You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
yacy_search_server/source/de/anomic/kelondro/text/ReferenceContainerArray.java

285 lines
9.9 KiB

// ReferenceContainerArray.java
// (C) 2009 by Michael Peter Christen; mc@yacy.net, Frankfurt a. M., Germany
// first published 04.01.2009 on http://yacy.net
//
// $LastChangedDate$
// $LastChangedRevision$
// $LastChangedBy$
//
// LICENSE
//
// This program is free software; you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation; either version 2 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
package de.anomic.kelondro.text;
import java.io.File;
import java.io.IOException;
import java.util.Date;
import java.util.Iterator;
import java.util.List;
import de.anomic.kelondro.blob.BLOB;
import de.anomic.kelondro.blob.BLOBArray;
import de.anomic.kelondro.index.Row;
import de.anomic.kelondro.index.RowSet;
import de.anomic.kelondro.order.ByteOrder;
import de.anomic.kelondro.order.CloneableIterator;
import de.anomic.kelondro.util.Log;
public final class ReferenceContainerArray {
private final Row payloadrow;
private final BLOBArray array;
private final IODispatcher merger;
/**
* open a index container based on a BLOB dump. The content of the BLOB will not be read
* unless a .idx file exists. Only the .idx file is opened to get a fast read access to
* the BLOB. This class provides no write methods, because BLOB files should not be
* written in random access. To support deletion, a write access to the BLOB for deletion
* is still possible
* @param payloadrow
* @param log
* @throws IOException
*/
public ReferenceContainerArray(
final File heapLocation,
final ByteOrder wordOrder,
final Row payloadrow,
IODispatcher merger) throws IOException {
this.payloadrow = payloadrow;
this.array = new BLOBArray(
heapLocation,
"index",
payloadrow.primaryKeyLength,
wordOrder,
0);
assert merger != null;
this.merger = merger;
}
public synchronized void close() {
this.array.close(true);
}
public synchronized void clear() throws IOException {
this.array.clear();
}
public synchronized int size() {
return (this.array == null) ? 0 : this.array.size();
}
public ByteOrder ordering() {
return this.array.ordering();
}
public File newContainerBLOBFile() {
return this.array.newBLOB(new Date());
}
public void mountBLOBFile(File location) throws IOException {
this.array.mountBLOB(location, false);
}
public Row rowdef() {
return this.payloadrow;
}
/**
* return an iterator object that creates top-level-clones of the indexContainers
* in the cache, so that manipulations of the iterated objects do not change
* objects in the cache.
* @throws IOException
*/
public synchronized CloneableIterator<ReferenceContainer> wordContainerIterator(final String startWordHash, final boolean rot, final boolean ram) {
try {
return new heapCacheIterator(startWordHash, rot);
} catch (IOException e) {
e.printStackTrace();
return null;
}
}
/**
* cache iterator: iterates objects within the heap cache. This can only be used
* for write-enabled heaps, read-only heaps do not have a heap cache
*/
public class heapCacheIterator implements CloneableIterator<ReferenceContainer>, Iterable<ReferenceContainer> {
// this class exists, because the wCache cannot be iterated with rotation
// and because every indexContainer Object that is iterated must be returned as top-level-clone
// so this class simulates wCache.tailMap(startWordHash).values().iterator()
// plus the mentioned features
private final boolean rot;
private CloneableIterator<byte[]> iterator;
public heapCacheIterator(final String startWordHash, final boolean rot) throws IOException {
this.rot = rot;
this.iterator = array.keys(true, startWordHash.getBytes());
// The collection's iterator will return the values in the order that their corresponding keys appear in the tree.
