GenSynsets is a tool conceived in order to facilitate the development of WordNets for languages other than English. It implements the algorithms described in the paper "On the Semiautomatic Generation of WordNet Type  Synsets and Clusters with Special Reference to Romanian". GenSynsets can be used with reference to any foreign language for which you have bilingual dictionaries in electronic format. The program has been tested for the Romanian language and the corresponding output (an XML file) can be explored on the Web.

Installation

GenSynsets is written in the Java programming language and runs on the Java 2 platforms.

Requirements:

1. In order to run, GenSynsets requires the Java 2. Therefore the operating system of the computer on which GenSynsets is to be installed must be one for which an implementation of Java 2 exists (Windows 95/98/ME/NT/000, most Unix versions, MacOS X).

2. The system on which GenSynsets is to be installed must be sufficiently powerful (processor speed, memory). In the case of a PC, a minimum of 133 MHz and 32M RAM are necessary.

Installation:

1. Install Java 2 on your system. If Java 2 is already installed, skip this step. Corresponding to Windows, Linux, Solaris platforms, the necessary installation kits can be found at java.sun.com. The entire JDK development kit or only the JRE runtime environment can be installed. Using version 1.3 or a more recent version is recommended.

2. Make sure that the PATH variable contains the path to the java executable. Corresponding to Windows 95/98/NT this can be achieved by including in the autoexec.bat file a line of the following form:

SET PATH=c:\path; %PATH%

where c:\path is to be replaced with the actual path leading to the directory where the java executable is placed. Following this operation (and restarting your computer), as a result of the command

java -version

the answer of your system should look like:

C:\>java -version
java version "1.3.0"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.3.0-C)
Java HotSpot(TM) Client VM (build 1.3.0-C, mixed mode)

C:\>

no matter which directory the command was issued from.

3. Unzip the archive gensynsets.zip and place its content wherever you wish within the existing directory structure.

Remark: The above directions have concentrated especially on installation under the Windows operating system. In order to install and to run  GenSynsets on any system, in essence, you must be able to install the Java 2 environment on that system, to unzip the archive gensynsets.zip, and then to run the (Java) class GenSynsets. 

Usage

GenSynsets is designed to be used from the command line. The general form of the program command line is:

java -classpath .;jwordnet.jar GenSynsets -pos noun|adj [-enrich] [-cs charset] [-l SynList] WnDictPath E_F F_E OutFile

where:

• -pos specifies the part of speech to be considered : noun or adj
• -enrich is used only in the case of adjectives. Its occurrence triggers using the enrichment technique (see the paper).
• -cs sets the character set as it will appear in the output XML file (<?xml version="1.0" encoding="charset">). The default value is iso-8859-1.
• -l SynList; if this switch is present, foreign synsets will be generated only for English synsets whose offset is present in the SynList. If the switch is not present, the whole list of the English synsets (corresponding to the specified part of speech) is processed by the program.
• WnDictPath is the path to the WordNet Databases (c:\wn16\dict).
• E_F, F_E represent the English-Foreign (language), Foreign-English dictionaries, respectively.
• OutFile will contain the result of the processing (the corresponding foreign language synsets).

The files format

bilingual dictionaries

• the English-Foreign dictionary contains lines each of which includes the English word followed by a space and the corresponding translation equivalents: 

eword fword1;fword2,fword3;fword4,fword5

In order to distinguish among fword1,fword2, etc. two different separators are used. A semicolon separates different meanings of the given word (eword). A comma separates synonyms which refer to one and the same meaning of the word (eword).

• the Foreign-English dictionary contains lines each of which includes the foreign word followed by a space and the corresponding translation equivalents:

fword eword1;eword2,eword3;eword4,eword5

In order to distinguish among eword1,eword2, etc. two different separators are used. A semicolon separates different meanings of the given word (fword). A comma separates synonyms which refer to one and the same meaning of the word (fword).

the output file

The output is delivered as an XML file. Thus, the XML files produced by GenSynsets can be easily transformed, by means of XSLT, into other formats (XML, HTML, etc.) and can be used by other applications. The structure of the file is:

where:

• tags <synset>...</synset> delimit a certain synset
• <synset_en>...</synset_en> mark the information concerning the English synset
• <e-sets>...</e-sets> delimit the elementary sets obtained by the algorithm for the English synset; while <e-set>...</e-set> mark a single elementary set
• <fsynsets> ... </fsynsets> demarcate the proposed set of foreign synsets
• <fsynset>... </fsynset> set the limits of an unique foreign synset

For more details see the DTD file (fsynsets.dtd) on which the output XML file is based.