From 4a4ae6bf272fbe4e55f8eb006a9cd0b6691c56b9 Mon Sep 17 00:00:00 2001
From: David Huynh <dfhuynh@gmail.com>
Date: Mon, 8 Mar 2010 19:40:51 +0000
Subject: [PATCH] Fixed toTitlecase to handle parentheses and other delimiters.

git-svn-id: http://google-refine.googlecode.com/svn/trunk@240 7d457c2a-affb-35e4-300a-418c747d4874
---
 .../expr/functions/strings/ToTitlecase.java   | 27 +++++++++++++------
 1 file changed, 19 insertions(+), 8 deletions(-)

diff --git a/src/main/java/com/metaweb/gridworks/expr/functions/strings/ToTitlecase.java b/src/main/java/com/metaweb/gridworks/expr/functions/strings/ToTitlecase.java
index c7b073d0d..9c3954ddf 100644
--- a/src/main/java/com/metaweb/gridworks/expr/functions/strings/ToTitlecase.java
+++ b/src/main/java/com/metaweb/gridworks/expr/functions/strings/ToTitlecase.java
@@ -2,6 +2,7 @@ package com.metaweb.gridworks.expr.functions.strings;
 
 import java.util.Properties;
 
+import org.apache.commons.lang.StringUtils;
 import org.json.JSONException;
 import org.json.JSONWriter;
 
@@ -15,17 +16,27 @@ public class ToTitlecase implements Function {
         if (args.length == 1 && args[0] != null) {
             Object o = args[0];
             String s = o instanceof String ? (String) o : o.toString();
-            String[] words = s.split("\\s+");
+            String[] segments = StringUtils.splitByCharacterType(s);
             
             StringBuffer sb = new StringBuffer();
-            for (int i = 0; i < words.length; i++) {
-                String word = words[i];
-                if (word.length() > 0) {
-                    if (sb.length() > 0) {
-                        sb.append(' ');
-                    }
-                    sb.append(word.substring(0, 1).toUpperCase() + word.substring(1).toLowerCase());
+            boolean startOfWord = true;
+            for (int i = 0; i < segments.length; i++) {
+                String segment = segments[i];
+                char c = segment.charAt(0);
+                
+                if (Character.isWhitespace(c)) {
+                	startOfWord = true;
+                } else if (c == '(' || c == '[' || c == '{' || c == '"' || c == '\'') {
+                	startOfWord = true;
+                } else if (Character.isLetter(c)) {
+                	if (startOfWord) {
+                		segment = StringUtils.capitalize(segment);
+                	}
+                	startOfWord = false;
+                } else {
+                	startOfWord = false;
                 }
+                sb.append(segment);
             }
             
             return sb.toString();