Package org.apache.lucene.analysis.hi

Class HindiNormalizer

java.lang.Object
org.apache.lucene.analysis.hi.HindiNormalizer
public class HindiNormalizer extends Object
Normalizer for Hindi.

Normalizes text to remove some differences in spelling variations.

Implements the Hindi-language specific algorithm specified in: Word normalization in Indian languages Prasad Pingali and Vasudeva Varma. http://web2py.iiit.ac.in/publications/default/download/inproceedings.pdf.3fe5b38c-02ee-41ce-9a8f-3e745670be32.pdf

with the following additions from Hindi CLIR in Thirty Days Leah S. Larkey, Margaret E. Connell, and Nasreen AbdulJaleel. http://maroo.cs.umass.edu/pub/web/getpdf.php?id=454:

  • Internal Zero-width joiner and Zero-width non-joiners are removed
  • In addition to chandrabindu, NA+halant is normalized to anusvara

  • Constructor Details

    • HindiNormalizer

      public HindiNormalizer()

  • Method Details

    • normalize

      public int normalize(char[] s, int len)
      Normalize an input buffer of Hindi text
      Parameters:
      s - input buffer
      len - length of input buffer
      Returns:
      length of input buffer after normalization