llm-app: Support for Assistant APIs?

[paulcoghlan]

It would be great if the llmclient could support more endpoints to enable the assistant features, namely:

        CreateAssistant(ctx context.Context, request openai.AssistantRequest) (openai.Assistant, error)
	CreateThread(ctx context.Context, request openai.ThreadRequest) (openai.Thread, error)
	CreateMessage(ctx context.Context, threadID string, request openai.MessageRequest) (msg openai.Message, err error)
	CreateRun(ctx context.Context, threadID string, request openai.RunRequest) (run openai.Run, err error)
	RetrieveRun(ctx context.Context, threadID string, runID string) (run openai.Run, err error)
	SubmitToolOutputs(ctx context.Context, threadID string, runID string, request openai.SubmitToolOutputsRequest) (response openai.Run, err error)
	ListMessage(ctx context.Context, threadID string) (openai.MessagesList, error)

Happy to help with a PR if necessary.

[gerboland]

Would be worth following up with the team to see what they've done since. Are they using OpenAI's api's directly? If so, we should try include the functionality in the llm-app...

[jtheory]

( following up in Slack with @paulcoghlan )