# Author: Jordon Smith # URL: http://code.google.com/p/sickbeard/ # # This file is part of SickRage. # # SickRage 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. # # SickRage 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 SickRage. If not, see . import urllib import sickbeard import generic from sickbeard import tvcache from sickbeard import helpers from sickbeard import classes from sickbeard import logger from sickbeard.exceptions import ex, AuthException from sickbeard import show_name_helpers from datetime import datetime try: import xml.etree.cElementTree as etree except ImportError: import elementtree.ElementTree as etree try: import json except ImportError: from lib import simplejson as json class OmgwtfnzbsProvider(generic.NZBProvider): def __init__(self): generic.NZBProvider.__init__(self, "omgwtfnzbs") self.enabled = False self.username = None self.api_key = None self.cache = OmgwtfnzbsCache(self) self.urls = {'base_url': 'https://omgwtfnzbs.org/'} self.url = self.urls['base_url'] self.supportsBacklog = True def isEnabled(self): return self.enabled def _checkAuth(self): if not self.username or not self.api_key: raise AuthException("Your authentication credentials for " + self.name + " are missing, check your config.") return True def _checkAuthFromData(self, parsed_data, is_XML=True): if parsed_data is None: return self._checkAuth() if is_XML: # provider doesn't return xml on error return True else: parsedJSON = parsed_data if 'notice' in parsedJSON: description_text = parsedJSON.get('notice') if 'information is incorrect' in parsedJSON.get('notice'): logger.log(u"Incorrect authentication credentials for " + self.name + " : " + str(description_text), logger.DEBUG) raise AuthException( "Your authentication credentials for " + self.name + " are incorrect, check your config.") elif '0 results matched your terms' in parsedJSON.get('notice'): return True else: logger.log(u"Unknown error given from " + self.name + " : " + str(description_text), logger.DEBUG) return False return True def _get_season_search_strings(self, ep_obj): return [x for x in show_name_helpers.makeSceneSeasonSearchString(self.show, ep_obj)] def _get_episode_search_strings(self, ep_obj, add_string=''): return [x for x in show_name_helpers.makeSceneSearchString(self.show, ep_obj)] def _get_title_and_url(self, item): return (item['release'], item['getnzb']) def _doSearch(self, search, search_mode='eponly', epcount=0, retention=0): self._checkAuth() params = {'user': self.username, 'api': self.api_key, 'eng': 1, 'catid': '19,20', # SD,HD 'retention': sickbeard.USENET_RETENTION, 'search': search} if retention or not params['retention']: params['retention'] = retention search_url = 'https://api.omgwtfnzbs.org/json/?' + urllib.urlencode(params) logger.log(u"Search url: " + search_url, logger.DEBUG) parsedJSON = self.getURL(search_url, json=True) if not parsedJSON: return [] if self._checkAuthFromData(parsedJSON, is_XML=False): results = [] for item in parsedJSON: if 'release' in item and 'getnzb' in item: results.append(item) return results return [] def findPropers(self, search_date=None): search_terms = ['.PROPER.', '.REPACK.'] results = [] for term in search_terms: for item in self._doSearch(term, retention=4): if 'usenetage' in item: title, url = self._get_title_and_url(item) try: result_date = datetime.fromtimestamp(int(item['usenetage'])) except: result_date = None if result_date: results.append(classes.Proper(title, url, result_date, self.show)) return results class OmgwtfnzbsCache(tvcache.TVCache): def __init__(self, provider): tvcache.TVCache.__init__(self, provider) self.minTime = 20 def _get_title_and_url(self, item): """ Retrieves the title and URL data from the item XML node item: An elementtree.ElementTree element representing the tag of the RSS feed Returns: A tuple containing two strings representing title and URL respectively """ title = item.get('title') if title: title = u'' + title title = title.replace(' ', '.') url = item.get('link') if url: url = url.replace('&', '&') return (title, url) def _getRSSData(self): params = {'user': provider.username, 'api': provider.api_key, 'eng': 1, 'catid': '19,20'} # SD,HD rss_url = 'https://rss.omgwtfnzbs.org/rss-download.php?' + urllib.urlencode(params) logger.log(self.provider.name + u" cache update URL: " + rss_url, logger.DEBUG) return self.getRSSFeed(rss_url, items=['entries', 'feed']) provider = OmgwtfnzbsProvider()