ShowdownLLMPlayer

ShowdownLLMPlayer is an AI-powered Pokémon battle bot designed to play Pokémon Showdown's Random Battles format using advanced language models.

Features

• Utilizes a large language model for strategic decision-making
• Supports Generation 9 Random Battles format
• Analyzes type effectiveness, move impacts, and team compositions
• Provides detailed reasoning for each move selection

Requirements

• Python 3.8+
• OptiLLM for efficient language model inference

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/ShowdownLLMPlayer.git
   cd ShowdownLLMPlayer
   

2. Install the required dependencies:

   pip install -r requirements.txt
   

3. Set up your environment variables: Create a .env file in the project root and add your Pokémon Showdown account credentials and Azure OpenAI API key:

   ACCT_PASSWORD=your_password_here
   INCOMING_USERNAME=your_username_here
   AZURE_OPENAI_API_KEY=your_api_key_here
   

Usage

To start the bot and connect it to Pokémon Showdown:

1. Ensure OptiLLM is running on your local machine (default: http://localhost:8000)

2. Run the main script:

   python play_showdown.py
   

3. The bot will automatically log in and start laddering or accepting challenges based on the configuration in play_showdown.py.

How It Works

ShowdownLLMPlayer uses OptiLLM to efficiently leverage a large language model for analyzing the current battle state, including:

• Team compositions
• Type advantages/disadvantages
• Available moves and their potential impacts
• Battle history

Based on this analysis, the model generates a detailed reasoning process and selects the most strategic move for each turn.

Customization

You can modify the ShowdownLLMPlayer class in ShowdownLLMPlayer.py to adjust the bot's behavior or integrate different language models. The use_local_llm parameter allows switching between local and Azure-hosted models.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is open source and available under the Apache License 2.0.