chatter.py

"""Build a simple LLM application"""

import os
from dotenv import load_dotenv

import groq


load_dotenv()
models = [
    "llama-3.1-405b-reasoning",
    "llama-3.1-70b-versatile",
    "llama-3.1-8b-instant",
    "mixtral-8x7b-32768",
]

def get_chatter():
    GROQ_API_KEY = os.environ.get("GROQ_API_KEY")
    groq_client = groq.Groq(api_key=GROQ_API_KEY)

    def send_chat_request(model, query, temperature=0):
        sys_prompt = """You are a helpful virtual assistant. Your name is Ada.
        Your goal is to provide useful and relevant responses to my requests."""

        response = groq_client.chat.completions.create(
            model = model,
            messages = [
                {"role": "system", "content": sys_prompt},
                {"role": "user", "content": query},
            ],
            response_format = {"type": "text"},
            temperature = temperature # accuracy is higher with temp closer to 0. Creativity is higher as value tends towards 1. Higher chance of hallucination
        )
        answer = response.choices[0].message.content
        return answer
    return send_chat_request