- enhanced postprocessing speed and memory footprint (by using HashMaps

instead of TreeMaps)
- enhanced memory footprint of database indexes (by introduction of
optimize calls)
- optimize calls shrink the amount of used memory for index sets if they
are not changed afterwards any more

htroot/HostBrowser.java

@@ -559,7 +559,7 @@ public class HostBrowser {
             if (fetchReferences) {
                 // get the references from the citation index
                 try {
-                    ReferenceReport rr = rrCache.getReferenceReport(ASCII.getBytes(urlhash), false);
+                    ReferenceReport rr = rrCache.getReferenceReport(urlhash, false);
                     List<String> internalIDs = new ArrayList<String>();
                     List<String> externalIDs = new ArrayList<String>();
                     HandleSet iids = rr.getInternallIDs();

htroot/api/webstructure.java

@@ -141,7 +141,7 @@ public class webstructure {
                 prop.put("citations", 1);
                 ReferenceReportCache rrc = sb.index.getReferenceReportCache();
                 ReferenceReport rr = null;
-                try {rr = rrc.getReferenceReport(urlhash, true);} catch (IOException e) {}
+                try {rr = rrc.getReferenceReport(ASCII.String(urlhash), true);} catch (IOException e) {}
             	if (rr != null && rr.getInternalCount() > 0 && rr.getExternalCount() > 0) {
                     prop.put("citations_count", 1);
                     prop.put("citations_documents", 1);

source/net/yacy/cora/federate/solr/SchemaConfiguration.java

@@ -204,7 +204,7 @@ public class SchemaConfiguration extends Configuration implements Serializable {
         Integer exthosts_old = sid == null ? null : (Integer) sid.getFieldValue(CollectionSchema.references_exthosts_i.getSolrFieldName());
         Integer hostextc_old = sid == null ? null : (Integer) sid.getFieldValue(CollectionSchema.host_extent_i.getSolrFieldName());
         try {
-            ReferenceReport rr = rrCache.getReferenceReport(url.hash(), false);
+            ReferenceReport rr = rrCache.getReferenceReport(ASCII.String(url.hash()), false);
             List<String> internalIDs = new ArrayList<String>();
             HandleSet iids = rr.getInternallIDs();
             for (byte[] b: iids) internalIDs.add(ASCII.String(b));

source/net/yacy/cora/storage/HandleMap.java

@@ -32,7 +32,7 @@ public interface HandleMap extends Iterable<Map.Entry<byte[], Long>> {
 
     public long mem();
 
-    public void trim();
+    public void optimize();
 
     /**
      * write a dump of the index to a file. All entries are written in order

source/net/yacy/crawler/Balancer.java

@@ -331,7 +331,7 @@ public class Balancer {
         HostHandles hh = this.domainStacks.get(host);
         if (hh == null) {
             // create new list
-            HandleSet domainList = new RowHandleSet(12, Base64Order.enhancedCoder, 1);
+            HandleSet domainList = new RowHandleSet(Word.commonHashLength, Base64Order.enhancedCoder, 1);
             domainList.put(urlhash);
             this.domainStacks.put(host, new HostHandles(hosthash, domainList));
         } else {

source/net/yacy/crawler/data/NoticedURL.java

@@ -42,6 +42,7 @@ import net.yacy.crawler.Balancer;
 import net.yacy.crawler.CrawlSwitchboard;
 import net.yacy.crawler.retrieval.Request;
 import net.yacy.crawler.robots.RobotsTxt;
+import net.yacy.kelondro.data.word.Word;
 import net.yacy.kelondro.index.RowHandleSet;
 
 public class NoticedURL {
@@ -184,7 +185,7 @@ public class NoticedURL {
      */
     public boolean removeByURLHash(final byte[] urlhashBytes) {
         try {
-            final HandleSet urlHashes = new RowHandleSet(12, Base64Order.enhancedCoder, 1);
+            final HandleSet urlHashes = new RowHandleSet(Word.commonHashLength, Base64Order.enhancedCoder, 1);
             urlHashes.put(urlhashBytes);
             boolean ret = false;
             try {ret |= this.noloadStack.remove(urlHashes) > 0;} catch (final IOException e) {}

