main.py

import requests
from pydantic_settings import BaseSettings

class Settings(BaseSettings):
    llm_url: str
    llm: str

settings = Settings()

print(settings.llm_url)

def get_llm_endpoint():
    """Returns the complete LLM API endpoint URL"""
    base_url = settings.llm_url
    return f"{base_url}/chat/completions"

def get_model_name():
    """Returns the model name to use for API requests"""
    return settings.llm

def call_llm_api(user_message):
    """Calls the LLM API and returns the response with caching"""
    chat_request = {
        "model": get_model_name(),
        "messages": [
            {
                "role": "system",
                "content": "You are a helpful assistant. Please provide structured responses using markdown formatting. Use headers (# for main points), bullet points (- for lists), bold (**text**) for emphasis, and code blocks (```code```) for code examples. Organize your responses with clear sections and concise explanations."
            },
            {
                "role": "user",
                "content": user_message
            }
        ]
    }
    
    headers = {"Content-Type": "application/json"}
    
    # Send request to LLM API
    response = requests.post(
        get_llm_endpoint(),
        headers=headers,
        json=chat_request,
        timeout=30
    )
    
    # Check if the status code is not 200 OK
    if response.status_code != 200:
        raise Exception(f"API returned status code {response.status_code}: {response.text}")
    
    # Parse the response
    chat_response = response.json()
    
    # Extract the assistant's message
    if chat_response.get('choices') and len(chat_response['choices']) > 0:
        return chat_response['choices'][0]['message']['content'].strip()
    
    raise Exception("No response choices returned from API")

if __name__ == '__main__':
    response = call_llm_api("Do you know poet, Ronelda Kamfer?")
    print(response)