config.py

import requests
from bs4 import BeautifulSoup
import os
import cv2
import numpy as np
from PIL import Image

# Define the directories
LINE_ART_DIR = "train_images/line_arts"
COLORED_DIR = "train_images/colored"

# Ensure directories exist
os.makedirs(LINE_ART_DIR, exist_ok=True)
os.makedirs(COLORED_DIR, exist_ok=True)

# Function to download and process images
def download_and_process_images(manga_title, url):
    response = requests.get(url)
    soup = BeautifulSoup(response.content, 'html.parser')

    # Find image tags (this will depend on the structure of the webpage)
    image_tags = soup.find_all('img')

    for img in image_tags:
        img_url = img['src']
        img_data = requests.get(img_url).content
        img_array = np.frombuffer(img_data, np.uint8)
        img = cv2.imdecode(img_array, cv2.IMREAD_COLOR)

        # Check if the image is colored or line art
        if len(img.shape) == 3 and img.shape[2] == 3:
            # Colored image
            colored_path = os.path.join(COLORED_DIR, f"{manga_title}_colored.png")
            cv2.imwrite(colored_path, img)
        else:
            # Line art image
            line_art_path = os.path.join(LINE_ART_DIR, f"{manga_title}_line_art.png")
            cv2.imwrite(line_art_path, img)

# Example usage
manga_title = "example_manga"
nhentai_url = "https://nhentai.net/g/your_manga_id/"
hitomi_url = "https://hitomi.la/galleries/your_manga_id.html"

download_and_process_images(manga_title, nhentai_url)
download_and_process_images(manga_title, hitomi_url)