Dependency of WS4J on some configuration files and WordNet (200Mb)

Question

I am using WS4J API for calculating semantic similarity between words:

ILexicalDatabase db = new NictWordNet();
RelatednessCalculator lin = new Lin(db);
RelatednessCalculator wup = new WuPalmer(db);

String w1 = "science";
String w2 = "university";
System.out.println(lin.calcRelatednessOfWords(w1, w2));
System.out.println(wup.calcRelatednessOfWords(w1, w2));

The problem is that this API depends on the following configuration files which must be placed into the project's directory (I use /resources for this purpose):

jaw.jaw.conf
similarity.conf
wordnet folder

Moreover it's a pity that this library is unavailable in Maven repository.

Is there any way to avoid putting the above-mentioned files into my local project's folder? These files occupy over 100Mb....

I also checked the library DISCO, but it doesn't seem to be so powerful as WS4J.

score 0 · Answer 1 · answered Apr 20 '17 at 14:50

Apparently, do to so you have to modify WS4J.

For instance, the similarity.conf file is loaded by the class WS4JConfiguration through an InputStream:

final public class WS4JConfiguration {

    private final static String CONF = "/similarity.conf";

     ...

    private WS4JConfiguration(){
        InputStream stream = null;
        try {
            stream = WS4JConfiguration.class.getResourceAsStream( CONF );

So, WS4JConfiguration loads resources using the same class loader that loaded your application and it should have access to resources in your JARs.

Dependency of WS4J on some configuration files and WordNet (200Mb)

1 Answers1