app.py

import gradio as gr
from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
import torch

# ---------------- Agent 1: Intent Classifier ----------------
intent_classifier = pipeline("zero-shot-classification", model="MoritzLaurer/DeBERTa-v3-base-mnli-fever-anli")


def detect_intent(text):
    labels = {
        "weather": "The user wants to know the weather.",
        "faq": "The user is asking for help.",
        "smalltalk": "The user is making casual conversation."
    }
    best_intent = "smalltalk"
    best_score = 0
    for label, hypothesis in labels.items():
        result = intent_classifier(text=text, text_pair=hypothesis)[0]
        if result['label'] == 'ENTAILMENT' and result['score'] > best_score:
            best_score = result['score']
            best_intent = label
    return best_intent

# ---------------- Agent 2: Domain Logic ----------------
def handle_logic(intent):
    if intent == "weather":
        return "It's sunny and 26°C today."
    elif intent == "faq":
        return "To reset your password, use the 'Forgot Password' option."
    else:
        return "That's great! Anything else you'd like to talk about?"

# ---------------- Agent 3: Natural Language Generation ----------------
tokenizer = AutoTokenizer.from_pretrained("microsoft/DialoGPT-small")
model = AutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-small")

def generate_reply(prompt):
    input_ids = tokenizer.encode(prompt + tokenizer.eos_token, return_tensors='pt')
    output_ids = model.generate(input_ids, max_length=100, pad_token_id=tokenizer.eos_token_id)
    response = tokenizer.decode(output_ids[:, input_ids.shape[-1]:][0], skip_special_tokens=True)
    return response

# ---------------- Chatbot Pipeline ----------------
def chatbot(user_input):
    intent = detect_intent(user_input)
    logic = handle_logic(intent)
    response = generate_reply(logic)
    return response

# ---------------- Gradio UI ----------------
gr.Interface(
    fn=chatbot,
    inputs=gr.Textbox(label="User Input"),
    outputs=gr.Textbox(label="Chatbot Response"),
    title="3-Agent Chatbot",
    description="Intent Detection → Domain Logic → Natural Language Generation"
).launch()