Package org.apache.lucene.index.memory

Source Code of org.apache.lucene.index.memory.MemoryIndex$MemoryIndexReader$MemoryTermsEnum

package org.apache.lucene.index.memory;

/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements.  See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License.  You may obtain a copy of the License at
*
*     http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/

import java.io.IOException;
import java.util.Arrays;
import java.util.Collection;
import java.util.Comparator;
import java.util.HashMap;
import java.util.Iterator;
import java.util.Map;
import java.util.NoSuchElementException;

import org.apache.lucene.analysis.Analyzer;
import org.apache.lucene.analysis.TokenStream;
import org.apache.lucene.analysis.tokenattributes.CharTermAttribute;
import org.apache.lucene.analysis.tokenattributes.OffsetAttribute;
import org.apache.lucene.analysis.tokenattributes.PositionIncrementAttribute;
import org.apache.lucene.analysis.tokenattributes.TermToBytesRefAttribute;
import org.apache.lucene.index.AtomicReader;
import org.apache.lucene.index.AtomicReaderContext;
import org.apache.lucene.index.BinaryDocValues;
import org.apache.lucene.index.DocsAndPositionsEnum;
import org.apache.lucene.index.DocsEnum;
import org.apache.lucene.index.FieldInfo.IndexOptions;
import org.apache.lucene.index.FieldInfo;
import org.apache.lucene.index.FieldInfos;
import org.apache.lucene.index.FieldInvertState;
import org.apache.lucene.index.Fields;
import org.apache.lucene.index.NumericDocValues;
import org.apache.lucene.index.OrdTermState;
import org.apache.lucene.index.SortedDocValues;
import org.apache.lucene.index.SortedSetDocValues;
import org.apache.lucene.index.StoredFieldVisitor;
import org.apache.lucene.index.TermState;
import org.apache.lucene.index.Terms;
import org.apache.lucene.index.TermsEnum;
import org.apache.lucene.search.Collector;
import org.apache.lucene.search.IndexSearcher;
import org.apache.lucene.search.Query;
import org.apache.lucene.search.Scorer;
import org.apache.lucene.search.similarities.Similarity;
import org.apache.lucene.store.RAMDirectory; // for javadocs
import org.apache.lucene.util.ArrayUtil;
import org.apache.lucene.util.Bits;
import org.apache.lucene.util.ByteBlockPool;
import org.apache.lucene.util.BytesRef;
import org.apache.lucene.util.BytesRefHash.DirectBytesStartArray;
import org.apache.lucene.util.BytesRefHash;
import org.apache.lucene.util.Counter;
import org.apache.lucene.util.IntBlockPool.SliceReader;
import org.apache.lucene.util.IntBlockPool.SliceWriter;
import org.apache.lucene.util.IntBlockPool;
import org.apache.lucene.util.RamUsageEstimator;
import org.apache.lucene.util.RecyclingByteBlockAllocator;
import org.apache.lucene.util.RecyclingIntBlockAllocator;


