Arabic Light Stemming Python

Arabic Light stemming with Python

This code performs light stemming for Arabic words. The implementation is based on ISRI Arabic Stemmer, which is a rooting algorithm for Arabic text. ISRI Arabic Stemmer is described in:

Taghva, K., Elkoury, R., and Coombs, J. 2005. Arabic Stemming without a root dictionary. Information Science Research Institute. University of Nevada, Las Vegas, USA.

Arabic Light Stemming Steps

1. remove diacritics which representing Arabic short vowels
2. remove length three and length two prefixes in this order
3. remove length three and length two suffixes in this order
4. remove connective ‘و’ if it precedes a word beginning with ‘و’
5. normalize initial hamza to bare alif

For more information, please refer to http://www.nltk.org/_modules/nltk/stem/isri.html

Please note steps are derived from http://www.nltk.org/_modules/nltk/stem/isri.html#ISRIStemmer.stem

Install requirements

pip install -r requirements

Arguments

usage: light_stem.py [-h] -i INFILE -o OUTFILE

performs light stemming for Arabic words.

optional arguments:
  -h, --help            show this help message and exit
  -i INFILE, --infile INFILE
                        input file.
  -o OUTFILE, --outfile OUTFILE
                        out file.