mirror of
https://github.com/moparisthebest/SickRage
synced 2024-12-12 19:12:26 -05:00
9a3e7ab0a9
Added fanzub anime nzb provider. Fixed NyaaTorrents anime provider. This is in testing phase so bugs are to be expected.
356 lines
12 KiB
Python
356 lines
12 KiB
Python
# Author: Mr_Orange <mr_orange@hotmail.it>
|
|
# 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/>.
|
|
|
|
from __future__ import with_statement
|
|
|
|
import time
|
|
import sys
|
|
import os
|
|
import traceback
|
|
import urllib, urlparse
|
|
import re
|
|
import datetime
|
|
|
|
import sickbeard
|
|
import generic
|
|
|
|
from sickbeard.common import Quality, cpu_presets
|
|
from sickbeard.name_parser.parser import NameParser, InvalidNameException
|
|
from sickbeard import logger
|
|
from sickbeard import tvcache
|
|
from sickbeard import helpers
|
|
from sickbeard import db
|
|
from sickbeard import classes
|
|
from sickbeard.show_name_helpers import allPossibleShowNames, sanitizeSceneName
|
|
from sickbeard.exceptions import ex
|
|
from sickbeard import encodingKludge as ek
|
|
from sickbeard import clients
|
|
|
|
from lib import requests
|
|
from lib.requests import exceptions
|
|
from bs4 import BeautifulSoup
|
|
from lib.unidecode import unidecode
|
|
|
|
|
|
class PublicHDProvider(generic.TorrentProvider):
|
|
def __init__(self):
|
|
|
|
generic.TorrentProvider.__init__(self, "PublicHD")
|
|
|
|
self.supportsBacklog = True
|
|
|
|
self.enabled = False
|
|
self.ratio = None
|
|
self.minseed = None
|
|
self.minleech = None
|
|
|
|
self.cache = PublicHDCache(self)
|
|
|
|
self.url = 'http://phdproxy.com/'
|
|
|
|
self.searchurl = self.url + 'index.php?page=torrents&search=%s&active=0&category=%s&order=5&by=2' #order by seed
|
|
|
|
self.categories = {'Season': ['23'], 'Episode': ['7', '14', '24'], 'RSS': ['7', '14', '23', '24']}
|
|
|
|
def isEnabled(self):
|
|
return self.enabled
|
|
|
|
def imageName(self):
|
|
return 'publichd.png'
|
|
|
|
def getQuality(self, item, anime=False):
|
|
|
|
quality = Quality.sceneQuality(item[0], anime)
|
|
return quality
|
|
|
|
def _get_season_search_strings(self, ep_obj):
|
|
search_string = {'Season': []}
|
|
|
|
for show_name in set(allPossibleShowNames(self.show)):
|
|
if ep_obj.show.air_by_date or ep_obj.show.sports:
|
|
ep_string = show_name + str(ep_obj.airdate).split('-')[0]
|
|
else:
|
|
ep_string = show_name + ' S%02d' % int(ep_obj.scene_season) #1) showName SXX -SXXE
|
|
search_string['Season'].append(ep_string)
|
|
|
|
if ep_obj.show.air_by_date or ep_obj.show.sports:
|
|
ep_string = show_name + ' Season ' + str(ep_obj.airdate).split('-')[0]
|
|
else:
|
|
ep_string = show_name + ' Season ' + str(ep_obj.scene_season) #2) showName Season X
|
|
search_string['Season'].append(ep_string)
|
|
|
|
return [search_string]
|
|
|
|
def _get_episode_search_strings(self, ep_obj, add_string=''):
|
|
|
|
search_string = {'Episode': []}
|
|
|
|
if not ep_obj:
|
|
return []
|
|
|
|
if self.show.air_by_date:
|
|
for show_name in set(allPossibleShowNames(self.show)):
|
|
ep_string = sanitizeSceneName(show_name) + ' ' + \
|
|
str(ep_obj.airdate).replace('-', '|')
|
|
search_string['Episode'].append(ep_string)
|
|
elif self.show.sports:
|
|
for show_name in set(allPossibleShowNames(self.show)):
|
|
ep_string = sanitizeSceneName(show_name) + ' ' + \
|
|
str(ep_obj.airdate).replace('-', '|') + '|' + \
|
|
ep_obj.airdate.strftime('%b')
|
|
search_string['Episode'].append(ep_string)
|
|
else:
|
|
for show_name in set(allPossibleShowNames(self.show)):
|
|
ep_string = sanitizeSceneName(show_name) + ' ' + \
|
|
sickbeard.config.naming_ep_type[2] % {'seasonnumber': ep_obj.scene_season,
