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
driver = None

def status():
 global 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, period=5, init = False):
 global driver

 if not (init or status()):
  login()
 driver.get(url)
 time.sleep(period)

def loginHandler(user = setuptools.user(), password = setuptools.password()):
 loadPage("https://scriptzbase.org/login/", 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()

class LoginError(Exception):
 pass

def login():
 global driver, session

 if not (status() or loginHandler()):
  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, requests.exceptions.ChunkedEncodingError):
  if not retry:
   time.sleep(10)
   fileHandler(url, True)
  else:
   raise

def pageHandler():
 global 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(p):
 global driver

 loadPage("https://scriptzbase.org/nulled_scripts/?page=" + str(p))

 if driver.current_url[-len(str(p)):] == str(p):
  pageHandler()
  return True

 return False

def mainHandler():
 os.makedirs("files", exist_ok=True)
 p = 1

 while True:
  if not siteHandler(p):
   break
  p += 1

def makeDriver():
 caps = webdriver.DesiredCapabilities().PHANTOMJS.copy()
 caps["phantoms.page.settings.userAgent"] = useragent
 return webdriver.PhantomJS(desired_capabilities=caps)

if __name__ == "__main__":
 driver = makeDriver()
 mainHandler()
 driver.close()