/**
* High-performance single-document main memory Apache Lucene fulltext search index.
*
* <h4>Overview</h4>
*
* This class is a replacement/substitute for a large subset of
* {@link RAMDirectory} functionality. It is designed to
* enable maximum efficiency for on-the-fly matchmaking combining structured and
* fuzzy fulltext search in realtime streaming applications such as Nux XQuery based XML
* message queues, publish-subscribe systems for Blogs/newsfeeds, text chat, data acquisition and
* distribution systems, application level routers, firewalls, classifiers, etc.
* Rather than targeting fulltext search of infrequent queries over huge persistent
* data archives (historic search), this class targets fulltext search of huge
* numbers of queries over comparatively small transient realtime data (prospective
* search).
* For example as in
* <pre class="prettyprint">
* float score = search(String text, Query query)
* </pre>
* <p>
* Each instance can hold at most one Lucene "document", with a document containing
* zero or more "fields", each field having a name and a fulltext value. The
* fulltext value is tokenized (split and transformed) into zero or more index terms
* (aka words) on <code>addField()</code>, according to the policy implemented by an
* Analyzer. For example, Lucene analyzers can split on whitespace, normalize to lower case
* for case insensitivity, ignore common terms with little discriminatory value such as "he", "in", "and" (stop
* words), reduce the terms to their natural linguistic root form such as "fishing"
* being reduced to "fish" (stemming), resolve synonyms/inflexions/thesauri
* (upon indexing and/or querying), etc. For details, see
* <a target="_blank" href="http://today.java.net/pub/a/today/2003/07/30/LuceneIntro.html">Lucene Analyzer Intro</a>.
* <p>
* Arbitrary Lucene queries can be run against this class - see <a target="_blank"
* href="{@docRoot}/../queryparser/org/apache/lucene/queryparser/classic/package-summary.html#package_description">
* Lucene Query Syntax</a>
* as well as <a target="_blank"
* href="http://today.java.net/pub/a/today/2003/11/07/QueryParserRules.html">Query Parser Rules</a>.
* Note that a Lucene query selects on the field names and associated (indexed)
* tokenized terms, not on the original fulltext(s) - the latter are not stored
* but rather thrown away immediately after tokenization.
* <p>
* For some interesting background information on search technology, see Bob Wyman's
* <a target="_blank"
* href="http://bobwyman.pubsub.com/main/2005/05/mary_hodder_poi.html">Prospective Search</a>,
* Jim Gray's
* <a target="_blank" href="http://www.acmqueue.org/modules.php?name=Content&pa=showpage&pid=293&page=4">
* A Call to Arms - Custom subscriptions</a>, and Tim Bray's
* <a target="_blank"
* href="http://www.tbray.org/ongoing/When/200x/2003/07/30/OnSearchTOC">On Search, the Series</a>.
*
*
* <h4>Example Usage</h4>
*
* <pre class="prettyprint">
* Analyzer analyzer = new SimpleAnalyzer(version);
* MemoryIndex index = new MemoryIndex();
* index.addField("content", "Readings about Salmons and other select Alaska fishing Manuals", analyzer);
* index.addField("author", "Tales of James", analyzer);
* QueryParser parser = new QueryParser(version, "content", analyzer);
* float score = index.search(parser.parse("+author:james +salmon~ +fish* manual~"));
* if (score &gt; 0.0f) {
*     System.out.println("it's a match");
* } else {
*     System.out.println("no match found");
* }
* System.out.println("indexData=" + index.toString());
* </pre>
*
*
* <h4>Example XQuery Usage</h4>
*
* <pre class="prettyprint">
* (: An XQuery that finds all books authored by James that have something to do with "salmon fishing manuals", sorted by relevance :)
* declare namespace lucene = "java:nux.xom.pool.FullTextUtil";
* declare variable $query := "+salmon~ +fish* manual~"; (: any arbitrary Lucene query can go here :)
*
* for $book in /books/book[author="James" and lucene:match(abstract, $query) > 0.0]
* let $score := lucene:match($book/abstract, $query)
* order by $score descending
* return $book
* </pre>
*
*
* <h4>No thread safety guarantees</h4>
*
* An instance can be queried multiple times with the same or different queries,
* but an instance is not thread-safe. If desired use idioms such as:
* <pre class="prettyprint">
* MemoryIndex index = ...
* synchronized (index) {
*    // read and/or write index (i.e. add fields and/or query)
* }
* </pre>
*
*
* <h4>Performance Notes</h4>
*
* Internally there's a new data structure geared towards efficient indexing
* and searching, plus the necessary support code to seamlessly plug into the Lucene
* framework.
* <p>
* This class performs very well for very small texts (e.g. 10 chars)
* as well as for large texts (e.g. 10 MB) and everything in between.
* Typically, it is about 10-100 times faster than <code>RAMDirectory</code>.
* Note that <code>RAMDirectory</code> has particularly
* large efficiency overheads for small to medium sized texts, both in time and space.
* Indexing a field with N tokens takes O(N) in the best case, and O(N logN) in the worst
* case. Memory consumption is probably larger than for <code>RAMDirectory</code>.
* <p>
* Example throughput of many simple term queries over a single MemoryIndex:
* ~500000 queries/sec on a MacBook Pro, jdk 1.5.0_06, server VM.
* As always, your mileage may vary.
* <p>
* If you're curious about
* the whereabouts of bottlenecks, run java 1.5 with the non-perturbing '-server
* -agentlib:hprof=cpu=samples,depth=10' flags, then study the trace log and
* correlate its hotspot trailer with its call stack headers (see <a
* target="_blank"
* href="http://java.sun.com/developer/technicalArticles/Programming/HPROF.html">
* hprof tracing </a>).
*
*/
public class MemoryIndex {

  /** info for each field: Map<String fieldName, Info field> */
  private final HashMap<String,Info> fields = new HashMap<String,Info>();
 
  /** fields sorted ascending by fieldName; lazily computed on demand */
  private transient Map.Entry<String,Info>[] sortedFields;
 
  private final boolean storeOffsets;
 
  private static final boolean DEBUG = false;

  private final ByteBlockPool byteBlockPool;
  private final IntBlockPool intBlockPool;
//  private final IntBlockPool.SliceReader postingsReader;
  private final IntBlockPool.SliceWriter postingsWriter;
 
  private HashMap<String,FieldInfo> fieldInfos = new HashMap<String,FieldInfo>();

  private Counter bytesUsed;
 
  /**
   * Sorts term entries into ascending order; also works for
   * Arrays.binarySearch() and Arrays.sort()
   */
  private static final Comparator<Object> termComparator = new Comparator<Object>() {
    @Override
    @SuppressWarnings({"unchecked","rawtypes"})
    public int compare(Object o1, Object o2) {
      if (o1 instanceof Map.Entry<?,?>) o1 = ((Map.Entry<?,?>) o1).getKey();
      if (o2 instanceof Map.Entry<?,?>) o2 = ((Map.Entry<?,?>) o2).getKey();
      if (o1 == o2) return 0;
      return ((Comparable) o1).compareTo((Comparable) o2);
    }
  };

  /**
   * Constructs an empty instance.
   */
  public MemoryIndex() {
    this(false);
  }
 
