app.py

import os
import gradio as gr
from ultralytics import YOLO
from PIL import Image, ImageDraw
from openai import AzureOpenAI

# Load YOLOv8 model
model = YOLO("yolov8m.pt")

# Azure OpenAI configuration
endpoint = "https://kendrickfilbert-9824-resource.cognitiveservices.azure.com/"
deployment = "gpt-5.3-chat-ken"
api_version = "2024-12-01-preview"

subscription_key = os.environ.get("AZURE_OPENAI_API_KEY")
if not subscription_key:
    raise ValueError(
        "AZURE_OPENAI_API_KEY is not set! "
        "Add it as a secret in your Hugging Face Space settings."
    )

client = AzureOpenAI(
    api_version=api_version,
    azure_endpoint=endpoint,
    api_key=subscription_key,
)

SYSTEM_PROMPT = (
    "You are Ken Chatbot, a helpful and friendly AI assistant. "
    "You can answer questions on any topic. When the user uploads an image, "
    "you may also receive object detection results to discuss."
)


def build_api_messages(chat_history):
    """Convert Gradio chat history to OpenAI API message format."""
    messages = [{"role": "system", "content": SYSTEM_PROMPT}]
    for msg in chat_history:
        messages.append({"role": msg["role"], "content": msg["content"]})
    return messages


def chat_with_bot(message, chat_history):
    """Send a text message to GPT-5.3 and return updated chat history."""
    if not message.strip():
        return chat_history, ""

    # Add user message to Gradio history
    chat_history.append({"role": "user", "content": message})

    # Call Azure OpenAI with full conversation context
    api_messages = build_api_messages(chat_history)
    response = client.chat.completions.create(
        model=deployment,
        messages=api_messages,
        max_completion_tokens=4096,
    )
    bot_reply = response.choices[0].message.content

    # Add assistant reply to Gradio history
    chat_history.append({"role": "assistant", "content": bot_reply})

    return chat_history, ""


def analyze_image(image_path, chat_history):
    """Run YOLOv8 detection, then ask GPT-5.3 to comment on the results."""
    if image_path is None:
        return None, chat_history

    try:
        image = Image.open(image_path).convert("RGB")
        results = model(image)

        detected_objects = []
        image_draw = image.copy()
        draw = ImageDraw.Draw(image_draw)

        for result in results:
            for box in result.boxes.data:
                x1, y1, x2, y2, score, class_id = box.tolist()
                if score > 0.5:
                    class_name = model.names[int(class_id)]
                    detected_objects.append(f"{class_name} ({score:.0%})")
                    draw.rectangle([x1, y1, x2, y2], outline="red", width=3)
                    draw.text((x1, y1 - 12), class_name, fill="red")

        # Build a summary and feed it to GPT-5.3
        if detected_objects:
            detection_summary = ", ".join(detected_objects)
            user_msg = (
                f"[Image uploaded] Objects detected by YOLO: {detection_summary}. "
                "Please describe what you see and any interesting observations."
            )
        else:
            user_msg = (
                "[Image uploaded] No objects were detected. "
                "Let the user know and suggest they try a clearer image."
            )

        chat_history.append({"role": "user", "content": user_msg})

        api_messages = build_api_messages(chat_history)
        response = client.chat.completions.create(
            model=deployment,
            messages=api_messages,
            max_completion_tokens=2048,
        )
        bot_reply = response.choices[0].message.content

        chat_history.append({"role": "assistant", "content": bot_reply})

        return image_draw, chat_history

    except Exception as e:
        error_msg = f"Error processing the image: {str(e)}"
        chat_history.append({"role": "user", "content": "Uploaded an image for analysis"})
        chat_history.append({"role": "assistant", "content": error_msg})
        return None, chat_history


def clear_all():
    """Reset conversation."""
    return None, [], ""


#Gradio UI 
with gr.Blocks(title="Ken Chatbot") as demo:
    gr.Markdown("# Ken Chatbot")
    gr.Markdown("Ask me anything or upload an image for analysis! Powered by **GPT-5.3** + **YOLOv8**.")

    chatbot = gr.Chatbot(elem_id="chatbot", height=480)

    with gr.Row():
        msg = gr.Textbox(
            placeholder="Type your message here...",
            show_label=False,
            scale=4,
        )
        send_btn = gr.Button("Send", variant="primary", scale=1)

    with gr.Row():
        img_upload = gr.Image(type="filepath", label="Upload an image for analysis")
        img_output = gr.Image(label="Detected Objects")

    clear_btn = gr.Button("Clear conversation")

    # Event bindings — pass chat_history as state via chatbot component
    msg.submit(chat_with_bot, [msg, chatbot], [chatbot, msg])
    send_btn.click(chat_with_bot, [msg, chatbot], [chatbot, msg])
    img_upload.change(analyze_image, [img_upload, chatbot], [img_output, chatbot])
    clear_btn.click(clear_all, None, [img_output, chatbot, msg])

demo.launch()