1
0
mirror of https://github.com/moparisthebest/SickRage synced 2024-12-12 11:02:21 -05:00
SickRage/sickbeard/tvcache.py

381 lines
14 KiB
Python
Raw Normal View History

# Author: Nic Wolfe <nic@wolfeden.ca>
# URL: http://code.google.com/p/sickbeard/
#
2014-05-23 08:37:22 -04:00
# This file is part of SickRage.
#
2014-05-23 08:37:22 -04:00
# 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.
#
2014-05-23 08:37:22 -04:00
# 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
2014-05-23 08:37:22 -04:00
# along with SickRage. If not, see <http://www.gnu.org/licenses/>.
from __future__ import with_statement
import time
import datetime
import sickbeard
from sickbeard import db
from sickbeard import logger
from sickbeard.common import Quality
from sickbeard import helpers, show_name_helpers
from sickbeard.exceptions import MultipleShowObjectsException
from sickbeard.exceptions import AuthException
from name_parser.parser import NameParser, InvalidNameException, InvalidShowException
from sickbeard.rssfeeds import RSSFeeds
from sickbeard import clients
import itertools
class CacheDBConnection(db.DBConnection):
def __init__(self, providerName):
db.DBConnection.__init__(self, "cache.db")
# Create the table if it's not already there
try:
if not self.hasTable(providerName):
self.action(
"CREATE TABLE [" + providerName + "] (name TEXT, season NUMERIC, episodes TEXT, indexerid NUMERIC, url TEXT, time NUMERIC, quality TEXT, release_group TEXT)")
else:
2014-06-27 23:33:31 -04:00
sqlResults = self.select(
"SELECT url, COUNT(url) as count FROM [" + providerName + "] GROUP BY url HAVING count > 1")
for cur_dupe in sqlResults:
self.action("DELETE FROM [" + providerName + "] WHERE url = ?", [cur_dupe["url"]])
# add unique index to prevent further dupes from happening if one does not exist
self.action("CREATE UNIQUE INDEX IF NOT EXISTS idx_url ON [" + providerName + "] (url)")
# add release_group column to table if missing
if not self.hasColumn(providerName, 'release_group'):
self.addColumn(providerName, 'release_group', "TEXT", "")
Fixed issues with editing/saving custom scene exceptions. Fixed charmap issues for anime show names. Fixed issues with display show page and epCat key errors. Fixed duplicate log messages for clearing provider caches. Fixed issues with email notifier ep names not properly being encoded to UTF-8. TVDB<->TVRAGE Indexer ID mapping is now performed on demand to be used when needed such as newznab providers can be searched with tvrage_id's and some will return tvrage_id's that later can be used to create show objects from for faster and more accurate name parsing, mapping is done via Trakt API calls. Added stop event signals to schedualed tasks, SR now waits indefinate till task has been fully stopped before completing a restart or shutdown event. NameParserCache is now persistent and stores 200 parsed results at any given time for quicker lookups and better performance, this helps maintain results between updates or shutdown/startup events. Black and White lists for anime now only get used for anime shows as intended, performance gain for non-anime shows that dont need to load these lists. Internal name cache now builds it self on demand when needed per show request plus checks if show is already in cache and if true exits routine to save time. Schedualer and QueueItems classes are now a sub-class of threading.Thread and a stop threading event signal has been added to each. If I forgot to list something it doesn't mean its not fixed so please test and report back if anything is wrong or has been corrected by this new release.