source/net/yacy/kelondro/blob/ArrayStack.java

@@ -206,7 +206,7 @@ public class ArrayStack implements BLOB {
                            oneBlob = new Heap(f, keylength, ordering, buffersize);
                        } else {
                            oneBlob = new HeapModifier(f, keylength, ordering);
-                           oneBlob.trim(); // no writings here, can be used with minimum memory
+                           oneBlob.optimize(); // no writings here, can be used with minimum memory
                        }
                        sortedItems.put(Long.valueOf(time), new blobItem(d, f, oneBlob));
                    } catch (final IOException e) {
@@ -236,7 +236,7 @@ public class ArrayStack implements BLOB {
     }
 
     @Override
-    public void trim() {
+    public void optimize() {
         // trim shall not be called for ArrayStacks because the characteristics of an ArrayStack is that the 'topmost' BLOB on the stack
         // is used for write operations and all other shall be trimmed automatically since they are not used for writing. And the
         // topmost BLOB must not be trimmed to support fast writings.
@@ -261,7 +261,7 @@ public class ArrayStack implements BLOB {
             oneBlob = new Heap(location, this.keylength, this.ordering, this.buffersize);
         } else {
             oneBlob = new HeapModifier(location, this.keylength, this.ordering);
-            oneBlob.trim();
+            oneBlob.optimize();
         }
         this.blobs.add(new blobItem(d, location, oneBlob));
     }

source/net/yacy/kelondro/blob/BLOB.java

@@ -63,7 +63,7 @@ public interface BLOB {
      * trim the index of the database: this releases memory not currently used
      * @throws IOException
      */
-    public void trim();
+    public void optimize();
     
     /**
      * calculate the memory in RAM that the BLOB occupies

source/net/yacy/kelondro/blob/Compressor.java

@@ -68,8 +68,8 @@ public class Compressor implements BLOB, Iterable<byte[]> {
     }
 
     @Override
-    public void trim() {
-        this.backend.trim();
+    public void optimize() {
+        this.backend.optimize();
     }
 
     @Override

source/net/yacy/kelondro/blob/HeapReader.java

@@ -131,8 +131,8 @@ public class HeapReader {
         return this.index.mem(); // don't add the memory for free here since then the asserts for memory management don't work
     }
 
-    public void trim() {
-        this.index.trim();
+    public void optimize() {
+        this.index.optimize();
     }
 
     protected byte[] normalizeKey(byte[] key) {

source/net/yacy/kelondro/data/word/Word.java

@@ -151,6 +151,7 @@ public class Word {
                 ConcurrentLog.logException(e);
                 return hashes;
             }
+        hashes.optimize();
         return hashes;
     }
 
@@ -163,6 +164,7 @@ public class Word {
                 ConcurrentLog.logException(e);
                 return hashes;
             }
+        hashes.optimize();
         return hashes;
     }
 }

source/net/yacy/kelondro/index/BufferedObjectIndex.java

@@ -81,6 +81,12 @@ public class BufferedObjectIndex implements Index, Iterable<Row.Entry> {
         }
     }
 
+    @Override
+    public void optimize() {
+        this.backend.optimize();
+        this.buffer.optimize();
+    }
+    
     @Override
     public long mem() {
         return this.backend.mem() + this.buffer.mem();
@@ -356,6 +362,7 @@ public class BufferedObjectIndex implements Index, Iterable<Row.Entry> {
                     break;
                 }
             }
+            handles.optimize();
             return handles;
         }
     }

source/net/yacy/kelondro/index/Cache.java

@@ -101,6 +101,13 @@ public final class Cache implements Index, Iterable<Row.Entry> {
         return this.index.mem() + this.readHitCache.mem() + this.readMissCache.mem();
     }
 
+    @Override
+    public void optimize() {
+        this.index.optimize();
+        this.readHitCache.optimize();
+        this.readMissCache.optimize();
+    }
+
     public final int writeBufferSize() {
         return 0;
     }

source/net/yacy/kelondro/index/Index.java

@@ -38,6 +38,7 @@ public interface Index extends Iterable<Row.Entry> {
 
     public String filename(); // returns a unique identified for this index; can be a real or artificial file name
     public int size();
+    public void optimize();
     public long mem();
     public boolean isEmpty();
     public Row row();

source/net/yacy/kelondro/index/RAMIndex.java

@@ -82,9 +82,10 @@ public final class RAMIndex implements Index, Iterable<Row.Entry> {
 		reset();
 	}
 
