app.py

import gradio as gr
from smolagents import CodeAgent, InferenceClientModel, Tool
import os
from PIL import Image
import tempfile
import base64
from io import BytesIO






# Initialize the image generation tool
image_generation_tool = Tool.from_space(
    "black-forest-labs/FLUX.1-schnell",
    name="image_generator",
    description="Generate an image from a prompt"
)

# Initialize the model and agent
model = InferenceClientModel("Qwen/Qwen2.5-Coder-32B-Instruct")
agent = CodeAgent(tools=[image_generation_tool], model=model)

def process_message(message, history):
    """
    Process user message with the SmolagentsAI agent
    
    Args:
        message: User input message
        history: Chat history
    
    Returns:
        Updated chat history
    """
    try:
        # Run the agent with the user's message
        response = agent.run(message)
        
        # Check if response contains image information
        if hasattr(response, 'content') and isinstance(response.content, list):
            # Handle multi-modal response (text + image)
            text_parts = []
            images = []
            
            for item in response.content:
                if hasattr(item, 'type'):
                    if item.type == 'text':
                        text_parts.append(item.text)
                    elif item.type == 'image':
                        images.append(item)
            
            # Combine text parts
            text_response = ' '.join(text_parts) if text_parts else str(response)
            
            # Add images to the response if any
            if images:
                text_response += f"\n\n🖼️ Generated {len(images)} image(s)"
        else:
            # Handle text-only response
            text_response = str(response)
        
        # Update chat history
        history.append((message, text_response))
        return history
        
    except Exception as e:
        error_message = f"❌ Error: {str(e)}"
        history.append((message, error_message))
        return history

def clear_chat():
    """Clear the chat history"""
    return []

def create_interface():
    """Create the Gradio interface"""
    
    # Custom CSS for better styling
    custom_css = """
    .gradio-container {
        max-width: 900px !important;
        margin: auto !important;
    }
    .chat-message {
        padding: 10px !important;
        margin: 5px 0 !important;
        border-radius: 10px !important;
    }
    .user-message {
        background-color: #e3f2fd !important;
        margin-left: 20% !important;
    }
    .bot-message {
        background-color: #f5f5f5 !important;
        margin-right: 20% !important;
    }
    """
    
    with gr.Blocks(
        title="🤖 AI Agent with Image Generation",
        theme=gr.themes.Soft(),
        css=custom_css
    ) as demo:
        
        # Header
        gr.Markdown("""
        # 🤖 AI Agent with Image Generation
        
        This intelligent agent can help you with various tasks and generate images using FLUX.1-schnell!
        
        **What you can do:**
        - Ask questions and get intelligent responses
        - Request image generation with detailed prompts
        - Combine text and image requests in natural language
        
        **Example prompts:**
        - "Generate an image of a sunset over mountains"
        - "Create a logo for a tech startup"
        - "Show me a futuristic city"
        - "Help me write code and create an illustration for it"
        """)
        
        # Chat interface
        with gr.Row():
            with gr.Column(scale=1):
                # Chat history
                chatbot = gr.Chatbot(
                    label="💬 Chat with AI Agent",
                    height=500,
                    show_copy_button=True,
                    show_share_button=True,
                    avatar_images=("👤", "🤖"),
                    bubble_full_width=False
                )
                
                # Input area
                with gr.Row():
                    msg_input = gr.Textbox(
                        label="Your Message",
                        placeholder="Ask me anything or request an image generation...",
                        lines=2,
                        scale=4
                    )
                    send_btn = gr.Button("Send 🚀", variant="primary", scale=1)
                
                # Control buttons
                with gr.Row():
                    clear_btn = gr.Button("Clear Chat 🗑️", variant="secondary")
                    examples_btn = gr.Button("Show Examples 💡", variant="secondary")
        
        # Examples section (initially hidden)
        examples_section = gr.Markdown(
            """
            ### 💡 Example Prompts:
            
            **Image Generation:**
            - "Generate a realistic photo of a golden retriever playing in a park"
            - "Create a minimalist logo design for a coffee shop"
            - "Show me an abstract art piece with vibrant colors"
            - "Generate a cyberpunk-style illustration of a neon city"
            
            **Text + Image Combination:**
            - "Explain quantum computing and create a visual representation"
            - "Write a short story about space exploration and generate an accompanying image"
            - "Help me design a website layout and show me a mockup"
            
            **General AI Tasks:**
            - "Help me write a Python function to sort a list"
            - "Explain the concept of machine learning in simple terms"
            - "Create a meal plan for the week"
            """,
            visible=False
        )
        
        # Event handlers
        def send_message(message, history):
            if not message.strip():
                return history, ""
            return process_message(message, history), ""
        
        def toggle_examples(examples_visible):
            return gr.update(visible=not examples_visible)
        
        # Wire up the events
        send_btn.click(
            send_message,
            inputs=[msg_input, chatbot],
            outputs=[chatbot, msg_input]
        )
        
        msg_input.submit(
            send_message,
            inputs=[msg_input, chatbot],
            outputs=[chatbot, msg_input]
        )
        
        clear_btn.click(
            clear_chat,
            outputs=chatbot
        )
        
        # Examples toggle
        examples_visible = gr.State(False)
        examples_btn.click(
            lambda visible: (gr.update(visible=not visible), not visible),
            inputs=examples_visible,
            outputs=[examples_section, examples_visible]
        )
        
        # Footer
        gr.Markdown("""
        ---
        **Powered by:**
        - 🧠 **Model**: Qwen2.5-Coder-32B-Instruct
        - 🎨 **Image Generation**: FLUX.1-schnell
        - 🤖 **Framework**: SmolagentsAI
        - 🌐 **Interface**: Gradio
        
        *Built for Hugging Face Spaces*
        """)
    
    return demo

# Launch the interface
if __name__ == "__main__":
    demo = create_interface()
    demo.launch(
        server_name="0.0.0.0",
        server_port=7860,
        share=False,  # Set to False for HF Spaces
        show_error=True
    )