Medical GraphRAG Assistant

""" Unit tests for src/search/scoring.py Tests similarity score calculations and display logic for medical image search. Following TDD: These tests should FAIL initially until scoring.py is implemented. """ import pytest import numpy as np from src.search.scoring import ( calculate_similarity, get_score_color, get_confidence_level, get_hex_color ) class TestCalculateSimilarity: """Test cosine similarity calculations.""" def test_identical_vectors_return_one(self): """Identical vectors should have similarity of 1.0""" vec = [1.0, 2.0, 3.0, 4.0, 5.0] similarity = calculate_similarity(vec, vec) assert similarity == pytest.approx(1.0, abs=0.01) def test_orthogonal_vectors_return_zero(self): """Orthogonal vectors should have similarity near 0.0""" vec1 = [1.0, 0.0, 0.0] vec2 = [0.0, 1.0, 0.0] similarity = calculate_similarity(vec1, vec2) assert similarity == pytest.approx(0.0, abs=0.01) def test_opposite_vectors_return_negative_one(self): """Opposite vectors should have similarity of -1.0""" vec1 = [1.0, 2.0, 3.0] vec2 = [-1.0, -2.0, -3.0] similarity = calculate_similarity(vec1, vec2) assert similarity == pytest.approx(-1.0, abs=0.01) def test_similar_vectors_return_high_score(self): """Similar vectors should have high positive similarity""" vec1 = [1.0, 2.0, 3.0, 4.0, 5.0] vec2 = [1.1, 2.1, 2.9, 4.1, 4.9] # Slightly perturbed similarity = calculate_similarity(vec1, vec2) assert similarity > 0.95 assert similarity <= 1.0 def test_returns_float_type(self): """Similarity should return float, not numpy type""" vec1 = [1.0, 2.0, 3.0] vec2 = [1.0, 2.0, 3.0] similarity = calculate_similarity(vec1, vec2) assert isinstance(similarity, float) assert not isinstance(similarity, np.floating) def test_handles_numpy_arrays(self): """Should accept numpy arrays as input""" vec1 = np.array([1.0, 2.0, 3.0]) vec2 = np.array([1.0, 2.0, 3.0]) similarity = calculate_similarity(vec1, vec2) assert similarity == pytest.approx(1.0, abs=0.01) def test_handles_lists(self): """Should accept Python lists as input""" vec1 = [1.0, 2.0, 3.0] vec2 = [1.0, 2.0, 3.0] similarity = calculate_similarity(vec1, vec2) assert similarity == pytest.approx(1.0, abs=0.01) def test_handles_high_dimensional_vectors(self): """Should work with 1024-dim vectors (NV-CLIP size)""" vec1 = np.random.rand(1024).tolist() vec2 = vec1.copy() # Identical similarity = calculate_similarity(vec1, vec2) assert similarity == pytest.approx(1.0, abs=0.01) def test_normalized_vectors(self): """Should work with pre-normalized vectors""" # Unit vectors vec1 = [0.707, 0.707, 0.0] # sqrt(2)/2 vec2 = [0.707, 0.0, 0.707] similarity = calculate_similarity(vec1, vec2) assert -1.0 <= similarity <= 1.0 def test_zero_vector_raises_error(self): """Zero vectors should raise ValueError or return NaN""" vec1 = [0.0, 0.0, 0.0] vec2 = [1.0, 2.0, 3.0] with pytest.raises((ValueError, ZeroDivisionError)): calculate_similarity(vec1, vec2) class TestGetScoreColor: """Test score color code mapping.""" def test_strong_score_returns_green(self): """Scores ≥ 0.7 should return 'green'""" assert get_score_color(0.7) == 'green' assert get_score_color(0.75) == 'green' assert get_score_color(0.9) == 'green' assert get_score_color(1.0) == 'green' def test_moderate_score_returns_yellow(self): """Scores 0.5-0.7 should return 'yellow'""" assert get_score_color(0.5) == 'yellow' assert get_score_color(0.6) == 'yellow' assert get_score_color(0.69) == 'yellow' def test_weak_score_returns_gray(self): """Scores < 0.5 should return 'gray'""" assert get_score_color(0.0) == 'gray' assert get_score_color(0.3) == 'gray' assert get_score_color(0.49) == 'gray' def test_boundary_values(self): """Test exact boundary values""" # 0.7 boundary (strong vs moderate) assert get_score_color(0.699999) == 'yellow' assert get_score_color(0.7) == 'green' # 0.5 boundary (moderate vs weak) assert get_score_color(0.499999) == 'gray' assert get_score_color(0.5) == 'yellow' def test_negative_scores_return_gray(self): """Negative scores (dissimilar) should return 'gray'""" assert get_score_color(-0.5) == 'gray' assert get_score_color(-1.0) == 'gray' def test_returns_string(self): """Should return string type""" result = get_score_color(0.8) assert isinstance(result, str) def test_lowercase_output(self): """Color strings should be lowercase""" assert get_score_color(0.8).islower() assert get_score_color(0.6).islower() assert get_score_color(0.3).islower() class TestGetConfidenceLevel: """Test confidence level label generation.""" def test_strong_confidence(self): """Scores ≥ 0.7 should return 'strong'""" assert get_confidence_level(0.7) == 'strong' assert get_confidence_level(0.8) == 'strong' assert get_confidence_level(1.0) == 'strong' def test_moderate_confidence(self): """Scores 0.5-0.7 should return 'moderate'""" assert get_confidence_level(0.5) == 'moderate' assert get_confidence_level(0.6) == 'moderate' assert get_confidence_level(0.69) == 'moderate' def test_weak_confidence(self): """Scores < 0.5 should return 'weak'""" assert