import gradio as gr from PIL import Image, ImageDraw, ImageFont import random import time # Mock function to simulate Google DeepMind Gemini 2.0 Flash # In a production environment, you would replace this logic with the actual Google Generative AI SDK def mock_gemini_generation(prompt, img1, img2, img3, history): """ Simulates the image generation/editing process. Since this is a 'free demo', we generate a placeholder image based on the prompt. """ # 1. Construct the User Message for the Chat History user_content = [] # Add text prompt if present if prompt and prompt.strip(): user_content.append({"type": "text", "text": prompt}) # Add images if present (up to 3 spots) images = [img1, img2, img3] for i, img in enumerate(images): if img is not None: user_content.append({"type": "image", "image": img}) if not user_content: # If empty, just add a placeholder text user_content.append({"type": "text", "text": "Generate something..."}) user_message = {"role": "user", "content": user_content} history.append(user_message) # 2. Simulate "Thinking" / Processing time time.sleep(1.5) # 3. Generate a Result (Mock) # Create a dynamic image based on the prompt length/content to simulate generation width, height = 512, 512 # Create a random background color bg_color = (random.randint(50, 255), random.randint(50, 255), random.randint(50, 255)) result_image = Image.new('RGB', (width, height), color=bg_color) draw = ImageDraw.Draw(result_image) # Try to use a default font, otherwise load a simple one try: # Attempting to load a font, fallback to default if not found font = ImageFont.truetype("/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf", 40) except: font = ImageFont.load_default() # Draw some "AI Art" shapes for _ in range(10): x1 = random.randint(0, width) y1 = random.randint(0, height) x2 = x1 + random.randint(-100, 100) y2 = y1 + random.randint(-100, 100) color = (random.randint(0, 255), random.randint(0, 255), random.randint(0, 255)) draw.ellipse([x1, y1, x2, y2], fill=color, outline="white") # Add text overlay based on input text_overlay = f"Gemini 2.0 Flash\nSimulated Output\nPrompt: {prompt[:20]}..." # Calculate text position (centered) # Using textbbox for better centering calculation in newer Pillow versions try: bbox = draw.textbbox((0, 0), text_overlay, font=font) text_width = bbox[2] - bbox[0] text_height = bbox[3] - bbox[1] except AttributeError: # Fallback for older Pillow versions text_width, text_height = draw.textsize(text_overlay, font=font) position = ((width - text_width) / 2, (height - text_height) / 2) # Add a shadow for readability draw.text((position[0]+2, position[1]+2), text_overlay, fill="black", font=font) draw.text(position, text_overlay, fill="white", font=font) # 4. Construct the Assistant Message bot_message = { "role": "assistant", "content": [ {"type": "text", "text": "Here is the generated image based on your inputs:"}, {"type": "image", "image": result_image} ] } history.append(bot_message) # 5. Return updated history and clear inputs return history, gr.Textbox(value=""), None, None, None def clear_chat(): return [], None, None, None, "" # Gradio 6 Application Structure # Note: In Gradio 6, theme, css, js go in demo.launch(), NOT in gr.Blocks() with gr.Blocks() as demo: # Header Section gr.HTML("""

🎨 Image Generation & Editing

Powered by Google DeepMind Gemini 2.0 Flash (Demo)

Built with anycoder

""") # Chat Interface Area chatbot = gr.Chatbot( label="Chat History", height=500, show_copy_button=True, avatar_images=(None, "https://www.google.com/favicon.ico") # Optional avatars ) # Input Section with gr.Row(): # The "Three Blank Spots" for dragging images with gr.Column(scale=1): img_spot_1 = gr.Image( label="Image Input 1", sources=["upload", "clipboard", "webcam"], type="pil", height=200 ) with gr.Column(scale=1): img_spot_2 = gr.Image( label="Image Input 2", sources=["upload", "clipboard", "webcam"], type="pil", height=200 ) with gr.Column(scale=1): img_spot_3 = gr.Image( label="Image Input 3", sources=["upload", "clipboard", "webcam"], type="pil", height=200 ) # Text Prompt Area with gr.Row(): text_prompt = gr.Textbox( label="Text Prompt", placeholder="Describe the image you want to generate or how to edit the uploaded images...", lines=2, scale=4, container=False ) submit_btn = gr.Button("Generate ✨", variant="primary", scale=1, size="lg") clear_btn = gr.Button("Clear 🗑️", variant="secondary", scale=1) # Event Listeners submit_btn.click( fn=mock_gemini_generation, inputs=[text_prompt, img_spot_1, img_spot_2, img_spot_3, chatbot], outputs=[chatbot, text_prompt, img_spot_1, img_spot_2, img_spot_3], api_visibility="public" # Gradio 6 syntax ) clear_btn.click( fn=clear_chat, inputs=None, outputs=[chatbot, img_spot_1, img_spot_2, img_spot_3, text_prompt], api_visibility="public" ) # Allow "Enter" key in textbox to trigger generation text_prompt.submit( fn=mock_gemini_generation, inputs=[text_prompt, img_spot_1, img_spot_2, img_spot_3, chatbot], outputs=[chatbot, text_prompt, img_spot_1, img_spot_2, img_spot_3], api_visibility="public" ) # Launch the App with Gradio 6 Syntax # All app-level parameters (theme, css, footer_links) go here! demo.launch( theme=gr.themes.Soft( primary_hue="blue", secondary_hue="cyan", neutral_hue="slate", font=gr.themes.GoogleFont("Inter") ), footer_links=[ {"label": "Google DeepMind", "url": "https://deepmind.google/"}, {"label": "Built with anycoder", "url": "https://huggingface.co/spaces/akhaliq/anycoder"} ] )