-    public void trim() {
-        if (this.index0 != null) this.index0.trim();
-        if (this.index1 != null) this.index1.trim();
+    @Override
+    public void optimize() {
+        if (this.index0 != null) this.index0.optimize();
+        if (this.index1 != null) this.index1.optimize();
     }
 
     public final synchronized void reset() {

source/net/yacy/kelondro/index/RAMIndexCluster.java

@@ -63,8 +63,8 @@ public final class RAMIndexCluster implements Index, Iterable<Row.Entry>, Clonea
         this.rowdef = rowdef;
     }
 
-    public void trim() {
-        for (final RAMIndex i: this.cluster) if (i != null)  i.trim();
+    public void optimize() {
+        for (final RAMIndex i: this.cluster) if (i != null) i.optimize();
     }
 
     @Override

source/net/yacy/kelondro/index/RowCollection.java

@@ -676,6 +676,11 @@ public class RowCollection implements Sortable<Row.Entry>, Iterable<Row.Entry>,
 
     }
 
+    public void optimize() {
+        sort();
+        trim();
+    }
+    
     public final void sort() {
         if (this.sortBound == this.chunkcount) return; // this is sorted
         synchronized (this) {

source/net/yacy/kelondro/index/RowHandleMap.java

@@ -104,12 +104,12 @@ public final class RowHandleMap implements HandleMap, Iterable<Map.Entry<byte[],
         is.close();
         is = null;
         assert this.index.size() == file.length() / (keylength + idxbytes);
-        trim();
+        optimize();
     }
 
     @Override
-    public void trim() {
-        this.index.trim();
+    public void optimize() {
+        this.index.optimize();
     }
 
     public long mem() {

source/net/yacy/kelondro/index/RowHandleSet.java

@@ -81,6 +81,7 @@ public final class RowHandleSet implements HandleSet, Iterable<byte[]>, Cloneabl
 
     @Override
     public RowHandleSet clone() {
+        optimize();
         return new RowHandleSet(this.rowdef, this.index.clone());
     }
 
@@ -92,6 +93,7 @@ public final class RowHandleSet implements HandleSet, Iterable<byte[]>, Cloneabl
     @Override
     public void optimize() {
         this.index.sort();
+        this.index.trim();
     }
     
     /**
@@ -305,6 +307,7 @@ public final class RowHandleSet implements HandleSet, Iterable<byte[]>, Cloneabl
             o = mi.next();
             if (large.has(o)) result.put(o);
         }
+        result.optimize();
         return result;
     }
 
@@ -331,6 +334,7 @@ public final class RowHandleSet implements HandleSet, Iterable<byte[]>, Cloneabl
                 }
             }
         }
+        result.optimize();
         return result;
     }
 

source/net/yacy/kelondro/table/SQLTable.java

@@ -103,6 +103,10 @@ public class SQLTable implements Index, Iterable<Row.Entry> {
         }
 
     }
+    
+    @Override
+    public void optimize() {
+    }
 
     @Override
     public long mem() {

source/net/yacy/kelondro/table/SplitTable.java

@@ -107,6 +107,11 @@ public class SplitTable implements Index, Iterable<Row.Entry> {
         init();
     }
 
+    @Override
+    public void optimize() {
+        for (Index table: tables.values()) table.optimize();
+    }
+    
     @Override
     public long mem() {
         long m = 0;

source/net/yacy/kelondro/table/Table.java

@@ -196,7 +196,7 @@ public class Table implements Index, Iterable<Row.Entry> {
                     this.table = null;
                 }
             }
-            this.index.trim();
+            optimize();
 
             // open the file
             this.file = new BufferedRecords(new Records(tablefile, rowdef.objectsize), this.buffersize);
@@ -270,6 +270,13 @@ public class Table implements Index, Iterable<Row.Entry> {
         } catch (final IOException e) {
             ConcurrentLog.severe("Table", "", e);
         }
+        optimize();
+    }
+
+    @Override
+    public void optimize() {
+        this.index.optimize();
+        if (this.table != null) this.table.optimize();
     }
 
