stemmer.js

/**
 * Porter Stemmer - see http://tartarus.org/~martin/PorterStemmer/
 * usage: Stemmer.stem( word )
 */
var Stemmer = {
  stem: function( word ) {
    word = word.toLowerCase();

    this.log( "Stemming", word );

    // skip anything with length 1 or 2
    if ( word.length <= 2 ) {
      return word;
    }

    var stem = word;

    var contains_vowel = /[aeiou]|[^aeiou]y/;
    var m_gt_0 = /^([^aeiou][^aeiouy]*([aeiouy]+[^aeiou][^aeiouy]*)|([aeiou]+[^aeiou][^aeiouy]*))([aeiouy]+[^aeiou][^aeiouy]*)*([aeiouy][aeiou]*)?$/;
    var m_eq_1 = /^([^aeiou][^aeiouy]*([aeiouy]+[^aeiou][^aeiouy]*)|([aeiou]+[^aeiou][^aeiouy]*))([aeiouy][aeiou]*)?$/;
    var m_gt_1 = /^([^aeiou][^aeiouy]*([aeiouy]+[^aeiou][^aeiouy]*)|([aeiou]+[^aeiou][^aeiouy]*))([aeiouy]+[^aeiou][^aeiouy]*)+([aeiouy][aeiou]*)?$/;

    var ends_o = /[^aeiou][aeiouy][^aeiouwxy]$/;

    // step 1a
    var es_end = /(ss|i)es$/;
    var s_end = /([^s])s$/;

    if ( es_end.test( stem ) ) {
      stem = stem.replace( es_end, "$1" );
      this.log( "Matched es_end =>", stem );
    } else if ( s_end.test( stem ) ) {
      stem = stem.replace( s_end, "$1" );
      this.log( "Matched s_end =>", stem );
    }

    // step 1b
    var eed_end = /^(.*)eed$/;
    var eding_end = /^(.*)(?:ed|ing)$/;

    if ( eed_end.test( stem ) ) {
      var prefix = eed_end.exec( stem )[ 1 ];
      if ( m_gt_0.test( prefix ) ) {
        stem = prefix + "ee";
        this.log( "Matched eed_end and m_gt_0 =>", stem );
      }
      else this.log( "Matched eed_end but not m_gt_0" );
    } else if ( eding_end.test( stem ) ) {
      var prefix = eding_end.exec( stem )[ 1 ];
      if ( contains_vowel.test( prefix ) ) {
        stem = prefix;

        this.log( "Matched eding_end and contains_vowel =>", stem );

        // step 1b followup
        var fix_suffix = /(at|bl|iz|is)$/;
        var double_consonant = /([^aeioulsz])\1$/;

        if ( fix_suffix.test( stem ) ) {
          stem = stem + "e";
          this.log( "Matched fix_suffix =>", stem );
        } else if ( double_consonant.test( stem ) ) {
          stem = stem.replace( double_consonant, "$1" );
          this.log( "Matched double_consonant =>", stem );
        } else if ( m_eq_1.test( stem ) && ends_o.test( stem ) ) {
          stem = stem + "e";
          this.log( "Matched m_eq_1 and ends_o =>", stem );
        }
      }
      else this.log( "Matched eding_end but not contains_vowel" );
    }

    // step 1c
    var y_end = /^(.*)y$/;

    if ( y_end.test( stem ) ) {
      var prefix = y_end.exec( stem )[ 1 ];
      if ( contains_vowel.test( prefix ) ) {
        stem = prefix + "i";
        this.log( "Matched y_end and contains_vowel =>", stem );
      }
      else this.log( "Matched y_end but not contains_vowel" );
    }

