import gradio as gr
import cv2
import numpy as np 
 
# Function to order points in a consistent manner
def order_points(pts):
    rect = np.zeros((4, 2), dtype='float32')
    pts = np.array(pts)
    s = pts.sum(axis=1)
    rect[0] = pts[np.argmin(s)]
    rect[2] = pts[np.argmax(s)]
    diff = np.diff(pts, axis=1)
    rect[1] = pts[np.argmin(diff)]
    rect[3] = pts[np.argmax(diff)]
    return rect.astype('int').tolist()

# Function to find the destination points for perspective transform
def find_dest(pts):
    (tl, tr, br, bl) = pts
    widthA = np.sqrt(((br[0] - bl[0]) ** 2) + ((br[1] - bl[1]) ** 2))
    widthB = np.sqrt(((tr[0] - tl[0]) ** 2) + ((tr[1] - tl[1]) ** 2))
    maxWidth = max(int(widthA), int(widthB))
    heightA = np.sqrt(((tr[0] - br[0]) ** 2) + ((tr[1] - br[1]) ** 2))
    heightB = np.sqrt(((tl[0] - bl[0]) ** 2) + ((tl[1] - bl[1]) ** 2))
    maxHeight = max(int(heightA), int(heightB))
    destination_corners = [[0, 0], [maxWidth, 0], [maxWidth, maxHeight], [0, maxHeight]]
    return order_points(destination_corners)

def scan(img):
    # Resize image if dimensions exceed limit
    dim_limit = 1080
    max_dim = max(img.shape)
    
    if max_dim > dim_limit:
        resize_scale = dim_limit / max_dim
        img = cv2.resize(img, None, fx=resize_scale, fy=resize_scale)
        
    orig_img = img.copy()
    kernel = np.ones((5, 5), np.uint8)
    
    # Perform morphological closing
    img = cv2.morphologyEx(img, cv2.MORPH_CLOSE, kernel, iterations=3)
    
    # Initialize mask for GrabCut
    mask = np.zeros(img.shape[:2], np.uint8)
    bgdModel = np.zeros((1, 65), np.float64)
    fgdModel = np.zeros((1, 65), np.float64)
    rect = (20, 20, img.shape[1] - 20, img.shape[0] - 20)
    cv2.grabCut(img, mask, rect, bgdModel, fgdModel, 5, cv2.GC_INIT_WITH_RECT)
    mask2 = np.where((mask == 2) | (mask == 0), 0, 1).astype('uint8')
    img = img * mask2[:, :, np.newaxis]
    
    # Convert image to grayscale and apply Gaussian blur
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    gray = cv2.GaussianBlur(gray, (11, 11), 0)
    
    # Perform Canny edge detection
    canny = cv2.Canny(gray, 0, 200)
    canny = cv2.dilate(canny, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5, 5)))
    
    # Find contours and sort them
    contours, hierarchy = cv2.findContours(canny, cv2.RETR_LIST, cv2.CHAIN_APPROX_NONE)
    page = sorted(contours, key=cv2.contourArea, reverse=True)[:5]
    
    if len(page) == 0:
        return orig_img
    
    for c in page:
        epsilon = 0.02 * cv2.arcLength(c, True)
        corners = cv2.approxPolyDP(c, epsilon, True)
        if len(corners) == 4:
            break
    
    # Order the corners and find destination points for perspective transform
    corners = sorted(np.concatenate(corners).tolist())
    corners = order_points(corners)
    destination_corners = find_dest(corners)
    h, w = orig_img.shape[:2]
    
    # Perform perspective transform to obtain top-down view
    M = cv2.getPerspectiveTransform(np.float32(corners), np.float32(destination_corners))
    final = cv2.warpPerspective(orig_img, M, (destination_corners[2][0], destination_corners[2][1]), flags=cv2.INTER_LINEAR)
    
    return final

def gradio_interface(input_image):
    if input_image is None:
        return None
    img = cv2.cvtColor(input_image, cv2.COLOR_RGB2BGR)
    processed_img = scan(img)
    return cv2.cvtColor(processed_img, cv2.COLOR_BGR2RGB)

with gr.Blocks() as demo:
    gr.Markdown("# Document Scanner using OpenCV")
    gr.Markdown("## [Reference: 4-point OpenCV getPerspective Transform Example](https://pyimagesearch.com/2014/08/25/4-point-opencv-getperspective-transform-example/)")
    gr.Markdown("""
    ### Image Processing Flow:
    - Resize the image if its dimensions exceed the limit.
    - Apply morphological transformations to enhance document boundaries.
    - Perform GrabCut for foreground extraction.
    - Convert the image to grayscale.
    - Apply Gaussian blur.
    - Perform Canny edge detection.
    - Dilate the edges to close gaps.
    - Find contours and identify the largest ones likely to be the document edges.
    - If a contour with four corners is found, transform the perspective to obtain a top-down view of the document.
    """)
    image_input = gr.Image(type="numpy", label="Upload Image")
    image_output = gr.Image(type="numpy", label="Processed Image")
    gr.Interface(fn=gradio_interface, inputs=image_input, outputs=image_output, allow_flagging="never")

demo.launch(debug=True)