import subprocess
  import os
  from BaseProcessor import baseProcessor
  import nltk
  import re

  class Tagger(baseProcessor):
      """ a calling to lia_tagg class"""
      def clean(self,dirtyString):

  	""" Clean string for using it into lia_tagg
  
          Change text to iso  and clean it  one word by line and separate sentences with <s> </s>"""

          p=subprocess.Popen([os.environ["LIA_TAGG"]+'/script/lia_clean'],stdin=subprocess.PIPE,stdout=subprocess.PIPE)
          (cleanString, err) = p.communicate(input=dirtyString.encode('iso8859-1','backslashreplace'))
          return cleanString
  
      def tagg(self,cleanString):

  	"""POS Tagg and lemm a string which come  from clean""" 

          p2=subprocess.Popen([os.environ["LIA_TAGG"]+'/script/lia_tagg+lemm','-guess'],stdin=subprocess.PIPE,stdout=subprocess.PIPE)
      	(taggedString,err) =p2.communicate(input=cleanString)
      # This is used beceause lia_tagg deal with iso8859 only
      	return taggedString.decode('iso8859').encode("utf8")

      def lemm(self,cleanString):

  	""" use the pos tagger to lemm word and return lemm only"""

   	taggedString = self.tagg(cleanString)       

  	# sub the string to get only lemm ( cut markup and origin word ) Can be Delete with better use of lia_tagg

  	sub = re.sub(r' </s>','',re.sub(r'<s> ',''," ".join([ x.rstrip().split().pop(2) for x in taggedString.rstrip().split("
  ") if x])))

          return sub

      def isReady(self):

          """ Check if the Tagger can be used ( depends on LIA_TAGG )"""

          os.environ["LIA_TAGG"]

          return True

  
  class Phoner(baseProcessor):
      """ a class which call the lia phoner """
      def clean(self,dirtyString):
          p=subprocess.Popen([os.environ["LIA_PHON_REP"]+'/script/lia_nett'],stdin=subprocess.PIPE,stdout=subprocess.PIPE)
          (cleanString, err) = p.communicate(input=dirtyString.encode('iso8859-1','backslashreplace'))
          return cleanString
      def phon(self,cleanString):
          p2=subprocess.Popen([os.environ["LIA_PHON_REP"]+'/script/lia_lex2phon'],stdin=subprocess.PIPE,stdout=subprocess.PIPE)
      	(taggedString,err) =p2.communicate(input=cleanString)
      # This is used beceause lia_phon deal with iso8859 only
      # We reconverte the output to utf8 back
      	return taggedString.decode('iso8859').encode("utf8")
      def isReady(self):
  	os.environ["LIA_PHON_REP"]

          return True

  class StopWord(baseProcessor):
      def isReady(self):

          return True

      def RemoveStopList(self,rowstring):
          """ Remove from set of word (splited String ) each words in the stoplist and join all of the other in a string """

          return u" ".join(unicode(value) for value in list(set(rowstring.split()) - set(nltk.corpus.stopwords.words("french"))))