1
0
mirror of https://github.com/moparisthebest/SickRage synced 2024-11-13 04:45:08 -05:00
SickRage/sickbeard/providers/ezrss.py
echel0n 0d9fbc1ad7 Welcome to our SickBeard-TVRage Edition ...
This version of SickBeard uses both TVDB and TVRage to search and gather it's series data from allowing you to now have access to and download shows that you couldn't before because of being locked into only what TheTVDB had to offer.

Also this edition is based off the code we used in our XEM editon so it does come with scene numbering support as well as all the other features our XEM edition has to offer.

Please before using this with your existing database (sickbeard.db) please make a backup copy of it and delete any other database files such as cache.db and failed.db if present, we HIGHLY recommend starting out with no database files at all to make this a fresh start but the choice is at your own risk!

Enjoy!
2014-03-09 22:39:12 -07:00

199 lines
5.7 KiB
Python

# Author: Nic Wolfe <nic@wolfeden.ca>
# URL: http://code.google.com/p/sickbeard/
#
# This file is part of Sick Beard.
#
# Sick Beard 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.
#
# Sick Beard 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 Sick Beard. If not, see <http://www.gnu.org/licenses/>.
import urllib
import re
try:
import xml.etree.cElementTree as etree
except ImportError:
import elementtree.ElementTree as etree
import sickbeard
import generic
from sickbeard.common import Quality
from sickbeard import logger
from sickbeard import tvcache
from sickbeard import helpers
class EZRSSProvider(generic.TorrentProvider):
def __init__(self):
generic.TorrentProvider.__init__(self, "EZRSS")
self.supportsBacklog = True
self.cache = EZRSSCache(self)
self.url = 'https://www.ezrss.it/'
def isEnabled(self):
return sickbeard.EZRSS
def imageName(self):
return 'ezrss.png'
def getQuality(self, item):
filename = helpers.get_xml_text(item.find('{http://xmlns.ezrss.it/0.1/}torrent/{http://xmlns.ezrss.it/0.1/}fileName'))
quality = Quality.nameQuality(filename)
return quality
def findSeasonResults(self, show, season):
results = {}
if show.air_by_date:
logger.log(self.name + u" doesn't support air-by-date backlog because of limitations on their RSS search.", logger.WARNING)
return results
results = generic.TorrentProvider.findSeasonResults(self, show, season)
return results
def _get_season_search_strings(self, show, season, wantedEp, searchSeason=False):
params = {}
if not show:
return params
params['show_name'] = helpers.sanitizeSceneName(show.name, ezrss=True).replace('.', ' ').encode('utf-8')
params['season'] = season
return [params]
def _get_episode_search_strings(self, ep_obj):
params = {}
if not ep_obj:
return params
params['show_name'] = helpers.sanitizeSceneName(ep_obj.show.name, ezrss=True).replace('.', ' ').encode('utf-8')
if ep_obj.show.air_by_date:
params['date'] = str(ep_obj.airdate)
else:
params['season'] = ep_obj.season
params['episode'] = ep_obj.episode
return [params]
def _doSearch(self, search_params, show=None):
params = {"mode": "rss"}
if search_params:
params.update(search_params)
search_url = self.url + 'search/index.php?' + urllib.urlencode(params)
logger.log(u"Search string: " + search_url, logger.DEBUG)
data = self.getURL(search_url)
if not data:
logger.log(u"No data returned from " + search_url, logger.ERROR)
return []
parsedXML = helpers.parse_xml(data)
if parsedXML is None:
logger.log(u"Error trying to load " + self.name + " RSS feed", logger.ERROR)
return []
items = parsedXML.findall('.//item')
results = []
for curItem in items:
(title, url) = self._get_title_and_url(curItem)
if title and url:
logger.log(u"Adding item from RSS to results: " + title, logger.DEBUG)
results.append(curItem)
else:
logger.log(u"The XML returned from the " + self.name + " RSS feed is incomplete, this result is unusable", logger.ERROR)
return results
def _get_title_and_url(self, item):
(title, url) = generic.TorrentProvider._get_title_and_url(self, item)
filename = helpers.get_xml_text(item.find('{http://xmlns.ezrss.it/0.1/}torrent/{http://xmlns.ezrss.it/0.1/}fileName'))
if filename:
new_title = self._extract_name_from_filename(filename)
if new_title:
title = new_title
logger.log(u"Extracted the name " + title + " from the torrent link", logger.DEBUG)
return (title, url)
def _extract_name_from_filename(self, filename):
name_regex = '(.*?)\.?(\[.*]|\d+\.TPB)\.torrent$'
logger.log(u"Comparing " + name_regex + " against " + filename, logger.DEBUG)
match = re.match(name_regex, filename, re.I)
if match:
return match.group(1)
return None
class EZRSSCache(tvcache.TVCache):
def __init__(self, provider):
tvcache.TVCache.__init__(self, provider)
# only poll EZRSS every 15 minutes max
self.minTime = 15
def _getRSSData(self):
rss_url = self.provider.url + 'feed/'
logger.log(self.provider.name + " cache update URL: " + rss_url, logger.DEBUG)
data = self.provider.getURL(rss_url)
if not data:
logger.log(u"No data returned from " + rss_url, logger.ERROR)
return None
return data
def _parseItem(self, item):
(title, url) = self.provider._get_title_and_url(item)
if title and url:
logger.log(u"Adding item from RSS to cache: " + title, logger.DEBUG)
url = self._translateLinkURL(url)
return self._addCacheEntry(title, url)
else:
logger.log(u"The XML returned from the " + self.provider.name + " feed is incomplete, this result is unusable", logger.ERROR)
return None
provider = EZRSSProvider()