💡 [Feature]: Add support for picking LLM model for GitHub Copilot chat participant

[Adam-it]

🎯 Aim of the feature

Currently in preview GitHub Copilot added the functionality to select the LLM model you want to use in the response

We should check if we may adapt SPFx Toolkit Chat participant to also use it. Currently we use GPT 4o

📷 Images (if possible) with expected result

No response

🤔 Additional remarks or comments

No response

[Adam-it]

so it is described here
https://code.visualstudio.com/api/extension-guides/language-model
"Once you've built the prompt for the language model, you first select the language model you want to use with the selectChatModels method. This method returns an array of language models that match the specified criteria. If you are implementing a chat participant, we recommend that you instead use the model that is passed as part of the request object in your chat request handler. This ensures that your extension respects the model that the user chose in the chat model dropdown. Then, you send the request to the language model by using the sendRequest method."

example: https://github.com/microsoft/vscode-extension-samples/blob/main/chat-sample/src/simple.ts

[Adam-it]

prototype

we should:

1. remove this

   vscode-viva/src/chat/PromptHandlers.ts

   Line 27 in 5919011

   const [model] = await vscode.lm.selectChatModels({ vendor: 'copilot', family: 'gpt-4o' });

   and take the model from request
2. for /manage we should still ensure gpt-4o LLM as it has most token possible and brings currently best results

[Adam-it]

Since I cannot seem to make it work with latest VS Code update lets wait a few weeks until it will get out of insiders and into standard VS Code version before we open it up