Change the default model: `chatlas.ChatGithub()` vs `chatlas.ChatAuto()`

[chendaniely]

querrychat.init() has this in the code that defaults the initial model to GPT 4.1

    # Default chat function if none provided
    create_chat_callback = create_chat_callback or partial(
        chatlas.ChatOpenAI,
        model="gpt-4.1",
    )

We should change this default behavior to either using either chatlas.ChatGithub() or chatlas.ChatAuto().
This would prevent having the user create a "callback" or partial() if they do not already pay and have OPENAI_API_KEY defined in their environment variables.

If we use chatlas.ChatGithub() they will need to create a github PAT and define GITHUB_TOKEN which is free.
Or we have them specify the model, keys, and anything else as the environment variable for chatlas.ChatAuto(). This way ChatGithub() can also be specified in ChatAuto() and we talk about setting all the environment variables during setup process.