/*******************************************************************************
* Copyright (c) 2000, 2003 IBM Corporation and others.
* All rights reserved. This program and the accompanying materials
* are made available under the terms of the Eclipse Public License v1.0
* which accompanies this distribution, and is available at
* https://www.eclipse.org/legal/epl-v10.html
*
* Contributors:
* IBM Corporation - initial API and implementation
* MetaMatrix, Inc - repackaging and updates for use as a metadata store
*******************************************************************************/
package org.teiid.internal.core.index;
import java.io.IOException;
import java.io.RandomAccessFile;
import java.util.ArrayList;
import java.util.HashMap;
import org.teiid.core.index.IDocument;
import org.teiid.core.index.IEntryResult;
import org.teiid.core.index.IQueryResult;
import org.teiid.core.util.LRUCache;
import org.teiid.metadata.VDBResource;
import org.teiid.metadata.index.RuntimeMetadataPlugin;
/**
* This input is used for reading indexes saved using a BlocksIndexOutput.
*/
public class BlocksIndexInput extends IndexInput {
public static final int CACHE_SIZE= 16; // Cache 16 blocks of 8K each, for a cache size of 128K
protected FileListBlock currentFileListBlock;
protected int currentFileListBlockNum;
protected int currentIndexBlockNum;
protected IndexBlock currentIndexBlock;
private RandomAccessFile raf;
private VirtualRandomAccessFile vraf;
protected VDBResource indexFile;
protected LRUCache blockCache;
protected IndexSummary summary;
public BlocksIndexInput(VDBResource inputFile) {
this.indexFile= inputFile;
blockCache= new LRUCache(CACHE_SIZE);
}
/**
* @see IndexInput#clearCache()
*/
public void clearCache() {
blockCache= new LRUCache(CACHE_SIZE);
}
/**
* @see IndexInput#close()
*/
public void close() throws IOException {
if (isOpen()) {
//System.out.println(" Closing Index File: " + this.indexFile.getName());
summary= null;
setOpen(false);
if (raf != null) {
raf.close();
}
if (vraf != null) {
vraf.close();
}
}
}
/**
* @see IndexInput#getCurrentFile()
*/
public IndexedFile getCurrentFile() throws IOException {
if (!hasMoreFiles())
return null;
IndexedFile file= null;
if ((file= currentFileListBlock.getFile(filePosition)) == null) {
currentFileListBlockNum= summary.getBlockNumForFileNum(filePosition);
currentFileListBlock= getFileListBlock(currentFileListBlockNum);
file= currentFileListBlock.getFile(filePosition);
}
return file;
}
/**
* Returns the entry corresponding to the given word.
*/
protected WordEntry getEntry(char[] word) throws IOException {
int blockNum= summary.getBlockNumForWord(word);
if (blockNum == -1) return null;
IndexBlock block= getIndexBlock(blockNum);
return block.findExactEntry(word);
}
/**
* Returns the FileListBlock with the given number.
*/
protected FileListBlock getFileListBlock(int blockNum) throws IOException {
Integer key= new Integer(blockNum);
Block block= (Block) blockCache.get(key);
if (block != null && block instanceof FileListBlock)
return (FileListBlock) block;
FileListBlock fileListBlock= new FileListBlock(IIndexConstants.BLOCK_SIZE);
fileListBlock.read(raf, blockNum);
blockCache.put(key, fileListBlock);
return fileListBlock;
}
/**
* Returns the IndexBlock (containing words) with the given number.
*/
protected IndexBlock getIndexBlock(int blockNum) throws IOException {
Integer key= new Integer(blockNum);
Block block= (Block) blockCache.get(key);
if (block != null && block instanceof IndexBlock)
return (IndexBlock) block;
IndexBlock indexBlock= new GammaCompressedIndexBlock(IIndexConstants.BLOCK_SIZE);
indexBlock.read(raf, blockNum);
blockCache.put(key, indexBlock);
return indexBlock;
}
/**
* @see IndexInput#getIndexedFile(int)
*/
public IndexedFile getIndexedFile(int fileNum) throws IOException {
int blockNum= summary.getBlockNumForFileNum(fileNum);
if (blockNum == -1)
return null;
FileListBlock block= getFileListBlock(blockNum);
return block.getFile(fileNum);
}
/**
* @see IndexInput#getIndexedFile(IDocument)
*/
public IndexedFile getIndexedFile(IDocument document) throws java.io.IOException {
setFirstFile();
String name= document.getName();
while (hasMoreFiles()) {
IndexedFile file= getCurrentFile();
String path= file.getPath();
if (path.equals(name))
return file;
moveToNextFile();
}
return null;
}
/**
* Returns the list of numbers of files containing the given word.