  /**
   * Constructs an empty instance that can optionally store the start and end
   * character offset of each token term in the text. This can be useful for
   * highlighting of hit locations with the Lucene highlighter package.
   * Protected until the highlighter package matures, so that this can actually
   * be meaningfully integrated.
   *
   * @param storeOffsets
   *            whether or not to store the start and end character offset of
   *            each token term in the text
   */
  public MemoryIndex(boolean storeOffsets) {
    this(storeOffsets, 0);
   
  }
 
  /**
   * Expert: This constructor accepts an upper limit for the number of bytes that should be reused if this instance is {@link #reset()}.
   * @param storeOffsets <code>true</code> if offsets should be stored
   * @param maxReusedBytes the number of bytes that should remain in the internal memory pools after {@link #reset()} is called
   */
  MemoryIndex(boolean storeOffsets, long maxReusedBytes) {
    this.storeOffsets = storeOffsets;
    this.bytesUsed = Counter.newCounter();
    final int maxBufferedByteBlocks = (int)((maxReusedBytes/2) / ByteBlockPool.BYTE_BLOCK_SIZE );
    final int maxBufferedIntBlocks = (int) ((maxReusedBytes - (maxBufferedByteBlocks*ByteBlockPool.BYTE_BLOCK_SIZE))/(IntBlockPool.INT_BLOCK_SIZE * RamUsageEstimator.NUM_BYTES_INT));
    assert (maxBufferedByteBlocks * ByteBlockPool.BYTE_BLOCK_SIZE) + (maxBufferedIntBlocks * IntBlockPool.INT_BLOCK_SIZE * RamUsageEstimator.NUM_BYTES_INT) <= maxReusedBytes;
    byteBlockPool = new ByteBlockPool(new RecyclingByteBlockAllocator(ByteBlockPool.BYTE_BLOCK_SIZE, maxBufferedByteBlocks, bytesUsed));
    intBlockPool = new IntBlockPool(new RecyclingIntBlockAllocator(IntBlockPool.INT_BLOCK_SIZE, maxBufferedIntBlocks, bytesUsed));
    postingsWriter = new SliceWriter(intBlockPool);
  }
 
  /**
   * Convenience method; Tokenizes the given field text and adds the resulting
   * terms to the index; Equivalent to adding an indexed non-keyword Lucene
   * {@link org.apache.lucene.document.Field} that is tokenized, not stored,
   * termVectorStored with positions (or termVectorStored with positions and offsets),
   *
   * @param fieldName
   *            a name to be associated with the text
   * @param text
   *            the text to tokenize and index.
   * @param analyzer
   *            the analyzer to use for tokenization
   */
  public void addField(String fieldName, String text, Analyzer analyzer) {
    if (fieldName == null)
      throw new IllegalArgumentException("fieldName must not be null");
    if (text == null)
      throw new IllegalArgumentException("text must not be null");
    if (analyzer == null)
      throw new IllegalArgumentException("analyzer must not be null");
   
    TokenStream stream;
    try {
      stream = analyzer.tokenStream(fieldName, text);
    } catch (IOException ex) {
      throw new RuntimeException(ex);
    }

    addField(fieldName, stream, 1.0f, analyzer.getPositionIncrementGap(fieldName));
  }
 
  /**
   * Convenience method; Creates and returns a token stream that generates a
   * token for each keyword in the given collection, "as is", without any
   * transforming text analysis. The resulting token stream can be fed into
   * {@link #addField(String, TokenStream)}, perhaps wrapped into another
   * {@link org.apache.lucene.analysis.TokenFilter}, as desired.
   *
   * @param keywords
   *            the keywords to generate tokens for
   * @return the corresponding token stream
   */
  public <T> TokenStream keywordTokenStream(final Collection<T> keywords) {
    // TODO: deprecate & move this method into AnalyzerUtil?
    if (keywords == null)
      throw new IllegalArgumentException("keywords must not be null");
   
    return new TokenStream() {
      private Iterator<T> iter = keywords.iterator();
      private int start = 0;
      private final CharTermAttribute termAtt = addAttribute(CharTermAttribute.class);
      private final OffsetAttribute offsetAtt = addAttribute(OffsetAttribute.class);
     
      @Override
      public boolean incrementToken() {
        if (!iter.hasNext()) return false;
       
        T obj = iter.next();
        if (obj == null)
          throw new IllegalArgumentException("keyword must not be null");
       
        String term = obj.toString();
        clearAttributes();
        termAtt.setEmpty().append(term);
        offsetAtt.setOffset(start, start+termAtt.length());
        start += term.length() + 1; // separate words by 1 (blank) character
        return true;
      }
    };
  }
 
  /**
   * Equivalent to <code>addField(fieldName, stream, 1.0f)</code>.
   *
   * @param fieldName
   *            a name to be associated with the text
   * @param stream
   *            the token stream to retrieve tokens from
   */
  public void addField(String fieldName, TokenStream stream) {
    addField(fieldName, stream, 1.0f);
  }
 
