src/llm.py

import os
from dotenv import load_dotenv
from huggingface_hub import InferenceClient
from src.prompt import system_instruction

load_dotenv()

HF_TOKEN = os.getenv("HF_TOKEN")

client = InferenceClient(
    model="mistralai/Mistral-7B-Instruct-v0.2",
    token=HF_TOKEN
)

messages = [
    {"role": "system", "content": system_instruction}
]


def ask_order(messages):
    # Fixed: Changed chat_completions.create to chat_completion
    response = client.chat_completion(
        messages=messages,
        max_tokens=500
    )
    
    # Fixed: Accessing the attribute correctly for InferenceClient
    return response.choices[0].message.content