Metaphone


1   /*
2    * Copyright 1999,2004 The Apache Software Foundation.
3    * 
4    * Licensed under the Apache License, Version 2.0 (the "License");
5    * you may not use this file except in compliance with the License.
6    * You may obtain a copy of the License at
7    * 
8    *      http://www.apache.org/licenses/LICENSE-2.0
9    * 
10   * Unless required by applicable law or agreed to in writing, software
11   * distributed under the License is distributed on an "AS IS" BASIS,
12   * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13   * See the License for the specific language governing permissions and
14   * limitations under the License.
15   */
16  //package com.bga.wbrogden.metaphone;
17  package org.apache.taglibs.string.util;
18  
19  public class Metaphone extends Object   {
20  
21      static String   vowels = "AEIOU" ;
22      static String   frontv = "EIY"   ;
23      static String   varson = "CSPTG" ;
24  
25      static final int maxCodeLen = 4 ;
26  
27      static public String   metaPhone( String   txt ){
28        int mtsz = 0  ;
29        boolean hard = false ;
30        if(( txt == null ) ||
31           ( txt.length() == 0 )) return "" ;
32        // single character is itself
33        if( txt.length() == 1 ) return txt.toUpperCase() ;
34        //
35        char[] inwd = txt.toUpperCase().toCharArray() ;
36        //
37        String   tmpS ;
38        StringBuffer   local = new StringBuffer  ( 40 ); // manipulate
39        StringBuffer   code = new StringBuffer  ( 10 ) ; //   output
40        // handle initial 2 characters exceptions
41        switch( inwd[0] ){
42          case 'K': case 'G' : case 'P' : /* looking for KN, etc*/
43            if( inwd[1] == 'N')local.append(inwd, 1, inwd.length - 1 );
44            else local.append( inwd );
45            break;
46          case 'A': /* looking for AE */
47            if( inwd[1] == 'E' )local.append(inwd, 1, inwd.length - 1 );
48            else local.append( inwd );
49            break;
50          case 'W' : /* looking for WR or WH */
51            if( inwd[1] == 'R' ){   // WR -> R
52              local.append(inwd, 1, inwd.length - 1 ); break ;
53            }
54            if( inwd[1] == 'H'){
55              local.append(inwd, 1, inwd.length - 1 );
56              local.setCharAt( 0,'W'); // WH -> W
57            }
58            else local.append( inwd );
59            break;
60          case 'X' : /* initial X becomes S */
61            inwd[0] = 'S' ;local.append( inwd );
62            break ;
63          default :
64            local.append( inwd );
65        } // now local has working string with initials fixed
66        int wdsz = local.length();
67        int n = 0 ;
68        while((mtsz < maxCodeLen ) && // max code size of 4 works well
69              (n < wdsz ) ){
70          char symb = local.charAt(n) ;
71          // remove duplicate letters except C
72          if(( symb != 'C' ) &&
73             (n > 0 ) && ( local.charAt(n - 1 ) == symb )) n++ ;
74          else{ // not dup
75            switch( symb ){
76              case 'A' : case 'E' : case 'I' : case 'O' : case 'U' :
77                if( n == 0 ) { code.append(symb );mtsz++;
78                }
79                break ; // only use vowel if leading char
80              case 'B' :
81                if( (n > 0 ) &&
82                    !(n + 1 == wdsz ) && // not MB at end of word
83                    ( local.charAt(n - 1) == 'M')) {
84                      code.append(symb);
85                    }
86                else code.append(symb);
87                mtsz++ ;
88                break ;
89              case 'C' : // lots of C special cases
90                /* discard if SCI, SCE or SCY */
91                if( ( n > 0 ) &&
92                    ( local.charAt(n-1) == 'S' ) &&
93                    ( n + 1 < wdsz ) &&
94                    ( frontv.indexOf( local.charAt(n + 1)) >= 0 )){ break ;}
95                tmpS = local.toString();
96                if( tmpS.indexOf("CIA", n ) == n ) { // "CIA" -> X
97                   code.append('X' ); mtsz++; break ;
98                }
99                if( ( n + 1 < wdsz ) &&
100                   (frontv.indexOf( local.charAt(n+1) )>= 0 )){
101                  code.append('S');mtsz++; break ; // CI,CE,CY -> S
102               }
103               if(( n > 0) &&
104                  ( tmpS.indexOf("SCH",n-1 )== n-1 )){ // SCH->sk
105                  code.append('K') ; mtsz++;break ;
106               }
107               if( tmpS.indexOf("CH", n ) == n ){ // detect CH
108                 if((n == 0 ) &&
