Worldwide Lexicon System Documentation


    /mt
    
    This web API interface is a general purpose machine translation proxy server.
    It front ends one or many different machine translation services, and presents
    a common interface to the clients using this interface, so the client can just
    submit one query to WWL, which proxies the request through to upstream services.
    
    It expects the following parameters:
    
    sl = source language (ISO code)
    tl = target language (ISO code)
    st = source text (Unicode UTF-8 encoding)
    mtengine = machine translation engine to use (to override automatic selection)
    
    It replies with a plain text response with the translated text, or an empty
    string if the query fails. The proxy server will memcache queries so frequently
    requested texts are cached in memory (such as in a page that is being reloaded
    frequently). Unless otherwise noted, the returned text will be Unicode with
    UTF-8 encoding.
    
    The proxy server currently supports the following machine translation
    services (more are being added for the June OSS release):
    
      Google : about 30 languages supported
      Apertium : Spanish, Catalan, French, Portuguese, Galician, Basque, Bretton
      Moses : open source statistical machine translation system
      World Lingo: commercial machine translation service (SaaS)
    
    We are planning to add connectors to several popular enterprise machine
    translation systems, as soon as we receive documentation from vendors
    on their SaaS offerings. If you have a machine translation system that
    you would like to add to WWL, just email brian@worldwidelexicon.org
    with instructions for querying your web API. The service should be
    accessible via a standard HTTP GET/POST CGI query and return Unicode
    compliant (UTF-8 encoded) text in response. You can also extend the
    mt.py module to create your own custom MT connectors for private or
    intranet applications.
     
    NOTE: to request a combined recordset containing both human and
    machine translations, call the /q (get.py) web service instead, as it
    returns both types of translations depending on query settings. This
    method is provided specifically for MT queries, and for backward
    compatibility with the version 1 web API.
    
    

Test Form

Source Language Code
Target Language Code
Machine Translation Engine