  /**
   * Iterates over the given token stream and adds the resulting terms to the index;
   * Equivalent to adding a tokenized, indexed, termVectorStored, unstored,
   * Lucene {@link org.apache.lucene.document.Field}.
   * Finally closes the token stream. Note that untokenized keywords can be added with this method via
   * {@link #keywordTokenStream(Collection)}, the Lucene <code>KeywordTokenizer</code> or similar utilities.
   *
   * @param fieldName
   *            a name to be associated with the text
   * @param stream
   *            the token stream to retrieve tokens from.
   * @param boost
   *            the boost factor for hits for this field
   * 
   * @see org.apache.lucene.document.Field#setBoost(float)
   */
  public void addField(String fieldName, TokenStream stream, float boost) {
    addField(fieldName, stream, boost, 0);
  }

  /**
   * Iterates over the given token stream and adds the resulting terms to the index;
   * Equivalent to adding a tokenized, indexed, termVectorStored, unstored,
   * Lucene {@link org.apache.lucene.document.Field}.
   * Finally closes the token stream. Note that untokenized keywords can be added with this method via
   * {@link #keywordTokenStream(Collection)}, the Lucene <code>KeywordTokenizer</code> or similar utilities.
   *
   * @param fieldName
   *            a name to be associated with the text
   * @param stream
   *            the token stream to retrieve tokens from.
   * @param boost
   *            the boost factor for hits for this field
   * @param positionIncrementGap
   *            the position increment gap if fields with the same name are added more than once
   *
   * @see org.apache.lucene.document.Field#setBoost(float)
   */
  public void addField(String fieldName, TokenStream stream, float boost, int positionIncrementGap) {
    try {
      if (fieldName == null)
        throw new IllegalArgumentException("fieldName must not be null");
      if (stream == null)
          throw new IllegalArgumentException("token stream must not be null");
      if (boost <= 0.0f)
          throw new IllegalArgumentException("boost factor must be greater than 0.0");
      int numTokens = 0;
      int numOverlapTokens = 0;
      int pos = -1;
      final BytesRefHash terms;
      final SliceByteStartArray sliceArray;
      Info info = null;
      long sumTotalTermFreq = 0;
      if ((info = fields.get(fieldName)) != null) {
        numTokens = info.numTokens;
        numOverlapTokens = info.numOverlapTokens;
        pos = info.lastPosition + positionIncrementGap;
        terms = info.terms;
        boost *= info.boost;
        sliceArray = info.sliceArray;
        sumTotalTermFreq = info.sumTotalTermFreq;
      } else {
        sliceArray = new SliceByteStartArray(BytesRefHash.DEFAULT_CAPACITY);
        terms = new BytesRefHash(byteBlockPool, BytesRefHash.DEFAULT_CAPACITY, sliceArray);
      }

      if (!fieldInfos.containsKey(fieldName)) {
        fieldInfos.put(fieldName,
            new FieldInfo(fieldName, true, fieldInfos.size(), false, false, false, this.storeOffsets ? IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS : IndexOptions.DOCS_AND_FREQS_AND_POSITIONS , null, null, null));
      }
      TermToBytesRefAttribute termAtt = stream.getAttribute(TermToBytesRefAttribute.class);
      PositionIncrementAttribute posIncrAttribute = stream.addAttribute(PositionIncrementAttribute.class);
      OffsetAttribute offsetAtt = stream.addAttribute(OffsetAttribute.class);
      BytesRef ref = termAtt.getBytesRef();
      stream.reset();
     
      while (stream.incrementToken()) {
        termAtt.fillBytesRef();
//        if (DEBUG) System.err.println("token='" + term + "'");
        numTokens++;
        final int posIncr = posIncrAttribute.getPositionIncrement();
        if (posIncr == 0)
          numOverlapTokens++;
        pos += posIncr;
        int ord = terms.add(ref);
        if (ord < 0) {
          ord = (-ord) - 1;
          postingsWriter.reset(sliceArray.end[ord]);
        } else {
          sliceArray.start[ord] = postingsWriter.startNewSlice();
        }
        sliceArray.freq[ord]++;
        sumTotalTermFreq++;
        if (!storeOffsets) {
          postingsWriter.writeInt(pos);
        } else {
          postingsWriter.writeInt(pos);
          postingsWriter.writeInt(offsetAtt.startOffset());
          postingsWriter.writeInt(offsetAtt.endOffset());
        }
        sliceArray.end[ord] = postingsWriter.getCurrentOffset();
      }
      stream.end();

      // ensure infos.numTokens > 0 invariant; needed for correct operation of terms()
      if (numTokens > 0) {
        fields.put(fieldName, new Info(terms, sliceArray, numTokens, numOverlapTokens, boost, pos, sumTotalTermFreq));
        sortedFields = null;    // invalidate sorted view, if any
      }
    } catch (Exception e) { // can never happen
      throw new RuntimeException(e);
    } finally {
      try {
        if (stream != null) {
          stream.close();
        }
      } catch (IOException e2) {
        throw new RuntimeException(e2);
      }
    }
  }

  /**
   * Creates and returns a searcher that can be used to execute arbitrary
   * Lucene queries and to collect the resulting query results as hits.
   *
   * @return a searcher
   */
  public IndexSearcher createSearcher() {
    MemoryIndexReader reader = new MemoryIndexReader();
    IndexSearcher searcher = new IndexSearcher(reader); // ensures no auto-close !!
    reader.setSearcher(searcher); // to later get hold of searcher.getSimilarity()
    return searcher;
  }
 
