should_stop({"iteration": 10})-- assert should_stop is True-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_quality_improvement_tracking(self, mock_strategy_cls):- """Test quality improvement is tracked"""- mock_strategy = MagicMock()- mock_strategy._track_improvement.return_value = {- "improvement_rate": 0.1,- "iterations": [0.3, 0.5, 0.7, 0.8],- }-- tracking = mock_strategy._track_improvement([0.3, 0.5, 0.7, 0.8])-- assert tracking["improvement_rate"] > 0-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_chain_building(self, mock_strategy_cls):- """Test reasoning chain is built"""- mock_strategy = MagicMock()- mock_strategy._build_reasoning_chain.return_value = [- {"step": 1, "reasoning": "Initial observation"},- {"step": 2, "reasoning": "Further analysis"},- {"step": 3, "reasoning": "Conclusion"},- ]-- chain = mock_strategy._build_reasoning_chain(- "query", [{"text": "evidence"}]- )-- assert len(chain) == 3-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_step_validation(self, mock_strategy_cls):- """Test reasoning steps are validated"""- mock_strategy = MagicMock()- mock_strategy._validate_reasoning_step.return_value = {- "valid": True,- "issues": [],- }-- validation = mock_strategy._validate_reasoning_step(- {"step": 1, "reasoning": "Valid reasoning"}- )-- assert validation["valid"] is True-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_contradiction_resolution(self, mock_strategy_cls):- """Test contradictions are resolved"""- mock_strategy = MagicMock()- mock_strategy._resolve_contradiction.return_value = {- "resolution": "Claim A is correct based on newer evidence",- "discarded": "Claim B",- }-- resolution = mock_strategy._resolve_contradiction("Claim A", "Claim B")-- assert "resolution" in resolution-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_gap_filling(self, mock_strategy_cls):- """Test reasoning gaps are filled"""- mock_strategy = MagicMock()- mock_strategy._fill_gaps.return_value = {- "gaps_identified": ["Gap 1"],- "gaps_filled": ["Gap 1"],- "remaining_gaps": [],- }-- filling = mock_strategy._fill_gaps({"reasoning_chain": []})-- assert len(filling["remaining_gaps"]) == 0-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_depth_control(self, mock_strategy_cls):- """Test reasoning depth is controlled"""- mock_strategy = MagicMock()- mock_strategy._control_depth.return_value = {- "current_depth": 3,- "max_depth": 5,- "should_go_deeper": True,- }-- control = mock_strategy._control_depth({"depth": 3})-- assert control["should_go_deeper"] is True-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_breadth_control(self, mock_strategy_cls):- """Test reasoning breadth is controlled"""- mock_strategy = MagicMock()- mock_strategy._control_breadth.return_value = {- "topics_explored": 5,- "max_topics": 10,- "should_explore_more": True,- }-- control = mock_strategy._control_breadth({"topics": 5})-- assert control["should_explore_more"] is True-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_priority_ordering(self, mock_strategy_cls):- """Test reasoning priorities are ordered"""- mock_strategy = MagicMock()- mock_strategy._prioritize_reasoning.return_value = [- {"topic": "High priority", "score": 0.9},- {"topic": "Medium priority", "score": 0.6},- {"topic": "Low priority", "score": 0.3},- ]-- priorities = mock_strategy._prioritize_reasoning(- [- {"topic": "Low priority"},- {"topic": "High priority"},- {"topic": "Medium priority"},- ]- )-- assert priorities[0]["score"] > priorities[1]["score"]-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_pruning_strategy(self, mock_strategy_cls):- """Test irrelevant reasoning is pruned"""- mock_strategy = MagicMock()- mock_strategy._prune_reasoning.return_value = {- "kept": 5,- "pruned": 3,- "remaining": ["r1", "r2", "r3", "r4", "r5"],- }-- pruning = mock_strategy._prune_reasoning(- [{"relevance": 0.9}, {"relevance": 0.8}, {"relevance": 0.1}]- )-- assert pruning["kept"] > pruning["pruned"]-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_multi_path_reasoning(self, mock_strategy_cls):- """Test multiple reasoning paths are explored"""- mock_strategy = MagicMock()- mock_strategy._explore_paths.return_value = {- "paths": [- {"path": "A -> B -> C", "confidence": 0.8},- {"path": "A -> D -> C", "confidence": 0.7},- ]- }-- paths = mock_strategy._explore_paths("query")-- assert len(paths["paths"]) == 2-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_merge_paths(self, mock_strategy_cls):- """Test reasoning paths are merged"""- mock_strategy = MagicMock()- mock_strategy._merge_paths.return_value = {- "merged_conclusion": "Combined conclusion",- "paths_merged": 2,- }-- merged = mock_strategy._merge_paths(- [{"conclusion": "C1"}, {"conclusion": "C2"}]- )-- assert "merged_conclusion" in merged-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_confidence_propagation(self, mock_strategy_cls):- """Test confidence propagates through reasoning"""- mock_strategy = MagicMock()- mock_strategy._propagate_confidence.return_value = {- "initial_confidence": 0.9,- "propagated_confidence": 0.8,- "decay_applied": True,- }-- propagation = mock_strategy._propagate_confidence(- {"confidence": 0.9}, steps=3- )-- assert (- propagation["propagated_confidence"]- < propagation["initial_confidence"]- )-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_uncertainty_handling(self, mock_strategy_cls):- """Test uncertainty is handled in reasoning"""- mock_strategy = MagicMock()- mock_strategy._handle_uncertainty.return_value = {- "uncertainties": ["U1", "U2"],- "mitigation": "Additional search needed",- }-- handling = mock_strategy._handle_uncertainty(["U1", "U2"])-- assert "mitigation" in handling-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_assumption_tracking(self, mock_strategy_cls):- """Test assumptions are tracked"""- mock_strategy = MagicMock()- mock_strategy._track_assumptions.return_value = {- "assumptions": ["A1", "A2"],- "validated": ["A1"],- "unvalidated": ["A2"],- }-- tracking = mock_strategy._track_assumptions(["A1", "A2"])-- assert len(tracking["assumptions"]) == 2-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_conclusion_extraction(self, mock_strategy_cls):- """Test conclusions are extracted"""- mock_strategy = MagicMock()- mock_strategy._extract_conclusion.return_value = {- "conclusion": "Final answer",- "confidence": 0.85,- "supporting_facts": 5,- }-- conclusion = mock_strategy._extract_conclusion(- {"reasoning_chain": [], "key_facts": []}- )-- assert "conclusion" in conclusion-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_supporting_evidence(self, mock_strategy_cls):- """Test supporting evidence is collected"""- mock_strategy = MagicMock()- mock_strategy._collect_supporting_evidence.return_value = [- {"text": "Evidence 1", "relevance": 0.9},- {"text": "Evidence 2", "relevance": 0.8},- ]-- evidence = mock_strategy._collect_supporting_evidence("conclusion")-- assert len(evidence) == 2-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_counterargument_handling(self, mock_strategy_cls):- """Test counterarguments are handled"""- mock_strategy = MagicMock()- mock_strategy._handle_counterarguments.return_value = {- "counterarguments": ["CA1"],- "refutations": ["Refutation of CA1"],- "unaddressed": [],- }-- handling = mock_strategy._handle_counterarguments(["CA1"])-- assert len(handling["unaddressed"]) == 0-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_synthesis_generation(self, mock_strategy_cls):- """Test synthesis is generated"""- mock_strategy = MagicMock()- mock_strategy._generate_synthesis.return_value = {- "synthesis": "Comprehensive answer",- "components_used": 5,- }-- synthesis = mock_strategy._generate_synthesis(- {"key_facts": [], "reasoning_chain": []}- )-- assert "synthesis" in synthesis-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_summary_creation(self, mock_strategy_cls):- """Test summary is created"""- mock_strategy = MagicMock()- mock_strategy._create_summary.return_value = {- "summary": "Brief summary of findings",- "word_count": 50,- }-- summary = mock_strategy._create_summary(- {"full_answer": "Long detailed answer..."}- )-- assert summary["word_count"] < 100-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_quality_assessment(self, mock_strategy_cls):- """Test reasoning quality is assessed"""- mock_strategy = MagicMock()- mock_strategy._assess_reasoning_quality.return_value = {- "quality_score": 0.85,- "strengths": ["Well-supported"],- "weaknesses": [],- }-- assessment = mock_strategy._assess_reasoning_quality(- {"reasoning_chain": []}- )-- assert assessment["quality_score"] >= 0.8-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_feedback_integration(self, mock_strategy_cls):- """Test feedback is integrated"""- mock_strategy = MagicMock()- mock_strategy._integrate_feedback.return_value = {- "adjustments_made": True,- "new_confidence": 0.75,- }-- integration = mock_strategy._integrate_feedback(- {"rating": 4, "comment": "Good but needs more depth"}- )-- assert integration["adjustments_made"] is