autogen-aider-docker

How to use

• Requires Docker to be installed.
• If using as-is, you will also need an OpenAI API key (but autogen can be configured to use other models).
• Optionally, build the docker image yourself:
  • Run docker build -t autogen-aider ..
  • Replace all instances of danum/autogen-aider in docker-compose.yml with autogen-aider.
• Create .env and add OpenAI API key: echo OPENAI_API_KEY=my-api-key > .env.
• Copy my-project and rename with your project name: cp -r my-project my-new-project.
• Specify your new project folder as bind-mount source via env var: export MOUNT_SOURCE=./my-new-project.
  • Alternatively, simply include -e MOUNT_SOURCE=./my-new-project after docker compose run, below.
• Optionally, edit 'main.py' in my-new-project to describe the task that you want autogen to work on.
• Run docker compose run autogen-aider.
  • Alternatively, run docker compose run --rm autogen-aider /bin/bash to prevent auto-run of autogen.
    • From within the container, execute autogen like this: python main.py
• The code generated by autogen can be found in generated-content, in my-new-project.
• Optionally, if you haven't exited the docker container yet, run aider --model gpt-4-1106-preview main.py (or any other file) to continue coding, with aider as your pair programmer.