  /**
   * Convenience method that efficiently returns the relevance score by
   * matching this index against the given Lucene query expression.
   *
   * @param query
   *            an arbitrary Lucene query to run against this index
   * @return the relevance score of the matchmaking; A number in the range
   *         [0.0 .. 1.0], with 0.0 indicating no match. The higher the number
   *         the better the match.
   *
   */
  public float search(Query query) {
    if (query == null)
      throw new IllegalArgumentException("query must not be null");
   
    IndexSearcher searcher = createSearcher();
    try {
      final float[] scores = new float[1]; // inits to 0.0f (no match)
      searcher.search(query, new Collector() {
        private Scorer scorer;

        @Override
        public void collect(int doc) throws IOException {
          scores[0] = scorer.score();
        }

        @Override
        public void setScorer(Scorer scorer) {
          this.scorer = scorer;
        }

        @Override
        public boolean acceptsDocsOutOfOrder() {
          return true;
        }

        @Override
        public void setNextReader(AtomicReaderContext context) { }
      });
      float score = scores[0];
      return score;
    } catch (IOException e) { // can never happen (RAMDirectory)
      throw new RuntimeException(e);
    } finally {
      // searcher.close();
      /*
       * Note that it is harmless and important for good performance to
       * NOT close the index reader!!! This avoids all sorts of
       * unnecessary baggage and locking in the Lucene IndexReader
       * superclass, all of which is completely unnecessary for this main
       * memory index data structure without thread-safety claims.
       *
       * Wishing IndexReader would be an interface...
       *
       * Actually with the new tight createSearcher() API auto-closing is now
       * made impossible, hence searcher.close() would be harmless and also
       * would not degrade performance...
       */
    }  
  }
 
  /**
   * Returns a reasonable approximation of the main memory [bytes] consumed by
   * this instance. Useful for smart memory sensititive caches/pools.
   * @return the main memory consumption
   */
  public long getMemorySize() {
    return RamUsageEstimator.sizeOf(this);
  }

  /** sorts into ascending order (on demand), reusing memory along the way */
  private void sortFields() {
    if (sortedFields == null) sortedFields = sort(fields);
  }
 
  /** returns a view of the given map's entries, sorted ascending by key */
  private static <K,V> Map.Entry<K,V>[] sort(HashMap<K,V> map) {
    int size = map.size();
    @SuppressWarnings("unchecked")
    Map.Entry<K,V>[] entries = new Map.Entry[size];
   
    Iterator<Map.Entry<K,V>> iter = map.entrySet().iterator();
    for (int i=0; i < size; i++) {
      entries[i] = iter.next();
    }
   
    if (size > 1) ArrayUtil.introSort(entries, termComparator);
    return entries;
  }
 
  /**
   * Returns a String representation of the index data for debugging purposes.
   *
   * @return the string representation
   */
  @Override
  public String toString() {
    StringBuilder result = new StringBuilder(256);   
    sortFields();  
    int sumPositions = 0;
    int sumTerms = 0;
    final BytesRef spare = new BytesRef();
    for (int i=0; i < sortedFields.length; i++) {
      Map.Entry<String,Info> entry = sortedFields[i];
      String fieldName = entry.getKey();
      Info info = entry.getValue();
      info.sortTerms();
      result.append(fieldName + ":\n");
      SliceByteStartArray sliceArray = info.sliceArray;
      int numPositions = 0;
      SliceReader postingsReader = new SliceReader(intBlockPool);
      for (int j=0; j < info.terms.size(); j++) {
        int ord = info.sortedTerms[j];
        info.terms.get(ord, spare);
        int freq = sliceArray.freq[ord];
        result.append("\t'" + spare + "':" + freq + ":");
        postingsReader.reset(sliceArray.start[ord], sliceArray.end[ord]);
        result.append(" [");
        final int iters = storeOffsets ? 3 : 1;
        while(!postingsReader.endOfSlice()) {
          result.append("(");
         
          for (int k = 0; k < iters; k++) {
            result.append(postingsReader.readInt());
            if (k < iters-1) {
              result.append(", ");
            }
          }
          result.append(")");
          if (!postingsReader.endOfSlice()) {
            result.append(",");
          }
         
        }
        result.append("]");
        result.append("\n");
        numPositions += freq;
      }
     
      result.append("\tterms=" + info.terms.size());
      result.append(", positions=" + numPositions);
      result.append(", memory=" + RamUsageEstimator.humanReadableUnits(RamUsageEstimator.sizeOf(info)));
      result.append("\n");
      sumPositions += numPositions;
      sumTerms += info.terms.size();
    }
   
    result.append("\nfields=" + sortedFields.length);
    result.append(", terms=" + sumTerms);
    result.append(", positions=" + sumPositions);
    result.append(", memory=" + RamUsageEstimator.humanReadableUnits(getMemorySize()));
    return result.toString();
  }
 
  /**
   * Index data structure for a field; Contains the tokenized term texts and
   * their positions.
   */
  private static final class Info {
   
    /**
     * Term strings and their positions for this field: Map <String
     * termText, ArrayIntList positions>
     */
    private final BytesRefHash terms;
   
    private final SliceByteStartArray sliceArray;
   
    /** Terms sorted ascending by term text; computed on demand */
    private transient int[] sortedTerms;
   