     @Override

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

@@ -30,12 +30,12 @@ import java.io.File;
 import java.io.IOException;
 import java.net.MalformedURLException;
 import java.util.Date;
+import java.util.HashMap;
+import java.util.HashSet;
 import java.util.Iterator;
 import java.util.List;
 import java.util.Map;
 import java.util.Set;
-import java.util.TreeMap;
-import java.util.TreeSet;
 import java.util.concurrent.BlockingQueue;
 import java.util.regex.Pattern;
 
@@ -52,7 +52,6 @@ import net.yacy.cora.federate.solr.connector.SolrConnector;
 import net.yacy.cora.federate.yacy.CacheStrategy;
 import net.yacy.cora.order.Base64Order;
 import net.yacy.cora.order.ByteOrder;
-import net.yacy.cora.order.NaturalOrder;
 import net.yacy.cora.protocol.ClientIdentification;
 import net.yacy.cora.protocol.ResponseHeader;
 import net.yacy.cora.storage.HandleSet;
@@ -215,9 +214,9 @@ public class Segment {
         final byte[] searchhash = url.hash();
         RowHandleSet rootCandidates = getPossibleRootHashes(url);
         
-        Set<byte[]> ignore = new TreeSet<byte[]>(NaturalOrder.naturalOrder); // a set of urlhashes to be ignored. This is generated from all hashes that are seen during recursion to prevent enless loops
-        Set<byte[]> levelhashes = new TreeSet<byte[]>(NaturalOrder.naturalOrder); // all hashes of a clickdepth. The first call contains the target hash only and therefore just one entry
-        levelhashes.add(searchhash);
+        Set<String> ignore = new HashSet<String>(); // a set of urlhashes to be ignored. This is generated from all hashes that are seen during recursion to prevent enless loops
+        Set<String> levelhashes = new HashSet<String>(); // all hashes of a clickdepth. The first call contains the target hash only and therefore just one entry
+        levelhashes.add(ASCII.String(searchhash));
         int leveldepth = 0; // the recursion depth and therefore the result depth-1. Shall be 0 for the first call
         final byte[] hosthash = new byte[6]; // the host of the url to be checked
         System.arraycopy(searchhash, 6, hosthash, 0, 6);
@@ -225,13 +224,13 @@ public class Segment {
         long timeout = System.currentTimeMillis() + maxtime;
         mainloop: for (int maxdepth = 0; maxdepth < 6 && System.currentTimeMillis() < timeout; maxdepth++) {
             
-            Set<byte[]> checknext = new TreeSet<byte[]>(NaturalOrder.naturalOrder);
+            Set<String> checknext = new HashSet<String>();
             
             // loop over all hashes at this clickdepth; the first call to this loop should contain only one hash and a leveldepth = 0
-            checkloop: for (byte[] urlhash: levelhashes) {
+            checkloop: for (String urlhashs: levelhashes) {
     
                 // get all the citations for this url and iterate
-                ReferenceReport rr = rrc.getReferenceReport(urlhash, false);
+                ReferenceReport rr = rrc.getReferenceReport(urlhashs, false);
                 //ReferenceContainer<CitationReference> references = this.urlCitationIndex.get(urlhash, null);
                 if (rr == null || rr.getInternalCount() == 0) continue checkloop; // don't know
                 Iterator<byte[]> i = rr.getInternallIDs().iterator();
@@ -241,17 +240,17 @@ public class Segment {
                     
                     // check if this is from the same host
                     assert (ByteBuffer.equals(u, 6, hosthash, 0, 6));
-                    
+                    String us = ASCII.String(u);
                     // check ignore
-                    if (ignore.contains(u)) continue nextloop;
+                    if (ignore.contains(us)) continue nextloop;
                     
                     // check if the url is a root url
                     if (rootCandidates.has(u)) {
                         return leveldepth + 1;
                     }
                     