    // steps 2, 3 and 4
    var steps = [
      [ this.step2_suffixes, -2, m_gt_0 ],
      [ this.step3_suffixes, -1, m_gt_0 ],
      [ this.step4_suffixes, -2, m_gt_1 ]
    ];

    for each ( [suffixes, char_pos, measure] in steps ) {
      suffixes = suffixes[ stem.substr( char_pos, 1 ) ];
      for ( var suf in suffixes ) {
        var suffix_replace = new RegExp( suf + "$" );
        if ( suffix_replace.test( stem ) ) {
          var prefix = suffix_replace.exec( stem )[ 1 ];
          if ( measure.test( prefix ) ) {
            stem = prefix + suffixes[ suf ];
            this.log( "Matched", suf, "and", measure, "=>", stem );
          }
          else this.log( "Matched", suf, "but not", measure );
          break;
        }
      }
    }

    // step 5a
    var e_end = /(.*)e$/;
    if ( e_end.test( stem ) ) {
      var prefix = e_end.exec( stem )[ 1 ];
      if ( m_gt_1.test( prefix ) ) {
        stem = prefix;
        this.log( "Matched e_end and m_gt_1 =>", stem );
      }
      else if ( m_eq_1.test( prefix ) && !ends_o.test( prefix ) ) {
        stem = prefix;
        this.log( "Matched e_end and m_eq_1 and !ends_o =>", stem );
      }
      else this.log( "Matched e_end but not m_gt_1 or m_eq_1 + *o$" );
    }

    // step 5b
    var double_l = /(.*l)l$/;
    if ( double_l.test( stem ) ) {
      var prefix = double_l.exec( stem )[ 1 ];
      if ( m_gt_1.test( prefix ) ) {
        stem = prefix;
        this.log( "Matched double_l and m_gt_1 =>", stem );
      }
      else this.log( "Matched double_l but not m_gt_1" );
    }

    return stem;
  },

  // grouped by penulimate letter
  step2_suffixes: {
    a: { "(.*)ational": "ate",
         "(.*)tional": "tion" },
    c: { "(.*)enci": "ence",
         "(.*)anci": "ance" },
    e: { "(.*)izer": "ize",
         "(.*)iser": "ise" },
    g: { "(.*)logi": "log" },
    l: { "(.*)bli": "ble",
         "(.*)alli": "al",
         "(.*)entli": "ent",
         "(.*)eli": "e",
         "(.*)ousli": "ous" },
    i: { "(.*)ization": "ize" },
    o: { "(.*)ation": "ate",
         "(.*)ator": "ate" },
    s: { "(.*)alism": "al",
         "(.*)iveness": "ive",
         "(.*)fulness": "ful",
         "(.*)ousness": "ous" },
    t: { "(.*)aliti": "al",
         "(.*)iviti": "ive",
         "(.*)biliti": "ble" }
  },

  // grouped by final letter
  step3_suffixes: {
    e: { "(.*)icate": "ic",
         "(.*)ative": "",
         "(.*)alize": "al",
         "(.*)alise": "al" },
    i: { "(.*)iciti": "ic" },
    l: { "(.*)ical": "ic",
         "(.*)ful": "" },
    s: { "(.*)ness": "" }
  },

  // grouped by penultimate letter
  step4_suffixes: {
    a: { "(.*)al": "" },
    c: { "(.*)ance": "",
         "(.*)ence": "" },
    e: { "(.*)er": "" },
    i: { "(.*)ic": "" },
    l: { "(.*)able": "",
         "(.*)ible": "" },
    n: { "(.*)ant": "",
         "(.*)ement": "",
         "(.*)ment": "",
         "(.*)ent": "" },
    o: { "(.*s)ion": "",
         "(.*t)ion": "",
         "(.*)ou": "" },
    s: { "(.*)ism": "" },
    t: { "(.*)ate": "",
         "(.*)iti": "" },
    u: { "(.*)ous": "" },
    v: { "(.*)ive": "" },
    z: { "(.*)ize": "" }
  },

  log: function() {
    // print.apply( this, arguments );
  }
}

// CommonJS export
if ( exports ) {
  exports.Stemmer = Stemmer;
}