2014-07-14 22:00:53 -04:00
2014-07-22 00:53:32 -04:00
# add version column to table if missing
if not self.hasColumn(providerName, 'version'):
self.addColumn(providerName, 'version', "NUMERIC", "-1")
except Exception, e:
if str(e) != "table [" + providerName + "] already exists":
raise
# Create the table if it's not already there
try:
if not self.hasTable('lastUpdate'):
self.action("CREATE TABLE lastUpdate (provider TEXT, time NUMERIC)")
except Exception, e:
if str(e) != "table lastUpdate already exists":
raise
class TVCache():
def __init__(self, provider):
self.provider = provider
self.providerID = self.provider.getID()
self.providerDB = None
self.minTime = 10
def _getDB(self):
# init provider database if not done already
if not self.providerDB:
self.providerDB = CacheDBConnection(self.providerID)
return self.providerDB
def _clearCache(self):
if self.shouldClearCache():
myDB = self._getDB()
myDB.action("DELETE FROM [" + self.providerID + "] WHERE 1")
def _get_title_and_url(self, item):
# override this in the provider if daily search has a different data layout to backlog searches
return self.provider._get_title_and_url(item)
def _getRSSData(self):
data = None
return data
def _checkAuth(self):
return self.provider._checkAuth()
def _checkItemAuth(self, title, url):
return True
def updateCache(self):
if self.shouldUpdate() and self._checkAuth():
# as long as the http request worked we count this as an update
data = self._getRSSData()
if not data:
return []
# clear cache
self._clearCache()
# set updated
self.setLastUpdate()
# parse data
cl = []
for item in data:
title, url = self._get_title_and_url(item)
ci = self._parseItem(title, url)
if ci is not None:
cl.append(ci)
if len(cl) > 0:
myDB = self._getDB()
myDB.mass_action(cl)
return []
2014-06-29 23:50:12 -04:00
def getRSSFeed(self, url, post_data=None, request_headers=None):
return RSSFeeds(self.providerID).getFeed(url, post_data, request_headers)
def _translateTitle(self, title):
return u'' + title.replace(' ', '.')
def _translateLinkURL(self, url):
return url.replace('&amp;', '&')
2014-08-05 12:19:54 -04:00
def _parseItem(self, title, url):
self._checkItemAuth(title, url)
if title and url:
title = self._translateTitle(title)
url = self._translateLinkURL(url)
2014-08-05 12:19:54 -04:00
logger.log(u"Attempting to add item to cache: " + title, logger.DEBUG)
return self._addCacheEntry(title, url)
else:
logger.log(
2014-04-26 00:08:27 -04:00
u"The data 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 _getLastSearch(self):
myDB = self._getDB()
sqlResults = myDB.select("SELECT time FROM lastSearch 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})
def setLastSearch(self, toDate=None):
if not toDate:
toDate = datetime.datetime.today()
myDB = self._getDB()
myDB.upsert("lastSearch",
{'time': int(time.mktime(toDate.timetuple()))},
{'provider': self.providerID})
lastUpdate = property(_getLastUpdate)
lastSearch = property(_getLastSearch)
def shouldUpdate(self):
# if we've updated recently then skip the update
2014-06-30 11:57:32 -04:00
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 shouldClearCache(self):
# if daily search hasn't used our previous results yet then don't clear the cache
if self.lastUpdate > self.lastSearch:
return False
return True
def _addCacheEntry(self, name, url, parse_result=None, indexer_id=0):
Fixed issues with editing/saving custom scene exceptions. Fixed charmap issues for anime show names. Fixed issues with display show page and epCat key errors. Fixed duplicate log messages for clearing provider caches. Fixed issues with email notifier ep names not properly being encoded to UTF-8. TVDB<->TVRAGE Indexer ID mapping is now performed on demand to be used when needed such as newznab providers can be searched with tvrage_id's and some will return tvrage_id's that later can be used to create show objects from for faster and more accurate name parsing, mapping is done via Trakt API calls. Added stop event signals to schedualed tasks, SR now waits indefinate till task has been fully stopped before completing a restart or shutdown event. NameParserCache is now persistent and stores 200 parsed results at any given time for quicker lookups and better performance, this helps maintain results between updates or shutdown/startup events. Black and White lists for anime now only get used for anime shows as intended, performance gain for non-anime shows that dont need to load these lists. Internal name cache now builds it self on demand when needed per show request plus checks if show is already in cache and if true exits routine to save time. Schedualer and QueueItems classes are now a sub-class of threading.Thread and a stop threading event signal has been added to each. If I forgot to list something it doesn't mean its not fixed so please test and report back if anything is wrong or has been corrected by this new release.
