Dan Urist

NAME

WWW::Babelfish - Perl extension for translation via Babelfish or Google

SYNOPSIS

  use WWW::Babelfish;
  $obj = new WWW::Babelfish( service => 'Babelfish', agent => 'Mozilla/8.0', proxy => 'myproxy' );
  die( "Babelfish server unavailable\n" ) unless defined($obj);

  $french_text = $obj->translate( 'source' => 'English',
                                  'destination' => 'French',
                                  'text' => 'My hovercraft is full of eels',
                                  'delimiter' => "\n\t",
                                  'ofh' => \*STDOUT );
  die("Could not translate: " . $obj->error) unless defined($french_text);

  @languages = $obj->languages;

DESCRIPTION

Perl interface to the WWW babelfish translation server.

METHODS

new

Creates a new WWW::Babelfish object.

Parameters:

 service:        Babelfish, Google or Yahoo; default is Babelfish
 agent:          user agent string
 proxy:          proxy in the form of host:port
services

Returns a plain array of the services available (currently Babelfish, Google or Yahoo).

languages

Returns a plain array of the languages available for translation.

languagepairs

Returns a reference to a hash of hashes. The keys of the outer hash reflect all available languages. The hashes the corresponding values reference contain one (key) entry for each destination language that the particular source language can be translated to. The values of these inner hashes contain the Babelfish option name for the language pair. You should not modify the returned structure unless you really know what you're doing.

Here's an example of a possible return value:

        {
          'Chinese' => {
                         'English' => 'zh_en'
                       },
          'English' => {
                         'Chinese' => 'en_zh',
                         'French' => 'en_fr',
                         'German' => 'en_de',
                         'Italian' => 'en_it',
                         'Japanese' => 'en_ja',
                         'Korean' => 'en_ko',
                         'Portuguese' => 'en_pt',
                         'Spanish' => 'en_es'
                       },
          'French' => {
                        'English' => 'fr_en',
                        'German' => 'fr_de'
                      },
          'German' => {
                        'English' => 'de_en',
                        'French' => 'de_fr'
                      },
          'Italian' => {
                         'English' => 'it_en'
                       },
          'Japanese' => {
                          'English' => 'ja_en'
                        },
          'Korean' => {
                        'English' => 'ko_en'
                      },
          'Portuguese' => {
                            'English' => 'pt_en'
                          },
          'Russian' => {
                         'English' => 'ru_en'
                       },
          'Spanish' => {
                         'English' => 'es_en'
                       }
        };
translate

Translates some text using Babelfish.

Parameters:

 source:      Source language
 destination: Destination language
 text:        If this is a reference, translate interprets it as an 
              open filehandle to read from. Otherwise, it is treated 
              as a string to translate.
 delimiter:   Paragraph delimiter for the text; the default is "\n\n".
              Note that this is a string, not a regexp.
 ofh:         Output filehandle; if provided, the translation will be 
              written to this filehandle.

If no ofh parameter is given, translate will return the text; otherwise it will return 1. On failure it returns undef.

error

Returns a (hopefully) meaningful error string.

NOTES

Babelfish translates 1000 characters at a time. This module tries to break the source text into reasonable logical chunks of less than 1000 characters, feeds them to Babelfish and then reassembles them. Formatting may get lost in the process; also it's doubtful this will work for non-Western languages since it tries to key on punctuation. What would make this work is if perl had properly localized regexps for sentence/clause boundaries.

Support for Google is preliminary and hasn't been extensively tested (by me). Google's translations used to be suspiciously similar to Babelfish's, but now some people tell me they're superior.

AUTHOR

Dan Urist, durist@frii.com

SEE ALSO

perl(1).