model_api.py

import requests
import os
import time

API_URL = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct-v0.2"

HF_TOKEN = os.getenv("HF_TOKEN")

headers = {
    "Authorization": f"Bearer {HF_TOKEN}"
}

def query_model(prompt):
    payload = {
        "inputs": prompt,
        "parameters": {
            "max_new_tokens": 800,
            "temperature": 0.7
        }
    }

    while True:
        response = requests.post(API_URL, headers=headers, json=payload)
        result = response.json()

        if isinstance(result, list):
            return result[0]["generated_text"]

        if "estimated_time" in result:
            time.sleep(result["estimated_time"])
        else:
            return f"Error: {result}"