109                    (wdsz >= 3 ) &&    // CH consonant -> K consonant
110                    (vowels.indexOf( local.charAt( 2) ) < 0 )){
111                      code.append('K');
112                 }
113                 else { code.append('X'); // CHvowel -> X
114                 }
115                 mtsz++;
116               }
117               else { code.append('K' );mtsz++;
118               }
119               break ;
120             case 'D' :
121               if(( n + 2 < wdsz )&&  // DGE DGI DGY -> J
122                  ( local.charAt(n+1) == 'G' )&&
123                  (frontv.indexOf( local.charAt(n+2) )>= 0)){
124                     code.append('J' ); n += 2 ;
125               }
126               else { code.append( 'T' );
127               }
128               mtsz++;
129               break ;
130             case 'G' : // GH silent at end or before consonant
131               if(( n + 2 == wdsz )&&
132                  (local.charAt(n+1) == 'H' )) break ;
133               if(( n + 2 < wdsz ) &&
134                  (local.charAt(n+1) == 'H' )&&
135                  (vowels.indexOf( local.charAt(n+2)) < 0 )) break ;
136               tmpS = local.toString();
137               if((n > 0) &&
138                  ( tmpS.indexOf("GN", n ) == n)||
139                  ( tmpS.indexOf("GNED",n) == n )) break ; // silent G
140               if(( n > 0 ) &&
141                  (local.charAt(n-1) == 'G')) hard = true ;
142               else hard = false ;
143               if((n+1 < wdsz) &&
144                  (frontv.indexOf( local.charAt(n+1) ) >= 0 )&&
145                  (!hard) ) code.append( 'J' );
146               else code.append('K');
147               mtsz++;
148               break ;
149             case 'H':
150               if( n + 1 == wdsz ) break ; // terminal H
151               if((n > 0) &&
152                  (varson.indexOf( local.charAt(n-1)) >= 0)) break ;
153               if( vowels.indexOf( local.charAt(n+1)) >=0 ){
154                   code.append('H') ; mtsz++;// Hvowel
155               }
156               break;
157             case 'F': case 'J' : case 'L' :
158             case 'M': case 'N' : case 'R' :
159               code.append( symb ); mtsz++; break ;
160             case 'K' :
161               if( n > 0 ){ // not initial
162                 if( local.charAt( n -1) != 'C' ) {
163                      code.append(symb );
164                 }
165               }
166               else   code.append( symb ); // initial K
167               mtsz++ ;
168               break ;
169             case 'P' :
170               if((n + 1 < wdsz) &&  // PH -> F
171                  (local.charAt( n+1) == 'H'))code.append('F');
172               else code.append( symb );
173               mtsz++;
174               break ;
175             case 'Q' :
176               code.append('K' );mtsz++; break ;
177             case 'S' :
178               tmpS = local.toString();
179               if((tmpS.indexOf("SH", n )== n) ||
180                  (tmpS.indexOf("SIO",n )== n) ||
181                  (tmpS.indexOf("SIA",n )== n)) code.append('X');
182               else code.append( 'S' );
183               mtsz++ ;
184               break ;
185             case 'T' :
186               tmpS = local.toString(); // TIA TIO -> X
187               if((tmpS.indexOf("TIA",n )== n)||
188                  (tmpS.indexOf("TIO",n )== n) ){
189                     code.append('X'); mtsz++; break;
190               }
191               if( tmpS.indexOf("TCH",n )==n) break;
192               // substitute numeral 0 for TH (resembles theta after all)
193               if( tmpS.indexOf("TH", n )==n) code.append('0');
194               else code.append( 'T' );
195               mtsz++ ;
196               break ;
197             case 'V' :
198               code.append('F'); mtsz++;break ;
199             case 'W' : case 'Y' : // silent if not followed by vowel
200               if((n+1 < wdsz) &&
201                  (vowels.indexOf( local.charAt(n+1))>=0)){
202                     code.append( symb );mtsz++;
203               }
204               break ;
205             case 'X' :
206               code.append('K'); code.append('S');mtsz += 2;
207               break ;
208             case 'Z' :
209               code.append('S'); mtsz++; break ;
210           } // end switch
211           n++ ;
212         } // end else from symb != 'C'
213         if( mtsz > 4 )code.setLength( 4);
214       }
215       return code.toString();
216     } // end static method metaPhone()
217 
218 }
219
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags