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.
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.
16 Redistribution and use in source and binary forms, with or without
17 modification, are permitted provided that the following conditions are
19 1. Redistributions of source code must retain the above copyright
20 notice, the list of contributors, this list of conditions, and the
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.
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/"
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.
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).
55 package org.egothor.stemmer;
57 import java.io.BufferedOutputStream;
58 import java.io.BufferedReader;
59 import java.io.DataOutputStream;
60 import java.io.FileInputStream;
61 import java.io.FileNotFoundException;
62 import java.io.FileOutputStream;
63 import java.io.IOException;
64 import java.io.InputStreamReader;
65 import java.io.LineNumberReader;
66 import java.util.StringTokenizer;
69 * The Compile class is used to compile a stemmer table.
71 public class Compile {
73 static boolean backward;
78 * Entry point to the Compile application.
80 * This program takes any number of arguments: the first is the name of the
81 * desired stemming algorithm to use (a list is available in the package
82 * description) , all of the rest should be the path or paths to a file or
83 * files containing a stemmer table to compile.
85 * @param args the command line arguments
87 public static void main(java.lang.String[] args) {
88 if (args.length < 1) {
92 args[0].toUpperCase();
94 backward = args[0].charAt(0) == '-';
95 int qq = (backward) ? 1 : 0;
96 boolean storeorig = false;
98 if (args[0].charAt(qq) == '0') {
103 multi = args[0].charAt(qq) == 'M';
108 String charset = System.getProperty("egothor.stemmer.charset", "UTF-8");
110 char optimizer[] = new char[args[0].length() - qq];
111 for (int i = 0; i < optimizer.length; i++) {
112 optimizer[i] = args[0].charAt(qq + i);
115 for (int i = 1; i < args.length; i++) {
117 // System.out.println("[" + args[i] + "]");
118 Diff diff = new Diff();
125 System.out.println(args[i]);
126 in = new LineNumberReader(new BufferedReader(new InputStreamReader(
127 new FileInputStream(args[i]), charset)));
128 for (String line = in.readLine(); line != null; line = in.readLine()) {
130 line = line.toLowerCase();
131 StringTokenizer st = new StringTokenizer(line);
132 String stem = st.nextToken();
134 trie.add(stem, "-a");
137 while (st.hasMoreTokens()) {
138 String token = st.nextToken();
139 if (token.equals(stem) == false) {
140 trie.add(token, diff.exec(token, stem));
144 } catch (java.util.NoSuchElementException x) {
145 // no base token (stem) on a line
149 Optimizer o = new Optimizer();
150 Optimizer2 o2 = new Optimizer2();
151 Lift l = new Lift(true);
152 Lift e = new Lift(false);
153 Gener g = new Gener();
155 for (int j = 0; j < optimizer.length; j++) {
157 switch (optimizer[j]) {
159 trie = trie.reduce(g);
163 trie = trie.reduce(l);
167 trie = trie.reduce(e);
171 trie = trie.reduce(o2);
175 trie = trie.reduce(o);
181 trie.printInfo(prefix + " ");
184 DataOutputStream os = new DataOutputStream(new BufferedOutputStream(
185 new FileOutputStream(args[i] + ".out")));
186 os.writeUTF(args[0]);
190 } catch (FileNotFoundException x) {
192 } catch (IOException x) {
198 static void allocTrie() {
200 trie = new MultiTrie2(!backward);
202 trie = new Trie(!backward);