    /** Number of added tokens for this field */
    private final int numTokens;
   
    /** Number of overlapping tokens for this field */
    private final int numOverlapTokens;
   
    /** Boost factor for hits for this field */
    private final float boost;

    private final long sumTotalTermFreq;

    /** the last position encountered in this field for multi field support*/
    private int lastPosition;

    public Info(BytesRefHash terms, SliceByteStartArray sliceArray, int numTokens, int numOverlapTokens, float boost, int lastPosition, long sumTotalTermFreq) {
      this.terms = terms;
      this.sliceArray = sliceArray;
      this.numTokens = numTokens;
      this.numOverlapTokens = numOverlapTokens;
      this.boost = boost;
      this.sumTotalTermFreq = sumTotalTermFreq;
      this.lastPosition = lastPosition;
    }

    public long getSumTotalTermFreq() {
      return sumTotalTermFreq;
    }
   
    /**
     * Sorts hashed terms into ascending order, reusing memory along the
     * way. Note that sorting is lazily delayed until required (often it's
     * not required at all). If a sorted view is required then hashing +
     * sort + binary search is still faster and smaller than TreeMap usage
     * (which would be an alternative and somewhat more elegant approach,
     * apart from more sophisticated Tries / prefix trees).
     */
    public void sortTerms() {
      if (sortedTerms == null)
        sortedTerms = terms.sort(BytesRef.getUTF8SortedAsUnicodeComparator());
    }
       
    public float getBoost() {
      return boost;
    }
  }
 
  ///////////////////////////////////////////////////////////////////////////////
  // Nested classes:
  ///////////////////////////////////////////////////////////////////////////////
   
  /**
   * Search support for Lucene framework integration; implements all methods
   * required by the Lucene IndexReader contracts.
   */
  private final class MemoryIndexReader extends AtomicReader {
   
    private IndexSearcher searcher; // needed to find searcher.getSimilarity()
   
    private MemoryIndexReader() {
      super(); // avoid as much superclass baggage as possible
    }
   
    private Info getInfo(String fieldName) {
      return fields.get(fieldName);
    }
   
    private Info getInfo(int pos) {
      return sortedFields[pos].getValue();
    }

    @Override
    public Bits getLiveDocs() {
      return null;
    }
   
    @Override
    public FieldInfos getFieldInfos() {
      return new FieldInfos(fieldInfos.values().toArray(new FieldInfo[fieldInfos.size()]));
    }

    @Override
    public NumericDocValues getNumericDocValues(String field) {
      return null;
    }

    @Override
    public BinaryDocValues getBinaryDocValues(String field) {
      return null;
    }

    @Override
    public SortedDocValues getSortedDocValues(String field) {
      return null;
    }
   
    @Override
    public SortedSetDocValues getSortedSetDocValues(String field) {
      return null;
    }

    @Override
    public Bits getDocsWithField(String field) throws IOException {
      return null;
    }

    private class MemoryFields extends Fields {
      @Override
      public Iterator<String> iterator() {
        return new Iterator<String>() {
          int upto = -1;

          @Override
          public String next() {
            upto++;
            if (upto >= sortedFields.length) {
              throw new NoSuchElementException();
            }
            return sortedFields[upto].getKey();
          }

          @Override
          public boolean hasNext() {
            return upto+1 < sortedFields.length;
          }

          @Override
          public void remove() {
            throw new UnsupportedOperationException();
          }
        };
      }

      @Override
      public Terms terms(final String field) {
        int i = Arrays.binarySearch(sortedFields, field, termComparator);
        if (i < 0) {
          return null;
        } else {
          final Info info = getInfo(i);
          info.sortTerms();

          return new Terms() {
            @Override
            public TermsEnum iterator(TermsEnum reuse) {
              return new MemoryTermsEnum(info);
            }

            @Override
            public Comparator<BytesRef> getComparator() {
              return BytesRef.getUTF8SortedAsUnicodeComparator();
            }

            @Override
            public long size() {
              return info.terms.size();
            }

            @Override
            public long getSumTotalTermFreq() {
              return info.getSumTotalTermFreq();
            }

            @Override
            public long getSumDocFreq() {
              // each term has df=1
              return info.terms.size();
            }

            @Override
            public int getDocCount() {
              return info.terms.size() > 0 ? 1 : 0;
            }

            @Override
            public boolean hasFreqs() {
              return true;
            }

            @Override
            public boolean hasOffsets() {
              return storeOffsets;
            }

            @Override
            public boolean hasPositions() {
              return true;
            }
           
            @Override
            public boolean hasPayloads() {
              return false;
            }
          };
        }
      }

      @Override
      public int size() {
        return sortedFields.length;
      }
    }
 
