1069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project/*
2069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * Copyright 2001-2004 The Apache Software Foundation.
3069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project *
4069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * Licensed under the Apache License, Version 2.0 (the "License");
5069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * you may not use this file except in compliance with the License.
6069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * You may obtain a copy of the License at
7069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project *
8069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project *      http://www.apache.org/licenses/LICENSE-2.0
9069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project *
10069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * Unless required by applicable law or agreed to in writing, software
11069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * distributed under the License is distributed on an "AS IS" BASIS,
12069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * See the License for the specific language governing permissions and
14069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * limitations under the License.
15069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project */
16069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project
17069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Projectpackage org.apache.commons.codec.language;
18069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project
19069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Projectimport org.apache.commons.codec.EncoderException;
20069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Projectimport org.apache.commons.codec.StringEncoder;
21069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project
22069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project/**
23069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * Utility methods for {@link Soundex} and {@link RefinedSoundex} classes.
24069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project *
25069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * @author Apache Software Foundation
26069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * @version $Id: SoundexUtils.java,v 1.5 2004/03/17 18:31:35 ggregory Exp $
27069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project * @since 1.3
28069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project */
29069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Projectfinal class SoundexUtils {
30069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project
31069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project    /**
32069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * Cleans up the input string before Soundex processing by only returning
33069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * upper case letters.
34069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *
35069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @param str
36069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *                  The String to clean.
37069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @return A clean String.
38069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     */
39069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project    static String clean(String str) {
40069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        if (str == null || str.length() == 0) {
41069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project            return str;
42069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        }
43069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        int len = str.length();
44069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        char[] chars = new char[len];
45069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        int count = 0;
46069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        for (int i = 0; i < len; i++) {
47069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project            if (Character.isLetter(str.charAt(i))) {
48069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project                chars[count++] = str.charAt(i);
49069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project            }
50069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        }
51069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        if (count == len) {
52069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project            return str.toUpperCase();
53069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        }
54069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        return new String(chars, 0, count).toUpperCase();
55069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project    }
56069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project
57069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project    /**
58069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * Encodes the Strings and returns the number of characters in the two
59069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * encoded Strings that are the same.
60069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * <ul>
61069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * <li>For Soundex, this return value ranges from 0 through 4: 0 indicates
62069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * little or no similarity, and 4 indicates strong similarity or identical
63069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * values.</li>
64069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * <li>For refined Soundex, the return value can be greater than 4.</li>
65069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * </ul>
66069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *
67069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @param encoder
68069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *                  The encoder to use to encode the Strings.
69069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @param s1
70069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *                  A String that will be encoded and compared.
71069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @param s2
72069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *                  A String that will be encoded and compared.
73069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @return The number of characters in the two Soundex encoded Strings that
74069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *             are the same.
75069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *
76069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @see #differenceEncoded(String,String)
77069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @see <a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/tsqlref/ts_de-dz_8co5.asp">
78069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *          MS T-SQL DIFFERENCE</a>
79069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *
80069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @throws EncoderException
81069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *                  if an error occurs encoding one of the strings
82069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     */
83069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project    static int difference(StringEncoder encoder, String s1, String s2) throws EncoderException {
84069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        return differenceEncoded(encoder.encode(s1), encoder.encode(s2));
85069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project    }
86069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project
87069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project    /**
88069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * Returns the number of characters in the two Soundex encoded Strings that
89069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * are the same.
90069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * <ul>
91069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * <li>For Soundex, this return value ranges from 0 through 4: 0 indicates
92069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * little or no similarity, and 4 indicates strong similarity or identical
93069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * values.</li>
94069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * <li>For refined Soundex, the return value can be greater than 4.</li>
95069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * </ul>
96069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *
97069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @param es1
98069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *                  An encoded String.
99069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @param es2
100069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *                  An encoded String.
101069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @return The number of characters in the two Soundex encoded Strings that
102069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *             are the same.
103069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *
104069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     * @see <a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/tsqlref/ts_de-dz_8co5.asp">
105069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     *          MS T-SQL DIFFERENCE</a>
106069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project     */
107069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project    static int differenceEncoded(String es1, String es2) {
108069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project
109069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        if (es1 == null || es2 == null) {
110069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project            return 0;
111069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        }
112069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        int lengthToMatch = Math.min(es1.length(), es2.length());
113069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        int diff = 0;
114069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        for (int i = 0; i < lengthToMatch; i++) {
115069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project            if (es1.charAt(i) == es2.charAt(i)) {
116069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project                diff++;
117069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project            }
118069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        }
119069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project        return diff;
120069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project    }
121069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project
122069490a5ca2fd1988d29daf45d892f47ad665115The Android Open Source Project}
123