Für openTMS wollen wir unser TM als serverseitige (Java) Komponente einsetzen. Um Java in Postgres serverseitig einsetzen zu können, ist es notwendig das mitgelieferte pljava zu "aktivieren", da per default ausgeschaltet. Das erwies sich als komplizierter als gedacht. Hier mein Weg zu einer funktionsfähigen Installation (für Windows).

Ich bekam die verschiedensten Fehlermeldungen, von dll nicht gefunden, ulkige Pfadnamen etc. Wichtig war u.a. die korrekte Schreibweise der Pfadnamen in der conf datei (also \\ statt \ verwenden). Etwas mühsam zu eruieren, auch mit der Suche im Web.

Zu beachten ist, dass ich Postgres unter C:\ProgrammeDatenbanken installiert habe. Pfade sind entsprechend anzupassen!

OK, hier nun die Details:

Schritte install pljava bei Postgres
====================================

Annahme Postgres unter: C:\ProgrammeDatenbanken\PostgreSQL\8.4

1) Anpassen in Systemsteuerung - korrekten Java Pfad einsetzen!
PATH
C:\Program Files\IDM Computer Solutions\UltraEdit\;C:\Program Files\IDM Computer Solutions\UltraCompare\;c:\Program Files\Java\jre6\bin\client\

Wenn nicht gesetzt, kann Java von Postgres nicht ausgeführt werden.

2) postgresql.conf in C:\ProgrammeDatenbanken\PostgreSQL\8.4\data anpassen

Einträge am Ende "entkommentieren". Hier die Einträge:

WICHTIG: Bei den Pfadnamen mit \ mussmit \\ geschrieben werden!

custom_variable_classes = 'pljava'
pljava.classpath ='C:\\ProgrammeDatenbanken\\PostgreSQL\\8.4\lib\\pljava.jar'
pljava.statement_cache_size = 10
pljava.release_lingering_savepoints = true
pljava.vmoptions = '-Xmx512M'
pljava.debug = false
dynamic_library_path = '$libdir;C:\\ProgrammeDatenbanken\\PostgreSQL\\8.4\\lib'
log_min_messages = info

3) in pgadmin Script C:\ProgrammeDatenbanken\PostgreSQL\8.4\share\pljava install.sql ausführen; vorher auswählen einer Datenbank

4) Testen mit
SELECT sqlj.install_jar('file:///C:/ProgrammeDatenbanken/PostgreSQL/8.4/lib/examples.jar',
'samples', true);
SELECT sqlj.set_classpath('javatest', 'samples');

Dann z.B.
SELECT * FROM javatest.listSupers();


Dann sollte alles funktionieren..

Viel Erfolg!

Araya has been updated now with some additional useful plugins.

It contains now the following new plugins.

a) Merge several tmx files into one new tmx file
b) Split one tmx file into several tmx files
c) Count the number of TU / TUV elements in a tmx file
d) Validate xml files with dtds registered in Araya

In addition the documentation was updated. The documentation update and the plugin update applies to the bilingual term extractor too (the xml validation is not contained in this plugin).

What's coming next? The term extractor will allow in the next version running the extraction in buckets, meaning the the import tmx file can be split up into several pieces and the results will be combined at the end.

Other infos as usual at http://www.heartsome.de

There have been some changes done in Araya.

First Arya uses the in addition to its own tm features the full openTMS plugin. This includes both the translation part and the reverse conversion part. openTMS is still using the converters of Araya.

The plugin menu has been enhanced. It adds now the TMX split and merge dialog enabling the user to merge or splt tmx files.

This all can be downloaded from here:

The free Araya Version:

http://www.heartsome.de/de/freearaya.php
or
http://www.heartsome.de/en/freearaya.php

The openTMS Server + Data Source Editor can also be downloaded from here: http://www.heartsome.de/de/opentmsdownload.php.

If you are interested in the commercial version of Araya please contakt us: info@heartsome.de

A new version of the Araya Bilingual Term Extraction with new features is available.

After the extraction process one can filter entries based on word lists.

Checkout http://www.heartsome.de/de/termextraction.php or http://www.heartsome.de/en/termextraction.php

A video about the bilingual term extraction is also available here:

http://www.youtube.com/watch?v=EASlNtjaHi0

OpenTMS - the Open Source translation memory - is making its prototype software now available both on www.opentms and also on Source forge - https://sourceforge.net/projects/open-tms/

A first Ubuntu based version can be downladed there - based on Suns VirtualBox.

The source code is also available. BUT - and I can say this as the system architect and currently one of the main developers of the system - this code will undergo heavy changes and restructuring.

What is available?

a) a TM server - e.g. accessible thru XML-RPC
b) SQL database support based on Hibernate
c) TMX / XLIFF Import and export
d) translation functionality - the converters are provided from Heartsome Europe
e) a php based Web interface for testing OpenTMS.

Can you do real translations? The system is still in ints beta phase, but why not? xml works fine. And it uses OpenOffice to convert all kind of formats like RTF, WinWord etc. So just give it a try.

If you need help or have a question - just send me an e-mail. klemens dot waldhoer at heartsome dot de