updated game setup and messaging

increased max tokens

new messaging system tweaks

redid messaging system

fixed chameleon guess

readded player logging

reconstructed players using langchain

openai default

removing old testing

fixed chameleon not seeing all players responses

added ollama engine

added player logging

separate prompts file

added game and player level logging

fixed duplicate names

added chameleon guessing logic

vote counting + calc winner

added basic logging

added Kani agents

added output parsing for LLMs

updated voting prompts

added llm calls to ai

added validation functions, mapped advanced phases

basic game loop

adding game elements

Updated to langchain 0.1.0, restructured agents

added game output collecting

upgrade Herd to use ReAct reasoning

experimental animal match tool

using agent for chameleon responses, reworked get_likely_animals tool to "discuss"

get_likely_animals tool added, now using GPT-4 for some responses

initial commit