Duali For Linux

Duali is an Arabic spell-checker designed to accommodate to the Arabic language (and extensible to other non-Arab based languages as well). It was named after the legendary founder of the Arabic grammar (Abul Aswad al Du'ali - d. 688).

Below are the scripts that come with your Duali package and what they do:

duali

This is the main program. You will need to make sure the 'duali.conf' file points to the right location of your indexed dictionary data files.

duali.conf

Be sure to copy this file into the /etc/ dir (or whatever the variable 'conf' points to in the duali file). The setup script should perform this for you.

dict2db

This program will generated dictionary data files to be used by the 'duali' program. It is preferable that you use the pybsddb3 module, but it will use whatever dbm is available on your system.

arabic2trans

This program will simply transliterate any Arabic encoded file to the Buckwalter transliteration scheme.

trans2arabic

NOTE: This script is probably only of interest to people interested in the actual development of Duali, not its users.

Requirements:

� Python > 2.2.x � duali-data files � dbm python modules

Now that you have a current version of Python, you will need to generate the dictionary files.

This is done by using the 'dict2db.py' script:

Usage: duali -c filename [-h | --help ] outputs this usage message [-V | --version ] program version [-c | --check= filename] input file to spellcheck [-C | --charset ] character encoding (cp1256, utf-8) [-n | --normalize ] turn normalization off (default on) [-p | --path ] dictionaries path [-v | --verbose ] verbose out (for debugging)

Normalization - What this means is it essentially removes all the hamza's from words, converts all the TEH_MARBUTA's to HEH's, etc. It guarantees that the word is in its simplest written form.

What's New in This Release:

� pyduali/aradict.py: Fixed bug #80 -- different dbm's may or may not add extensions to generated db files. � dict2db: Fixed bug #80. Also added better handling of path names. � arabic2trans: Added new script to transliterate UTF-8 or CP1256 encoded files. � pyduali/aralex.py: Removed redundant SUKUN from stripDiacritics() � setup.py: added arabic2trans to installation and incremented version number. � debian/: Improved on debian packaging. � doc/man/: Slight modifications of manpages. � CHANGELOG: Renamed it to ChangeLog (this file) to comply with the new Arabeyes Developer Guide.