*/
protected int[] getMatchingFileNumbers(char[] word) throws IOException {
int blockNum= summary.getBlockNumForWord(word);
if (blockNum == -1)
return new int[0];
IndexBlock block= getIndexBlock(blockNum);
WordEntry entry= block.findExactEntry(word);
return entry == null ? new int[0] : entry.getRefs();
}
/**
* @see IndexInput#getNumFiles()
*/
public int getNumFiles() {
return summary.getNumFiles();
}
/**
* @see IndexInput#getNumWords()
*/
public int getNumWords() {
return summary.getNumWords();
}
/**
* @see IndexInput#getSource()
*/
public Object getSource() {
return indexFile;
}
/**
* Initialises the blocksIndexInput
*/
protected void init() throws IOException {
clearCache();
setFirstFile();
setFirstWord();
}
/**
* @see IndexInput#moveToNextFile()
*/
public void moveToNextFile() throws IOException {
filePosition++;
}
/**
* @see IndexInput#moveToNextWordEntry()
*/
public void moveToNextWordEntry() throws IOException {
wordPosition++;
if (!hasMoreWords()) {
return;
}
//if end of the current block, we load the next one.
boolean endOfBlock= !currentIndexBlock.nextEntry(currentWordEntry);
if (endOfBlock) {
currentIndexBlock= getIndexBlock(++currentIndexBlockNum);
currentIndexBlock.nextEntry(currentWordEntry);
}
}
/**
* @see IndexInput#open()
*/
public void open() throws IOException {
if (!isOpen()) {
vraf = new VirtualRandomAccessFile(indexFile, "r");
raf= vraf.getSafeRandomAccessFile();
String sig= raf.readUTF();
if (!sig.equals(IIndexConstants.SIGNATURE))
throw new IOException(RuntimeMetadataPlugin.Util.getString("exception.wrongFormat")); //$NON-NLS-1$
int summaryBlockNum= raf.readInt();
raf.seek(summaryBlockNum * (long) IIndexConstants.BLOCK_SIZE);
summary= new IndexSummary();
summary.read(raf);
init();
setOpen(true);
}
}
/**
* @see IndexInput#query(String)
*/
public IQueryResult[] query(String word) throws IOException {
open();
int[] fileNums= getMatchingFileNumbers(word.toCharArray());
int size= fileNums.length;
IQueryResult[] files= new IQueryResult[size];
for (int i= 0; i < size; ++i) {
files[i]= getIndexedFile(fileNums[i]);
}
return files;
}
/**
* Overloaded the method in BlocksIndexInput to allow a user to specify if the
* query should be case sensitive.
* @param pattern
* @param isCaseSensitive
* @return
* @throws IOException
*/
public IEntryResult[] queryEntriesMatching(char[] pattern, boolean isCaseSensitive) throws IOException {
open();
if (pattern == null || pattern.length == 0) return null;
int[] blockNums = null;
int firstStar = indexOf('*', pattern);
switch (firstStar){
case -1 :
WordEntry entry = getEntry(pattern);
if (entry == null) return null;
return new IEntryResult[]{ new EntryResult(entry.getWord(), entry.getRefs()) };
case 0 :
blockNums = summary.getAllBlockNums();
break;
default :
char[] prefix = new char[firstStar];
System.arraycopy(pattern, 0, prefix, 0, firstStar);
blockNums = summary.getBlockNumsForPrefix(prefix);
}
if (blockNums == null || blockNums.length == 0) return null;
IEntryResult[] entries = new IEntryResult[5];
int count = 0;
for (int i = 0, max = blockNums.length; i < max; i++) {
IndexBlock block = getIndexBlock(blockNums[i]);
block.reset();
WordEntry entry = new WordEntry();
while (block.nextEntry(entry)) {
// Switch argument order to fix bug in BlocksIndexInput
if (CharOperation.match(pattern, entry.getWord(), isCaseSensitive)) {
if (count == entries.length){
System.arraycopy(entries, 0, entries = new IEntryResult[count*2], 0, count);
}
entries[count++] = new EntryResult(entry.getWord(), entry.getRefs());
}
}
}
if (count != entries.length){
System.arraycopy(entries, 0, entries = new IEntryResult[count], 0, count);
}
return entries;
}
static final int indexOf(char toBeFound, char[] array) {
for (int i = 0; i < array.length; i++)
if (toBeFound == array[i])
return i;
return -1;
}
public IEntryResult[] queryEntriesPrefixedBy(char[] prefix) throws IOException {
open();
int blockLoc = summary.getFirstBlockLocationForPrefix(prefix);
if (blockLoc < 0) return null;
IEntryResult[] entries = new IEntryResult[5];
int count = 0;
while(blockLoc >= 0){
IndexBlock block = getIndexBlock(summary.getBlockNum(blockLoc));
block.reset();
boolean found = false;
WordEntry entry = new WordEntry();
while (block.nextEntry(entry)) {
if (CharOperation.prefixEquals(prefix, entry.getWord(), true)) {
if (count == entries.length){
System.arraycopy(entries, 0, entries = new IEntryResult[count*2], 0, count);
}
entries[count++] = new EntryResult(entry.getWord(), entry.getRefs());
found = true;
} else {
if (found) break;
}
}
/* consider next block ? */
blockLoc = summary.getNextBlockLocationForPrefix(prefix, blockLoc);
}
if (count == 0) return null;
if (count != entries.length){
System.arraycopy(entries, 0, entries = new IEntryResult[count], 0, count);
}
return entries;
}
/**
* Overloaded the method in BlocksIndexInput to allow a user to specify if the
* query should be case sensitive.
