Open Census MCP Server

# test_poverty_fix.py """ Test the poverty rate fix specifically """ import os from llm_mapper import LLMConceptMapper def test_poverty_fix(): """Test if poverty rate now maps correctly""" api_key = os.getenv('OPENAI_API_KEY') if not api_key: print("❌ Need OPENAI_API_KEY") return mapper = LLMConceptMapper(api_key=api_key) print("🧪 Testing Poverty Rate Fix") print("=" * 40) # Test poverty rate with enhanced prompting print("\n📊 Testing: PovertyRate") result = mapper.map_concept_to_variables( concept="PovertyRate", concept_definition="Percentage of population below the federal poverty line" ) print(f"Confidence: {result.confidence}") print(f"Variables: {result.census_variables}") print(f"Statistical Method: {result.statistical_method}") print(f"Universe: {result.universe}") print(f"Calculation Note: {result.calculation_note}") print(f"Reasoning: {result.reasoning}") # Check if we got rate-appropriate variables if result.confidence > 0: if len(result.census_variables) >= 2: print("\n✅ SUCCESS: Got multiple variables (likely numerator/denominator)") elif "rate" in str(result.statistical_method).lower(): print("\n✅ SUCCESS: Recognized as rate calculation") else: print("\n⚠️ PARTIAL: Mapped but may not understand rate calculation") else: print("\n❌ STILL FAILING") return result.confidence > 0.7 if __name__ == "__main__": success = test_poverty_fix() if success: print("\n🎉 Poverty rate issue fixed! Ready to proceed.") else: print("\n🔧 Still needs more work.")