app.py

import google.generativeai as genai
import torch
from diffusers import StableDiffusionPipeline
from PIL import Image
import gradio as gr

# Configure Google Generative AI
GOOGLE_API_KEY = "AIzaSyBn9Ehq5oIqkEov_fmAMT258X6imfNXfvg"  # Replace with your real API key
genai.configure(api_key=GOOGLE_API_KEY)
model = genai.GenerativeModel("models/gemini-2.0-flash")

# Load Stable Diffusion model
model_id = "prompthero/openjourney-v4"
device = "cuda" if torch.cuda.is_available() else "cpu"
print(f"✅ Using device: {device}")

try:
    pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float32)
    pipe.to(device)
    print("✅ Stable Diffusion Model Loaded Successfully!")
except Exception as e:
    print(f"❌ Error loading Stable Diffusion model: {e}")
    exit()

# Art styles (expand if needed)
art_styles = {
    "Cinematic": "cinematic, highly detailed, vivid colors, realistic style",
    "Anime": "anime style, vibrant colors, expressive characters",
    "Watercolor": "watercolor painting, soft colors, artistic style",
    "Cyberpunk": "cyberpunk, neon lights, futuristic, dystopian",
}


def get_summary_or_concept(prompt, num_panels):
    content_prompt = (
        f"Summarize the key concepts or moments of '{prompt}' into {num_panels} brief points, "
        "with 1 concise sentence for each panel, equally distributing the information."
    )

    try:
        response = model.generate_content(content_prompt)
        if response and response.text:
            text_response = response.text.strip()
            panels = [p.strip() for p in text_response.split("\n") if p.strip()]
            return panels[:num_panels] if len(panels) > 1 else [text_response] * num_panels
        else:
            return []
    except Exception as e:
        print(f"❌ Error generating content: {e}")
        return []


def generate_comic_images(story_prompts, style):
    images = []
    for i, panel in enumerate(story_prompts):
        print(f"🎨 Generating image for Panel {i+1}/{len(story_prompts)}: {panel.strip()[:50]}...")

        image_prompt = (
            f"Highly detailed, {art_styles[style]} scene depicting: {panel.strip()}, "
            "no text, no words"
        )

        try:
            image = pipe(
                image_prompt,
                num_inference_steps=20,
                guidance_scale=7.0,
                negative_prompt="text, words, letters, nsfw, explicit, unsafe"
            ).images[0]
            images.append(image)
        except Exception as e:
            print(f"❌ Error generating image: {e}")
            return None
    return images


def generate_comic(topic, num_panels, style):
    """Generates story and images"""
    story = get_summary_or_concept(topic, num_panels)
    if not story:
        return "⚠️ Error generating story.", None

    images = generate_comic_images(story, style)
    return story, images


def comic_interface(topic, num_panels, style):
    """Generate and display comic"""
    story, images = generate_comic(topic, num_panels, style)
    if images:
        return story, images
    else:
        return "⚠️ Error generating comic.", None


# Gradio UI
interface = gr.Interface(
    fn=comic_interface,
    inputs=[
        gr.Textbox(label="Enter a topic for the comic strip", value="Government of India"),
        gr.Slider(minimum=3, maximum=10, step=1, label="Number of Comic Panels", value=6),
        gr.Radio(list(art_styles.keys()), label="Choose an Art Style", value="Cinematic"),
    ],
    outputs=[
        gr.Textbox(label="Generated Story Prompts"),
        gr.Gallery(label="Generated Comic Strip"),
    ],
    title="🤖 ComicWala AI Generator",
    description="Generate AI-based comic strips using Gemini and Stable Diffusion.",
)

if __name__ == "__main__":
    interface.launch()