MCP Test MCP

#!/usr/bin/env python3 """Test script to execute a weather prompt with the LLM.""" import json import os import requests from dotenv import load_dotenv # Load environment variables load_dotenv() # Weather data from Boston, MA weather_data = { "location": "Boston, MA", "temperature": "48.2°F (9.0°C)", "conditions": "Clear", "forecast": "Clear, with a low around 39. Northwest wind around 6 mph.", "humidity": "42.8%", "wind": "9.2 mph from 330°", "timestamp": "2025-10-10T03:21:20.356262+00:00", "source": "Weather.gov" } # Prompt template from weather_report prompt prompt_template = """Given the following weather data: {weather_data} Generate a professional weather report that includes: - Current conditions summary in natural language - Temperature analysis with context (e.g., "cooler than average", "typical for this time of year") - Wind and humidity details with practical implications - Visibility and air quality if mentioned in conditions - Practical recommendations for outdoor activities - Appropriate clothing suggestions Output format: {output_format} The report should be informative yet accessible to a general audience. Use clear, concise language and avoid excessive technical jargon. Include relevant safety advisories if severe conditions are present.""" # Fill in the template filled_prompt = prompt_template.format( weather_data=json.dumps(weather_data, indent=2), output_format="JSON" ) # Prepare LLM request llm_url = os.getenv("LLM_URL") llm_model = os.getenv("LLM_MODEL_NAME") llm_api_key = os.getenv("LLM_API_KEY") if not all([llm_url, llm_model, llm_api_key]): print("Error: Missing LLM configuration in .env file") exit(1) headers = { "Content-Type": "application/json", "Authorization": f"Bearer {llm_api_key}" } payload = { "model": llm_model, "messages": [ { "role": "user", "content": filled_prompt } ], "max_tokens": 1000, "temperature": 0.7 } print("=" * 80) print("WEATHER DATA:") print("=" * 80) print(json.dumps(weather_data, indent=2)) print() print("=" * 80) print("FILLED PROMPT:") print("=" * 80) print(filled_prompt) print() print("=" * 80) print("SENDING TO LLM...") print("=" * 80) # Send request to LLM response = requests.post( f"{llm_url}/chat/completions", headers=headers, json=payload ) if response.status_code == 200: result = response.json() llm_response = result["choices"][0]["message"]["content"] print("=" * 80) print("LLM RESPONSE:") print("=" * 80) print(llm_response) print() # Try to parse as JSON if it looks like JSON if llm_response.strip().startswith("{"): try: parsed = json.loads(llm_response) print("=" * 80) print("PARSED JSON STRUCTURE:") print("=" * 80) print(json.dumps(parsed, indent=2)) print() # Validate expected fields print("=" * 80) print("VALIDATION:") print("=" * 80) expected_fields = ["summary", "temperature", "conditions", "recommendations"] for field in expected_fields: if field in parsed: print(f"✓ {field}: Present") else: print(f"✗ {field}: Missing") except json.JSONDecodeError as e: print(f"Warning: Response looks like JSON but failed to parse: {e}") print() print("=" * 80) print("TEST RESULT: SUCCESS") print("=" * 80) else: print(f"Error: LLM request failed with status {response.status_code}") print(response.text) print() print("=" * 80) print("TEST RESULT: FAILED") print("=" * 80)