|
|
'episodenumber': ep_obj.scene_episode}
|
|
|
|
for x in add_string.split('|'):
|
|
to_search = re.sub('\s+', ' ', ep_string + ' %s' % x)
|
|
search_string['Episode'].append(to_search)
|
|
|
|
return [search_string]
|
|
|
|
def _doSearch(self, search_params, epcount=0, age=0):
|
|
|
|
results = []
|
|
items = {'Season': [], 'Episode': [], 'RSS': []}
|
|
|
|
for mode in search_params.keys():
|
|
for search_string in search_params[mode]:
|
|
|
|
if mode == 'RSS':
|
|
searchURL = self.url + 'index.php?page=torrents&active=1&category=%s' % (
|
|
';'.join(self.categories[mode]))
|
|
logger.log(u"PublicHD cache update URL: " + searchURL, logger.DEBUG)
|
|
else:
|
|
searchURL = self.searchurl % (
|
|
urllib.quote(unidecode(search_string)), ';'.join(self.categories[mode]))
|
|
logger.log(u"Search string: " + searchURL, logger.DEBUG)
|
|
|
|
html = self.getURL(searchURL)
|
|
|
|
if not html:
|
|
continue
|
|
|
|
#remove unneccecary <option> lines which are slowing down BeautifulSoup
|
|
optreg = re.compile(r'<option.*</option>')
|
|
html = os.linesep.join([s for s in html.splitlines() if not optreg.search(s)])
|
|
|
|
try:
|
|
soup = BeautifulSoup(html, features=["html5lib", "permissive"])
|
|
|
|
torrent_table = soup.find('table', attrs={'id': 'torrbg'})
|
|
torrent_rows = torrent_table.find_all('tr') if torrent_table else []
|
|
|
|
#Continue only if one Release is found
|
|
if len(torrent_rows) < 2:
|
|
logger.log(u"The Data returned from " + self.name + " do not contains any torrent",
|
|
logger.DEBUG)
|
|
continue
|
|
|
|
for tr in torrent_rows[1:]:
|
|
|
|
try:
|
|
link = self.url + tr.find(href=re.compile('page=torrent-details'))['href']
|
|
title = tr.find(lambda x: x.has_attr('title')).text.replace('_', '.')
|
|
url = tr.find(href=re.compile('magnet+'))['href']
|
|
seeders = int(tr.find_all('td', {'class': 'header'})[4].text)
|
|
leechers = int(tr.find_all('td', {'class': 'header'})[5].text)
|
|
except (AttributeError, TypeError):
|
|
continue
|
|
|
|
if mode != 'RSS' and (seeders == 0 or seeders < self.minseed or leechers < self.minleech):
|
|
continue
|
|
|
|
if not title or not url:
|
|
continue
|
|
|
|
item = title, url, link, seeders, leechers
|
|
|
|
items[mode].append(item)
|
|
|
|
except Exception, e:
|
|
logger.log(u"Failed to parsing " + self.name + " Traceback: " + traceback.format_exc(),
|
|
logger.ERROR)
|
|
|
|
#For each search mode sort all the items by seeders
|
|
items[mode].sort(key=lambda tup: tup[3], reverse=True)
|
|
|
|
results += items[mode]
|
|
|
|
return results
|
|
|
|
def _get_title_and_url(self, item):
|
|
|
|
title, url, id, seeders, leechers = item
|
|
|
|
if url:
|
|
url = url.replace('&', '&')
|
|
|
|
return (title, url)
|
|
|
|
def getURL(self, url, post_data=None, headers=None, json=False):
|
|
|
|
if not self.session:
|
|
self.session = requests.Session()
|
|
|
|
try:
|
|
# Remove double-slashes from url
|
|
parsed = list(urlparse.urlparse(url))
|
|
parsed[2] = re.sub("/{2,}", "/", parsed[2]) # replace two or more / with one
|
|
url = urlparse.urlunparse(parsed)
|
|
|
|
r = self.session.get(url, verify=False)
|
|
except (requests.exceptions.ConnectionError, requests.exceptions.HTTPError), e:
|
|
logger.log(u"Error loading " + self.name + " URL: " + str(sys.exc_info()) + " - " + ex(e), logger.ERROR)
|
|
return None
|
|
|
|
if r.status_code != 200:
|
|
logger.log(self.name + u" page requested with url " + url + " returned status code is " + str(
|
|
r.status_code) + ': ' + clients.http_error_code[r.status_code], logger.WARNING)
|
|
return None
|
|
|
|
return r.content
|
|
|
|
def downloadResult(self, result):
|
|
"""
|
|
Save the result to disk.
