All Downloads are FREE. Search and download functionalities are using the official Maven repository.

src.it.unimi.dsi.big.mg4j.tool.Merge Maven / Gradle / Ivy

Go to download

MG4J (Managing Gigabytes for Java) is a free full-text search engine for large document collections written in Java. The big version is a fork of the original MG4J that can handle more than 2^31 terms and documents.

The newest version!
package it.unimi.dsi.big.mg4j.tool;

/*		 
 * MG4J: Managing Gigabytes for Java (big)
 *
 * Copyright (C) 2005-2011 Sebastiano Vigna 
 *
 *  This library is free software; you can redistribute it and/or modify it
 *  under the terms of the GNU Lesser General Public License as published by the Free
 *  Software Foundation; either version 3 of the License, or (at your option)
 *  any later version.
 *
 *  This library 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 Lesser General Public License
 *  for more details.
 *
 *  You should have received a copy of the GNU Lesser General Public License
 *  along with this program; if not, see .
 *
 */

import it.unimi.dsi.Util;
import it.unimi.dsi.fastutil.ints.IntBigArrays;
import it.unimi.dsi.fastutil.ints.IntIterator;
import it.unimi.dsi.fastutil.longs.LongHeapSemiIndirectPriorityQueue;
import it.unimi.dsi.io.OutputBitStream;
import it.unimi.dsi.big.mg4j.index.CompressionFlags.Coding;
import it.unimi.dsi.big.mg4j.index.CompressionFlags.Component;
import it.unimi.dsi.big.mg4j.index.Index;
import it.unimi.dsi.big.mg4j.index.IndexIterator;

import java.io.Closeable;
import java.io.IOException;
import java.lang.reflect.InvocationTargetException;
import java.net.URISyntaxException;
import java.util.Map;

import org.apache.commons.configuration.ConfigurationException;
import org.apache.log4j.Logger;

import com.martiansoftware.jsap.JSAPException;

/** Merges several indices.
 * 
 * 

This class merges indices by performing a simple ordered list merge. Documents * appearing in two indices will cause an error. * * @author Sebastiano Vigna * @since 1.0 * */ public class Merge extends Combine { @SuppressWarnings("unused") private static final Logger LOGGER = Util.getLogger( Merge.class ); /** The reference array of the document queue. */ protected long[] doc; /** The queue containing document pointers (for remapped indices). */ protected LongHeapSemiIndirectPriorityQueue documentQueue; /** Merges several indices into one. * * @param outputBasename the basename of the combined index. * @param inputBasename the basenames of the input indices. * @param metadataOnly if true, we save only metadata (term list, frequencies, global counts). * @param bufferSize the buffer size for index readers. * @param writerFlags the flags for the index writer. * @param interleaved forces an interleaved index. * @param skips whether to insert skips in case interleaved is true. * @param quantum the quantum of skipping structures; if negative, a percentage of space for variable-quantum indices (irrelevant if skips is false). * @param height the height of skipping towers (irrelevant if skips is false). * @param skipBufferSize the size of the buffer used to hold temporarily inverted lists during the skipping structure construction. * @param logInterval how often we log. */ public Merge( final String outputBasename, final String[] inputBasename, final boolean metadataOnly, final int bufferSize, final Map writerFlags, final boolean interleaved, final boolean skips, final int quantum, final int height, final int skipBufferSize, final long logInterval ) throws IOException, ConfigurationException, URISyntaxException, ClassNotFoundException, SecurityException, InstantiationException, IllegalAccessException, InvocationTargetException, NoSuchMethodException { super( outputBasename, inputBasename, metadataOnly, false, bufferSize, writerFlags, interleaved, skips, quantum, height, skipBufferSize, logInterval ); doc = new long[ numIndices ]; documentQueue = new LongHeapSemiIndirectPriorityQueue( doc, numIndices ); } protected long combineNumberOfDocuments() { long n = 0; for( int i = 0; i < numIndices; i++ ) n = Math.max( n, index[ i ].numberOfDocuments ); return n; } protected int combineSizes( final OutputBitStream sizesOutputBitStream ) throws IOException { int curSize, s, maxDocSize = 0; if ( needsSizes ) size = IntBigArrays.newBigArray( numberOfDocuments ); final IntIterator[] sizes = new IntIterator[ numIndices ]; for( int i = 0; i < numIndices; i++ ) sizes[ i ] = sizes( i ); for( int d = 0; d < numberOfDocuments; d++ ) { curSize = 0; for( int i = 0; i < numIndices; i++ ) { if ( d < index[ i ].numberOfDocuments && ( s = sizes[ i ].nextInt() ) != 0 ) { if ( curSize != 0 ) throw new IllegalArgumentException( "Document " + d + " has nonzero length in two indices" ); curSize = s; } } if ( needsSizes ) IntBigArrays.set( size, d, curSize ); if ( curSize > maxDocSize ) maxDocSize = curSize; sizesOutputBitStream.writeGamma( curSize ); } for( int i = 0; i < numIndices; i++ ) if ( sizes[ i ] instanceof Closeable ) ((Closeable)sizes[ i ]).close(); return maxDocSize; } protected long combine( final int numUsedIndices ) throws IOException { // We gather the frequencies from the subindices and just add up. At the same time, we load the document queue. long totalFrequency = 0; int currIndex, lastIndex = -1; for( int k = numUsedIndices; k-- != 0; ) { currIndex = usedIndex[ k ]; totalFrequency += ( frequency[ currIndex ] = indexIterator[ currIndex ].frequency() ); if ( ! metadataOnly ) { doc[ currIndex ] = indexIterator[ currIndex ].nextDocument(); documentQueue.enqueue( currIndex ); } } if ( ! metadataOnly ) { if ( p != 0 ) variableQuantumIndexWriter.newInvertedList(totalFrequency, p, predictedSize, predictedLengthNumBits ); else indexWriter.newInvertedList(); indexWriter.writeFrequency( totalFrequency ); long currDoc = -1; int count; OutputBitStream obs; Index i; IndexIterator ir; while( ! documentQueue.isEmpty() ) { // We extract the smallest document pointer, and enqueue it in the new index. if ( currDoc == doc[ currIndex = documentQueue.first() ] ) throw new IllegalStateException( "The indices to be merged contain document " + currDoc + " at least twice (once in index " + inputBasename[ lastIndex ] + " and once in index " + inputBasename[ currIndex ] + ")" ); currDoc = doc[ currIndex ]; obs = indexWriter.newDocumentRecord(); indexWriter.writeDocumentPointer( obs, currDoc ); i = index[ currIndex ]; ir = indexIterator[ currIndex ]; if ( i.hasPayloads ) indexWriter.writePayload( obs, ir.payload() ); if ( i.hasCounts ) { count = ir.count(); if ( hasCounts ) indexWriter.writePositionCount( obs, count ); if ( hasPositions ) indexWriter.writeDocumentPositions( obs, ir.positionArray(), 0, count, size == null ? -1 : IntBigArrays.get( size, currDoc ) ); } // If we just wrote the last document pointer of this term in index j, we dequeue it. if ( --frequency[ currIndex ] == 0 ) documentQueue.dequeue(); else { doc[ currIndex ] = ir.nextDocument(); documentQueue.changed(); } lastIndex = currIndex; } } return totalFrequency; } public static void main( String arg[] ) throws ConfigurationException, SecurityException, JSAPException, IOException, URISyntaxException, ClassNotFoundException, InstantiationException, IllegalAccessException, InvocationTargetException, NoSuchMethodException { Combine.main( arg, Merge.class ); } }





© 2015 - 2025 Weber Informatics LLC | Privacy Policy