1
0
mirror of https://github.com/moparisthebest/SickRage synced 2024-11-17 06:45:05 -05:00
SickRage/sickbeard/tvcache.py
2014-03-11 22:30:49 -07:00

419 lines
16 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 time
import datetime
import sqlite3
import sickbeard
from sickbeard import db
from sickbeard import logger
from sickbeard.common import Quality
from sickbeard import helpers, show_name_helpers
from sickbeard import name_cache
from sickbeard.exceptions import MultipleShowObjectsException, ex
from sickbeard.exceptions import ex, AuthException
try:
import xml.etree.cElementTree as etree
except ImportError:
import elementtree.ElementTree as etree
from sickbeard.indexers import indexer_api, indexer_exceptions
from name_parser.parser import NameParser, InvalidNameException
class CacheDBConnection(db.DBConnection):
def __init__(self, providerName):
db.DBConnection.__init__(self, "cache.db")
# Create the table if it's not already there
try:
sql = "CREATE TABLE [" + providerName + "] (name TEXT, season NUMERIC, episodes TEXT, indexerid NUMERIC, url TEXT, time NUMERIC, quality TEXT);"
self.connection.execute(sql)
self.connection.commit()
except sqlite3.OperationalError, e:
if str(e) != "table [" + providerName + "] already exists":
raise
# Create the table if it's not already there
try:
sql = "CREATE TABLE lastUpdate (provider TEXT, time NUMERIC);"
self.connection.execute(sql)
self.connection.commit()
except sqlite3.OperationalError, e:
if str(e) != "table lastUpdate already exists":
raise
class TVCache():
def __init__(self, provider):
self.provider = provider
self.providerID = self.provider.getID()
self.minTime = 10
def _getDB(self):
return CacheDBConnection(self.providerID)
def _clearCache(self):
myDB = self._getDB()
myDB.action("DELETE FROM [" + self.providerID + "] WHERE 1")
def _getRSSData(self):
data = None
return data
def _checkAuth(self, parsedXML):
return True
def _checkItemAuth(self, title, url):
return True
def updateCache(self):
if not self.shouldUpdate():
return
if self._checkAuth(None):
data = self._getRSSData()
# as long as the http request worked we count this as an update
if data:
self.setLastUpdate()
else:
return []
# now that we've loaded the current RSS feed lets delete the old cache
logger.log(u"Clearing " + self.provider.name + " cache and updating with new information")
self._clearCache()
parsedXML = helpers.parse_xml(data)
if parsedXML is None:
logger.log(u"Error trying to load " + self.provider.name + " RSS feed", logger.ERROR)
return []
if self._checkAuth(parsedXML):
if parsedXML.tag == 'rss':
items = parsedXML.findall('.//item')
else:
logger.log(u"Resulting XML from " + self.provider.name + " isn't RSS, not parsing it", logger.ERROR)
return []
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 []
def _translateTitle(self, title):
return title.replace(' ', '.')