|
|
"""
|
|
|
|
if not self.session:
|
|
self.session = requests.Session()
|
|
|
|
torrent_hash = re.findall('urn:btih:([\w]{32,40})', result.url)[0].upper()
|
|
|
|
if not torrent_hash:
|
|
logger.log("Unable to extract torrent hash from link: " + ex(result.url), logger.ERROR)
|
|
return False
|
|
|
|
try:
|
|
r = self.session.get('http://torcache.net/torrent/' + torrent_hash + '.torrent', verify=False)
|
|
except Exception, e:
|
|
logger.log("Unable to connect to TORCACHE: " + ex(e), logger.ERROR)
|
|
try:
|
|
logger.log("Trying TORRAGE cache instead")
|
|
r = self.session.get('http://torrage.com/torrent/' + torrent_hash + '.torrent', verify=False)
|
|
except Exception, e:
|
|
logger.log("Unable to connect to TORRAGE: " + ex(e), logger.ERROR)
|
|
return False
|
|
|
|
if not r.status_code == 200:
|
|
return False
|
|
|
|
magnetFileName = ek.ek(os.path.join, sickbeard.TORRENT_DIR,
|
|
helpers.sanitizeFileName(result.name) + '.' + self.providerType)
|
|
magnetFileContent = r.content
|
|
|
|
try:
|
|
with open(magnetFileName, 'wb') as fileOut:
|
|
fileOut.write(magnetFileContent)
|
|
|
|
helpers.chmodAsParent(magnetFileName)
|
|
|
|
except EnvironmentError, e:
|
|
logger.log("Unable to save the file: " + ex(e), logger.ERROR)
|
|
return False
|
|
|
|
logger.log(u"Saved magnet link to " + magnetFileName + " ", logger.MESSAGE)
|
|
return True
|
|
|
|
def findPropers(self, search_date=datetime.datetime.today()):
|
|
|
|
results = []
|
|
|
|
sqlResults = db.DBConnection().select(
|
|
'SELECT s.show_name, e.showid, e.season, e.episode, e.status, e.airdate FROM tv_episodes AS e' +
|
|
' INNER JOIN tv_shows AS s ON (e.showid = s.indexer_id)' +
|
|
' WHERE e.airdate >= ' + str(search_date.toordinal()) +
|
|
' AND (e.status IN (' + ','.join([str(x) for x in Quality.DOWNLOADED]) + ')' +
|
|
' OR (e.status IN (' + ','.join([str(x) for x in Quality.SNATCHED]) + ')))'
|
|
)
|
|
if not sqlResults:
|
|
return []
|
|
|
|
for sqlshow in sqlResults:
|
|
self.show = curshow = helpers.findCertainShow(sickbeard.showList, int(sqlshow["showid"]))
|
|
if not self.show: continue
|
|
curEp = curshow.getEpisode(int(sqlshow["season"]), int(sqlshow["episode"]))
|
|
|
|
searchString = self._get_episode_search_strings(curEp, add_string='PROPER|REPACK')
|
|
|
|
for item in self._doSearch(searchString[0]):
|
|
title, url = self._get_title_and_url(item)
|
|
results.append(classes.Proper(title, url, datetime.datetime.today()))
|
|
|
|
return results
|
|
|
|
def seedRatio(self):
|
|
return self.ratio
|
|
|
|
|
|
class PublicHDCache(tvcache.TVCache):
|
|
def __init__(self, provider):
|
|
|
|
tvcache.TVCache.__init__(self, provider)
|
|
|
|
# only poll ThePirateBay every 10 minutes max
|
|
self.minTime = 20
|
|
|
|
def updateCache(self):
|
|
|
|
# delete anything older then 7 days
|
|
logger.log(u"Clearing " + self.provider.name + " cache")
|
|
self._clearCache()
|
|
|
|
if not self.shouldUpdate():
|
|
return
|
|
|
|
search_params = {'RSS': ['rss']}
|
|
rss_results = self.provider._doSearch(search_params)
|
|
|
|
if rss_results:
|
|
self.setLastUpdate()
|
|
else:
|
|
return []
|
|
|
|
ql = []
|
|
for result in rss_results:
|
|
|
|
item = (result[0], result[1])
|
|
ci = self._parseItem(item)
|
|
if ci is not None:
|
|
ql.append(ci)
|
|
|
|
myDB = self._getDB()
|
|
myDB.mass_action(ql)
|
|
|
|
def _parseItem(self, item):
|
|
|
|
(title, url) = item
|
|
|
|
if not title or not url:
|
|
return None
|
|
|
|
logger.log(u"Attempting to cache item:[" + title +"]", logger.DEBUG)
|
|
|
|
return self._addCacheEntry(title, url)
|
|
|
|
|
|
provider = PublicHDProvider()
|