Assignment

In this assignment, we’ll take all our knowledge of adding sampling implementations to the client and server and apply it to a fun area, namely, a mystery game, and the conversational part. The idea is to create a character that’s interesting to talk to over pre-programmed responses.

Here’s how it should work:

• User: Talk to character N
• Server: Retrieve character info
• Server: Send sampling request with character info
• Client: Receive sampling request and produce a response using the LLM
• Server: Store response for caching and logging

To help you, imagine that a character is defined in a JSON file like so:

[
  {
    "id": "1",
    "name": "Monsieur Lestrange",
    "description": "a 600 year old vampire",
    "personality": "very polite and will tell you a great deal of what
      it's like paying the electricity bill of a 1200...