From thinkit8@lycos.com Tue Sep 18 00:39:57 2001
Return-Path: <thinkit8@lycos.com>
X-Sender: thinkit8@lycos.com
X-Apparently-To: lojban@yahoogroups.com
Received: (EGP: mail-7_3_2_2); 18 Sep 2001 07:39:56 -0000
Received: (qmail 48904 invoked from network); 18 Sep 2001 03:29:29 -0000
Received: from unknown (10.1.10.142)
  by l10.egroups.com with QMQP; 18 Sep 2001 03:29:29 -0000
Received: from unknown (HELO n25.groups.yahoo.com) (216.115.96.75)
  by mta3 with SMTP; 18 Sep 2001 03:29:29 -0000
X-eGroups-Return: thinkit8@lycos.com
Received: from [10.1.10.107] by mv.egroups.com with NNFMP; 18 Sep 2001 03:29:29 -0000
Date: Tue, 18 Sep 2001 03:29:28 -0000
To: lojban@yahoogroups.com
Subject: lujvo banro...python programing to expand lujvo
Message-ID: <9o6f2o+a270@eGroups.com>
User-Agent: eGroups-EW/0.82
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Length: 1155
X-Mailer: eGroups Message Poster
X-Originating-IP: 24.5.121.32
From: thinkit8@lycos.com
X-Yahoo-Message-Num: 10823

ok, i went ahead and hacked 2 python programs to expand all lujvo 
into compenent cmavo/gismu, using vlatai from jbofi'e.

first, the batch file, %1 is infile, %2 is outfile:

python space.py < %1 > temp1.txt
vlatai -el < temp1.txt > temp2.txt
python sub.py < %1 > %2

note that sub.py expects "temp2.txt" to have the vlatai output.

next space.py:

import re
while 1:
 try:
  s=raw_input()
  a=re.split("[^a-zA-Z\',]+",s);
  for x in a:
   print x
 except EOFError:
  break

next sub.py:

import re,string
qb=[]
qa=[]
f=open("temp2.txt")
s=f.readline()
while s!="":
 if re.search(": lujvo :",s) is not None:
  res=re.match("[a-z\',]+",s)
  qb.append(res.group(0))
  res=re.search("\[[a-z\',\+\?]+",s)
  s2=res.group(0)
  qa.append(string.replace(s2,"[",""))
 s=f.readline()
s2=""
while 1:
 try:
  s2+=raw_input()
  s2+="\n"
 except EOFError:
  break
for x in range(len(qa)):
  s2=re.sub(qb[x],qa[x],s2)
print s2

much thanks to richard curnow for providing vlatai.  it's a bit 
roundabout because all i have is a binary.  hopefully i'll get 
working with the source and can do something more elegant, and/or 
(gi'a) make it web accessible.