Multi-user conversations in a thread about gptdiscord

[Kav-K]

This has been deprioritized, I will do async for requests and image concat before this gets any thought :), although, this is very easy so it'd be great if someone wants to just drop in a quick PR

[Kav-K]

I take back what I said about this being very easy earlier, this is super tough with the current architecture, everything is keyed to 1 user, and not keyed on threads... will continue looking into it

[sudoskys]

Multiple users in one session context?
This could be achieved by inserting other users' records, perhaps using a mixed message bucket approach. But it would require an additional shared message bucket to record the group's chat.

[Kav-K]

Can you elaborate a bit more on what you mean by mixed message buckets? :)

[sudoskys]

Can you elaborate a bit more on what you mean by mixed message buckets? :)

Each user has a message bucket
{ask reply time}
This is used to remember chat context.

Then each group also has a public bucket of messages.
This allows you to combine them, filter them, match and mix content.

like https://github.com/sudoskys/openai-kira/blob/main/src/openai_kira/utils/data.py#L175

Content can be produced dynamically each time a request is made to the Api.

I've included in https://github.com/sudoskys/openai-kira/blob/main/src/openai_kira/Chat/__init__.py#L184
Multi-dimensionally calculate the similarity of each entry to the prompt to achieve a match.

The message bucket is not infinitely long, preventing the traversal calculation from taking too long.
Although it may seem primitive, this is the method I am using to retrieve matches.