Test multiple agents about smallville HOT 2 OPEN

Need to test interactions with multiple agents, how they react to observations of seeing other agents, planning with other agents, and conversations.

Multiple agents seem to be able to plan coherently together (as long as their current locations or the state of the environment is sent to the agents memory)

I will run further tests on this, still need to fix some issues regarding conversations and reactions. I'm experimenting with the functions call api which should help with those things

from smallville.

Need to test interactions with multiple agents, how they react to observations of seeing other agents, planning with other agents, and conversations.

Multiple agents seem to be able to plan coherently together (as long as their current locations or the state of the environment is sent to the agents memory)

I will run further tests on this, still need to fix some issues regarding conversations and reactions. I'm experimenting with the functions call api which should help with those things

I tested multiple agents further and there is some room for improvement.

I fixed agent conversations and reactions (#14)
I moved the agent locations and states from the agents memory to the world description

However the example game still needs to use the client to add reactable observations and display conversations from the server. Also, I tested out the function calling API and it didn't seem to improve the results so I reverted back to the original without function calling. Emojis were often missing from the function calling although they were marked as required. I wasn't getting this problem when using the old prompts.

from smallville.