arcoder.py implements an abstract base class called Encoder along with two concrete implementations: ARCoder and Holmes. Each has an encode method that takes a string and encodes it into a series of symbols designed to be used for similarity measurements.

    >>> from arcoder import ARCoder, Holmes
    >>> a = ARCoder()
    >>> a.encode("Sohaib")
    ['suhaeb', 'suhib']
    >>> h = Holmes()
    >>> h.encode("Sohaib")
    ['sohayb']

The ARCoder algorithm is described more fully in Moore, J., Hamid, S., and Bromberger, S.: "An Evaluation of Transliterated Arabic Name Matching Methods".

The Holmes implementation is derived from Holmes, D., Kashfi, S., Aqeel, S. U.: "Transliterated arabic name search", Communications, Internet, and Information Technology, pp. 267-273. (2004).