    @Override
    public Fields fields() {
      sortFields();
      return new MemoryFields();
    }

    private class MemoryTermsEnum extends TermsEnum {
      private final Info info;
      private final BytesRef br = new BytesRef();
      int termUpto = -1;

      public MemoryTermsEnum(Info info) {
        this.info = info;
        info.sortTerms();
      }
     
      private final int binarySearch(BytesRef b, BytesRef bytesRef, int low,
          int high, BytesRefHash hash, int[] ords, Comparator<BytesRef> comparator) {
        int mid = 0;
        while (low <= high) {
          mid = (low + high) >>> 1;
          hash.get(ords[mid], bytesRef);
          final int cmp = comparator.compare(bytesRef, b);
          if (cmp < 0) {
            low = mid + 1;
          } else if (cmp > 0) {
            high = mid - 1;
          } else {
            return mid;
          }
        }
        assert comparator.compare(bytesRef, b) != 0;
        return -(low + 1);
      }
   

      @Override
      public boolean seekExact(BytesRef text) {
        termUpto = binarySearch(text, br, 0, info.terms.size()-1, info.terms, info.sortedTerms, BytesRef.getUTF8SortedAsUnicodeComparator());
        return termUpto >= 0;
      }

      @Override
      public SeekStatus seekCeil(BytesRef text) {
        termUpto = binarySearch(text, br, 0, info.terms.size()-1, info.terms, info.sortedTerms, BytesRef.getUTF8SortedAsUnicodeComparator());
        if (termUpto < 0) { // not found; choose successor
          termUpto = -termUpto-1;
          if (termUpto >= info.terms.size()) {
            return SeekStatus.END;
          } else {
            info.terms.get(info.sortedTerms[termUpto], br);
            return SeekStatus.NOT_FOUND;
          }
        } else {
          return SeekStatus.FOUND;
        }
      }

      @Override
      public void seekExact(long ord) {
        assert ord < info.terms.size();
        termUpto = (int) ord;
      }
     
      @Override
      public BytesRef next() {
        termUpto++;
        if (termUpto >= info.terms.size()) {
          return null;
        } else {
          info.terms.get(info.sortedTerms[termUpto], br);
          return br;
        }
      }

      @Override
      public BytesRef term() {
        return br;
      }

      @Override
      public long ord() {
        return termUpto;
      }

      @Override
      public int docFreq() {
        return 1;
      }

      @Override
      public long totalTermFreq() {
        return info.sliceArray.freq[info.sortedTerms[termUpto]];
      }

      @Override
      public DocsEnum docs(Bits liveDocs, DocsEnum reuse, int flags) {
        if (reuse == null || !(reuse instanceof MemoryDocsEnum)) {
          reuse = new MemoryDocsEnum();
        }
        return ((MemoryDocsEnum) reuse).reset(liveDocs, info.sliceArray.freq[info.sortedTerms[termUpto]]);
      }

      @Override
      public DocsAndPositionsEnum docsAndPositions(Bits liveDocs, DocsAndPositionsEnum reuse, int flags) {
        if (reuse == null || !(reuse instanceof MemoryDocsAndPositionsEnum)) {
          reuse = new MemoryDocsAndPositionsEnum();
        }
        final int ord = info.sortedTerms[termUpto];
        return ((MemoryDocsAndPositionsEnum) reuse).reset(liveDocs, info.sliceArray.start[ord], info.sliceArray.end[ord], info.sliceArray.freq[ord]);
      }

      @Override
      public Comparator<BytesRef> getComparator() {
        return BytesRef.getUTF8SortedAsUnicodeComparator();
      }

      @Override
      public void seekExact(BytesRef term, TermState state) throws IOException {
        assert state != null;
        this.seekExact(((OrdTermState)state).ord);
      }

      @Override
      public TermState termState() throws IOException {
        OrdTermState ts = new OrdTermState();
        ts.ord = termUpto;
        return ts;
      }
    }
   
    private class MemoryDocsEnum extends DocsEnum {
      private boolean hasNext;
      private Bits liveDocs;
      private int doc = -1;
      private int freq;

      public DocsEnum reset(Bits liveDocs, int freq) {
        this.liveDocs = liveDocs;
        hasNext = true;
        doc = -1;
        this.freq = freq;
        return this;
      }

      @Override
      public int docID() {
        return doc;
      }

      @Override
      public int nextDoc() {
        if (hasNext && (liveDocs == null || liveDocs.get(0))) {
          hasNext = false;
          return doc = 0;
        } else {
          return doc = NO_MORE_DOCS;
        }
      }

      @Override
      public int advance(int target) throws IOException {
        return slowAdvance(target);
      }

      @Override
      public int freq() throws IOException {
        return freq;
      }

      @Override
      public long cost() {
        return 1;
      }
    }
   
    private class MemoryDocsAndPositionsEnum extends DocsAndPositionsEnum {
      private int posUpto; // for assert
      private boolean hasNext;
      private Bits liveDocs;
      private int doc = -1;
      private SliceReader sliceReader;
      private int freq;
      private int startOffset;
      private int endOffset;
     
      public MemoryDocsAndPositionsEnum() {
        this.sliceReader = new SliceReader(intBlockPool);
      }

      public DocsAndPositionsEnum reset(Bits liveDocs, int start, int end, int freq) {
        this.liveDocs = liveDocs;
        this.sliceReader.reset(start, end);
        posUpto = 0; // for assert
        hasNext = true;
        doc = -1;
        this.freq = freq;
        return this;
      }


      @Override
      public int docID() {
        return doc;
      }

      @Override
      public int nextDoc() {
        if (hasNext && (liveDocs == null || liveDocs.get(0))) {
          hasNext = false;
          return doc = 0;
        } else {
          return doc = NO_MORE_DOCS;
        }
      }

