1
0
mirror of https://github.com/moparisthebest/SickRage synced 2024-11-11 03:45:01 -05:00
SickRage/sickbeard/providers/newznab.py
echel0n a32de0d523 Merge pull request #769 from KontiSR/dev
Added gui for for choosing newznab (tv) categories
2014-09-05 19:16:46 -07:00

456 lines
16 KiB
Python
Executable File

# Author: Nic Wolfe <nic@wolfeden.ca>
# 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 <http://www.gnu.org/licenses/>.
import urllib
import time
import datetime
import os
try:
import xml.etree.cElementTree as etree
except ImportError:
import elementtree.ElementTree as etree
import sickbeard
import generic
from sickbeard import classes
from sickbeard import helpers
from sickbeard import scene_exceptions
from sickbeard import encodingKludge as ek
from sickbeard import logger
from sickbeard import tvcache
from sickbeard.exceptions import ex, AuthException
from lib import requests
from lib.requests import exceptions
from lib.bencode import bdecode
class NewznabProvider(generic.NZBProvider):
def __init__(self, name, url, key='', catIDs='5030,5040', search_mode='eponly', search_fallback=False,
enable_daily=False, enable_backlog=False):
generic.NZBProvider.__init__(self, name)
self.cache = NewznabCache(self)
self.url = url
self.key = key
self.search_mode = search_mode
self.search_fallback = search_fallback
self.enable_daily = enable_daily
self.enable_backlog = enable_backlog
# a 0 in the key spot indicates that no key is needed
if self.key == '0':
self.needs_auth = False
else:
self.needs_auth = True
if catIDs:
self.catIDs = catIDs
else:
self.catIDs = '5030,5040'
self.enabled = True
self.supportsBacklog = True
self.default = False
def configStr(self):
return self.name + '|' + self.url + '|' + self.key + '|' + self.catIDs + '|' + str(
int(self.enabled)) + '|' + self.search_mode + '|' + str(int(self.search_fallback)) + '|' + str(
int(self.enable_daily)) + '|' + str(int(self.enable_backlog))
def imageName(self):
if ek.ek(os.path.isfile,
ek.ek(os.path.join, sickbeard.PROG_DIR, 'gui', sickbeard.GUI_NAME, 'images', 'providers',
self.getID() + '.png')):
return self.getID() + '.png'
return 'newznab.png'
def isEnabled(self):
return self.enabled
def _getURL(self, url, post_data=None, params=None, timeout=30, json=False):
"""
By default this is just a simple urlopen call but this method should be overridden
for providers with special URL requirements (like cookies)
Not really changed much from the superclass, can be used in future.
"""
# check for auth
if not self._doLogin():
return
return helpers.getURL(url, post_data=post_data, params=params, headers=self.headers, timeout=timeout,
session=self.session, json=json)
def get_newznab_categories(self):
"""
Uses the newznab provider url and apikey to get the capabilities.
Makes use of the default newznab caps param. e.a. http://yournewznab/api?t=caps&apikey=skdfiw7823sdkdsfjsfk
Returns a tuple with (succes or not, array with dicts [{"id": "5070", "name": "Anime"},
{"id": "5080", "name": "Documentary"}, {"id": "5020", "name": "Foreign"}...etc}], error message)
"""
return_categories = []
self._checkAuth()
params = {"t": "caps"}
if self.needs_auth and self.key:
params['apikey'] = self.key
categories = self.getURL("%s/api" % (self.url), params=params)
xml_categories = helpers.parse_xml(categories)
if not xml_categories:
return (False, return_categories, "Error parsing xml for [%s]" % (self.name))
try:
for category in xml_categories.iter('category'):
if category.get('name') == 'TV':
for subcat in category.findall('subcat'):
return_categories.append(subcat.attrib)
except:
return (False, return_categories, "Error parsing result for [%s]" % (self.name))
return (True, return_categories, "")
def _get_season_search_strings(self, ep_obj):
to_return = []
cur_params = {}
# season
if ep_obj.show.air_by_date or ep_obj.show.sports:
date_str = str(ep_obj.airdate).split('-')[0]
cur_params['season'] = date_str
cur_params['q'] = date_str.replace('-', '.')
