/interpreter/tags/at2dist041108/src/edu/vub/util/Regexp.java

 1/**
 2 * AmbientTalk/2 Project
 3 * Regexp.java created on 09-apr-2008 at 09:50:03
 4 * (c) Programming Technology Lab, 2006 - 2007
 5 * Authors: Tom Van Cutsem & Stijn Mostinckx
 6 * 
 7 * Permission is hereby granted, free of charge, to any person
 8 * obtaining a copy of this software and associated documentation
 9 * files (the "Software"), to deal in the Software without
10 * restriction, including without limitation the rights to use,
11 * copy, modify, merge, publish, distribute, sublicense, and/or
12 * sell copies of the Software, and to permit persons to whom the
13 * Software is furnished to do so, subject to the following
14 * conditions:
15 *
16 * The above copyright notice and this permission notice shall be
17 * included in all copies or substantial portions of the Software.
18 *
19 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
20 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
21 * OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
22 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
23 * HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
24 * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
25 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
26 * OTHER DEALINGS IN THE SOFTWARE.
27 */
28package edu.vub.util;
29
30import edu.vub.at.exceptions.InterpreterException;
31
32import org.apache.regexp.RE;
33import org.apache.regexp.RECompiler;
34import org.apache.regexp.REProgram;
35
36/**
37 * A thin wrapper around the Apache Regexp library. Primarily provides support for
38 * finding or replacing all occurences of a regular expression in a string.
39 * 
40 * This is the only class that should explicitly manage the compilation of regexp patterns.
41 * Instances of the class {@link RE} should never be shared between threads. However,
42 * {@link REProgram} instances can be shared among threads. Compilation of an REProgram
43 * into an RE however, must be synchronised between threads.
44 * 
45 * @author tvcutsem
46 */
47public class Regexp {
48	
49    private static final RECompiler _COMPILER_ = new RECompiler();
50    
51    // serve as "closures" passed to the findAll and replaceAll methods
52    
53	public interface StringCallable {
54		public String call(String input) throws InterpreterException;
55	}
56	public interface StringRunnable {
57		public void run(String input) throws InterpreterException;
58	}
59	
60	public static REProgram compile(String regex) {
61		// synchronize access to the compiler because it is not multiple-thread safe
62		synchronized (_COMPILER_) {
63			return _COMPILER_.compile(regex);
64		}
65	}
66	
67	public static void findAll(RE regex, String input, StringRunnable consumer) throws InterpreterException {
68		while (regex.match(input)) {
69			consumer.run(regex.getParen(0));
70			input = input.substring(regex.getParenEnd(0), input.length());
71		}
72	}
73	
74	public static String replaceAll(RE regex, String input, StringCallable replacer) throws InterpreterException {
75		StringBuffer out = new StringBuffer(input);
76		int ofs = 0;
77		while (regex.match(input)) {
78			int  start = regex.getParenStart(0);
79			String matched = regex.getParen(0);
80			int end = regex.getParenEnd(0);
81			String substitute = replacer.call(matched);
82			// advance input string to the end of the current match
83			// e.g. if input = "xxabyab" is matched against "(a*)b", then
84			// input is set to "yab" after the first match
85			input = input.substring(end, input.length());
86			// in the output string, replace the matching part (start - end) by the substitute
87			// i.e. if out = "xxabyab" and every matching string is replaced by * then out
88			// is changed to "xx*yab"
89			out.replace(ofs + start, ofs + end, substitute);
90			// move offset in which to write to out to the end of the newly substituted part
91			// in the example above, ofs is set to 0 + 2 + 1 = 3 such that the next replace
92			// will occur relative to the string "yab"
93			ofs = ofs + start + substitute.length();
94		}
95		return out.toString();
96	}
97	
98}