myPREP is a text aligner software, a tool which makes possible to automatically align two by two the documents in a multilingual corpus. The outcome of the alignment is a translation memory in TMX format. The alignment is done at the sentence level.
Thanks to myPREP, the production of training corpus for statistical translation is also possible (in Moses format). Such corpora are divided into several parts for the training, the tuning and the evaluation.
myPREP also makes possible the alignment of comparable corpora. The outcome of the alignment is a set of pair of sentences associated with a score, the number of aligned terms, and the length of sentences. These functions can control the alignments.
myPREP requires segmented documents corpora in UTF-8 format. The converter and the segmentation tool of the myCAT software are included in the installation of myPREP.
myPREP is available for both Windows and for GNU/Linux (Ubuntu 12.04); please find the links to the respective installation files below. As for the sources, they are the same for both versions.
myPREP 2.0 for Windows and GNU/Linux (tested on Windows 7, Windows 2008 Server, Ubuntu 12.04 LTS)
The myPREP installation files can be downloaded download (62 Mo).
- SHA1 key: ed84e05ec106527e1c557836a9a817a8
Tools for Windows
The Java, Tomcat and OpenOffice applications are to be downloaded here download (232 Mb). Please use only these distribution of the applications so as to make sure you have the required version.
- SHA1 key: 2B7DE1A6590E04B241A679451952256834A6C21B
Tools for GNU/Linux
The OpenOffice file is to be downloaded here (171 Mb). Please use only this distribution of the application so as to make sure you have the required version
- SHA1 key: 2B8071CF3F26C202BB7B0BE1AFD5DC2F21221614
The procedure to perform a standard installation of myPREP is described here.
Additional Alignment Maps
myCAT originally comes with the following alignment maps: English-Arabic, English-French, English-Portuguese, English-Russian, English-Spanish. It is possible to add the following language pairs:
Please read the documentation: Adding Alignment Maps in myCAT documentation.
This document details the bug fixes and improvements in each new release of the myPREP software.
Version 2.0 Date: 25 June 2013 * Initial release
source & licence
Licence AGPL v3 - sources @GitHub
Installation of myPREP
- myPREP for Windows (prerequisites here) The procedure to perform a standard installation of myPREP on a dedicated server running Windows is described here install.
- myPREP for GNU/Linux (prerequisites here) The procedure to perform a standard installation of myPREP on a dedicated server running Ubuntu 12.04 LTS is described here install.
Adding Alignment Maps
myCAT originally comes with the following alignment maps:
It is possible to add the following language pairs: * English-German * English-Chinese(experimental, all comments welcome)
Simply download them and unzip them to: * for Windows: C:\MYCAT\map * for GNU/Linux: ~/MYCAT/map
Document Naming Rules
The corpus of bi-text documents to be used with myMT should comply with the following specifications. This myMT distribution comes with a very small test corpus of 48 documents allowing to test the following six languages: English, French, Spanish, Arabic, Russian and Chinese. These documents are organized in three collections: UNO, WIPO and WTO. They are all public documents which were downloaded from those organization’s websites.