import cv2
import numpy as np
import gradio as gr
import random

def apply_cartoon_filter(frame):
    """Cartoon Filter"""
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    gray = cv2.medianBlur(gray, 5)
    edges = cv2.adaptiveThreshold(gray, 255, 
                                   cv2.ADAPTIVE_THRESH_MEAN_C, 
                                   cv2.THRESH_BINARY, 11, 7)
    color = cv2.bilateralFilter(frame, 9, 300, 300)
    cartoon = cv2.bitwise_and(color, color, mask=edges)
    return cartoon

def apply_neon_effect(frame):
    """Neon Light Filter"""
    # Intensify colors
    frame_neon = frame.copy().astype(np.float32)
    frame_neon = np.clip(frame_neon * 1.5, 0, 255).astype(np.uint8)
    
    # Highlight edges
    edges = cv2.Canny(cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY), 100, 200)
    edges_colored = cv2.applyColorMap(edges, cv2.COLORMAP_JET)
    
    # Blend
    result = cv2.addWeighted(frame_neon, 0.7, edges_colored, 0.3, 0)
    return result

def apply_pixelate_effect(frame, pixel_size=15):
    """Pixelate Effect"""
    h, w = frame.shape[:2]
    small = cv2.resize(frame, (w//pixel_size, h//pixel_size), interpolation=cv2.INTER_LINEAR)
    return cv2.resize(small, (w, h), interpolation=cv2.INTER_NEAREST)

def apply_glitch_effect(frame):
    """Glitch Filter"""
    glitched = frame.copy()
    
    # Randomly shift color channels
    glitched[:, :, 0] = np.roll(glitched[:, :, 0], random.randint(-50, 50), axis=0)
    glitched[:, :, 1] = np.roll(glitched[:, :, 1], random.randint(-50, 50), axis=1)
    
    # Add noise to random areas
    noise = np.random.randint(0, 255, frame.shape, dtype=np.uint8)
    glitched = cv2.addWeighted(glitched, 0.7, noise, 0.3, 0)
    
    return glitched

def apply_watercolor_effect(frame):
    """Watercolor Effect"""
    # Smooth using bilateral filtering
    frame_soft = cv2.bilateralFilter(frame, 9, 75, 75)
    
    # Highlight edges
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    edges = cv2.Canny(gray, 100, 200)
    edges = cv2.cvtColor(edges, cv2.COLOR_GRAY2BGR)
    
    # Blend
    result = cv2.addWeighted(frame_soft, 0.8, edges, 0.2, 0)
    return result

def apply_upscale(frame, scale_factor=1.5):
    """
    Upscaling Effect
    
    Args:
        frame (numpy.ndarray): Input Image
        scale_factor (float): Scaling Factor (default 1.5)
    
    Returns:
        numpy.ndarray: Upscaled Image
    """
    interpolation_methods = [
        cv2.INTER_CUBIC,     
        cv2.INTER_LANCZOS4   
    ]
    
    method = random.choice(interpolation_methods)
    
    height, width = frame.shape[:2]
    new_height = int(height * scale_factor)
    new_width = int(width * scale_factor)
    
    upscaled = cv2.resize(frame, (new_width, new_height), interpolation=method)
    
    kernel = np.array([[-1,-1,-1], [-1,9,-1], [-1,-1,-1]])
    sharpened = cv2.filter2D(upscaled, -1, kernel)
    
    return sharpened

def apply_filter(filter_type, input_image=None):
    if input_image is None:
        cap = cv2.VideoCapture(0)
        ret, frame = cap.read()
        cap.release()
        if not ret:
            return "Failed to capture image from webcam"
    else:
        frame = input_image

    if filter_type == "Upscale":
        return apply_upscale(frame)
    elif filter_type == "Cartoon":
        return apply_cartoon_filter(frame)
    elif filter_type == "Neon Light":
        return apply_neon_effect(frame)
    elif filter_type == "Pixelate":
        return apply_pixelate_effect(frame)
    elif filter_type == "Glitch":
        return apply_glitch_effect(frame)
    elif filter_type == "Watercolor":
        return apply_watercolor_effect(frame)

# Gradio interface
with gr.Blocks() as demo:
    gr.Markdown('# <p align="center"> OpenCV Image Effects </p>')

    # Filter options
    filter_type = gr.Dropdown(
        label="Select Filter",
        choices=["Upscale","Cartoon", "Neon Light", "Pixelate", "Glitch", "Watercolor"],
        value="Upscale"
    )

    with gr.Row():
        input_image = gr.Image(label="Upload Image", type="numpy")
        output_image = gr.Image(label="Filtered Image")

    # Apply filter button
    apply_button = gr.Button("Apply Filter")

    # Apply filter function on button click
    apply_button.click(fn=apply_filter, inputs=[filter_type, input_image], outputs=output_image)

demo.launch()