twistedbrodymangaflow / converter.py
codebybrody's picture
Upload 105 files
14b0378 verified
Raw History Blame Contribute Delete
1.87 kB
import img2pdf
import requests
import sys
import json
import os
from concurrent.futures import ThreadPoolExecutor
def download_image(args):
url, index, total = args
try:
response = requests.get(url, timeout=15, headers={'User-Agent': 'Mozilla/5.0'})
if response.status_code == 200:
# Imprimimos progreso para Node.js
print(f"PROGRESS:IMAGE_DONE", flush=True)
return index, response.content
except Exception as e:
print(f"DEBUG: Error en imagen {index}: {e}", flush=True)
return index, None
def create_pdf(urls, output_name):
total = len(urls)
print(f"DEBUG: Iniciando descarga paralela de {total} imágenes...", flush=True)
# Descarga en paralelo (50 hilos a la vez para saturar la conexión)
# Esto es lo que lo hace igual de rápido que el ZIP
with ThreadPoolExecutor(max_workers=50) as executor:
results = list(executor.map(download_image, [(url, i, total) for i, url in enumerate(urls)]))
# Ordenamos las imágenes por su índice original para que el manga no esté desordenado
results.sort(key=lambda x: x[0])
img_data = [data for index, data in results if data is not None]
if img_data:
print("DEBUG: Empaquetando PDF...", flush=True)
with open(output_name, "wb") as f:
# img2pdf es ultra rápido porque no re-comprime
f.write(img2pdf.convert(img_data))
print("STATUS:SUCCESS", flush=True)
else:
print("STATUS:ERROR", flush=True)
if __name__ == "__main__":
try:
input_data = sys.stdin.read()
if not input_data:
sys.exit(1)
data = json.loads(input_data)
create_pdf(data['urls'], data['output'])
except Exception as e:
print(f"DEBUG: Error crítico: {e}", flush=True)