True-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_learning_from_outcome(self, mock_strategy_cls):- """Test learning from outcomes"""- mock_strategy = MagicMock()- mock_strategy._learn_from_outcome.return_value = {- "patterns_learned": ["Pattern 1"],- "success_rate_updated": True,- }-- learning = mock_strategy._learn_from_outcome(- {"success": True, "user_rating": 5}- )-- assert len(learning["patterns_learned"]) > 0-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_context_management(self, mock_strategy_cls):- """Test context is managed"""- mock_strategy = MagicMock()- mock_strategy._manage_context.return_value = {- "context_size": 2000,- "truncated": False,- }-- management = mock_strategy._manage_context(- {"accumulated_context": "..." * 500}- )-- assert not management["truncated"]-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_reasoning_resource_optimization(self, mock_strategy_cls):- """Test resources are optimized"""- mock_strategy = MagicMock()- mock_strategy._optimize_resources.return_value = {- "llm_calls_reduced": 2,- "search_calls_optimized": True,- }-- optimization = mock_strategy._optimize_resources(- {"budget_remaining": 0.5}- )-- assert optimization["search_calls_optimized"] is True---class TestKnowledgeState:- """Tests for KnowledgeState dataclass"""-- def test_knowledge_state_creation(self):- """Test KnowledgeState can be created"""- from dataclasses import dataclass-- @dataclass- class MockKnowledgeState:- original_query: str- key_facts: list- uncertainties: list- search_history: list- candidate_answers: list- confidence: float- iteration: int-- state = MockKnowledgeState(- original_query="test query",- key_facts=["fact1"],- uncertainties=["uncertainty1"],- search_history=[],- candidate_answers=[],- confidence=0.5,- iteration=1,- )-- assert state.original_query == "test query"- assert state.confidence == 0.5-- def test_knowledge_state_to_string(self):- """Test KnowledgeState string representation"""- from dataclasses import dataclass-- @dataclass- class MockKnowledgeState:- original_query: str- key_facts: list-- def to_string(self):- return f"Query: {self.original_query}, Facts: {self.key_facts}"-- state = MockKnowledgeState(- original_query="test", key_facts=["fact1", "fact2"]- )-- string_repr = state.to_string()-- assert "test" in string_repr- assert "fact1" in string_repr---class TestSearchDecision:- """Tests for search decision logic"""-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_decide_next_search(self, mock_strategy_cls):- """Test next search decision"""- mock_strategy = MagicMock()- mock_strategy._decide_next_search.return_value = {- "search_query": "refined query",- "strategy": "targeted",- }-- decision = mock_strategy._decide_next_search(- {"key_facts": [], "uncertainties": ["What is X?"]}- )-- assert "search_query" in decision-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_execute_search(self, mock_strategy_cls):- """Test search execution"""- mock_strategy = MagicMock()- mock_strategy._execute_search.return_value = {- "results": [{"title": "Result 1"}],- "count": 1,- }-- results = mock_strategy._execute_search("search query")-- assert results["count"] == 1-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_update_knowledge(self, mock_strategy_cls):- """Test knowledge update"""- mock_strategy = MagicMock()- mock_strategy._update_knowledge.return_value = {- "key_facts": ["new_fact"],- "uncertainties_resolved": 1,- }-- update = mock_strategy._update_knowledge([{"text": "New information"}])-- assert len(update["key_facts"]) > 0-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_assess_answer(self, mock_strategy_cls):- """Test answer assessment"""- mock_strategy = MagicMock()- mock_strategy._assess_answer.return_value = {- "confidence": 0.85,- "complete": True,- }-- assessment = mock_strategy._assess_answer(- {"candidate_answers": [{"answer": "test", "confidence": 0.85}]}- )-- assert assessment["confidence"] >= 0.8-- @patch(- "local_deep_research.advanced_search_system.strategies.iterative_reasoning_strategy.IterativeReasoningStrategy"- )- def test_synthesize_final_answer(self, mock_strategy_cls):- """Test final answer synthesis"""- mock_strategy = MagicMock()- mock_strategy._synthesize_final_answer.return_value = {- "answer": "Final synthesized answer",- "sources": ["source1", "source2"],- }-- answer = mock_strategy._synthesize_final_answer(- {- "key_facts": ["fact1"],- "candidate_answers": [{"answer": "candidate"}],- }- )-- assert "answer" in answer