Modifications to how the API is accepted

[jswope00]

I'd like to modify how this application uses the API.

1. I don't want to show the "Enter OpenAI API Key" screen at the start. The opening screen should be a working version of the app.
   1. FYI, right now you don't need to enter a real key in that first screen. You can enter anything to get rid of the screen and it will use the key saved in the environment variables.
2. By default, i want people to be able to use my key to generate responses with the GPT Turbo-3.5 model.
3. I want to allow users to enter their own key to also use the GPT-4 model.
4. I want the sidebar's "OpenAI Models" selector to allow the user to select which model(s) they'd like to get responses from.

[jswope00]

@yasir1brahim, good progress.

I made and committed a couple tweaks.

1 is done.

2 is done.

3 seems done too.

4 still needs some work.

If a user hasn't entered a valid OpenAI token, then the system should only run the GPT3.5 model. I just added functionality that waits until a user enters a token to be able to have the option to add GPT-4. You should build on that to validate that the key is valid before adding the gpt-4 option. Once a user has put in a valid token, then the system can run 3.5, 4 (on their token), or both. If it makes things simpler, then once a valid token it input it can "override" my token and both 3.5 and 4 run on it. The output should be responsive to which Models are selected. So the app only runs prompts and shows answers for the model(s) that are selected.