app.py

import gradio as gr
import os
import requests
from dotenv import load_dotenv

# Load GROQ API key from environment
load_dotenv()
GROQ_API_KEY = os.environ.get("GROQ_API_KEY")
GROQ_API_URL = "https://api.groq.com/openai/v1/chat/completions"
MODEL_NAME = "llama3-8b-8192"

SYSTEM_PROMPT = """
You are an intelligent and helpful machine learning assistant. 
Your task is to help users choose appropriate machine learning models based on their problem type 
(classification, regression, clustering, etc.), dataset characteristics, accuracy needs, and resource constraints. 
Explain recommendations clearly and concisely, and offer guidance on why a model is suitable.
"""

def query_groq(message, chat_history):
    headers = {
        "Authorization": f"Bearer {GROQ_API_KEY}",
        "Content-Type": "application/json"
    }
    messages = [{"role": "system", "content": SYSTEM_PROMPT}]
    for user, bot in chat_history:
        messages.append({"role": "user", "content": user})
        messages.append({"role": "assistant", "content": bot})
    messages.append({"role": "user", "content": message})

    response = requests.post(GROQ_API_URL, headers=headers, json={
        "model": MODEL_NAME,
        "messages": messages,
        "temperature": 0.7
    })

    if response.status_code == 200:
        reply = response.json()["choices"][0]["message"]["content"]
        return reply
    else:
        return f"Error {response.status_code}: {response.text}"

def generate_question(problem_type, notes):
    return f"I am working on a {problem_type} problem. {notes} What machine learning model would be suitable?"

def respond(message, chat_history):
    bot_reply = query_groq(message, chat_history)
    chat_history.append((message, bot_reply))
    return "", chat_history

with gr.Blocks() as demo:
    gr.Markdown("## 🤖 ML Model Selector Chatbot (Powered by GROQ LLM)")
    gr.Markdown("Use the dropdown to specify your ML problem and get recommendations!")

    with gr.Row():
        problem_type = gr.Dropdown(
            choices=["Classification", "Regression", "Clustering", "Anomaly Detection", "Dimensionality Reduction"],
            label="Select your ML Problem Type",
            value="Classification"
        )
        notes = gr.Textbox(label="Add extra details (optional)")

    generate_btn = gr.Button("Generate Question")
    chatbot = gr.Chatbot()
    msg = gr.Textbox(label="Ask your question or use generated one above")
    clear = gr.Button("Clear Chat")
    state = gr.State([])

    def fill_generated_question(ptype, detail):
        return generate_question(ptype, detail)

    generate_btn.click(fill_generated_question, [problem_type, notes], msg)
    msg.submit(respond, [msg, state], [msg, chatbot])
    clear.click(lambda: ([], []), None, [chatbot, state])

demo.launch()