From 37efcdb77039f34e2602b0ccbe57278b4b95d62c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?J=C3=A9r=C3=B4me=20Euzenat?= <Jerome.Euzenat@inria.fr>
Date: Sat, 28 Feb 2009 21:32:43 +0000
Subject: [PATCH] - corrected quotation character in WordNet 3.0
---
.../align/impl/method/StringDistances.java | 26 +++++++++++++++----
1 file changed, 21 insertions(+), 5 deletions(-)
diff --git a/src/fr/inrialpes/exmo/align/impl/method/StringDistances.java b/src/fr/inrialpes/exmo/align/impl/method/StringDistances.java
index d5faf26d..cd6cb0da 100644
--- a/src/fr/inrialpes/exmo/align/impl/method/StringDistances.java
+++ b/src/fr/inrialpes/exmo/align/impl/method/StringDistances.java
@@ -327,6 +327,23 @@ public class StringDistances {
return 1.0 - (double)metrics.score( s1, s2 );
}
+ /**
+ * @param s a String
+ * @return s without included quotations between ' or "
+ */
+ public static String stripQuotations( String s ) {
+ int sLength = s.length();
+ String result = "";
+ int sStart = 0;
+ int sEnd = sStart;
+ while ( sStart < sLength ) {
+ while ( sEnd < sLength && s.charAt(sStart) != '\"' ) sEnd++;
+ if ( sEnd < sLength ) result += s.substring(sStart, sEnd);
+ while ( sEnd < sLength && s.charAt(sStart) != '\"' ) sEnd++;
+ sStart = sEnd;
+ }
+ return result;
+ }
/**
* JE//: This is independent from WordNet and should go to StringDistances
@@ -340,8 +357,10 @@ public class StringDistances {
* if it is a suffix
* otherwise the last letter will be taken as the new token
* start
+ *
+ * Would be useful to parameterise with stop words as well
*/
- public static Vector<String> tokenize(String s) {
+ public static Vector<String> tokenize( String s ) {
String str1 = s;
int sLength = s.length();
Vector<String> vTokens = new Vector<String>();
@@ -394,10 +413,7 @@ public class StringDistances {
}
tkStart=tkEnd;
}
- }
-
- // else the standard naming convention will be used
- else{
+ } else { // else the standard naming convention will be used
// start at the beginning of the string
tkStart = 0;
tkEnd = tkStart;
--
GitLab