Aerospace MCP

"""Agent tools for helping users interact with aerospace-mcp tools effectively.""" import json import logging import os from typing import Any import litellm from pydantic import BaseModel logger = logging.getLogger(__name__) # Check if LLM tools are enabled via environment variable LLM_TOOLS_ENABLED = os.environ.get("LLM_TOOLS_ENABLED", "false").lower() == "true" # Configure LiteLLM for OpenAI GPT-5-Medium litellm.set_verbose = False # Log status of LLM tools if not LLM_TOOLS_ENABLED: logger.info("LLM tools disabled via LLM_TOOLS_ENABLED environment variable.") elif "OPENAI_API_KEY" not in os.environ: logger.warning( "LLM_TOOLS_ENABLED=true but OPENAI_API_KEY not set. Agent tools will not function without it." ) class ToolReference(BaseModel): """Reference to an aerospace-mcp tool with its schema.""" name: str description: str parameters: dict[str, Any] examples: list[str] = [] # Define all available aerospace-mcp tools with their schemas AEROSPACE_TOOLS = [ ToolReference( name="search_airports", description="Search for airports by IATA code or city name", parameters={ "query": "str - IATA code (e.g., 'SJC') or city name (e.g., 'San Jose')", "country": "str | None - Optional ISO country code filter (e.g., 'US', 'JP')", "query_type": "Literal['iata', 'city', 'auto'] - Type of query, defaults to 'auto'", }, examples=[ 'search_airports("SFO")', 'search_airports("London", "GB")', 'search_airports("Tokyo")', ], ), ToolReference( name="plan_flight", description="Generate complete flight plan between airports", parameters={ "departure": "dict - Airport info with city, iata (optional), country (optional)", "arrival": "dict - Airport info with city, iata (optional), country (optional)", "aircraft": "dict - Aircraft config with type, cruise_alt_ft, mass_kg (optional)", "route_options": "dict - Route config with step_km (optional, default 25.0)", }, examples=[ 'plan_flight({"city": "San Francisco"}, {"city": "New York"}, {"type": "A320", "cruise_alt_ft": 37000}, {})', 'plan_flight({"city": "London", "iata": "LHR"}, {"city": "Dubai", "iata": "DXB"}, {"type": "B777", "cruise_alt_ft": 39000, "mass_kg": 220000}, {"step_km": 50.0})', ], ), ToolReference( name="calculate_distance", description="Calculate great-circle distance between airports", parameters={ "origin": "dict - Origin airport with city and optional iata/country", "destination": "dict - Destination airport with city and optional iata/country", "step_km": "float - Optional step size for route polyline generation (default 25.0)", }, examples=[ 'calculate_distance({"city": "New York"}, {"city": "Los Angeles"})', 'calculate_distance({"city": "Paris", "iata": "CDG"}, {"city": "Tokyo", "iata": "NRT"}, 100.0)', ], ), ToolReference( name="get_aircraft_performance", description="Get performance estimates for aircraft", parameters={ "aircraft_type": "str - Aircraft type code (e.g., 'A320', 'B737', 'B777')", "distance_km": "float - Flight distance in kilometers", "cruise_altitude_ft": "float - Cruise altitude in feet (optional, default 35000)", "mass_kg": "float - Aircraft mass in kg (optional, uses 85% MTOW if not provided)", }, examples=[ 'get_aircraft_performance("A320", 2500.0, 37000)', 'get_aircraft_performance("B777", 5500.0, 39000, 250000)', ], ), ToolReference( name="get_atmosphere_profile", description="Calculate atmospheric conditions at various altitudes", parameters={ "altitudes_m": "List[float] - List of altitudes in meters", "model_type": "Literal['isa', 'enhanced'] - Atmospheric model type (default 'isa')", }, examples=[ "get_atmosphere_profile([0, 1000, 5000, 10000])", 'get_atmosphere_profile([0, 2000, 4000, 6000, 8000, 10000], "enhanced")', ], ), ToolReference( name="wind_model_simple", description="Calculate wind profiles at various