}
public heapCacheIterator clone(final Object secondWordHash) {
try {
return new heapCacheIterator((String) secondWordHash, rot);
} catch (IOException e) {
e.printStackTrace();
return null;
}
}
public boolean hasNext() {
if (this.iterator == null) return false;
if (rot) return true;
return iterator.hasNext();
}
public ReferenceContainer next() {
try {
if (iterator.hasNext()) {
return get(new String(iterator.next()));
}
// rotation iteration
if (!rot) {
return null;
}
iterator = array.keys(true, null);
return get(new String(iterator.next()));
} catch (IOException e) {
e.printStackTrace();
return null;
}
}
public void remove() {
iterator.remove();
}
public Iterator<ReferenceContainer> iterator() {
return this;
}
}
/**
* test if a given key is in the heap
* this works with heaps in write- and read-mode
* @param key
* @return true, if the key is used in the heap; false othervise
* @throws IOException
*/
public synchronized boolean has(final String key) {
return this.array.has(key.getBytes());
}
/**
* get a indexContainer from a heap
* @param key
* @return the indexContainer if one exist, null otherwise
* @throws IOException
*/
public synchronized ReferenceContainer get(final String key) throws IOException {
List<byte[]> entries = this.array.getAll(key.getBytes());
if (entries == null || entries.size() == 0) return null;
byte[] a = entries.remove(0);
ReferenceContainer c = new ReferenceContainer(key, RowSet.importRowSet(a, payloadrow));
while (entries.size() > 0) {
c = c.merge(new ReferenceContainer(key, RowSet.importRowSet(entries.remove(0), payloadrow)));
}
return c;
}
/**
* delete a indexContainer from the heap cache. This can only be used for write-enabled heaps
* @param wordHash
* @return the indexContainer if the cache contained the container, null othervise
* @throws IOException
*/
public synchronized void delete(final String wordHash) throws IOException {
// returns the index that had been deleted
array.remove(wordHash.getBytes());
}
public synchronized int replace(final String wordHash, ContainerRewriter rewriter) throws IOException {
return array.replace(wordHash.getBytes(), new BLOBRewriter(wordHash, rewriter));
}
public class BLOBRewriter implements BLOB.Rewriter {
ContainerRewriter rewriter;
String wordHash;
public BLOBRewriter(String wordHash, ContainerRewriter rewriter) {
this.rewriter = rewriter;
this.wordHash = wordHash;
}
public byte[] rewrite(byte[] b) {
if (b == null) return null;
ReferenceContainer c = rewriter.rewrite(new ReferenceContainer(this.wordHash, RowSet.importRowSet(b, payloadrow)));
if (c == null) return null;
return c.exportCollection();
}
}
public interface ContainerRewriter {
public ReferenceContainer rewrite(ReferenceContainer container);
}
public int entries() {
return this.array.entries();
}
public synchronized boolean shrink(long targetFileSize, long maxFileSize) {
if (this.array.entries() < 2) return false;
boolean donesomething = false;
// first try to merge small files that match
while (this.merger.queueLength() < 3) {
File[] ff = this.array.unmountBestMatch(2.0, targetFileSize);
if (ff == null) break;
Log.logInfo("RICELL-shrink", "unmountBestMatch(2.0, " + targetFileSize + ")");
merger.merge(ff[0], ff[1], this.array, this.payloadrow, newContainerBLOBFile());
donesomething = true;
}
// then try to merge simply any small file
while (this.merger.queueLength() < 2) {
File[] ff = this.array.unmountSmallest(targetFileSize);
if (ff == null) break;
Log.logInfo("RICELL-shrink", "unmountSmallest(" + targetFileSize + ")");
merger.merge(ff[0], ff[1], this.array, this.payloadrow, newContainerBLOBFile());
donesomething = true;
}
// if there is no small file, then merge matching files up to limit
while (this.merger.queueLength() < 1) {
File[] ff = this.array.unmountBestMatch(2.0, maxFileSize);
if (ff == null) break;
Log.logInfo("RICELL-shrink", "unmountBestMatch(2.0, " + maxFileSize + ")");
merger.merge(ff[0], ff[1], this.array, this.payloadrow, newContainerBLOBFile());
donesomething = true;
}
return donesomething;
}
}