#!/usr/bin/python
# Python Google Translator v1.1, by Macie
# Source code: www.linuxfusion.net/macie/Scripts/gtranslate.py
#
# Purpose: translate various languages, whether single word or
# complete sentence, without the need of a localized language
# database. All data is queried from translate.google.com

# Credits: Google for providing their translation site. Please
# show support for thir service by using their translation 
# website directly. This will not only show your support but
# also allow you to make improvements to their translations
# by providing suggestions and corrections.

# START CONFIG
# 0=off / 1=on

# Default language. Good if you plan to use this script to
# only translate a specific pair of languages.
# Otherwise set to 0.
default_lang = "0"

# Print translation type. This will simply print language
# titles of what is being done.
print_lang = "1"

# Should translations be logged? Also, define what file you
# want to save the translations to.
# Example:
# Unix systems: /var/log/translation.log
# windows systems: c:\Log\translation.txt
logging = "0"
log_file = "/tmp/translation.log"

# Query language availability in real time? This will query
# translate.google.com to check "what" can be translated to "what"
# rather then using the locally defined options. Enable this if
# this script is not allowing you to translate a language that you
# can plainly see is "now" available via google.
update_langdb = "0"

# Google translation page URL. Probably NOT a good idea to modify.
url = "http://translate.google.com/"

# END CONFIG

import sys
import urllib
import urllib2
import re
import string
from time import gmtime, strftime

global lang_defs, lang_allow

# Default language definitions
lang_defs = {'ar':'Arabic','zh-CN':'Chinese','nl':'Dutch',    \
             'en':'English','fr':'French','de':'German',      \
             'it':'Italian','ja':'Japanese','ko':'Korean',    \
             'pt':'Portuguese','ru':'Russian','es':'Spanish', \
             'zh-TW':'Taiwan','el':'Greek'}
lang_allow = ['en:ar','en:nl','en:fr','en:de','en:el',    \
              'en:it','en:ja','en:ko','en:pt','en:ru',    \
              'en:es','fr:en','fr:de','de:en','de:fr',    \
              'el:en','it:en','ja:en','ko:en','pt:en',    \
              'ru:en','es:en', 'ar:en','nl:en',           \
              'zh-CN:en','en:zh-TW','en:zh-CN']

# Queries lang_defs and lang_allow to show what is available
def print_avail_lang():
    for lang in lang_allow:
        from_to = give_lang_name(lang)
        show_lang = "  "+lang+"  "+'['+from_to[0]+'->'+from_to[1]+']'
        print show_lang

# Assigns name values to desired language pair
def give_lang_name(lang_name):
    from_to = ["?", "?"]
    parse_div = string.find(lang_name, ':')
    lang_from = lang_name[0 : parse_div]
    lang_to = lang_name[parse_div+1 : len(lang_name)]
    for (lkey, lname) in lang_defs.items():
        if lkey == lang_from:
           from_to[0] = lname
        if lkey == lang_to:
           from_to[1] = lname
    return from_to

# URL download/form submit function
def pull_url(url, data):
    data = urllib.urlencode(data)
    header = { 'User-Agent' : 'Mozilla/5.0 (X11; U; Linux i686; en-US)'}
    send_request = urllib2.Request(url, data, header)
    html_data = urllib2.urlopen(send_request)
    html_text = html_data.read()
    return html_text

# If something failes, print usage and exit
def usage():
    print 'Usage:'
    print ' ',sys.argv[0],'<from:to> "<text to translate>"'
    print 'If default_lang var is set, then you can get away with:'
    print ' ',sys.argv[0],'"<text to translate>"'
    print 'Use double quotations for more than one word, sentences, etc.\n'
    print 'Language translations available: '
    print_avail_lang()
    print 'Example:',sys.argv[0],'en:fr "How are you?"'
    sys.exit(0)

# Get an updated language list
if update_langdb == "1":
   print "Retreaving updated language list..."
   form_val = {}
   html_text = pull_url(url, form_val)
   # Hopefully google won't change their code anytime soon
   # otherwise the filters blow will break badly
   split_forms = html_text.replace('</select>', '</select>\n')
   filter_html = re.search("<select name=langpair>"+"(.*)"+"</select>" ,split_forms).group()
   filter_html = filter_html.replace('</option>', '\n')
   filter_vars = ['<select name=langpair>','</select>','selected>','<option value=','>','class="line-above', '"']
   for filter in filter_vars:
       filter_html = filter_html.replace(filter, ' ')
   lang_allow = []
   for single_lang in filter_html.splitlines():
       single_lang = single_lang.strip()
       parse_div = string.find(single_lang, ' ')
       new_entry = single_lang[0 : parse_div]
       lang_allow.append(new_entry.replace('|',':'))
   lang_allow.remove('')

# Make sure given arguments are valid
if len(sys.argv)==1:
   usage()
elif len(sys.argv)==2:
     if default_lang != 0 and lang_allow.count(default_lang) > 0:
        lang_arg = default_lang
        lang_prc = sys.argv[1]
        if lang_allow.count(lang_prc) > 0:
           usage()
     else:
        print "Not enough arugments passed."
        usage()
else:
   lang_arg = sys.argv[1]
   lang_prc = sys.argv[2]

#if len(sys.argv)<3:
#   usage()
if lang_allow.count(lang_arg) > 0:
   pass
else:
   usage()

# Google form values
form_val = {}
form_val['hl'] = 'EN'
form_val['ie'] = 'UTF-8'
form_val['text'] = lang_prc

#if re.match("^[a-zA-Z0-9]+$", values['text']):

# Set URL to desired lang and pull html data
langpair = lang_arg.replace(':','|')
url = url+"translate_t?langpair="+langpair
html_text = pull_url(url, form_val)

# Parse output
parse_string = "<div id=result_box dir="
make_breaks = html_text.replace('<',"\n<")
single_line = re.search(parse_string+"(.*)", make_breaks)
if single_line:
   trans_output = single_line.group()
   parse_div = string.find(trans_output, '>')
   remove_from = trans_output[0 : parse_div]
   trans_output = trans_output.replace(remove_from+">", '')
   if trans_output:
      full_output = trans_output
   else:
      full_output = "No Translation Found!"
else:
   full_output = "Problem parsing information! The \"parse_string\" probably needs adjustment since google sometimes changes that."

# Log request
if logging == "1":
   from_to = give_lang_name(lang_arg)
   log_stamp = strftime("%b %d %Y %H:%M:%S", gmtime())
   f = open(log_file, "aw")
   f.write(log_stamp+" Translated "+from_to[0]+" to "+from_to[1]+" | \""+lang_prc+"\" to \""+full_output+"\"\n")
   f.close()

# Print translation type
if print_lang == "1":
   from_to = give_lang_name(lang_arg)
   if default_lang == lang_arg:
      default_set = "(set as default)"
   else:
      default_set = ""
   print "Translation from",from_to[0],"to",from_to[1],default_set

# Print result to console
print  full_output