app.py

import gradio as gr
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
from fastapi import FastAPI, Request
from fastapi.middleware.cors import CORSMiddleware
from gradio.routes import mount_gradio_app
import uvicorn

# -------------------------------------------------
# 1. Load model
# -------------------------------------------------
print("Initializing DialoGPT-medium model...")
model_name = "microsoft/DialoGPT-medium"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

if tokenizer.pad_token is None:
    tokenizer.pad_token = tokenizer.eos_token

print("DialoGPT-medium loaded successfully!")

# -------------------------------------------------
# 2. Helper: Generate a response
# -------------------------------------------------
def generate_response(message: str, chat_history: list):
    if not message.strip():
        return "Please enter a message."

    # Build the conversation context
    conversation = ""
    for user, bot in chat_history:
        conversation += f"User: {user}\nBot: {bot}\n"
    conversation += f"User: {message}\nBot:"

    inputs = tokenizer.encode(conversation, return_tensors="pt", max_length=1024, truncation=True)

    with torch.no_grad():
        outputs = model.generate(
            inputs,
            max_length=inputs.shape[1] + 128,
            pad_token_id=tokenizer.eos_token_id,
            do_sample=True,
            temperature=0.7,
            top_k=50,
            top_p=0.95,
            repetition_penalty=1.2,
        )

    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    response = response.split("Bot:")[-1].strip()
    if "\nUser:" in response:
        response = response.split("\nUser:")[0]

    return response

# -------------------------------------------------
# 3. Gradio chat handler
# -------------------------------------------------
def chat_fn(message: str, history: list):
    response = generate_response(message, history or [])
    history.append((message, response))
    return "", history  # clear textbox, update chat

# -------------------------------------------------
# 4. Build the Gradio UI
# -------------------------------------------------
example_questions = [
    "Hello! How are you today?",
    "What can you help me with?",
    "Tell me about artificial intelligence.",
    "What's your favorite programming language?",
    "Can you explain machine learning?",
    "How does a neural network work?"
]

with gr.Blocks(
    theme=gr.themes.Soft(primary_hue="blue", secondary_hue="green"),
    title="GihonTech - AI Conversation Assistant"
) as demo:

    gr.Markdown("# 🤖 GihonTech AI Conversation Assistant")
    gr.Markdown("Chat with an AI powered by **DialoGPT-medium**.")

    with gr.Row():
        with gr.Column(scale=3):
            chatbot = gr.Chatbot(label="Conversation", height=500)

            with gr.Row():
                msg = gr.Textbox(
                    label="Your Message",
                    placeholder="Type your message here...",
                    lines=2,
                    scale=4,
                )
                send = gr.Button("Send", variant="primary", scale=1)

            clear = gr.Button("Clear Chat", variant="secondary")

        with gr.Column(scale=1):
            gr.Markdown("### Example Questions")
            for q in example_questions:
                gr.Button(q[:40] + ("..." if len(q) > 40 else ""), size="sm").click(
                    lambda x=q: x, outputs=msg
                )
            gr.Markdown("---")
            gr.Markdown("### Model Info")
            gr.Textbox(
                value="DialoGPT-medium: Loaded ✅",
                label="Model Status",
                interactive=False,
            )
            gr.Markdown(
                """
                **Features**  
                - Context-aware replies  
                - Conversation memory  

                **Tips**  
                - Ask clear, simple questions  
                - Use *Clear Chat* to start over  
                """
            )

    # Wire up events
    send.click(chat_fn, inputs=[msg, chatbot], outputs=[msg, chatbot])
    msg.submit(chat_fn, inputs=[msg, chatbot], outputs=[msg, chatbot])
    clear.click(lambda: ([], ""), outputs=[chatbot, msg])

# -------------------------------------------------
# 5. FastAPI app + Lambda route
# -------------------------------------------------
fastapi_app = FastAPI()

# Allow AnythingLLM / frontend CORS access
fastapi_app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_methods=["*"],
    allow_headers=["*"],
)

@fastapi_app.post("/lambda")
async def lambda_endpoint(req: Request):
    payload = await req.json()
    user_msg = payload.get("data", [""])[0]
    response = generate_response(user_msg, [])
    return {"data": [response]}

# Mount Gradio app inside FastAPI
mount_gradio_app(fastapi_app, demo, path="/")

# -------------------------------------------------
# 6. Run the combined FastAPI + Gradio app
# -------------------------------------------------
if __name__ == "__main__":
    uvicorn.run(fastapi_app, host="0.0.0.0", port=7860)