app.py

import gradio as gr
import cv2
import numpy as np
from PIL import Image

def hex_to_rgb(hex_code):
    # Remove the '#' symbol if present
    hex_code = hex_code.lstrip('#')
    
    # Convert the hex code to RGB
    return tuple(int(hex_code[i:i+2], 16) for i in (0, 2, 4))

def crop_image(image, crop_top_percentage, crop_bottom_percentage):
    height, width = image.shape[:2]

    # Calculate the number of pixels to crop from the top and bottom as a percentage of the image height
    crop_top = int(height * crop_top_percentage / 100)
    crop_bottom = int(height * crop_bottom_percentage / 100)

    # Crop the input image
    cropped_image = image[crop_top:height - crop_bottom, :]

    return cropped_image

def resize_and_overlay_image(input_image, reduction_percentage, shift_percentage_lr, shift_percentage_ud, background_color, crop_top_percentage, crop_bottom_percentage):
    # Check if the input image is empty
    if input_image.size == 0:
        return None

    img = np.array(input_image)

    # Check if the image has shape information
    if img.ndim < 2:
        return None

    # Crop the input image
    cropped_img = crop_image(img, crop_top_percentage, crop_bottom_percentage)

    # Get the dimensions of the cropped image
    height, width = cropped_img.shape[:2]

    # Calculate the new dimensions based on the reduction percentage
    new_height = int(height * reduction_percentage / 100)
    new_width = int(width * reduction_percentage / 100)

    # Calculate the shift values as a percentage of the image dimensions
    shift_x = int(width * shift_percentage_lr / 100)
    shift_y = int(height * shift_percentage_ud / 100)

    # Create a new background image with the same dimensions as the input image and specified color
    background_rgb = hex_to_rgb(background_color)
    background_img = np.ones_like(img) * background_rgb

    # Resize the cropped image while maintaining aspect ratio
    aspect_ratio = width / height
    if new_width / new_height > aspect_ratio:
        resized_height = new_height
        resized_width = int(new_height * aspect_ratio)
    else:
        resized_width = new_width
        resized_height = int(new_width / aspect_ratio)
    resized_img = cv2.resize(cropped_img, (resized_width, resized_height))

    # Calculate the position to overlay the resized image on the background image
    x = int((width - resized_width) / 2) + shift_x
    y = int((height - resized_height) / 2) + shift_y

    # Overlay the resized image on the background image
    background_img[y:y + resized_height, x:x + resized_width] = resized_img

    # Return the resulting image as a NumPy array
    return background_img, f"Input Image Dimensions: {width}x{height}"

# Create the Gradio interface
iface = gr.Interface(
    fn=resize_and_overlay_image,
    inputs=[
        gr.inputs.Image(type="pil", label="Input Image"),
        gr.inputs.Slider(minimum=0, maximum=100, step=10, default=80, label="Percentage of Original"),
        gr.inputs.Slider(minimum=-50, maximum=50, step=10, default=0, label="Shift Left / Right (%)"),
        gr.inputs.Slider(minimum=-50, maximum=50, step=10, default=0, label="Shift Up / Down (%)"),
        gr.inputs.Textbox(default="#ffffff", label="Background Color (Hex Code)"),
        gr.inputs.Slider(minimum=0, maximum=100, step=10, default=0, label="Crop Top (%)"),
        gr.inputs.Slider(minimum=0, maximum=100, step=10, default=0, label="Crop Bottom (%)")
    ],
    outputs=[
        gr.outputs.Image(type="numpy", label="Result"),
        gr.outputs.Textbox(label="Image Info")
    ],
    title="Image Resizer",
    description="Crop the input image, overlay it on a new background of the specified color, start centered, and shift it as a percentage without stretching."
)

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