added X handling

NaturalNode · Feb 13, 2012 · dd4c205 · dd4c205
1 parent 85109cf
commit dd4c205
Show file tree

Hide file tree

Showing 2 changed files with 59 additions and 26 deletions.
diff --git a/lib/natural/phonetics/double_metaphone.js b/lib/natural/phonetics/double_metaphone.js
@@ -54,6 +54,30 @@ function process(token, spyCallback) {
     	add(encoded || c);
     }
 
+    function handleD() {
+    	if(token.substring(pos, pos + 2) == 'DG') {
+    		if(['I', 'E', 'Y'].indexOf(token[pos + 2]) > -1)  {
+    			add('J');
+    			pos += 2;
+    		} else {
+    			add('TK');
+    			pos++;
+    		}
+	    } else if(token.substring(pos, pos + 2) == 'DT') {
+    		add('T');
+	    	pos++;    		
+    	} else
+    		addCompressedDouble('D', 'T');
+    }
+
+    function handleH() {
+		// keep if starts a word or is surrounded by vowels
+		if((pos == 0 || isVowel(token[pos - 1])) && isVowel(token[pos + 1])) {
+			add('H');
+			pos++;
+		}    	
+    }    
+
     function handleL() {
     	if(token[pos + 1] == 'L') {
     		if(pos == token.length - 3 && (
@@ -90,14 +114,6 @@ function process(token, spyCallback) {
     	}
     }
 
-    function handleH() {
-		// keep if starts a word or is surrounded by vowels
-		if((pos == 0 || isVowel(token[pos - 1])) && isVowel(token[pos + 1])) {
-			add('H');
-			pos++;
-		}    	
-    }
-
     function handleR() {
     	if(pos == token.length - 1 && !slavoGermanic
     			&& token.substring(pos - 2, pos) == 'IE'
@@ -107,28 +123,22 @@ function process(token, spyCallback) {
 	    	addCompressedDouble('R');    		
     }
 
-    function handleD() {
-    	if(token.substring(pos, pos + 2) == 'DG') {
-    		if(['I', 'E', 'Y'].indexOf(token[pos + 2]) > -1)  {
-    			add('J');
-    			pos += 2;
-    		} else {
-    			add('TK');
-    			pos++;
-    		}
-	    } else if(token.substring(pos, pos + 2) == 'DT') {
-    		add('T');
-	    	pos++;    		
-    	} else
-    		addCompressedDouble('D', 'T');
+    function handleX() {
+    	if(pos == 0) {
+    		add('S');
+    	} else if(!(pos == token.length - 1 
+	    		&& (['IAU', 'EAU', 'IEU'].indexOf(token.substring(pos - 3, pos)) > -1
+	    			|| ['AU', 'OU'].indexOf(token.substring(pos - 2, pos)) > -1))) {
+    		add('KS');
+    	}
     }
 
     function handleZ() {
     	if(token[pos + 1] == 'H') {
     		add('J');
     		pos++;
-    	} else if(['ZO', 'ZI', 'ZA'].indexOf(token.substring(pos + 1, pos + 3)) > -1 || 
-    			(slavoGermanic && pos > 0 && token[pos - 1] != 'T')) {
+    	} else if(['ZO', 'ZI', 'ZA'].indexOf(token.substring(pos + 1, pos + 3)) > -1 
+    			|| (slavoGermanic && pos > 0 && token[pos - 1] != 'T')) {
     		addSecondary('S', 'TS');
     		pos++; 
     	} else
@@ -186,6 +196,9 @@ function process(token, spyCallback) {
 	        case 'V':
 	        	addCompressedDouble('V', 'F');
 	        	break;
+	        case 'X':
+	        	handleX();
+	        	break;
 	        case 'Z':
 	        	handleZ();
 	        	break;

diff --git a/spec/double_metaphone_spec.js b/spec/double_metaphone_spec.js
@@ -249,7 +249,7 @@ describe('double metaphone', function() {
 	    	expect(encodings[1]).toMatch(/^R.*/);
     	});
 
-    	it('should ignore trailing french Rs', function() {
+    	it('should ignore trailing French Rs', function() {
 	    	var encodings = doubleMetaphone.process('papier');
 	    	expect(encodings[0]).toMatch(/.*[^R]$/);
 	    	expect(encodings[1]).toMatch(/.*R$/);    		
@@ -274,14 +274,34 @@ describe('double metaphone', function() {
 	    });
 	});
 
+	describe('X', function() {
+		it('should encode X as S at start', function() {
+	    	var encodings = doubleMetaphone.process('xenophobia');
+	    	expect(encodings[0]).toMatch(/^S.*/);
+	    	expect(encodings[1]).toMatch(/^S.*/);
+		});
+
+		it('should encode X as KS at end for non-French words', function() {
+	    	var encodings = doubleMetaphone.process('box');
+	    	expect(encodings[0]).toMatch(/.*KS$/);
+	    	expect(encodings[1]).toMatch(/.*KS$/);
+		});
+
+		it('should skip X end for French words', function() {
+	    	var encodings = doubleMetaphone.process('lemieux');
+	    	expect(encodings[0]).toNotMatch(/.*KS$/);
+	    	expect(encodings[1]).toNotMatch(/.*KS$/);
+		});
+	});
+
 	describe('Z', function() {
 		it('should encode Z to S', function() {
 	    	var encodings = doubleMetaphone.process('zookeeper');
 	    	expect(encodings[0]).toMatch(/^S.*$/);
 	    	expect(encodings[1]).toMatch(/^S.*$/);
 		});
 
-		it('should encode chinese ZH to J', function() {
+		it('should encode Chinese ZH to J', function() {
 	    	var encodings = doubleMetaphone.process('zheng');
 	    	expect(encodings[0]).toMatch(/^J.*$/);
 	    	expect(encodings[1]).toMatch(/^J.*$/);