council-python-agent

A general purpose interative Python coding assistant implemented with Council. Runnable as a standalone Python script or as a Flask/JavaScript app with a UI.

There is a custom controller, which is a development version of LLMInstructController

• src/python_agent/controller.py

Demo Video and Files

If you just want to see how this Agent works, check out the following video where we use the Python Agent to help build a particle simulator!

• 📺 Using the Python Agent to Create a Particle Simulator
• 📄 Demo Code and Transcripts

Prerequisites

To run the demo as-is, you'll need an OpenAI API Key. Though it's possible to use GPT 3.5 Turbo, we highly recommend using GPT-4 for this demo.

Council Components

We define an Agent with the following structure:

graph TB
    User[User] -->|Messages| LLMInstructController
    subgraph Agent[Agent]
        LLMInstructController[LLMInstructController]
        BasicEvaluator[BasicEvaluator]
        subgraph Chains
            Chain1[code_generation_chain]
            Chain2[code_execution_chain]
            Chain3[error_correction_chain]
            Chain4[general_chain]
            Chain5[direct_to_user_chain]
        end
        LLMInstructController -->|"chain;score;instructions"| Chains
        Chains --> BasicEvaluator
        BasicEvaluator --> LLMInstructController
    end

Loading

Demo setup

Before running the demo, you'll need to:

1. Install Council - pip install council-ai==0.0.9
2. Create a Python virtual environment for code execution (see below)
3. Populate your .env file with your virtual environment path and your OpenAI API key

Virtual environment setup

You need to create a Python virtual environment.

1. In your Terminal, cd to src/python_agent
2. Run python -m venv code_sandbox
3. Run source code_sandbox/bin/activate
4. Install any packages you plan to use in your interactive coding, e.g. pip install pygame
5. Take note of the full path to your virtual environment's bin directory (e.g. your-home-directory/council-python-agent/src/python_agent/code_sandbox/bin)

Running the demo

To run the demo,

1. Navigate to the src directory in your Terminal
2. Run python app.py
3. Open the src directory in your file browser (e.g. Finder) then open index.html in your browser.
4. If you see "No code to display." in the code window and "Ready." in the log window below, your app is working and ready.

Troubleshooting

• Keep the Terminal window running app.py open and visible. If there are unhandled errors, it will let you know.
• If you get unhandled errors (e.g. 500 errors) you may need to restart both app.py and your browser. We're working on improving this in future revisions.