Add AI Chatbot for Interactive Scenario Control

[rfoo1250]

Summary

Introduce an AI chatbot interface that enables users to modify scenarios and control the simulation through natural language commands. The chatbot will execute function calls to modify game state, generate new scenarios, and change/update scenario files dynamically.

Motivation

Setup and changes could be complex. An AI chatbot interface would:

1. Improve Usability – Users can describe scenario changes in plain English instead of manual configuration
2. Accelerate Scenario Building – Quickly add units, modify formations, adjust timing through conversation
3. Reduce Friction – Natural language is faster than clicking through multiple menus
4. Enable Scripted Training – Researchers can programmatically build complex scenarios via chat commands

Proposed Architecture

Function Call Execution Model

The chatbot can directly execute available simulation functions:

User: "Add 3 aircraft to the Blue side at coordinates 40.7128, -74.0060"
↓
Chatbot understands intent and calls: addAircraft(side='blue', quantity=3, lat=40.7128, lon=-74.0060)
↓
Game state updates with new units

Available Functions:

• addAircraft(sideId, quantity, latitude, longitude, altitude)
• addShip(sideId, quantity, latitude, longitude)
• addFacility(sideId, name, latitude, longitude, weaponIds)
• deleteUnit(unitId)
• moveUnit(unitId, newLatitude, newLongitude)
• assignMission(unitId, missionType, targetArea)
• modifyWeapon(unitId, weaponId, parameters)
• setScenarioTime(startTime, duration)

Scenario File Generation Model

The chatbot modifies the scenario JSON structure directly:

User: "Replace all aircraft with ships and add 5 new facilities in the red zone"
↓
Chatbot analyzes current scenario JSON
↓
Generates updated scenario file with modifications
↓
Game loads new scenario configuration

Advantages:

• More flexible for complex multi-step changes
• Can export/reuse modified scenarios
• Version control friendly (scenario files as git artifacts)

Implementation Details

Chatbot Capabilities Required

1. Simulation Function Registry – Chatbot needs access to:

   • All available game functions and their signatures
   • Parameter constraints (valid ranges, required fields)
   • Side/unit/weapon options and their properties

2. Context Awareness – Maintain understanding of:

   • Current scenario state (existing units, timeline, objectives)
   • User intent and implied references ("those aircraft" → list of active aircraft)
   • Undo/redo state for experimentation

3. Natural Language Processing – Parse commands like:

   • "Move my aircraft north 50 nautical miles"
   • "Create a patrol mission over the eastern sector"
   • "Give the red side better weapons"
   • "Add 2 enemy ships near the coast"

4. Real Data Integration – Access to geospatial and contextual datasets:

   • Geospatial Data – Map place names to coordinates (e.g., "deploy over Baghdad" → 33.3128°N, 44.3615°E)
   • Terrain & Geography – Understand terrain features, coastlines, water bodies, and topography
   • Infrastructure Data – Identify real-world airports, ports, strategic locations, and population centers
   • Terrain Elevation – Altitude and terrain intelligence for realistic flight paths and tactical planning
   • Political Boundaries – Territorial and regional knowledge for scenario context
   • Weather & Environmental Data – Historical or simulated weather patterns affecting operations

Integration Points

• Game Engine – Direct calls to Game.ts methods or scenario mutations
• Scenario Manager – Read/write access to scenario files
• Validation – Ensure all chatbot-generated changes maintain game state consistency

Technical Stack

• LLM Integration – LLM API or a local LLM with function calling support
• Function Schema – Formal definition of all callable simulation functions
• Geospatial Data Sources – APIs and datasets for:
  • Geocoding (OpenStreetMap Nominatim, Google Maps API)
  • Terrain & Elevation (USGS DEM, SRTM data)
  • Real-world Infrastructure (airports, ports, military bases)
  • Coastlines & Water Bodies (Natural Earth, GEBCO bathymetry)
• Error Handling – Graceful rejection of invalid commands with helpful suggestions
• Rollback – Undo chatbot actions if scenario becomes invalid

Considerations

• Validation – Chatbot must validate all changes before applying (e.g., no units outside map bounds)
• Permissions – In multi-player scenarios, control who can modify what
• Transparency – Show users exactly what changes the chatbot will make before executing
• Learning Curve – Document what commands the chatbot understands

[harn3ss]

i'm looking to fork this into an updated repo, do you want to team up?