sbtools/handler.py

#!/usr/bin/env python3

from bs4 import BeautifulSoup
from selenium import webdriver
from selenium.webdriver.common.keys import Keys
from selenium.common.exceptions import NoSuchElementException
import glob, multiprocessing, re, requests, urllib.request, urllib.error, urllib.parse, time, os
import setuptools

useragent = "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:54.0) Gecko/20100101 Firefox/54.0"
session = None

def status(driver):
 if "Benutzername oder E-Mail-Adresse:" not in driver.page_source and 'href="login/"' not in driver.page_source:
  return True
 else:
  return False

def loadPage(url, driver, period=5,init=False):
 if not (init or status(driver)):
  login(driver)
 driver.get(url)
 time.sleep(period)

def loginHandler(driver, user = setuptools.user(), password = setuptools.password()):
 loadPage("https://scriptzbase.org/login/", driver, 3, True)
  
 curfield = driver.find_element_by_name("login")
 curfield.send_keys(user)
  
 curfield = driver.find_element_by_name("password")
 curfield.send_keys(password)
  
 curfield.send_keys(Keys.RETURN)
 time.sleep(3)

 return status(driver)

class LoginError(Exception):
 pass

def login(driver):
 global session

 if not (status(driver) or loginHandler(driver)):
  raise LoginError("Login failed.")

 session = requests.Session()
 kekse = driver.get_cookies()

 for keks in kekse:
  session.cookies.set(keks["name"], keks["value"])

 session.headers.update({"User-Agent": useragent})

 return True

def fileHandler(url, retry = False):
 global session

 try:
  fid = url.split("=")[1]
  if not glob.glob("files/sbd%s*" % fid):
   res = session.get("https://scriptzbase.org/%s" % url)
   fname = re.findall("filename=(.+)", res.headers["content-disposition"])[0].split(";")[0].strip('"')

   with open("files/sbd%s_%s" % (fid, fname), "wb") as out:
    out.write(res.content)

 except requests.exceptions.ConnectionError:
  if not retry:
   time.sleep(10)
   fileHandler(url, True)
  else:
   raise

def pageHandler(driver):
 juha = BeautifulSoup(driver.page_source, "html5lib")

 for a in juha.findAll("a"):
  try:
   if "/download?version=" in a["href"]:
    fileHandler(a["href"])
  except KeyError:
   pass

def siteHandler(driver, p = 1):
 loadPage("https://scriptzbase.org/nulled_scripts/?page=" + str(p), driver)

 if driver.current_url[-len(str(p)):] == str(p):
  pageHandler(driver)
  siteHandler(driver, p+1)

def mainHandler(driver):
 os.makedirs("files", exist_ok=True)
 siteHandler(driver)

if __name__ == "__main__":
 caps = webdriver.DesiredCapabilities().PHANTOMJS.copy()
 caps["phantoms.page.settings.userAgent"] = useragent
 driver = webdriver.PhantomJS(desired_capabilities=caps)
 mainHandler(driver)
 driver.close()