get_confidence_level(0.0) == 'weak' assert get_confidence_level(0.3) == 'weak' assert get_confidence_level(0.49) == 'weak' def test_boundary_values(self): """Test exact boundary values""" assert get_confidence_level(0.699999) == 'moderate' assert get_confidence_level(0.7) == 'strong' assert get_confidence_level(0.499999) == 'weak' assert get_confidence_level(0.5) == 'moderate' def test_returns_string(self): """Should return string type""" result = get_confidence_level(0.8) assert isinstance(result, str) def test_lowercase_output(self): """Confidence levels should be lowercase""" assert get_confidence_level(0.8).islower() class TestGetHexColor: """Test hex color code generation for UI.""" def test_green_hex(self): """Green score should return green hex color""" hex_color = get_hex_color(0.8) assert hex_color.startswith('#') assert len(hex_color) == 7 # Should be greenish (high G value) assert hex_color.lower() in ['#28a745', '#2ecc71', '#27ae60', '#00ff00'] # Common greens def test_yellow_hex(self): """Yellow score should return yellow/orange hex color""" hex_color = get_hex_color(0.6) assert hex_color.startswith('#') assert len(hex_color) == 7 # Should be yellowish/orange assert hex_color.lower() in ['#ffc107', '#f39c12', '#ffeb3b', '#ffa500'] # Common yellows def test_gray_hex(self): """Gray score should return gray hex color""" hex_color = get_hex_color(0.3) assert hex_color.startswith('#') assert len(hex_color) == 7 # Should be grayish assert hex_color.lower() in ['#6c757d', '#95a5a6', '#7f8c8d', '#808080'] # Common grays def test_returns_valid_hex_format(self): """All results should be valid hex colors""" for score in [0.0, 0.3, 0.5, 0.7, 0.9, 1.0]: hex_color = get_hex_color(score) assert hex_color.startswith('#') assert len(hex_color) == 7 # Verify it's valid hex try: int(hex_color[1:], 16) except ValueError: pytest.fail(f"Invalid hex color: {hex_color}") class TestIntegration: """Integration tests combining multiple functions.""" def test_complete_scoring_workflow(self): """Test full workflow: similarity → color → confidence""" # Setup: Two similar vectors vec1 = [1.0, 2.0, 3.0, 4.0, 5.0] vec2 = [1.1, 2.0, 3.1, 3.9, 5.0] # Step 1: Calculate similarity score = calculate_similarity(vec1, vec2) assert 0.0 <= score <= 1.0 # Step 2: Get color color = get_score_color(score) assert color in ['green', 'yellow', 'gray'] # Step 3: Get confidence confidence = get_confidence_level(score) assert confidence in ['strong', 'moderate', 'weak'] # Step 4: Get hex color hex_color = get_hex_color(score) assert hex_color.startswith('#') # Validate consistency if score >= 0.7: assert color == 'green' assert confidence == 'strong' elif score >= 0.5: assert color == 'yellow' assert confidence == 'moderate' else: assert color == 'gray' assert confidence == 'weak' def test_medical_image_search_scenario(self): """Simulate realistic medical image search scores""" # Simulate NV-CLIP embeddings (1024-dim) query_embedding = np.random.rand(1024) # Simulate 5 image embeddings with varying similarity image_embeddings = [ query_embedding + np.random.rand(1024) * 0.1, # Very similar query_embedding + np.random.rand(1024) * 0.3, # Moderately similar query_embedding + np.random.rand(1024) * 0.5, # Less similar np.random.rand(1024), # Random (dissimilar) -query_embedding # Opposite ] results = [] for img_emb in image_embeddings: score = calculate_similarity(query_embedding.tolist(), img_emb.tolist()) color = get_score_color(score) confidence = get_confidence_level(score) results.append({ 'score': score, 'color': color, 'confidence': confidence }) # Validate all results are valid for result in results: assert -1.0 <= result['score'] <= 1.0 assert result['color'] in ['green', 'yellow', 'gray'] assert result['confidence'] in ['strong', 'moderate', 'weak'] # Fixtures for reusable test data @pytest.fixture def sample_embeddings(): """Provide sample NV-CLIP-like embeddings.""" return { 'query': np.random.rand(1024).tolist(), 'similar': (np.random.rand(1024) * 1.1).tolist(), 'different': np.random.rand(1024).tolist(), } @pytest.fixture def score_examples(): """Provide example scores for each category.""" return { 'strong': [0.7, 0.8, 0.9, 1.0], 'moderate': [0.5, 0.55, 0.6, 0.65], 'weak': [0.0, 0.1, 0.3, 0.49] } # Parametrized tests for comprehensive coverage @pytest.mark.parametrize("score,expected_color", [ (1.0, 'green'), (0.85, 'green'), (0.7, 'green'), (0.69, 'yellow'), (0.6, 'yellow'), (0.5, 'yellow'), (0.49, 'gray'), (0.3, 'gray'), (0.0, 'gray'), (-0.5, 'gray'), ]) def test_score_color_mapping(score, expected_color): """Test all score-to-color mappings.""" assert get_score_color(score) == expected_color @pytest.mark.parametrize("score,expected_confidence", [ (1.0, 'strong'), (0.75, 'strong'), (0.7, 'strong'), (0.65, 'moderate'), (0.5, 'moderate'), (0.45, 'weak'), (0.0, 'weak'), ]) def test_score_confidence_mapping(score, expected_confidence): """Test all score-to-confidence mappings.""" assert get_confidence_level(score) == expected_confidence