def _translateLinkURL(self, url):
return url.replace('&amp;', '&')
def _parseItem(self, item):
title = helpers.get_xml_text(item.find('title'))
url = helpers.get_xml_text(item.find('link'))
self._checkItemAuth(title, url)
if title and url:
title = self._translateTitle(title)
url = self._translateLinkURL(url)
logger.log(u"Adding item from RSS to cache: " + title, logger.DEBUG)
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.DEBUG)
return None
def _getLastUpdate(self):
myDB = self._getDB()
sqlResults = myDB.select("SELECT time FROM lastUpdate WHERE provider = ?", [self.providerID])
if sqlResults:
lastTime = int(sqlResults[0]["time"])
if lastTime > int(time.mktime(datetime.datetime.today().timetuple())):
lastTime = 0
else:
lastTime = 0
return datetime.datetime.fromtimestamp(lastTime)
def setLastUpdate(self, toDate=None):
if not toDate:
toDate = datetime.datetime.today()
myDB = self._getDB()
myDB.upsert("lastUpdate",
{'time': int(time.mktime(toDate.timetuple()))},
{'provider': self.providerID})
lastUpdate = property(_getLastUpdate)
def shouldUpdate(self):
# if we've updated recently then skip the update
if datetime.datetime.today() - self.lastUpdate < datetime.timedelta(minutes=self.minTime):
logger.log(u"Last update was too soon, using old cache: today()-" + str(self.lastUpdate) + "<" + str(datetime.timedelta(minutes=self.minTime)), logger.DEBUG)
return False
return True
def _addCacheEntry(self, name, url, season=None, episodes=None, indexer_id=0, quality=None, extraNames=[]):
myDB = self._getDB()
parse_result = None
# if we don't have complete info then parse the filename to get it
for curName in [name] + extraNames:
try:
myParser = NameParser()
parse_result = myParser.parse(curName, True)
except InvalidNameException:
logger.log(u"Unable to parse the filename " + curName + " into a valid episode", logger.DEBUG)
continue
if not parse_result:
logger.log(u"Giving up because I'm unable to parse this name: "+name, logger.DEBUG)
return None
if not parse_result.series_name:
logger.log(u"No series name retrieved from " + name + ", unable to cache it", logger.DEBUG)
return None
indexer_lang = None
# if we need indexer_id then search the DB for them
if not indexer_id:
# if we have only the indexer_id, use the database
if indexer_id:
try:
showObj = helpers.findCertainShow(sickbeard.showList, indexer_id)
except (MultipleShowObjectsException):
showObj = None
if showObj:
indexer_lang = showObj.lang
# if they're both empty then fill out as much info as possible by searching the show name
else:
# check the name cache and see if we already know what show this is
logger.log(u"Checking the cache to see if we already know the indexer id of "+parse_result.series_name, logger.DEBUG)
indexer_id = name_cache.retrieveNameFromCache(parse_result.series_name)
# remember if the cache lookup worked or not so we know whether we should bother updating it later
if indexer_id == None:
logger.log(u"No cache results returned, continuing on with the search", logger.DEBUG)
from_cache = False
else:
logger.log(u"Cache lookup found " + repr(indexer_id) + ", using that", logger.DEBUG)
from_cache = True
# if the cache failed, try looking up the show name in the database
if indexer_id == None:
logger.log(u"Trying to look the show up in the show database", logger.DEBUG)
showResult = helpers.searchDBForShow(parse_result.series_name)
if showResult:
logger.log(parse_result.series_name + " was found to be show " + showResult[2] + " ("+str(showResult[1])+") in our DB.", logger.DEBUG)
indexer_id = showResult[1]
# if the DB lookup fails then do a comprehensive regex search
if indexer_id == None:
logger.log(u"Couldn't figure out a show name straight from the DB, trying a regex search instead", logger.DEBUG)
for curShow in sickbeard.showList:
if show_name_helpers.isGoodResult(name, curShow, False):
logger.log(u"Successfully matched " + name + " to " + curShow.name + " with regex", logger.DEBUG)
indexer_id = curShow.indexerid
indexer_lang = curShow.lang
break
# if indexer_id was anything but None (0 or a number) then
if not from_cache:
name_cache.addNameToCache(parse_result.series_name, indexer_id)
# if we came out with indexer_id = None it means we couldn't figure it out at all, just use 0 for that
if indexer_id == None:
indexer_id = 0
# if we found the show then retrieve the show object
if indexer_id:
