AWN Browser

The AWN Browser facilitates browsing and searching of the WordNet, and viewing Arabic translations of concepts.

1. Browser features

• Browsing the WordNet - through a searchable and navigable tree representation of the Princeton WordNet. Colour-coding of items in the tree allows for easy identification of those concepts that have already been translated into Arabic, as well as those that have been assigned to be translated during the Arabic Wordnet project. Selecting items from the tree causes English synonyms and gloss to be displayed, as well as Arabic translations where they exist.
• Searching for concepts in the WordNet - possible using both English and Arabic search terms. Arabic searches may be carried out using either words (entered with or without diacritics) or roots. In all cases, the searches return a list of synsets containing the search term. Selecting a synset from the returned list causes it to be displayed in the tree, allowing relationships with other concepts to be viewed, as well as information about the corresponding English/Arabic synset.
• Updating Arabic data - The AWN Browser uses a locally-stored database of Arabic data, but it provides facilities to update this database automatically at regular intervals with the latest data from the Barcelona server.

2. Downloading the browser

Prerequisites

The AWN broswer is written in Java and so you must have the Java Runtime Environment (JRE) installed on your machine. The code has been developed and tested using Java 5.0. If you do not have the JRE installed on your machine, then please click here to download it.

Downloading and running the AWN Browser

1. Download AWNBrowser.zip
2. Unzip AWNBrowser.zip. The contents of the AWNBrowser directory should be as follows:
   • AWNBrowser.jar
   • Readme.txt
   • runAWN.bat
   • runAWN.sh
   • wordnet (directory) - This directory contains the database; its contents should not be modified manually.
3. Start up the AWN Browser interface:
   • Windows users should double click on runAWN.bat
   • Non-windows users should click on runAWN.sh

3. Getting started

Browsing the tree

The tree representation of the Wordnet is located at the bottom left of the AWN Browser.

Clicking on will expand the appropriate branch to reveal the top-level concepts for nouns or verbs, as follows:

Each concept in the tree is accompanied by a colour-coded icon to show the current status of the concept in the Arabic Wordnet. Green-coloured icons indicate that the concept has an Arabic translation.

Double clicking on one of the concepts will reveal if it has sub-concepts. If sub-concepts exist, then the symbol will appear beside the node. A subsequent click on this symbol will reveal the sub-concepts, as shown below, where the enity concept node has been expanded:

Clicking on the name of a concept in the tree will cause information about that concept to be displayed on other parts of the interface, including Arabic transations of the concept if they are available.

Searching for concepts

Concepts in the WordNet may be searched for in either English or Arabic.

• For searches in English, the word should be entered into the textbox labelled English Word.
• For searches in Arabic, it is possible to enter either a word into a text box in the row labelled Arabic Word or a root into a text box in the row labelled Arabic Root. On both rows, there are 2 text boxes. If Arabic character keyboard mappings are known, then the word or root can be entered into the box labelled Arabic Input. Otherwise, the word or root can be entered using the Buckwalter transliteration system into the box labelled Buckwalter Input.

For both English and Arabic searches, the default is to search for matching concepts belonging to any part of speech. However, the search may be constrained to a single part of speech by choosing an alternative value from the drop down menu labelled Part of Speech

In order to perform the search, the Find Senses button on the appropriate side of the interface should be clicked. If the search term is found in any synsets, then they will be displayed in the box below the search area. Each line displayed in the box corresponds to a synset, with synoyms separated by commas. Clicking over a synset causes further information about it to be displayed on other parts of the interface, including its position in the tree (nouns and verbs only), gloss and equivalent English or Arabic synset.

NOTE:It can take a few seconds for information to be displayed after a synset has been clicked upon.