# Wavelet Image Compression Algorithm
# Using PyWavelets and NumPy

import numpy as np
import pywt
from PIL import Image

def compress_image(input_path, output_path, compression_ratio=0.85):
    # Load image and convert to grayscale
    img = Image.open(input_path).convert('L')
    data = np.array(img)
    
    # Perform 2D wavelet transform
    coeffs = pywt.wavedec2(data, 'bior4.4', level=3)
    
    # Threshold coefficients based on compression ratio
    coeffs = [coeffs[0]] + [tuple(pywt.threshold(c, 0.5, mode='soft') for c in subband) for subband in coeffs[1:]]
    
    # Reconstruct image
    reconstructed = pywt.waverec2(coeffs, 'bior4.4')
    reconstructed = np.clip(reconstructed, 0, 255).astype('uint8')
    
    # Save compressed image
    Image.fromarray(reconstructed).save(output_path)
    
    return reconstructed