elif ep_obj.show.is_anime:
cur_params['season'] = "%d" % ep_obj.scene_absolute_number
else:
cur_params['season'] = str(ep_obj.scene_season)
# search
rid = helpers.mapIndexersToShow(ep_obj.show)[2]
if rid:
cur_return = cur_params.copy()
cur_return['rid'] = rid
to_return.append(cur_return)
# add new query strings for exceptions
name_exceptions = list(
set(scene_exceptions.get_scene_exceptions(ep_obj.show.indexerid) + [ep_obj.show.name]))
for cur_exception in name_exceptions:
if 'q' in cur_params:
cur_params['q'] = helpers.sanitizeSceneName(cur_exception) + '.' + cur_params['q']
to_return.append(cur_params)
return to_return
def _get_episode_search_strings(self, ep_obj, add_string=''):
to_return = []
params = {}
if not ep_obj:
return [params]
if ep_obj.show.air_by_date or ep_obj.show.sports:
date_str = str(ep_obj.airdate)
params['season'] = date_str.partition('-')[0]
params['ep'] = date_str.partition('-')[2].replace('-', '/')
elif ep_obj.show.anime:
params['ep'] = "%i" % int(ep_obj.scene_absolute_number)
else:
params['season'] = ep_obj.scene_season
params['ep'] = ep_obj.scene_episode
# search
rid = helpers.mapIndexersToShow(ep_obj.show)[2]
if rid:
cur_return = params.copy()
cur_return['rid'] = rid
to_return.append(cur_return)
# add new query strings for exceptions
name_exceptions = list(
set(scene_exceptions.get_scene_exceptions(ep_obj.show.indexerid) + [ep_obj.show.name]))
for cur_exception in name_exceptions:
params['q'] = helpers.sanitizeSceneName(cur_exception)
to_return.append(params)
return to_return
def _doGeneralSearch(self, search_string):
return self._doSearch({'q': search_string})
def _checkAuth(self):
if self.needs_auth and not self.key:
logger.log(u"Incorrect authentication credentials for " + self.name + " : " + "API key is missing",
logger.DEBUG)
raise AuthException("Your authentication credentials for " + self.name + " are missing, check your config.")
return True
def _checkAuthFromData(self, data):
if data is None:
return self._checkAuth()
if 'error' in data.feed:
code = data.feed['error']['code']
if code == '100':
raise AuthException("Your API key for " + self.name + " is incorrect, check your config.")
elif code == '101':
raise AuthException("Your account on " + self.name + " has been suspended, contact the administrator.")
elif code == '102':
raise AuthException(
"Your account isn't allowed to use the API on " + self.name + ", contact the administrator")
else:
logger.log(u"Unknown error given from " + self.name + ": " + data.feed['error']['description'],
logger.ERROR)
return False
return True
def _doSearch(self, search_params, search_mode='eponly', epcount=0, age=0):
self._checkAuth()
params = {"t": "tvsearch",
"maxage": sickbeard.USENET_RETENTION,
"limit": 100,
"attrs": "rageid",
"offset": 0}
# category ids
if self.show and self.show.is_sports:
params['cat'] = self.catIDs + ',5060'
elif self.show and self.show.is_anime:
params['cat'] = self.catIDs + ',5070'
else:
params['cat'] = self.catIDs
# if max_age is set, use it, don't allow it to be missing
if age or not params['maxage']:
params['maxage'] = age
if search_params:
params.update(search_params)
if self.needs_auth and self.key:
params['apikey'] = self.key
results = []
while True:
search_url = self.url + 'api?' + urllib.urlencode(params)
logger.log(u"Search url: " + search_url, logger.DEBUG)
data = self.cache.getRSSFeed(search_url)
if data and 'entries' in data and self._checkAuthFromData(data):
for item in data.entries:
(title, url) = self._get_title_and_url(item)
if title and url:
results.append(item)
else:
logger.log(
u"The data returned from the " + self.name + " is incomplete, this result is unusable",
logger.DEBUG)
# attempt to grab the total and offset newznab responses
try:
total = int(data.feed.newznab_response['total'])
offset = int(data.feed.newznab_response['offset'])
except (AttributeError, TypeError):
break
# sanity check - limiting at 10 at getting 1000 results in-case incorrect total parameter is reported
if params['limit'] > 1000:
logger.log("Excessive results for search, ending search", logger.WARNING)
break
# sanity check - total should remain constant
if offset != 0 and total != initial_total:
logger.log("Total number of items on newznab response changed, ending search", logger.DEBUG)
break
else:
initial_total = total
# if there are more items available then the amount given in one call, grab some more
if (total - params['limit']) > offset == params['offset']:
params['offset'] += params['limit']
logger.log(str(
total - params['offset']) + " more items to be fetched from provider. Fetching another " + str(
params['limit']) + " items.", logger.DEBUG)
else:
break
else:
break
return results
def findPropers(self, search_date=None):
search_terms = ['.proper.', '.repack.']
