2020-04-05-torrent

2020-04-05-torrent.py

from subprocess import PIPE
from subprocess import Popen
import re
import requests
from bs4 import BeautifulSoup
import re
import shlex
from pathlib import Path
from collections import namedtuple
from concurrent.futures import ThreadPoolExecutor
from os import environ as E
import os

HERE = Path(__file__).resolve().parent
Datum = namedtuple('Datum', ['id', 'name', 'magnet'])


def nyaanet():
    data = []
    for page in range(1, 10):
        url = f"https://nyaa.net/search/{page}?c=4_8&order=false&sort=2"
        with requests.get(url) as r:
            html = r.text
        soup = BeautifulSoup(html, 'lxml')

        for tr in soup.find_all('tr', {'class': 'torrent-info'}):
            id = tr.get('id')
            name = tr.find('td', {'class': 'tr-name'}).text.strip()
            magnet = tr.find(attrs={'class': 'tr-links'}).find('a', attrs={'href': re.compile(r'^magnet:')}).get('href')
            datum = Datum(id, name, magnet)
            data.append(datum)
    return data


def nyaasi():
    data = []
    for page in range(1, 10):
        url = f"https://nyaa.si/?c=3_3&s=seeders&o=desc&p={page}"
        with requests.get(url) as r:
            html = r.text
        # print(html)
        soup = BeautifulSoup(html, 'lxml')

        for tr in soup.find_all('tr', attrs={'class': 'default'}):
            name = ''.join([a.text for a in tr.find_all('a', {'href': True})])
            id = tr.find('a', {'href': re.compile(r'^/view/[0-9]{1,}$')}).get('href')
            magnet = tr.find('a', {'href': re.compile(r'^magnet:')}).get('href')
            datum = Datum(id, name.strip(), magnet.strip())
            data.append(datum)
    return data


def proc(datum):
    pid = os.getpid()
    if Path(f'{HERE}/Downloads/{datum.name}').exists():
        return

    with Popen(['aria2c', datum.magnet, '--dir', f'{HERE}/Downloads/{datum.name}', '--seed-time=0', '--dht-listen-port=63000-65535', '--listen-port=60000-62999'], stdout=None) as proc:
        try:
            proc.communicate(timeout=60 * 60)
        except Exception as exc:
            print(exc)


def run():
    data = nyaanet() + nyaasi()

    if E.get('SINGLE') == 'TRUE':
        for datum in data:
            proc(datum)
    else:
        with ThreadPoolExecutor(max_workers=1000) as exe:
            exe.map(proc, data)


if __name__ == "__main__":
    run()