# creates a file of word pairs from a simple thesaurus
# works on a plain text file with no naycount prefixes
# usage: python ths2prs.py thesaurus-file wordpair-file
#
# this is not an exact inverse to the related program
# prs2ths.py since the input to this program does not
# have to have sorted lines but the output will be sorted
# pairs, however they are inverses in the sense that
# running either on the output of the other will produce
# files that the progam can use to reproduce that output
#
# note: written to be read and maintained, not to be
# short and sweet -- as all good programs should be
#
# ths2prs.py is Copyright (C) 2010 Douglas Pardoe Wilson
# This program comes with ABSOLUTELY NO WARRANTY
# This is free software, and you are welcome to redistribute it
# subject to the conditions of the GNU General Public License (GPL)

import sys
import string

comlineargs = sys.argv
argcount = len(comlineargs)
# print "command line argument count was " + `argcount`
# print "command line: "
# print comlineargs
if len(comlineargs) != 3:
   print "usage: python ths2prs.py thesaurus-infile wordpair-outfile"
else:
   infile = open(comlineargs[1], "r")
   outfile = open(comlineargs[2], "w")
   wordpairs = list()
   while 1:  # loop indefinitely, but will break out when no line found
      thisline = infile.readline()
      if thisline:
         theslinelist = string.split(string.strip(thisline))
         if len(theslinelist) > 1:
            lemma = theslinelist[0]
	    syns = theslinelist[1:]
      	    for synonym in syns:
               pair = (lemma, synonym)
	       wordpairs.append(pair)
	       pair = (synonym, lemma)
	       wordpairs.append(pair)
      else:
         break
   infile.close()

   if len(wordpairs) > 0:
      wordpairs.sort()
      lastoutstring = ""
      for pair in wordpairs:
	  outstring = pair[0] + " " + pair[1] + "\n"
          if outstring != lastoutstring:
             outfile.write(outstring)
	     lastoutstring = outstring
   outfile.close()