*/
public IEntryResult[] queryEntriesPrefixedBy(char[] prefix, boolean isCaseSensitive) throws IOException {
open();
int blockLoc = summary.getFirstBlockLocationForPrefix(prefix);
if (blockLoc < 0) return null;
IEntryResult[] entries = new IEntryResult[5];
int count = 0;
while(blockLoc >= 0){
IndexBlock block = getIndexBlock(summary.getBlockNum(blockLoc));
block.reset();
boolean found = false;
WordEntry entry = new WordEntry();
while (block.nextEntry(entry)) {
if (CharOperation.prefixEquals(prefix, entry.getWord(), isCaseSensitive)) {
if (count == entries.length){
System.arraycopy(entries, 0, entries = new IEntryResult[count*2], 0, count);
}
entries[count++] = new EntryResult(entry.getWord(), entry.getRefs());
found = true;
} else {
if (found) break;
}
}
/* consider next block ? */
blockLoc = summary.getNextBlockLocationForPrefix(prefix, blockLoc);
}
if (count == 0) return null;
if (count != entries.length){
System.arraycopy(entries, 0, entries = new IEntryResult[count], 0, count);
}
return entries;
}
public IQueryResult[] queryFilesReferringToPrefix(char[] prefix) throws IOException {
open();
int blockLoc = summary.getFirstBlockLocationForPrefix(prefix);
if (blockLoc < 0) return null;
// each filename must be returned already once
HashMap<Integer, IndexedFile> fileMatches = new HashMap<Integer, IndexedFile>();
int count = 0;
while(blockLoc >= 0){
IndexBlock block = getIndexBlock(summary.getBlockNum(blockLoc));
block.reset();
boolean found = false;
WordEntry entry = new WordEntry();
while (block.nextEntry(entry)) {
if (CharOperation.prefixEquals(prefix, entry.getWord(), true)) {
int [] refs = entry.getRefs();
for (int i = 0, max = refs.length; i < max; i++){
int ref = refs[i];
Integer key = Integer.valueOf(ref);
if (!fileMatches.containsKey(key)){
count++;
fileMatches.put(key, getIndexedFile(ref));
}
}
found = true;
} else {
if (found) break;
}
}
/* consider next block ? */
blockLoc = summary.getNextBlockLocationForPrefix(prefix, blockLoc);
}
/* extract indexed files */
IQueryResult[] files = fileMatches.values().toArray(new IQueryResult[count]);
return files;
}
/**
* @see IndexInput#queryInDocumentNames(String)
*/
public IQueryResult[] queryInDocumentNames(String word) throws IOException {
open();
ArrayList matches= new ArrayList();
setFirstFile();
while (hasMoreFiles()) {
IndexedFile file= getCurrentFile();
if (file.getPath().indexOf(word) != -1)
matches.add(file);
moveToNextFile();
}
IQueryResult[] match= new IQueryResult[matches.size()];
matches.toArray(match);
return match;
}
/**
* @see IndexInput#setFirstFile()
*/
protected void setFirstFile() throws IOException {
filePosition= 1;
if (getNumFiles() > 0) {
currentFileListBlockNum= summary.getBlockNumForFileNum(1);
currentFileListBlock= getFileListBlock(currentFileListBlockNum);
}
}
/**
* @see IndexInput#setFirstWord()
*/
protected void setFirstWord() throws IOException {
wordPosition= 1;
if (getNumWords() > 0) {
currentIndexBlockNum= summary.getFirstWordBlockNum();
currentIndexBlock= getIndexBlock(currentIndexBlockNum);
currentWordEntry= new WordEntry();
currentIndexBlock.reset();
currentIndexBlock.nextEntry(currentWordEntry);
}
}
}