mirror of
https://github.com/SickGear/SickGear.git
synced 2024-12-11 05:33:37 +00:00
236 lines
10 KiB
Python
236 lines
10 KiB
Python
# coding=utf-8
|
|
#
|
|
# This file is part of SickGear.
|
|
#
|
|
# SickGear 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.
|
|
#
|
|
# SickGear 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 SickGear. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
from __future__ import with_statement
|
|
|
|
import re
|
|
import os
|
|
import datetime
|
|
import traceback
|
|
import urllib
|
|
|
|
from . import generic
|
|
from sickbeard import config, logger, tvcache, show_name_helpers, helpers
|
|
from sickbeard.bs4_parser import BS4Parser
|
|
from sickbeard.common import Quality, mediaExtensions
|
|
from sickbeard.name_parser.parser import NameParser, InvalidNameException, InvalidShowException
|
|
from lib.unidecode import unidecode
|
|
|
|
|
|
class KATProvider(generic.TorrentProvider):
|
|
|
|
def __init__(self):
|
|
generic.TorrentProvider.__init__(self, 'KickAssTorrents')
|
|
|
|
self.url_base = 'https://kat.ph/'
|
|
self.urls = {'config_provider_home_uri': self.url_base,
|
|
'search': [self.url_base, 'http://katproxy.com/'],
|
|
'cache_params': 'tv/?field=time_add&sorder=desc',
|
|
'search_params': 'usearch/%s/?field=seeders&sorder=desc'}
|
|
|
|
self.url = self.urls['config_provider_home_uri']
|
|
|
|
self.minseed, self.minleech = 2 * [None]
|
|
self.confirmed = False
|
|
self.cache = KATCache(self)
|
|
|
|
def _find_season_quality(self, title, torrent_link, ep_number):
|
|
""" Return the modified title of a Season Torrent with the quality found inspecting torrent file list """
|
|
|
|
quality = Quality.UNKNOWN
|
|
|
|
file_name = None
|
|
|
|
data = self.get_url(torrent_link)
|
|
if not data:
|
|
return None
|
|
|
|
try:
|
|
with BS4Parser(data, features=['html5lib', 'permissive']) as soup:
|
|
file_table = soup.find('table', attrs={'class': 'torrentFileList'})
|
|
|
|
if not file_table:
|
|
return None
|
|
|
|
files = [x.text for x in file_table.find_all('td', attrs={'class': 'torFileName'})]
|
|
video_files = filter(lambda i: i.rpartition('.')[2].lower() in mediaExtensions, files)
|
|
|
|
# Filtering SingleEpisode/MultiSeason Torrent
|
|
if len(video_files) < ep_number or len(video_files) > float(ep_number * 1.1):
|
|
logger.log(u'Result %s lists %s episodes with %s episodes retrieved in torrent'
|
|
% (title, ep_number, len(video_files)), logger.DEBUG)
|
|
logger.log(u'Result %s seem to be a single episode or multi-season torrent, skipping result...'
|
|
% title, logger.DEBUG)
|
|
return None
|
|
|
|
if Quality.UNKNOWN != Quality.sceneQuality(title):
|
|
return title
|
|
|
|
for file_name in video_files:
|
|
quality = Quality.sceneQuality(os.path.basename(file_name))
|
|
if Quality.UNKNOWN != quality:
|
|
break
|
|
|
|
if None is not file_name and Quality.UNKNOWN == quality:
|
|
quality = Quality.assumeQuality(os.path.basename(file_name))
|
|
|
|
if Quality.UNKNOWN == quality:
|
|
logger.log(u'Unable to obtain a Season Quality for ' + title, logger.DEBUG)
|
|
return None
|
|
|
|
try:
|
|
my_parser = NameParser(showObj=self.show)
|
|
parse_result = my_parser.parse(file_name)
|
|
except (InvalidNameException, InvalidShowException):
|
|
return None
|
|
|
|
logger.log(u'Season quality for %s is %s' % (title, Quality.qualityStrings[quality]), logger.DEBUG)
|
|
|
|
if parse_result.series_name and parse_result.season_number:
|
|
title = parse_result.series_name + ' S%02d %s' % (int(parse_result.season_number),
|
|
self._reverse_quality(quality))
|
|
return title
|
|
|
|
except Exception:
|
|
logger.log(u'Failed to quality parse ' + self.name + ' Traceback: ' + traceback.format_exc(), logger.ERROR)
|
|
|
|
def _get_season_search_strings(self, ep_obj, **kwargs):
|
|
|
|
if ep_obj.show.air_by_date or ep_obj.show.sports:
|
|
airdate = str(ep_obj.airdate).split('-')[0]
|
|
ep_detail = [airdate, 'Season ' + airdate]
|
|
elif ep_obj.show.anime:
|
|
ep_detail = '%02i' % ep_obj.scene_absolute_number
|
|
else:
|
|
ep_detail = ['S%(s)02i -S%(s)02iE' % {'s': ep_obj.scene_season},
|
|
'Season %s -Ep*' % ep_obj.scene_season]
|
|
|
|
