app.py

import gradio as gr
from PIL import Image
import clipGPT

# Define model loading functions (if needed)
def load_model_1():  # CLIP-GPT2
    # Load model components here if necessary
    return None 

# ... load_model_2(), load_model_3() - Define if and when needed

# Caption generation functions
def generate_caption_clipgpt(image):
    caption = clipGPT.generate_caption_clipgpt(image)
    return caption

# ... Add more caption generation functions for future models

# Sample image paths
sample_images = [
    "CXR191_IM-0591-1001.jpg",
    "CXR191_IM-0598-1001.jpg",
    "CXR191_IM-0601-1001.jpg",
    "CXR191_IM-0609-1001.jpg",
    "CXR191_IM-0618-1001.jpg"
]

# Gradio interface 
with gr.Blocks() as demo:
    with gr.Row():
        image = gr.Image(label="Upload Chest X-ray") 
        sample_images_gallery = gr.Gallery(sample_images, label="Sample Images") 
    with gr.Row():  
        model_choice = gr.Radio(["CLIP-GPT2", "ViT-GPT2", "ViT-CoAttention"], label="Select Model")
    with gr.Row():
        caption = gr.Textbox(label="Generated Caption") 

    def predict(img, model_name):
        if model_name == "CLIP-GPT2":
            return generate_caption_clipgpt(img)
        # Add elif blocks for "ViT-GPT2", "ViT-CoAttention"  as you implement them
        else:
            return "Caption generation for this model is not yet implemented."

    # Handle changes for both uploaded and sample images
    image.change(predict, [image, model_choice], caption)
    sample_images_gallery.change(predict, [sample_images_gallery, model_choice], caption)

demo.launch()