1 /* 2 Egothor Software License version 1.00 3 Copyright (C) 1997-2004 Leo Galambos. 4 Copyright (C) 2002-2004 "Egothor developers" 5 on behalf of the Egothor Project. 6 All rights reserved. 7 8 This software is copyrighted by the "Egothor developers". If this 9 license applies to a single file or document, the "Egothor developers" 10 are the people or entities mentioned as copyright holders in that file 11 or document. If this license applies to the Egothor project as a 12 whole, the copyright holders are the people or entities mentioned in 13 the file CREDITS. This file can be found in the same location as this 14 license in the distribution. 15 16 Redistribution and use in source and binary forms, with or without 17 modification, are permitted provided that the following conditions are 18 met: 19 1. Redistributions of source code must retain the above copyright 20 notice, the list of contributors, this list of conditions, and the 21 following disclaimer. 22 2. Redistributions in binary form must reproduce the above copyright 23 notice, the list of contributors, this list of conditions, and the 24 disclaimer that follows these conditions in the documentation 25 and/or other materials provided with the distribution. 26 3. The name "Egothor" must not be used to endorse or promote products 27 derived from this software without prior written permission. For 28 written permission, please contact Leo.G@seznam.cz 29 4. Products derived from this software may not be called "Egothor", 30 nor may "Egothor" appear in their name, without prior written 31 permission from Leo.G@seznam.cz. 32 33 In addition, we request that you include in the end-user documentation 34 provided with the redistribution and/or in the software itself an 35 acknowledgement equivalent to the following: 36 "This product includes software developed by the Egothor Project. 37 http://egothor.sf.net/" 38 39 THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED OR IMPLIED 40 WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF 41 MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 42 IN NO EVENT SHALL THE EGOTHOR PROJECT OR ITS CONTRIBUTORS BE LIABLE 43 FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 44 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 45 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR 46 BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, 47 WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE 48 OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN 49 IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 50 51 This software consists of voluntary contributions made by many 52 individuals on behalf of the Egothor Project and was originally 53 created by Leo Galambos (Leo.G@seznam.cz). 54 */ 55 package org.egothor.stemmer; 56 57 import java.util.ArrayList; 58 import java.util.Arrays; 59 import java.util.Iterator; 60 import java.util.List; 61 62 /** 63 * The Optimizer class is a Trie that will be reduced (have empty rows removed). 64 * 65 * <p>The reduction will be made by joining two rows where the first is a subset of the second. 66 */ 67 public class Optimizer extends Reduce { 68 /** Constructor for the Optimizer object. */ Optimizer()69 public Optimizer() {} 70 71 /** 72 * Optimize (remove empty rows) from the given Trie and return the resulting Trie. 73 * 74 * @param orig the Trie to consolidate 75 * @return the newly consolidated Trie 76 */ 77 @Override optimize(Trie orig)78 public Trie optimize(Trie orig) { 79 List<CharSequence> cmds = orig.cmds; 80 List<Row> rows = new ArrayList<>(); 81 List<Row> orows = orig.rows; 82 int[] remap = new int[orows.size()]; 83 84 for (int j = orows.size() - 1; j >= 0; j--) { 85 Row now = new Remap(orows.get(j), remap); 86 boolean merged = false; 87 88 for (int i = 0; i < rows.size(); i++) { 89 Row q = merge(now, rows.get(i)); 90 if (q != null) { 91 rows.set(i, q); 92 merged = true; 93 remap[j] = i; 94 break; 95 } 96 } 97 98 if (merged == false) { 99 remap[j] = rows.size(); 100 rows.add(now); 101 } 102 } 103 104 int root = remap[orig.root]; 105 Arrays.fill(remap, -1); 106 rows = removeGaps(root, rows, new ArrayList<Row>(), remap); 107 108 return new Trie(orig.forward, remap[root], cmds, rows); 109 } 110 111 /** 112 * Merge the given rows and return the resulting Row. 113 * 114 * @param master the master Row 115 * @param existing the existing Row 116 * @return the resulting Row, or <code>null</code> if the operation cannot be realized 117 */ merge(Row master, Row existing)118 public Row merge(Row master, Row existing) { 119 Iterator<Character> i = master.cells.keySet().iterator(); 120 Row n = new Row(); 121 for (; i.hasNext(); ) { 122 Character ch = i.next(); 123 // XXX also must handle Cnt and Skip !! 124 Cell a = master.cells.get(ch); 125 Cell b = existing.cells.get(ch); 126 127 Cell s = (b == null) ? new Cell(a) : merge(a, b); 128 if (s == null) { 129 return null; 130 } 131 n.cells.put(ch, s); 132 } 133 i = existing.cells.keySet().iterator(); 134 for (; i.hasNext(); ) { 135 Character ch = i.next(); 136 if (master.at(ch) != null) { 137 continue; 138 } 139 n.cells.put(ch, existing.at(ch)); 140 } 141 return n; 142 } 143 144 /** 145 * Merge the given Cells and return the resulting Cell. 146 * 147 * @param m the master Cell 148 * @param e the existing Cell 149 * @return the resulting Cell, or <code>null</code> if the operation cannot be realized 150 */ merge(Cell m, Cell e)151 public Cell merge(Cell m, Cell e) { 152 Cell n = new Cell(); 153 154 if (m.skip != e.skip) { 155 return null; 156 } 157 158 if (m.cmd >= 0) { 159 if (e.cmd >= 0) { 160 if (m.cmd == e.cmd) { 161 n.cmd = m.cmd; 162 } else { 163 return null; 164 } 165 } else { 166 n.cmd = m.cmd; 167 } 168 } else { 169 n.cmd = e.cmd; 170 } 171 if (m.ref >= 0) { 172 if (e.ref >= 0) { 173 if (m.ref == e.ref) { 174 if (m.skip == e.skip) { 175 n.ref = m.ref; 176 } else { 177 return null; 178 } 179 } else { 180 return null; 181 } 182 } else { 183 n.ref = m.ref; 184 } 185 } else { 186 n.ref = e.ref; 187 } 188 n.cnt = m.cnt + e.cnt; 189 n.skip = m.skip; 190 return n; 191 } 192 } 193