summarizer.py

import os
from dotenv import load_dotenv
from groq import Groq

load_dotenv()

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

MODEL = "meta-llama/llama-4-scout-17b-16e-instruct"  # You can change this to any Groq-supported model

async def summarize_text(text: str) -> str:
    prompt = (
        "You are a helpful assistant that summarizes text clearly and concisely.\n\n"
        f"Text to summarize:\n{text}\n\nSummary:"
    )

    completion = client.chat.completions.create(
        model=MODEL,
        messages=[
            {"role": "system", "content": "You summarize texts into clear, short bullet points."},
            {"role": "user", "content": prompt}
        ],
        temperature=0.5,
        max_tokens=512,
        top_p=1,
        stream=False,  # We keep this False for simplicity
    )

    return completion.choices[0].message.content.strip()