enhanced did-you-mean (a bit): can now remember previously searched

words (plus small enhancements)

htroot/suggest.java

@@ -58,7 +58,7 @@ public class suggest {
         final String ext = header.get("EXT", "");
         final boolean json = ext.equals("json");
         final boolean xml = ext.equals("xml");
-        final boolean more = post != null && post.containsKey("more");
+        final boolean more = sb.index.connectedRWI() || (post != null && post.containsKey("more")); // with RWIs connected the guessing is super-fast
 
         // get query
         final String originalquerystring = (post == null) ? "" : post.get("query", post.get("q", "")).trim();

source/net/yacy/cora/document/WordCache.java

@@ -26,6 +26,7 @@ import java.io.FileInputStream;
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.InputStreamReader;
+import java.util.Collection;
 import java.util.ConcurrentModificationException;
 import java.util.HashSet;
 import java.util.Map;
@@ -219,7 +220,7 @@ public class WordCache {
         }
     }
 
-    public static void learn(Set<String> wordset) {
+    public static void learn(Collection<String> wordset) {
         for (String s: wordset) {
             learn(new StringBuilder(s));
         }
@@ -320,8 +321,12 @@ public class WordCache {
         }
         return size;
     }
+
+    public static int sizeCommonWords() {
+        return commonWords.size();
+    }
     
-    public static void clear() {
+    public static void clearCommonWords() {
         commonWords.clear();
     }
 

source/net/yacy/data/DidYouMean.java

@@ -432,7 +432,7 @@ public class DidYouMean {
                 StringBuilder s;
                 try {
                     while ((s = DidYouMean.this.guessLib.take()) != POISON_STRING) {
-                        if (s.length() >= MinimumOutputWordLength && DidYouMean.this.segment.getWordCountGuess(s.toString()) > 0) {
+                        if (s.length() >= MinimumOutputWordLength && DidYouMean.this.segment.getWordCountGuess(s.toString()) > 2) {
                             DidYouMean.this.resultSet.add(s);
                         }
                         if (System.currentTimeMillis() > DidYouMean.this.timeLimit) {

source/net/yacy/search/ResourceObserver.java

@@ -93,7 +93,7 @@ public class ResourceObserver {
                 SearchEventCache.cleanupEvents(true);
                 this.sb.trail.clear();
                 Switchboard.urlBlacklist.clearblacklistCache();
-                WordCache.clear();
+                WordCache.clearCommonWords();
                 Domains.clear();
     		}
     	}

source/net/yacy/search/Switchboard.java

@@ -2045,7 +2045,7 @@ public final class Switchboard extends serverSwitch {
             PDFont.clearResources(); // eats up megabytes, see http://markmail.org/thread/quk5odee4hbsauhu
             
             // clear caches
-            WordCache.clear();
+            if (WordCache.sizeCommonWords() > 1000) WordCache.clearCommonWords();
             Domains.clear();
             
             // clean up image stack

source/net/yacy/search/index/Segment.java

@@ -287,15 +287,16 @@ public class Segment {
      */
     public int getWordCountGuess(String word) {
         if (word == null || word.indexOf(':') >= 0 || word.indexOf(' ') >= 0 || word.indexOf('/') >= 0) return 0;
-        if (this.termIndex == null) {
-            try {
-                return (int) this.fulltext.getDefaultConnector().getQueryCount(CollectionSchema.text_t.getSolrFieldName() + ':' + word);
-            } catch (Throwable e) {
-                Log.logException(e);
-                return 0;
-            }
+        if (this.termIndex != null) {
+            int count = this.termIndex.count(Word.word2hash(word));
+            if (count > 0) return count;
+        }
+        try {
+            return (int) this.fulltext.getDefaultConnector().getQueryCount(CollectionSchema.text_t.getSolrFieldName() + ':' + word);
+        } catch (Throwable e) {
+            Log.logException(e);
+            return 0;
         }
-        return this.termIndex.count(Word.word2hash(word));
     }
 
     public boolean exists(final String urlhash) {

source/net/yacy/search/query/QueryGoal.java

@@ -27,6 +27,7 @@ import java.util.ArrayList;
 import java.util.Map;
 import java.util.SortedSet;
 
+import net.yacy.cora.document.WordCache;
 import net.yacy.cora.federate.solr.Ranking;
 import net.yacy.cora.federate.solr.SchemaDeclaration;
 import net.yacy.cora.federate.solr.SolrType;
@@ -92,6 +93,9 @@ public class QueryGoal {
         for (String s: this.include_strings) parseQuery(s, this.include_words, this.include_words, this.all_words);
         for (String s: this.exclude_strings) parseQuery(s, this.exclude_words, this.exclude_words, this.all_words);
 
+        WordCache.learn(this.include_strings);
+        WordCache.learn(this.exclude_strings);
+        
         this.include_hashes = null;
         this.exclude_hashes = null;
         this.all_hashes = null;