altitudes", parameters={ "altitudes_m": "List[float] - List of altitudes in meters", "surface_wind_mps": "float - Surface wind speed in m/s", "model": "Literal['logarithmic', 'power_law'] - Wind profile model (default 'logarithmic')", "surface_roughness_m": "float - Surface roughness in meters (default 0.1)", }, examples=[ "wind_model_simple([0, 100, 500, 1000], 10.0)", 'wind_model_simple([0, 200, 1000, 3000], 15.0, "power_law", 0.05)', ], ), ToolReference( name="elements_to_state_vector", description="Convert orbital elements to state vector", parameters={ "elements": "dict - Orbital elements with semi_major_axis_m, eccentricity, inclination_deg, raan_deg, arg_periapsis_deg, true_anomaly_deg, epoch_utc" }, examples=[ 'elements_to_state_vector({"semi_major_axis_m": 6793000, "eccentricity": 0.001, "inclination_deg": 51.6, "raan_deg": 0.0, "arg_periapsis_deg": 0.0, "true_anomaly_deg": 0.0, "epoch_utc": "2024-01-01T12:00:00"})' ], ), ToolReference( name="propagate_orbit_j2", description="Propagate satellite orbit with J2 perturbations", parameters={ "initial_state": "dict - Initial orbital elements or state vector", "time_span_s": "float - Propagation time span in seconds", "time_step_s": "float - Time step for propagation in seconds (default 300)", }, examples=[ 'propagate_orbit_j2({"semi_major_axis_m": 6793000, "eccentricity": 0.001, "inclination_deg": 51.6, "raan_deg": 0.0, "arg_periapsis_deg": 0.0, "true_anomaly_deg": 0.0, "epoch_utc": "2024-01-01T12:00:00"}, 86400, 600)' ], ), ToolReference( name="hohmann_transfer", description="Calculate Hohmann transfer orbit between two circular orbits", parameters={ "r1_m": "float - Initial orbit radius in meters", "r2_m": "float - Final orbit radius in meters", }, examples=[ "hohmann_transfer(6778000, 42164000)", # LEO to GEO "hohmann_transfer(6578000, 6793000)", # Lower LEO to ISS altitude ], ), ToolReference( name="rocket_3dof_trajectory", description="Simulate 3DOF rocket trajectory with atmospheric effects", parameters={ "geometry": "dict - Rocket geometry with mass_kg, thrust_n, burn_time_s, drag_coeff, reference_area_m2", "dt_s": "float - Time step in seconds (default 0.1)", "max_time_s": "float - Maximum simulation time in seconds (default 300)", "launch_angle_deg": "float - Launch angle in degrees from vertical (default 0)", }, examples=[ 'rocket_3dof_trajectory({"mass_kg": 500, "thrust_n": 8000, "burn_time_s": 60, "drag_coeff": 0.3, "reference_area_m2": 0.5}, 0.1, 300, 15)' ], ), ] def format_data_for_tool( tool_name: str, user_requirements: str, raw_data: str = "" ) -> str: """ Help format data in the correct format for a specific aerospace-mcp tool. Uses GPT-5-Medium to analyze the user's requirements and raw data, then provides the correctly formatted parameters for the specified tool. Args: tool_name: Name of the aerospace-mcp tool to format data for user_requirements: Description of what the user wants to accomplish raw_data: Any raw data that needs to be formatted (optional) Returns: Formatted JSON string with the correct parameters for the tool """ # Check if LLM tools are enabled if not LLM_TOOLS_ENABLED: return "Error: LLM agent tools are disabled. Set LLM_TOOLS_ENABLED=true to enable them." # Find the tool reference first tool_ref = None for tool in AEROSPACE_TOOLS: if tool.name == tool_name: tool_ref = tool break if not tool_ref: available_tools = [t.name for t in AEROSPACE_TOOLS] return f"Error: Tool '{tool_name}' not found. Available tools: {', '.join(available_tools)}" if "OPENAI_API_KEY" not in os.environ: return "Error: OPENAI_API_KEY environment variable not set. Cannot use agent tools." # Build the prompt for GPT-5-Medium system_prompt = f"""You are a data formatting assistant for aerospace-mcp tools. Your job is to help format data