Add GPT response as text aswell

[yantao0527]

Support to add GPT response as text aswell

Just showing the idea, need to debug

[HHousen]

Hello. Thanks for working on this! The message text probably shouldn't be attached to the AudioMessage. Do you think you could use the EventsManager for this?

[yantao0527]

The idea of using event subscription is better than the one of extension audio message. I'm going to update the code via event today.

[yantao0527]

@HHousen I've finished update code and debug successfully.

• vocode-react-sdk
• vocode-react-demo

[HHousen]

Looks good! Just want to fix the scroll bar at vocodedev/vocode-react-demo#4 before merging.

[yantao0527]

• optimize the code to merge the transcript of same sender.
• show scroll bar, but it's ugly.

[shalabhsingh]

@yantao0527 I have been using the useConversation hook as shown here to receive transcripts on the frontend, but I only receive the transcript for the initial message (BaseMessage) set up in my client backend. I believe something has changed since this was merged?

Note-

1. I have already set subscribeTranscript to be true in the useConversation hook.
2. on debugging, I found out that message of type websocket_transcript is received only once for the initial message, which is causing the error

Edit (for others who might struggle with this)-
This error is resolved in the latest version of the code in the latest commit, but is not a part of the latest release v0.1.111. It's better to pip install vocode from the latest commit instead i believe