Simulate Current Model

Run a simulation on the currently loaded model and present results.

Steps

1. Save the model — the server picks a private writable path and returns it:

   save_osm_model(save_name="<descriptive_name>")   # response carries osm_path
   

2. Ask the user which weather file to use. Weather files are SERVER-side; list the available ones (remote clients can stage their own EPW with request_upload):

   list_weather_files()
   

3. Run the simulation with the path the save returned:

   run_simulation(osm_path=<osm_path from save>, epw_path="<epw_path>")
   

4. Poll until complete — first status check ~60 seconds after submitting, then no more than once per minute (every 2-3 minutes for long simulations):

   get_run_status(run_id=<id>)
   

   Each status check costs a full LLM turn. If your client can wait (e.g. a shell sleep), wait before checking rather than polling back-to-back.

5. Extract key results:

   extract_summary_metrics(run_id=<id>)
   extract_end_use_breakdown(run_id=<id>)
   

6. Present a summary with:

   • Total site energy (GJ and kBtu)
   • EUI (MJ/m2 and kBtu/ft2)
   • Unmet heating/cooling hours
   • End-use breakdown by category (heating, cooling, lighting, equipment, fans, pumps)

Error Handling

• If no model is loaded, tell the user to load one first
• If simulation fails, show get_run_logs(run_id=<id>) output
• Common failure: missing weather file or design days