cache_results = self.cache.listPropers(search_date)
results = [classes.Proper(x['name'], x['url'], datetime.datetime.fromtimestamp(x['time']), self.show) for x in
cache_results]
index = 0
alt_search = ('nzbs_org' == self.getID())
term_items_found = False
do_search_alt = False
while index < len(search_terms):
search_params = {'q': search_terms[index]}
if alt_search:
if do_search_alt:
index += 1
if term_items_found:
do_search_alt = True
term_items_found = False
else:
if do_search_alt:
search_params['t'] = "search"
do_search_alt = (True, False)[do_search_alt]
else:
index += 1
for item in self._doSearch(search_params, age=4):
(title, url) = self._get_title_and_url(item)
if item.has_key('published_parsed') and item['published_parsed']:
result_date = item.published_parsed
if result_date:
result_date = datetime.datetime(*result_date[0:6])
else:
logger.log(u"Unable to figure out the date for entry " + title + ", skipping it")
continue
if not search_date or result_date > search_date:
search_result = classes.Proper(title, url, result_date, self.show)
results.append(search_result)
term_items_found = True
do_search_alt = False
return results
class NewznabCache(tvcache.TVCache):
def __init__(self, provider):
tvcache.TVCache.__init__(self, provider)
# only poll newznab providers every 15 minutes max
self.minTime = 15
def _getRSSData(self):
params = {"t": "tvsearch",
"cat": self.provider.catIDs + ',5060,5070',
"attrs": "rageid"}
if self.provider.needs_auth and self.provider.key:
params['apikey'] = self.provider.key
rss_url = self.provider.url + 'api?' + urllib.urlencode(params)
logger.log(self.provider.name + " cache update URL: " + rss_url, logger.DEBUG)
return self.getRSSFeed(rss_url)
def _checkAuth(self, data):
return self.provider._checkAuthFromData(data)
def updateCache(self):
if self.shouldUpdate() and self._checkAuth(None):
data = self._getRSSData()
# as long as the http request worked we count this as an update
if not data:
return []
self.setLastUpdate()
# clear cache
self._clearCache()
if self._checkAuth(data):
items = data.entries
cl = []
for item in items:
ci = self._parseItem(item)
if ci is not None:
cl.append(ci)
if len(cl) > 0:
myDB = self._getDB()
myDB.mass_action(cl)
else:
raise AuthException(
u"Your authentication credentials for " + self.provider.name + " are incorrect, check your config")
return []
# overwrite method with that parses the rageid from the newznab feed
def _parseItem(self, item):
title = item.title
url = item.link
attrs = item.newznab_attr
if not isinstance(attrs, list):
attrs = [item.newznab_attr]
tvrageid = 0
for attr in attrs:
if attr['name'] == 'tvrageid':
tvrageid = int(attr['value'])
break
self._checkItemAuth(title, url)
if not title or not url:
logger.log(
u"The data returned from the " + self.provider.name + " feed is incomplete, this result is unusable",
logger.DEBUG)
return None
url = self._translateLinkURL(url)
logger.log(u"Attempting to add item from RSS to cache: " + title, logger.DEBUG)
return self._addCacheEntry(title, url, indexer_id=tvrageid)