try:
showObj = helpers.findCertainShow(sickbeard.showList, indexer_id)
except (MultipleShowObjectsException):
showObj = None
if showObj:
self.indexer = showObj.indexer
indexer_lang = showObj.lang
# if we weren't provided with season/episode information then get it from the name that we parsed
if not season:
season = parse_result.season_number if parse_result.season_number != None else 1
if not episodes:
episodes = parse_result.episode_numbers
# if we have an air-by-date show then get the real season/episode numbers
if parse_result.air_by_date and indexer_id:
try:
lINDEXER_API_PARMS = {'indexer': self.indexer}
if not (indexer_lang == "" or indexer_lang == "en" or indexer_lang == None):
lINDEXER_API_PARMS['language'] = indexer_lang
t = indexer_api.indexerApi(**lINDEXER_API_PARMS)
epObj = t[indexer_id].airedOn(parse_result.air_date)[0]
season = int(epObj["seasonnumber"])
episodes = [int(epObj["episodenumber"])]
except indexer_exceptions.indexer_episodenotfound:
logger.log(u"Unable to find episode with date " + str(parse_result.air_date) + " for show " + parse_result.series_name+", skipping", logger.WARNING)
return None
except indexer_exceptions.indexer_error, e:
logger.log(u"Unable to contact " + self.indexer + ": " + ex(e), logger.WARNING)
return None
episodeText = "|"+"|".join(map(str, episodes))+"|"
# get the current timestamp
curTimestamp = int(time.mktime(datetime.datetime.today().timetuple()))
if not quality:
quality = Quality.sceneQuality(name)
if not isinstance(name, unicode):
name = unicode(name, 'utf-8')
myDB.action("INSERT INTO [" + self.providerID + "] (name, season, episodes, indexerid, url, time, quality) VALUES (?,?,?,?,?,?,?)",
[name, season, episodeText, indexer_id, url, curTimestamp, quality])
def searchCache(self, episode, manualSearch=False):
neededEps = self.findNeededEpisodes(episode, manualSearch)
return neededEps[episode]
def listPropers(self, date=None, delimiter="."):
myDB = self._getDB()
sql = "SELECT * FROM [" + self.providerID + "] WHERE name LIKE '%.PROPER.%' OR name LIKE '%.REPACK.%'"
if date != None:
sql += " AND time >= " + str(int(time.mktime(date.timetuple())))
#return filter(lambda x: x['indexerid'] != 0, myDB.select(sql))
return myDB.select(sql)
def findNeededEpisodes(self, episode=None, manualSearch=False):
neededEps = {}
if episode:
neededEps[episode] = []
myDB = self._getDB()
if not episode:
sqlResults = myDB.select("SELECT * FROM [" + self.providerID + "]")
else:
sqlResults = myDB.select("SELECT * FROM [" + self.providerID + "] WHERE indexerid = ? AND season = ? AND episodes LIKE ?", [episode.show.indexerid, episode.season, "%|" + str(episode.episode) + "|%"])
# for each cache entry
for curResult in sqlResults:
# skip non-tv crap (but allow them for Newzbin cause we assume it's filtered well)
if self.providerID != 'newzbin' and not show_name_helpers.filterBadReleases(curResult["name"]):
continue
# get the show object, or if it's not one of our shows then ignore it
try:
showObj = helpers.findCertainShow(sickbeard.showList, int(curResult["indexerid"]))
except (MultipleShowObjectsException):
showObj = None
if not showObj:
continue
# get season and ep data (ignoring multi-eps for now)
curSeason = int(curResult["season"])
if curSeason == -1:
continue
curEp = curResult["episodes"].split("|")[1]
if not curEp:
continue
curEp = int(curEp)
curQuality = int(curResult["quality"])
# if the show says we want that episode then add it to the list
if not showObj.wantEpisode(curSeason, curEp, curQuality, manualSearch):
logger.log(u"Skipping " + curResult["name"] + " because we don't want an episode that's " + Quality.qualityStrings[curQuality], logger.DEBUG)
else:
if episode:
epObj = episode
else:
epObj = showObj.getEpisode(curSeason, curEp)
# build a result object
title = curResult["name"]
url = curResult["url"]
logger.log(u"Found result " + title + " at " + url)
result = self.provider.getResult([epObj])
result.url = url
result.name = title
result.quality = curQuality
result.content = self.provider.getURL(url) \
if self.provider.providerType == sickbeard.providers.generic.GenericProvider.TORRENT \
and not url.startswith('magnet') else None
# add it to the list
if epObj not in neededEps:
neededEps[epObj] = [result]
else:
neededEps[epObj].append(result)
return neededEps