correctly for the '{tool_name}' tool. Tool Information: - Name: {tool_ref.name} - Description: {tool_ref.description} - Parameters: {json.dumps(tool_ref.parameters, indent=2)} - Examples: {json.dumps(tool_ref.examples, indent=2)} User Requirements: {user_requirements} Raw Data (if provided): {raw_data} Please provide ONLY a valid JSON object with the correctly formatted parameters for this tool. Do not include any explanation or additional text - just the JSON object that can be directly used as input to the tool. If the user's requirements are unclear or insufficient data is provided, return a JSON object with an "error" field explaining what additional information is needed.""" try: # Call GPT-5-Medium via LiteLLM response = litellm.completion( model="gpt-5-medium", messages=[ {"role": "system", "content": system_prompt}, { "role": "user", "content": f"Format data for {tool_name}: {user_requirements}", }, ], temperature=0.1, max_tokens=1000, ) formatted_result = response.choices[0].message.content.strip() # Validate it's valid JSON try: json.loads(formatted_result) return formatted_result except json.JSONDecodeError: return f'{{"error": "Failed to generate valid JSON format. Raw response: {formatted_result}"}}' except Exception as e: return f'{{"error": "Failed to format data: {str(e)}"}}' def select_aerospace_tool(user_task: str, user_context: str = "") -> str: """ Help select the most appropriate aerospace-mcp tool for a given task. Uses GPT-5-Medium to analyze the user's task and recommend the best tool(s) along with guidance on how to use them. Args: user_task: Description of what the user wants to accomplish user_context: Additional context about the user's situation (optional) Returns: Recommendation with tool name(s) and usage guidance """ # Check if LLM tools are enabled if not LLM_TOOLS_ENABLED: return "Error: LLM agent tools are disabled. Set LLM_TOOLS_ENABLED=true to enable them." if "OPENAI_API_KEY" not in os.environ: return "Error: OPENAI_API_KEY environment variable not set. Cannot use agent tools." # Build comprehensive tool catalog for the AI tools_catalog = [] for tool in AEROSPACE_TOOLS: tools_catalog.append( { "name": tool.name, "description": tool.description, "parameters": tool.parameters, "examples": tool.examples, } ) system_prompt = f"""You are an aerospace engineering assistant specialized in helping users select the right tools for their aerospace calculations and analysis tasks. Available Aerospace Tools: {json.dumps(tools_catalog, indent=2)} User Task: {user_task} User Context: {user_context} Your job is to: 1. Analyze the user's task and context 2. Recommend the most appropriate tool(s) from the available aerospace tools 3. Provide clear guidance on how to use the recommended tool(s) 4. If multiple tools are needed, explain the workflow and order of operations 5. Highlight any prerequisites, limitations, or important considerations Respond in a clear, structured format with: - PRIMARY_TOOL: The main tool to use - SECONDARY_TOOLS: Any additional tools needed (if applicable) - WORKFLOW: Step-by-step guidance - CONSIDERATIONS: Important notes, limitations, or prerequisites - EXAMPLE: A concrete example of how to use the tool(s) for this task If the user's task cannot be accomplished with the available tools, clearly explain what's missing and suggest alternatives.""" try: # Call GPT-5-Medium via LiteLLM response = litellm.completion( model="gpt-5-medium", messages=[ {"role": "system", "content": system_prompt}, { "role": "user", "content": f"Help me with this task: {user_task}\n\nContext: {user_context}", }, ], temperature=0.2, max_tokens=1500, ) return response.choices[0].message.content.strip() except Exception as e: return f"Error: Failed to select appropriate tool: {str(e)}"