Open Census MCP Server

# test_improved_candidates.py """ Quick test to see if improved candidate selection works better """ from llm_mapper import LLMConceptMapper def test_candidate_improvements(): """Test improved candidate selection for problematic concepts""" mapper = LLMConceptMapper() test_cases = [ { "concept": "MedianHouseholdIncome", "definition": "The median income of all households in a geographic area", "expected_vars": ["B19013"] }, { "concept": "PovertyRate", "definition": "Percentage of population below the federal poverty line", "expected_vars": ["B17001"] }, { "concept": "EducationalAttainment", "definition": "Highest level of education completed by individuals", "expected_vars": ["B15003", "B15002"] } ] print("🔍 Testing Improved Candidate Selection") print("=" * 50) for test in test_cases: print(f"\n📊 Testing: {test['concept']}") candidates = mapper._find_candidate_variables( test["concept"], test["definition"] ) print(f" Found {len(candidates)} candidates:") # Show top 5 candidates for i, candidate in enumerate(candidates[:5]): var_id = candidate['variable_id'] score = candidate['score'] label = candidate['label'][:60] + "..." # Check if any expected variable patterns are found found_expected = any(exp in var_id for exp in test["expected_vars"]) marker = "✅" if found_expected else " " print(f" {marker} {i+1}. {var_id}: {label} (score: {score})") # Summary found_any = any( any(exp in c['variable_id'] for exp in test["expected_vars"]) for c in candidates[:10] ) if found_any: print(f" ✅ SUCCESS: Found expected variable patterns in top 10") else: print(f" ❌ ISSUE: No expected patterns found in top 10") if __name__ == "__main__": test_candidate_improvements()