--- /dev/null
+//\r
+// This software is now distributed according to\r
+// the Lesser Gnu Public License. Please see\r
+// http://www.gnu.org/copyleft/lesser.txt for\r
+// the details.\r
+// -- Happy Computing!\r
+//\r
+package com.stevesoft.pat;\r
+import java.util.*;\r
+/** \r
+ Shareware: package pat\r
+ <a href="copyright.html">Copyright 2001, Steven R. Brandt</a>\r
+*/ /**\r
+The RegexTokenizer is similar to the StringTokenizer class\r
+provided with java, but allows one to tokenize using\r
+regular expressions, rather than a simple list of characters.\r
+Tokens are any strings between the supplied regular expression,\r
+as well as any backreferences (things in parenthesis)\r
+contained within the regular expression. */\r
+public class RegexTokenizer implements Enumeration {\r
+ String toParse;\r
+ Regex r;\r
+ int count = 0;\r
+ Vector v = new Vector();\r
+ Vector vi = new Vector();\r
+ int pos=0;\r
+\r
+ int offset = 1;\r
+ void getMore() {\r
+ String s = r.right();\r
+ if(r.searchFrom(toParse,pos)) {\r
+ v.addElement(r.left().substring(pos));\r
+ vi.addElement(new Integer(r.matchFrom()+\r
+ r.charsMatched()));\r
+ for(int i=0;i<r.numSubs();i++)\r
+ if(r.substring() != null) {\r
+ v.addElement(r.substring(i+offset));\r
+ vi.addElement(\r
+ new Integer(r.matchFrom(i+offset)+\r
+ r.charsMatched(i+offset)));\r
+ }\r
+ pos = r.matchFrom()+r.charsMatched();\r
+ } else if(s != null) v.addElement(s);\r
+ }\r
+\r
+ /** Initialize the tokenizer with a string of text and a pattern */\r
+ public RegexTokenizer(String txt,String ptrn) {\r
+ toParse = txt;\r
+ r = new Regex(ptrn);\r
+ offset = r.BackRefOffset;\r
+ getMore();\r
+ }\r
+ /** Initialize the tokenizer with a Regex object. */\r
+ public RegexTokenizer(String txt,Regex r) {\r
+ toParse = txt;\r
+ this.r = r;\r
+ offset = r.BackRefOffset;\r
+ getMore();\r
+ }\r
+ /** This should always be cast to a String, as in StringTokenizer,\r
+ and as in StringTokenizer one can do this by calling\r
+ nextString(). */\r
+ public Object nextElement() {\r
+ if(count >= v.size()) getMore();\r
+ return v.elementAt(count++);\r
+ }\r
+ /** This is the equivalent (String)nextElement(). */\r
+ public String nextToken() { return (String)nextElement(); }\r
+ /** This asks for the next token, and changes the pattern\r
+ being used at the same time. */\r
+ public String nextToken(String newpat) {\r
+ try { r.compile(newpat); } catch (RegSyntax r_) {}\r
+ return nextToken(r);\r
+ }\r
+ /** This asks for the next token, and changes the pattern\r
+ being used at the same time. */\r
+ public String nextToken(Regex nr) {\r
+ r = nr;\r
+ if(vi.size() > count) {\r
+ pos = ((Integer)vi.elementAt(count)).intValue();\r
+ v.setSize(count);\r
+ vi.setSize(count);\r
+ }\r
+ getMore();\r
+ return nextToken();\r
+ }\r
+ /** Tells whether there are more tokens in the pattern. */\r
+ public boolean hasMoreElements() {\r
+ if(count >= v.size()) getMore();\r
+ return count < v.size();\r
+ }\r
+ /** Tells whether there are more tokens in the pattern, but\r
+ in the fashion of StringTokenizer. */\r
+ public boolean hasMoreTokens() { return hasMoreElements(); }\r
+ /** Determines the # of remaining tokens */\r
+ public int countTokens() {\r
+ int old_pos=pos,_count=count;\r
+ while(hasMoreTokens())\r
+ nextToken();\r
+ count=_count;\r
+ return v.size()-count;\r
+ }\r
+ /** Returns all tokens in the String */\r
+ public String[] allTokens() {\r
+ countTokens();\r
+ String[] ret = new String[v.size()];\r
+ v.copyInto(ret);\r
+ return ret;\r
+ }\r
+};\r