# -*- coding: utf-8 -*- """ This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, see <http://www.gnu.org/licenses/>. @author: zoidberg """ import re from module.plugins.internal.SimpleHoster import SimpleHoster, create_getInfo class BezvadataCz(SimpleHoster): __name__ = "BezvadataCz" __type__ = "hoster" __pattern__ = r'http://(?:www\.)?bezvadata.cz/stahnout/.*' __version__ = "0.24" __description__ = """BezvaData.cz hoster plugin""" __author_name__ = "zoidberg" __author_mail__ = "zoidberg@mujmail.cz" FILE_NAME_PATTERN = r'<p><b>Soubor: (?P<N>[^<]+)</b></p>' FILE_SIZE_PATTERN = r'<li><strong>Velikost:</strong> (?P<S>[^<]+)</li>' FILE_OFFLINE_PATTERN = r'<title>BezvaData \| Soubor nenalezen</title>' def setup(self): self.multiDL = self.resumeDownload = True def handleFree(self): #download button found = re.search(r'<a class="stahnoutSoubor".*?href="(.*?)"', self.html) if not found: self.parseError("page1 URL") url = "http://bezvadata.cz%s" % found.group(1) #captcha form self.html = self.load(url) self.checkErrors() for _ in xrange(5): action, inputs = self.parseHtmlForm('frm-stahnoutFreeForm') if not inputs: self.parseError("FreeForm") found = re.search(r'<img src="data:image/png;base64,(.*?)"', self.html) if not found: self.parseError("captcha img") #captcha image is contained in html page as base64encoded data but decryptCaptcha() expects image url self.load, proper_load = self.loadcaptcha, self.load try: inputs['captcha'] = self.decryptCaptcha(found.group(1), imgtype='png') finally: self.load = proper_load if '<img src="data:image/png;base64' in self.html: self.invalidCaptcha() else: self.correctCaptcha() break else: self.fail("No valid captcha code entered") #download url self.html = self.load("http://bezvadata.cz%s" % action, post=inputs) self.checkErrors() found = re.search(r'<a class="stahnoutSoubor2" href="(.*?)">', self.html) if not found: self.parseError("page2 URL") url = "http://bezvadata.cz%s" % found.group(1) self.logDebug("DL URL %s" % url) #countdown found = re.search(r'id="countdown">(\d\d):(\d\d)<', self.html) wait_time = (int(found.group(1)) * 60 + int(found.group(2)) + 1) if found else 120 self.wait(wait_time, False) self.download(url) def checkErrors(self): if 'images/button-download-disable.png' in self.html: self.longWait(5 * 60, 24) # parallel dl limit elif '<div class="infobox' in self.html: self.tempOffline() def loadcaptcha(self, data, *args, **kwargs): return data.decode("base64") getInfo = create_getInfo(BezvadataCz)