# Written by Jelle Roozenburg
# see LICENSE.txt for license information

import re, sys, os
from traceback import print_exc

from Tribler.__init__ import LIBRARYNAME

WORDS_REGEXP = re.compile('[a-zA-Z0-9]+')
DEBUG = False

class XXXFilter:
    def __init__(self, install_dir):
        termfilename = os.path.join(install_dir, LIBRARYNAME, 'Category','filter_terms.filter')
        self.xxx_terms, self.xxx_searchterms = self.initTerms(termfilename)
                
    def initTerms(self, filename):
        terms = set()
        searchterms = set()

        try:
            f = file(filename, 'r')
            lines = f.read().lower().splitlines()
    
            for line in lines:
                if line.startswith('*'):
                    searchterms.add(line[1:])
                else:
                    terms.add(line)
            f.close()
        except:
            if DEBUG:
                print_exc()
            
        if DEBUG:
            print 'Read %d XXX terms from file %s' % (len(terms)+len(searchterms), filename)
        return terms, searchterms
    
    def _getWords(self, string):
        return [a.lower() for a in WORDS_REGEXP.findall(string)]
    
        
    def isXXXTorrent(self, files_list, torrent_name, tracker, comment=None):
        if tracker:
            tracker = tracker.lower().replace('http://', '').replace('announce','')
        else:
            tracker = ''
        terms = [a[0].lower() for a in files_list]
        is_xxx = (len(filter(self.isXXX, terms)) > 0 or 
                  self.isXXX(torrent_name, False) or
                  self.isXXX(tracker, False) or
                  (comment and self.isXXX(comment, False))
                  )
        if DEBUG:
            if is_xxx:
                print 'Torrent is XXX: %s %s' % (torrent_name, tracker)
            else:
                print 'Torrent is NOT XXX: %s %s' % (torrent_name, tracker)
        return is_xxx
                
    
    def isXXX(self, s, isFilename=True):
        s = s.lower()
        if self.isXXXTerm(s): # We have also put some full titles in the filter file
            return True
        if not self.isAudio(s) and self.foundXXXTerm(s):
            return True
        words = self._getWords(s)
        words2 = [' '.join(words[i:i+2]) for i in xrange(0, len(words)-1)]
        num_xxx = len([w for w in words+words2 if self.isXXXTerm(w, s)])
        if isFilename and self.isAudio(s):
            return num_xxx > 2 # almost never classify mp3 as porn
        else:
            return num_xxx > 0
    
    def foundXXXTerm(self, s):
        for term in self.xxx_searchterms:
            if term in s:
                if DEBUG:
                    print 'XXXFilter: Found term "%s" in %s' % (term, s)
                return True
        return False
    
    def isXXXTerm(self, s, title=None):
        # check if term-(e)s is in xxx-terms
        s = s.lower()
        if s in self.xxx_terms:
            if DEBUG:
                print 'XXXFilter: "%s" is dirty%s' % (s, title and ' in %s' % title or '')
            return True
        if s.endswith('es'):
            if s[:-2] in self.xxx_terms:
                if DEBUG:
                    print 'XXXFilter: "%s" is dirty%s' % (s[:-2], title and ' in %s' % title or '')
                return True
        elif s.endswith('s') or s.endswith('n'):
            if s[:-1] in self.xxx_terms:
                if DEBUG:
                    print 'XXXFilter: "%s" is dirty%s' % (s[:-1], title and ' in %s' % title or '')
                return True
        
        return False

    audio_extensions = ['cda', 'flac', 'm3u', 'mp2', 'mp3', 'md5', 'vorbis', 'wav', 'wma', 'ogg']
    def isAudio(self, s):
        return s[s.rfind('.')+1:] in self.audio_extensions