      @Override
      public int advance(int target) throws IOException {
        return slowAdvance(target);
      }

      @Override
      public int freq() throws IOException {
        return freq;
      }

      @Override
      public int nextPosition() {
        assert posUpto++ < freq;
        assert !sliceReader.endOfSlice() : " stores offsets : " + startOffset;
        if (storeOffsets) {
          int pos = sliceReader.readInt();
          startOffset = sliceReader.readInt();
          endOffset = sliceReader.readInt();
          return pos;
        } else {
          return sliceReader.readInt();
        }
      }

      @Override
      public int startOffset() {
        return startOffset;
      }

      @Override
      public int endOffset() {
        return endOffset;
      }

      @Override
      public BytesRef getPayload() {
        return null;
      }
     
      @Override
      public long cost() {
        return 1;
      }
    }
   
    @Override
    public Fields getTermVectors(int docID) {
      if (docID == 0) {
        return fields();
      } else {
        return null;
      }
    }

    private Similarity getSimilarity() {
      if (searcher != null) return searcher.getSimilarity();
      return IndexSearcher.getDefaultSimilarity();
    }
   
    private void setSearcher(IndexSearcher searcher) {
      this.searcher = searcher;
    }
 
    @Override
    public int numDocs() {
      if (DEBUG) System.err.println("MemoryIndexReader.numDocs");
      return 1;
    }
 
    @Override
    public int maxDoc() {
      if (DEBUG) System.err.println("MemoryIndexReader.maxDoc");
      return 1;
    }
 
    @Override
    public void document(int docID, StoredFieldVisitor visitor) {
      if (DEBUG) System.err.println("MemoryIndexReader.document");
      // no-op: there are no stored fields
    }
 
    @Override
    protected void doClose() {
      if (DEBUG) System.err.println("MemoryIndexReader.doClose");
    }
   
    /** performance hack: cache norms to avoid repeated expensive calculations */
    private NumericDocValues cachedNormValues;
    private String cachedFieldName;
    private Similarity cachedSimilarity;
   
    @Override
    public NumericDocValues getNormValues(String field) {
      FieldInfo fieldInfo = fieldInfos.get(field);
      if (fieldInfo == null || fieldInfo.omitsNorms())
        return null;
      NumericDocValues norms = cachedNormValues;
      Similarity sim = getSimilarity();
      if (!field.equals(cachedFieldName) || sim != cachedSimilarity) { // not cached?
        Info info = getInfo(field);
        int numTokens = info != null ? info.numTokens : 0;
        int numOverlapTokens = info != null ? info.numOverlapTokens : 0;
        float boost = info != null ? info.getBoost() : 1.0f;
        FieldInvertState invertState = new FieldInvertState(field, 0, numTokens, numOverlapTokens, 0, boost);
        long value = sim.computeNorm(invertState);
        norms = new MemoryIndexNormDocValues(value);
        // cache it for future reuse
        cachedNormValues = norms;
        cachedFieldName = field;
        cachedSimilarity = sim;
        if (DEBUG) System.err.println("MemoryIndexReader.norms: " + field + ":" + value + ":" + numTokens);
      }
      return norms;
    }
  }
 
  /**
   * Resets the {@link MemoryIndex} to its initial state and recycles all internal buffers.
   */
  public void reset() {
    this.fieldInfos.clear();
    this.fields.clear();
    this.sortedFields = null;
    byteBlockPool.reset(false, false); // no need to 0-fill the buffers
    intBlockPool.reset(true, false); // here must must 0-fill since we use slices
  }
 
  private static final class SliceByteStartArray extends DirectBytesStartArray {
    int[] start; // the start offset in the IntBlockPool per term
    int[] end; // the end pointer in the IntBlockPool for the postings slice per term
    int[] freq; // the term frequency
   
    public SliceByteStartArray(int initSize) {
      super(initSize);
    }
   
    @Override
    public int[] init() {
      final int[] ord = super.init();
      start = new int[ArrayUtil.oversize(ord.length, RamUsageEstimator.NUM_BYTES_INT)];
      end = new int[ArrayUtil.oversize(ord.length, RamUsageEstimator.NUM_BYTES_INT)];
      freq = new int[ArrayUtil.oversize(ord.length, RamUsageEstimator.NUM_BYTES_INT)];
      assert start.length >= ord.length;
      assert end.length >= ord.length;
      assert freq.length >= ord.length;
      return ord;
    }

    @Override
    public int[] grow() {
      final int[] ord = super.grow();
      if (start.length < ord.length) {
        start = ArrayUtil.grow(start, ord.length);
        end = ArrayUtil.grow(end, ord.length);
        freq = ArrayUtil.grow(freq, ord.length);
      }     
      assert start.length >= ord.length;
      assert end.length >= ord.length;
      assert freq.length >= ord.length;
      return ord;
    }

    @Override
    public int[] clear() {
     start = end = null;
     return super.clear();
    }
   
  }
}
TOP

Related Classes of org.apache.lucene.index.memory.MemoryIndex$MemoryIndexReader$MemoryTermsEnum

TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.