"""
file: bookIndex.py
language: python3
author: Ivona Bezakova
author: Arthur Nunes-Harwitt
purpose: Builds and displays an index of all the words in a text file
note: our index includes all strings that are preceded and followed
      by a space or the beginning/end of line.
"""

from quickSort import quickSort

# An index is represented as a tuple ( L, D ), where L is 
# the list of words in order, and D is the dictionary
# mapping words to locations.

def withoutNonAlphabetic( word ):
   result = ''
   for c in word:
      if c.isalpha():
         result = result + c
   return result

def indexFromFile( file ):
   """
   indexFromFile: File -> Tuple( List( String ), Dict( String, List( NatNum ) ) )
   """
   lineNumber = 1
   wordList = []
   wordLocations = dict()
   for line in file:
      words = line.split()
      for word in words:
         word = withoutNonAlphabetic( word.lower() ) # Convert the word to lower-case, no punc
         if word != '':
            if word not in wordLocations:  # New word
               wordList.append( word )
               wordLocations[word] = [lineNumber]
            else:                          # Existing word
               wordLocations[word].append( lineNumber )
      lineNumber = lineNumber+1
   sortedWordList = quickSort( wordList ) # sorting occurs here
   return ( sortedWordList, wordLocations )
   

def indexFromFileName( filename ):
   """
   indexFromFileName: String -> Tuple( List( String ), Dict( String, List( NatNum ) ) )
   """
   return indexFromFile( open( filename ) )

def printIndex( index ):
   """
   printIndex: Tuple( List( String ), Dict( String, List( NatNum ) ) ) -> NoneType
   """
   ( words, locations ) = index
   for word in words:
      print( word,"appears at lines:", locations[word] )


def main():
   """
   main : Void -> None
   main prompts for a file name to index.
   """
   filename = input( "Enter file name: " )
   if filename == "":
      filename = "AliceInWonderlandChapter1.txt"

   printIndex( indexFromFileName( filename ) )
                

main()