-                    checknext.add(u);
-                    ignore.add(u);
+                    checknext.add(us);
+                    ignore.add(us);
                 }
                 if (System.currentTimeMillis() > timeout) break mainloop;
             }
@@ -286,16 +285,16 @@ public class Segment {
     }
     
     public class ReferenceReportCache {
-        Map<byte[], ReferenceReport> cache;
+        private final Map<String, ReferenceReport> cache;
         public ReferenceReportCache() {
-            this.cache = new TreeMap<byte[], ReferenceReport>(Base64Order.enhancedCoder);
+            this.cache = new HashMap<String, ReferenceReport>();
         }
-        public ReferenceReport getReferenceReport(final byte[] id, final boolean acceptSelfReference) throws IOException {
+        public ReferenceReport getReferenceReport(final String id, final boolean acceptSelfReference) throws IOException {
             ReferenceReport rr = cache.get(id);
             if (MemoryControl.shortStatus()) cache.clear();
             if (rr != null) return rr;
             try {
-                rr = new ReferenceReport(id, acceptSelfReference);
+                rr = new ReferenceReport(ASCII.getBytes(id), acceptSelfReference);
                 cache.put(id, rr);
                 return rr;
             } catch (final SpaceExceededException e) {
@@ -311,13 +310,13 @@ public class Segment {
     
     public class ClickdepthCache {
         ReferenceReportCache rrc;
-        Map<byte[], Integer> cache;
+        Map<String, Integer> cache;
         public ClickdepthCache(ReferenceReportCache rrc) {
             this.rrc = rrc;
-            this.cache = new TreeMap<byte[], Integer>(Base64Order.enhancedCoder);
+            this.cache = new HashMap<String, Integer>();
         }
         public int getClickdepth(final DigestURL url, int maxtime) throws IOException {
-            Integer clickdepth = cache.get(url.hash());
+            Integer clickdepth = cache.get(ASCII.String(url.hash()));
             if (MemoryControl.shortStatus()) cache.clear();
             if (clickdepth != null) {
                 //ConcurrentLog.info("Segment", "get clickdepth of url " + url.toNormalform(true) + ": " + clickdepth + " CACHE HIT");
@@ -325,7 +324,7 @@ public class Segment {
             }
             clickdepth = Segment.this.getClickDepth(this.rrc, url, maxtime);
             //ConcurrentLog.info("Segment", "get clickdepth of url " + url.toNormalform(true) + ": " + clickdepth);
-            this.cache.put(url.hash(), clickdepth);
+            this.cache.put(ASCII.String(url.hash()), clickdepth);
             return clickdepth.intValue();
         }
     }
@@ -343,8 +342,8 @@ public class Segment {
             this.internal = 0;
             this.external = 0;
             this.externalHosts = new RowHandleSet(6, Base64Order.enhancedCoder, 0);
-            this.internalIDs = new RowHandleSet(12, Base64Order.enhancedCoder, 0);
-            this.externalIDs = new RowHandleSet(12, Base64Order.enhancedCoder, 0);
+            this.internalIDs = new RowHandleSet(Word.commonHashLength, Base64Order.enhancedCoder, 0);
+            this.externalIDs = new RowHandleSet(Word.commonHashLength, Base64Order.enhancedCoder, 0);
             try {
                 if (connectedCitation()) {
                     // read the references from the citation index
@@ -397,6 +396,9 @@ public class Segment {
                     ConcurrentLog.logException(e);
                 }
             }
+            this.externalHosts.optimize();
+            this.internalIDs.optimize();
+            this.externalIDs.optimize();
         }
         public int getInternalCount() {
             return this.internal;

source/net/yacy/search/schema/CollectionConfiguration.java

@@ -1322,7 +1322,7 @@ public class CollectionConfiguration extends SchemaConfiguration implements Seri
             try {
                 for (Map.Entry<String, double[]> entry: this.crt.entrySet()) {
                     String id = entry.getKey();
-                    ReferenceReport rr = this.rrCache.getReferenceReport(ASCII.getBytes(id), false);
+                    ReferenceReport rr = this.rrCache.getReferenceReport(id, false);
                     // sum up the cr of the internal links
                     HandleSet iids = rr.getInternallIDs();
                     double ncr = 0.0d;

source/net/yacy/search/snippet/MediaSnippet.java

@@ -244,6 +244,7 @@ public class MediaSnippet implements Comparable<MediaSnippet>, Comparator<MediaS
                 }
             }
         }
+        remaininghashes.optimize();
         return remaininghashes;
     }