- improved "did you mean"

- added &meanCount= to query string
- &meanCount=0 ==> no suggestion, no performance loss
- sorting suggestions by sb.indexSegment.termIndex().count()

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@6059 6c8d7289-2bf4-0310-a012-ef5d649a1542

htroot/yacysearch.java

@@ -469,35 +469,41 @@ public class yacysearch {
             sb.localSearches.add(theQuery);
                         
             // check suggestions
-            DidYouMean didYouMean = new DidYouMean(sb);
+            int meanMax = 0;
-            Iterator<String> meanIt = didYouMean.getSuggestion(querystring).iterator();
+            if (post != null && post.containsKey("meanCount")) {
-            int meanCount = 0;
+            	meanMax = Integer.parseInt(post.get("meanCount"));
-            String suggestion;
+            }
-            prop.put("didYouMean", 0);
+            if(meanMax > 0) {
-            while(meanIt.hasNext()) {
+                DidYouMean didYouMean = new DidYouMean(sb);
-            	suggestion = meanIt.next();
+            	Iterator<String> meanIt = didYouMean.getSuggestion(querystring).iterator();
-            	prop.put("didYouMean_suggestions_"+meanCount+"_word", suggestion);
+                int meanCount = 0;
-            	prop.put("didYouMean_suggestions_"+meanCount+"_url",
+                String suggestion;
-            		"/yacysearch.html" + "?display=" + display +
+                while(meanCount<meanMax && meanIt.hasNext()) {
-	                "&search=" + suggestion +
+                	suggestion = meanIt.next();
-	                "&maximumRecords="+ theQuery.displayResults() +
+                	prop.put("didYouMean_suggestions_"+meanCount+"_word", suggestion);
-	                "&startRecord=" + (0 * theQuery.displayResults()) +
+                	prop.put("didYouMean_suggestions_"+meanCount+"_url",
-	                "&resource=" + ((theQuery.isLocal()) ? "local" : "global") +
+                		"/yacysearch.html" + "?display=" + display +
-	                "&verify=" + ((theQuery.onlineSnippetFetch) ? "true" : "false") +
+    	                "&search=" + suggestion +
-	                "&nav=" + theQuery.navigators +
+    	                "&maximumRecords="+ theQuery.displayResults() +
-	                "&urlmaskfilter=" + originalUrlMask +
+    	                "&startRecord=" + (0 * theQuery.displayResults()) +
-	                "&prefermaskfilter=" + theQuery.prefer +
+    	                "&resource=" + ((theQuery.isLocal()) ? "local" : "global") +
-	                "&cat=href&amp;constraint=" + ((theQuery.constraint == null) ? "" : theQuery.constraint.exportB64()) +
+    	                "&verify=" + ((theQuery.onlineSnippetFetch) ? "true" : "false") +
-	                "&contentdom=" + theQuery.contentdom() +
+    	                "&nav=" + theQuery.navigators +
-	                "&former=" + theQuery.queryString(true)
+    	                "&urlmaskfilter=" + originalUrlMask +
-	             );
+    	                "&prefermaskfilter=" + theQuery.prefer +
-            	prop.put("didYouMean_suggestions_"+meanCount+"_sep","|");
+    	                "&cat=href&amp;constraint=" + ((theQuery.constraint == null) ? "" : theQuery.constraint.exportB64()) +
-            	meanCount++;
+    	                "&contentdom=" + theQuery.contentdom() +
+    	                "&former=" + theQuery.queryString(true)
+    	             );
+                	prop.put("didYouMean_suggestions_"+meanCount+"_sep","|");
+                	meanCount++;
+                }
+                prop.put("didYouMean_suggestions_"+(meanCount-1)+"_sep","");
+                prop.put("didYouMean", 1);
+                prop.put("didYouMean_suggestions", meanCount);
+            } else {
+                prop.put("didYouMean", 0);
             }
-            prop.put("didYouMean_suggestions_"+(meanCount-1)+"_sep","");
-            if(meanCount > 0)
-            	prop.put("didYouMean", 1);
-            prop.put("didYouMean_suggestions", meanCount);
             
             // update the search tracker
             try {

source/de/anomic/tools/DidYouMean.java

@@ -42,15 +42,12 @@ public class DidYouMean {
 		ReversingTwoConsecutiveLetters();
 		
 		final Iterator<String> it = this.set.iterator();
-		// final TreeSet<String> rset = new TreeSet<String>(new wordSizeComparator());
+		final TreeSet<String> rset = new TreeSet<String>(new wordSizeComparator());
-		final TreeSet<String> rset = new TreeSet<String>();
 		String s;
-		int count = 0;
+		while(it.hasNext()) {
-		while(count<10 && it.hasNext()) {
 			s = it.next();			
 			if(sb.indexSegment.termIndex().has(Word.word2hash(s))) {
 				rset.add(s);
-				count++;
 			}			
 		}	
 		rset.remove(word.toLowerCase());
@@ -90,7 +87,7 @@ public class DidYouMean {
 		public int compare(final String o1, final String o2) {
     		final Integer i1 = sb.indexSegment.termIndex().count(Word.word2hash(o1));
     		final Integer i2 = sb.indexSegment.termIndex().count(Word.word2hash(o2));
-    		return i1.compareTo(i2);
+    		return i2.compareTo(i1);
     	}
     	
     }