return [{'Season': self._build_search_strings(ep_detail, append=(' category:tv', '')[self.show.anime])}]
|
|
|
|
def _get_episode_search_strings(self, ep_obj, add_string='', **kwargs):
|
|
|
|
if not ep_obj:
|
|
return []
|
|
|
|
if self.show.air_by_date or self.show.sports:
|
|
ep_detail = str(ep_obj.airdate).replace('-', ' ')
|
|
if self.show.sports:
|
|
ep_detail += '|' + ep_obj.airdate.strftime('%b')
|
|
elif self.show.anime:
|
|
ep_detail = '%02i' % ep_obj.scene_absolute_number
|
|
else:
|
|
ep_detail = '%s|%s' % (config.naming_ep_type[2] % {'seasonnumber': ep_obj.scene_season,
|
|
'episodenumber': ep_obj.scene_episode},
|
|
config.naming_ep_type[0] % {'seasonnumber': ep_obj.scene_season,
|
|
'episodenumber': ep_obj.scene_episode})
|
|
# include provider specific appends
|
|
if not isinstance(add_string, list):
|
|
add_string = [add_string]
|
|
add_string = [x + ' category:tv' for x in add_string]
|
|
|
|
return [{'Episode': self._build_search_strings(ep_detail, append=(add_string, '')[self.show.anime])}]
|
|
|
|
def _do_search(self, search_params, search_mode='eponly', epcount=0, age=0):
|
|
|
|
results = []
|
|
items = {'Season': [], 'Episode': [], 'Cache': []}
|
|
|
|
rc = dict((k, re.compile('(?i)' + v)) for (k, v) in {'link': 'normal'}.items())
|
|
url = 0
|
|
for mode in search_params.keys():
|
|
for search_string in search_params[mode]:
|
|
|
|
self.url = self.urls['search'][url]
|
|
search_args = ('search_params', 'cache_params')['Cache' == mode]
|
|
search_url = self.url + self.urls[search_args]
|
|
if 'Cache' != mode:
|
|
search_url %= urllib.quote(unidecode(search_string))
|
|
|
|
html = helpers.getURL(search_url)
|
|
|
|
cnt = len(items[mode])
|
|
try:
|
|
if not html or self._has_no_results(html) or re.search(r'did not match any documents', html):
|
|
if html and 'kastatic' not in html:
|
|
url += (1, 0)[url == len(self.urls['search'])]
|
|
raise generic.HaltParseException
|
|
|
|
with BS4Parser(html, features=['html5lib', 'permissive']) as soup:
|
|
torrent_table = soup.find('table', attrs={'class': 'data'})
|
|
torrent_rows = [] if not torrent_table else torrent_table.find_all('tr')
|
|
|
|
if 2 > len(torrent_rows):
|
|
raise generic.HaltParseException
|
|
|
|
for tr in torrent_rows[1:]:
|
|
try:
|
|
seeders, leechers = [int(tr.find_all('td')[x].get_text().strip()) for x in (-2, -1)]
|
|
if 'Cache' != mode and (seeders < self.minseed or leechers < self.minleech):
|
|
continue
|
|
|
|
info = tr.find('div', {'class': 'torrentname'})
|
|
title = (info.find_all('a')[1].get_text() or info.find('a', 'cellMainLink').get_text())\
|
|
.strip()
|
|
link = self.url + info.find('a', {'class': rc['link']})['href'].lstrip('/')
|
|
|
|
download_magnet = tr.find('a', 'imagnet')['href']
|
|
except (AttributeError, TypeError):
|
|
continue
|
|
|
|
if self.confirmed and not tr.find('a', 'iverify'):
|
|
logger.log(u'Skipping untrusted non-verified result: %s' % title, logger.DEBUG)
|
|
continue
|
|
|
|
# Check number video files = episode in season and find the real Quality for full season torrent analyzing files in torrent
|
|
if 'Season' == mode and 'sponly' == search_mode:
|
|
ep_number = int(epcount / len(set(show_name_helpers.allPossibleShowNames(self.show))))
|
|
title = self._find_season_quality(title, link, ep_number)
|
|
|
|
if title and download_magnet:
|
|
items[mode].append((title, download_magnet, seeders))
|
|
|
|
except generic.HaltParseException:
|
|
pass
|
|
except Exception:
|
|
logger.log(u'Failed to parse. Traceback: %s' % traceback.format_exc(), logger.ERROR)
|
|
self._log_result(mode, len(items[mode]) - cnt, search_url)
|
|
|
|
# For each search mode sort all the items by seeders
|
|
items[mode].sort(key=lambda tup: tup[2], reverse=True)
|
|
|
|
results += items[mode]
|
|
|
|
return results
|
|
|
|
def find_propers(self, search_date=datetime.datetime.today()):
|
|
|
|
return self._find_propers(search_date, '')
|
|
|
|
|
|
class KATCache(tvcache.TVCache):
|
|
|
|
def __init__(self, this_provider):
|
|
tvcache.TVCache.__init__(self, this_provider)
|
|
|
|
self.minTime = 20 # cache update frequency
|
|
|
|
def _getRSSData(self):
|
|
|
|
return self.provider.get_cache_data()
|
|
|
|
|
|
provider = KATProvider()
|