2014-07-14 22:00:53 -04:00
# check if we passed in a parsed result or should we try and create one
if not parse_result:
# create showObj from indexer_id if available
showObj=None
if indexer_id:
showObj = helpers.findCertainShow(sickbeard.showList, indexer_id)
try:
myParser = NameParser(showObj=showObj, convert=True)
parse_result = myParser.parse(name)
except InvalidNameException:
logger.log(u"Unable to parse the filename " + name + " into a valid episode", logger.DEBUG)
return None
except InvalidShowException:
logger.log(u"Unable to parse the filename " + name + " into a valid show", logger.DEBUG)
return None
if not parse_result or not parse_result.series_name:
return None
# if we made it this far then lets add the parsed result to cache for usager later on
season = parse_result.season_number if parse_result.season_number else 1
episodes = parse_result.episode_numbers
2014-04-26 01:42:40 -04:00
if season and episodes:
# store episodes as a seperated string
episodeText = "|" + "|".join(map(str, episodes)) + "|"
# get the current timestamp
curTimestamp = int(time.mktime(datetime.datetime.today().timetuple()))
# get quality of release
quality = parse_result.quality
if not isinstance(name, unicode):
Fixed issues with editing/saving custom scene exceptions. Fixed charmap issues for anime show names. Fixed issues with display show page and epCat key errors. Fixed duplicate log messages for clearing provider caches. Fixed issues with email notifier ep names not properly being encoded to UTF-8. TVDB<->TVRAGE Indexer ID mapping is now performed on demand to be used when needed such as newznab providers can be searched with tvrage_id's and some will return tvrage_id's that later can be used to create show objects from for faster and more accurate name parsing, mapping is done via Trakt API calls. Added stop event signals to schedualed tasks, SR now waits indefinate till task has been fully stopped before completing a restart or shutdown event. NameParserCache is now persistent and stores 200 parsed results at any given time for quicker lookups and better performance, this helps maintain results between updates or shutdown/startup events. Black and White lists for anime now only get used for anime shows as intended, performance gain for non-anime shows that dont need to load these lists. Internal name cache now builds it self on demand when needed per show request plus checks if show is already in cache and if true exits routine to save time. Schedualer and QueueItems classes are now a sub-class of threading.Thread and a stop threading event signal has been added to each. If I forgot to list something it doesn't mean its not fixed so please test and report back if anything is wrong or has been corrected by this new release.
2014-07-14 22:00:53 -04:00
name = unicode(name, 'utf-8', 'replace')
# get release group
release_group = parse_result.release_group
2014-07-22 00:53:32 -04:00
# get version
version = parse_result.version
logger.log(u"Added RSS item: [" + name + "] to cache: [" + self.providerID + "]", logger.DEBUG)
return [
2014-07-22 00:53:32 -04:00
"INSERT OR IGNORE INTO [" + self.providerID + "] (name, season, episodes, indexerid, url, time, quality, release_group, version) VALUES (?,?,?,?,?,?,?,?,?)",
[name, season, episodeText, parse_result.show.indexerid, url, curTimestamp, quality, release_group, version]]
def searchCache(self, episode, manualSearch=False):
neededEps = self.findNeededEpisodes(episode, manualSearch)
if len(neededEps) > 0:
return neededEps[episode]
else:
return []
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))
def findNeededEpisodes(self, episode, manualSearch=False):
neededEps = {}
cl = []
myDB = self._getDB()
if type(episode) != list:
sqlResults = myDB.select(
"SELECT * FROM [" + self.providerID + "] WHERE indexerid = ? AND season = ? AND episodes LIKE ?",
[episode.show.indexerid, episode.season, "%|" + str(episode.episode) + "|%"])
else:
for epObj in episode:
cl.append([
"SELECT * FROM [" + self.providerID + "] WHERE indexerid = ? AND season = ? AND episodes LIKE ? "
"AND quality IN (" + ",".join([str(x) for x in epObj.wantedQuality]) + ")",
[epObj.show.indexerid, epObj.season, "%|" + str(epObj.episode) + "|%"]])
sqlResults = myDB.mass_action(cl, fetchall=True)
sqlResults = list(itertools.chain(*sqlResults))
# for each cache entry
for curResult in sqlResults:
# skip non-tv crap
if not show_name_helpers.filterBadReleases(curResult["name"], parse=False):
continue
# get the show object, or if it's not one of our shows then ignore it
showObj = helpers.findCertainShow(sickbeard.showList, int(curResult["indexerid"]))
if not showObj:
continue
# skip if provider is anime only and show is not anime
if self.provider.anime_only and not showObj.is_anime:
logger.log(u"" + str(showObj.name) + " is not an anime, skiping", logger.DEBUG)
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"])
curReleaseGroup = curResult["release_group"]
curVersion = curResult["version"]
# 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)
continue
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.show = showObj
result.url = url
result.name = title
result.quality = curQuality
result.release_group = curReleaseGroup
result.version = curVersion
result.content = None
# add it to the list
if epObj not in neededEps:
neededEps[epObj] = [result]
else:
neededEps[epObj].append(result)
# datetime stamp this search so cache gets cleared
self.setLastSearch()
return neededEps