lucene/analysis/common/src/java/org/apache/lucene/analysis/lv/LatvianStemmer.java

/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
package org.apache.lucene.analysis.lv;

import static org.apache.lucene.analysis.util.StemmerUtil.*;

/**
 * Light stemmer for Latvian.
 *
 * <p>This is a light version of the algorithm in Karlis Kreslin's PhD thesis <i>A stemming
 * algorithm for Latvian</i> with the following modifications:
 *
 * <ul>
 *   <li>Only explicitly stems noun and adjective morphology
 *   <li>Stricter length/vowel checks for the resulting stems (verb etc suffix stripping is removed)
 *   <li>Removes only the primary inflectional suffixes: case and number for nouns ; case, number,
 *       gender, and definitiveness for adjectives.
 *   <li>Palatalization is only handled when a declension II,V,VI noun suffix is removed.
 * </ul>
 */
public class LatvianStemmer {
  /** Stem a latvian word. returns the new adjusted length. */
  public int stem(char s[], int len) {
    int numVowels = numVowels(s, len);

    for (int i = 0; i < affixes.length; i++) {
      Affix affix = affixes[i];
      if (numVowels > affix.vc && len >= affix.affix.length + 3 && endsWith(s, len, affix.affix)) {
        len -= affix.affix.length;
        return affix.palatalizes ? unpalatalize(s, len) : len;
      }
    }

    return len;
  }

  static final Affix affixes[] = {
    new Affix("ajiem", 3, false), new Affix("ajai", 3, false),
    new Affix("ajam", 2, false), new Affix("ajām", 2, false),
    new Affix("ajos", 2, false), new Affix("ajās", 2, false),
    new Affix("iem", 2, true), new Affix("ajā", 2, false),
    new Affix("ais", 2, false), new Affix("ai", 2, false),
    new Affix("ei", 2, false), new Affix("ām", 1, false),
    new Affix("am", 1, false), new Affix("ēm", 1, false),
    new Affix("īm", 1, false), new Affix("im", 1, false),
    new Affix("um", 1, false), new Affix("us", 1, true),
    new Affix("as", 1, false), new Affix("ās", 1, false),
    new Affix("es", 1, false), new Affix("os", 1, true),
    new Affix("ij", 1, false), new Affix("īs", 1, false),
    new Affix("ēs", 1, false), new Affix("is", 1, false),
    new Affix("ie", 1, false), new Affix("u", 1, true),
    new Affix("a", 1, true), new Affix("i", 1, true),
    new Affix("e", 1, false), new Affix("ā", 1, false),
    new Affix("ē", 1, false), new Affix("ī", 1, false),
    new Affix("ū", 1, false), new Affix("o", 1, false),
    new Affix("s", 0, false), new Affix("š", 0, false),
  };

  static class Affix {
    char affix[]; // suffix
    int vc; // vowel count of the suffix
    boolean palatalizes; // true if we should fire palatalization rules.

    Affix(String affix, int vc, boolean palatalizes) {
      this.affix = affix.toCharArray();
      this.vc = vc;
      this.palatalizes = palatalizes;
    }
  }

  /**
   * Most cases are handled except for the ambiguous ones:
   *
   * <ul>
   *   <li>s -&gt; š
   *   <li>t -&gt; š
   *   <li>d -&gt; ž
   *   <li>z -&gt; ž
   * </ul>
   */
  private int unpalatalize(char s[], int len) {
    // we check the character removed: if it's -u then
    // it's 2,5, or 6 gen pl., and these two can only apply then.
    if (s[len] == 'u') {
      // kš -> kst
      if (endsWith(s, len, "kš")) {
        len++;
        s[len - 2] = 's';
        s[len - 1] = 't';
        return len;
      }
      // ņņ -> nn
      if (endsWith(s, len, "ņņ")) {
        s[len - 2] = 'n';
        s[len - 1] = 'n';
        return len;
      }
    }

    // otherwise all other rules
    if (endsWith(s, len, "pj")
        || endsWith(s, len, "bj")
        || endsWith(s, len, "mj")
        || endsWith(s, len, "vj")) {
      // labial consonant
      return len - 1;
    } else if (endsWith(s, len, "šņ")) {
      s[len - 2] = 's';
      s[len - 1] = 'n';
      return len;
    } else if (endsWith(s, len, "žņ")) {
      s[len - 2] = 'z';
      s[len - 1] = 'n';
      return len;
    } else if (endsWith(s, len, "šļ")) {
      s[len - 2] = 's';
      s[len - 1] = 'l';
      return len;
    } else if (endsWith(s, len, "žļ")) {
      s[len - 2] = 'z';
      s[len - 1] = 'l';
      return len;
    } else if (endsWith(s, len, "ļņ")) {
      s[len - 2] = 'l';
      s[len - 1] = 'n';
      return len;
    } else if (endsWith(s, len, "ļļ")) {
      s[len - 2] = 'l';
      s[len - 1] = 'l';
      return len;
    } else if (s[len - 1] == 'č') {
      s[len - 1] = 'c';
      return len;
    } else if (s[len - 1] == 'ļ') {
      s[len - 1] = 'l';
      return len;
    } else if (s[len - 1] == 'ņ') {
      s[len - 1] = 'n';
      return len;
    }

    return len;
  }

  /**
   * Count the vowels in the string, we always require at least one in the remaining stem to accept
   * it.
   */
  private int numVowels(char s[], int len) {
    int n = 0;
    for (int i = 0; i < len; i++) {
      switch (s[i]) {
        case 'a':
        case 'e':
        case 'i':
        case 'o':
        case 'u':
        case 'ā':
        case 'ī':
        case 'ē':
        case 'ū':
          n++;
      }
    }
    return n;
  }
}