Any Ideas how i can achieve these things? Welcome Message/Multiple Personalities or Agents/Handling of Multilingual Documents/User Feedback

[cloudhunnter]

So, i do have some ideas how to improve my projects based on this repo. Does anyone have an idea, how i could approach these topics?

1. Welcome Message and Topic Focus/Multiple personalities?
   The chat is initialized by the chatbot with a brief introduction and a question about which Topic the user wants to talk about. The goal is that by answering the question, the user directly sets a filter for the index or a specific system prompt is getting set for the further conversation. This could also incloud a switch between different indexes based on the topic. Sounds somehow like a function call, i guess, which switches between different defined indexvariables and parameters?

2. How could i handle this problem? Access to multilingual documents
   The user should always receive documents in the language they are using. As a fallback solution, if no documents in the user’s language are found, documents in a specific language should be provided. The focus is only about the citation - the llm is able to translate, that's not the problem. My data is focused mainly on technical information about machines, so most of the time people need to check in the citations to verify correctness.

3. Integration of a user feedback system with thumbs-up, thumbs-down, and logback for both good and bad prompts and responses.
   I guess, i have seen an issue about this topic in the repo. I think this would be a huge improvement for the users and the admin, so based on the feedback the data or settings of the chatbot can be optimized.