digit to word in english language bzr revid: mga@mga-laptop-20081010075801-idqqkuzeztwhf3ya
150 lines
5.0 KiB
Python
Executable File
150 lines
5.0 KiB
Python
Executable File
# -*- coding: iso8859-1 -*-
|
|
##############################################################################
|
|
#
|
|
# Copyright (c) 2004 TINY SPRL. (http://tiny.be) All Rights Reserved.
|
|
# Fabien Pinckaers <fp@tiny.Be>
|
|
#
|
|
# WARNING: This program as such is intended to be used by professional
|
|
# programmers who take the whole responsability of assessing all potential
|
|
# consequences resulting from its eventual inadequacies and bugs
|
|
# End users who are looking for a ready-to-use solution with commercial
|
|
# garantees and support are strongly adviced to contract a Free Software
|
|
# Service Company
|
|
#
|
|
# This program is Free Software; you can redistribute it and/or
|
|
# modify it under the terms of the GNU General Public License
|
|
# as published by the Free Software Foundation; either version 2
|
|
# of the License, or (at your option) any later version.
|
|
#
|
|
# This program is distributed in the hope that it will be useful,
|
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
# GNU General Public License for more details.
|
|
#
|
|
# You should have received a copy of the GNU General Public License
|
|
# along with this program; if not, write to the Free Software
|
|
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
|
|
#
|
|
##############################################################################
|
|
|
|
#-------------------------------------------------------------
|
|
#ENGLISH
|
|
#-------------------------------------------------------------
|
|
|
|
ones = {
|
|
0: '', 1:'One', 2:'Two', 3:'Three', 4:'Four', 5:'Five', 6:'Six', 7:'Seven', 8:'Eight', 9:'Nine',
|
|
10:'Ten', 11:'Eleven', 12:'Twelve', 13:'Thirteen', 14:'Forteen', 15:'Fifteen', 16:'Sixteen', 17:"Seventeen",18:"Eighteen",19:"Nineteen",
|
|
}
|
|
|
|
tens = {
|
|
1: 'Ten', 2: 'Twenty ', 3:'Thirty', 4:'Forty', 5:'Fifty', 6: 'Sixty', 7 : 'Seventy', 8:'Eighty' ,9: 'Ninety'}
|
|
|
|
hundred = {
|
|
0:'',1: 'One Hundred', 2: 'Two Hundred', 3: 'Three Hundred', 4 :'Four Hundred', 5: 'Five Hundred', 6: 'Six Hundred', 7 :'Seven Hundred', 8:' Eight Hundred ', 9:'Nine Hundred '
|
|
}
|
|
|
|
thousands ={
|
|
0:'',1: 'One Thousand'
|
|
}
|
|
|
|
lacs = {
|
|
0:'',1: 'Lac'
|
|
}
|
|
|
|
def _100_to_text(number):
|
|
if number in ones:
|
|
return ones[number]
|
|
else:
|
|
if number%10>0:
|
|
return tens[number / 10]+'-'+ones[number % 10]
|
|
else:
|
|
return tens[number / 10]
|
|
|
|
def _1000_to_text(number):
|
|
d = _100_to_text(number % 100)
|
|
d2 = number/100
|
|
if d2>0 and d:
|
|
return hundred[d2]+' '+d
|
|
elif d2>1 and not(d):
|
|
return hundred[d2]+'s'
|
|
else:
|
|
return hundred[d2] or d
|
|
|
|
def _10000_to_text(number):
|
|
if number==0:
|
|
return 'zero'
|
|
part1 = _1000_to_text(number % 1000)
|
|
part2 = thousands.get(number / 1000, _1000_to_text(number / 1000)+' Thousands')
|
|
if part2 and part1:
|
|
part1 = ' '+part1
|
|
return part2+part1
|
|
|
|
def _1000000_to_text(number):
|
|
if number==0:
|
|
return 'zero'
|
|
part1 = _10000_to_text(number % 100000)
|
|
part2 = lacs.get(number / 100000, _10000_to_text(number / 100000)+' Lacs')
|
|
if part2 and part1:
|
|
part1 = ' '+part1
|
|
return part2+part1
|
|
|
|
|
|
def amount_to_text(number, currency):
|
|
lacs_number = int(number)
|
|
units_name = currency
|
|
if lacs_number > 1:
|
|
units_name += 's'
|
|
|
|
lacs = _1000000_to_text(lacs_number)
|
|
lacs = lacs_number and '%s %s' % (lacs, units_name) or ''
|
|
|
|
units_number = int(number * 10000) % 10000
|
|
units = _10000_to_text(units_number)
|
|
units = units_number and '%s %s' % (units, units_name) or ''
|
|
|
|
cents_number = int(number * 100) % 100
|
|
cents_name = (cents_number > 1) and 'cents' or 'cent'
|
|
cents = _100_to_text(cents_number)
|
|
cents = cents_number and '%s %s' % (cents, cents_name) or ''
|
|
return lacs
|
|
|
|
|
|
#-------------------------------------------------------------
|
|
# Generic functions
|
|
#-------------------------------------------------------------
|
|
|
|
_translate_funcs = {'en' : amount_to_text}
|
|
|
|
#TODO: we should use the country AND language (ex: septante VS soixante dix)
|
|
#TODO: we should use en by default, but the translation func is yet to be implemented
|
|
def amount_to_text(nbr, lang='en', currency='euro'):
|
|
"""
|
|
Converts an integer to its textual representation, using the language set in the context if any.
|
|
Example:
|
|
1654: thousands six cent cinquante-quatre.
|
|
"""
|
|
if nbr > 10000000:
|
|
#TODO: use logger
|
|
print "WARNING: number too large '%d', can't translate it!" % (nbr,)
|
|
return str(nbr)
|
|
|
|
if not _translate_funcs.has_key(lang):
|
|
#TODO: use logger
|
|
print "WARNING: no translation function found for lang: '%s'" % (lang,)
|
|
#TODO: (default should be en) same as above
|
|
lang = 'en'
|
|
return _translate_funcs[lang](nbr, currency)
|
|
|
|
if __name__=='__main__':
|
|
from sys import argv
|
|
|
|
lang = 'nl'
|
|
if len(argv) < 2:
|
|
for i in range(1,200):
|
|
print i, ">>", int_to_text(i, lang)
|
|
for i in range(200,999999,139):
|
|
print i, ">>", int_to_text(i, lang)
|
|
else:
|
|
print int_to_text(int(argv[1]), lang)
|
|
|