llm.py

from groq import Groq
import os
from dotenv import load_dotenv
load_dotenv()

client = Groq(api_key=os.getenv("GROQ_API_KEY"))

def generate_answer(context, question):
    prompt = f"""

You are an AI assistant. Answer ONLY from the given context.



Context:

{context}



Question:

{question}

"""

    response = client.chat.completions.create(
        model="openai/gpt-oss-20b",
        messages=[{"role": "user", "content": prompt}],
        temperature=0.3
    )

    return response.choices[0].message.content