Examples of org.apache.lucene.facet.taxonomy.directory.ParallelTaxonomyArrays

Package org.apache.lucene.facet.taxonomy.directory

Examples of org.apache.lucene.facet.taxonomy.directory.ParallelTaxonomyArrays

org.apache.lucene.facet.taxonomy.directory.ParallelTaxonomyArrays
Returns 3 arrays for traversing the taxonomy:
- {@code parents}: {@code parents[i]} denotes the parent of categoryordinal {@code i}.
- {@code children}: {@code children[i]} denotes the youngest child ofcategory ordinal {@code i}. The youngest child is defined as the category that was added last to the taxonomy as an immediate child of {@code i}.
- {@code siblings}: {@code siblings[i]} denotes the sibling of categoryordinal {@code i}. The sibling is defined as the previous youngest child of {@code parents[i]}.
To traverse the taxonomy tree, you typically start with {@code children[0]}(ordinal 0 is reserved for ROOT), and then depends if you want to do DFS or BFS, you call {@code children[children[0]]} or {@code siblings[children[0]]}and so forth, respectively.
NOTE: you are not expected to modify the values of the arrays, since the arrays are shared with other threads. @lucene.experimental

    Directory indexDir = newDirectory();
    TaxonomyWriter tw = new DirectoryTaxonomyWriter(indexDir);
    fillTaxonomy(tw);
    tw.close();
    TaxonomyReader tr = new DirectoryTaxonomyReader(indexDir);
    ParallelTaxonomyArrays ca = tr.getParallelTaxonomyArrays();
    int[] youngestChildArray = ca.children();
    assertEquals(tr.getSize(), youngestChildArray.length);
    int[] olderSiblingArray = ca.siblings();
    assertEquals(tr.getSize(), olderSiblingArray.length);
    for (int i=0; i<expectedCategories.length; i++) {
      // find expected children by looking at all expectedCategories
      // for children
      ArrayList<Integer> expectedChildren = new ArrayList<Integer>();

View Full Code Here

    Directory indexDir = newDirectory();
    TaxonomyWriter tw = new DirectoryTaxonomyWriter(indexDir);
    fillTaxonomy(tw);
    tw.close();
    TaxonomyReader tr = new DirectoryTaxonomyReader(indexDir);
    ParallelTaxonomyArrays ca = tr.getParallelTaxonomyArrays();
    int[] children = ca.children();
    assertEquals(tr.getSize(), children.length);
    int[] olderSiblingArray = ca.siblings();
    assertEquals(tr.getSize(), olderSiblingArray.length);
        
    // test that the "youngest child" of every category is indeed a child:
    for (int i=0; i<tr.getSize(); i++) {
      int youngestChild = children[i];

View Full Code Here

    Directory indexDir = newDirectory();
    TaxonomyWriter tw = new DirectoryTaxonomyWriter(indexDir);
    tw.addCategory(new CategoryPath("hi", "there"));
    tw.commit();
    TaxonomyReader tr = new DirectoryTaxonomyReader(indexDir);
    ParallelTaxonomyArrays ca = tr.getParallelTaxonomyArrays();
    assertEquals(3, tr.getSize());
    assertEquals(3, ca.siblings().length);
    assertEquals(3, ca.children().length);
    assertTrue(Arrays.equals(new int[] { 1, 2, -1 }, ca.children()));
    assertTrue(Arrays.equals(new int[] { -1, -1, -1 }, ca.siblings()));
    tw.addCategory(new CategoryPath("hi", "ho"));
    tw.addCategory(new CategoryPath("hello"));
    tw.commit();
    // Before refresh, nothing changed..
    ParallelTaxonomyArrays newca = tr.getParallelTaxonomyArrays();
    assertSame(newca, ca); // we got exactly the same object
    assertEquals(3, tr.getSize());
    assertEquals(3, ca.siblings().length);
    assertEquals(3, ca.children().length);
    // After the refresh, things change:

View Full Code Here

    final CategoryPath abPath = new CategoryPath("a", "b");
    twBase.addCategory(abPath);
    twBase.commit();
    TaxonomyReader trBase = new DirectoryTaxonomyReader(indexDirBase);


    final ParallelTaxonomyArrays ca1 = trBase.getParallelTaxonomyArrays();
    
    final int abOrd = trBase.getOrdinal(abPath);
    final int abYoungChildBase1 = ca1.children()[abOrd]; 
    
    final int numCategories = atLeast(800);
    for (int i = 0; i < numCategories; i++) {
      twBase.addCategory(new CategoryPath("a", "b", Integer.toString(i)));
    }
    twBase.close();
    
    TaxonomyReader newTaxoReader = TaxonomyReader.openIfChanged(trBase);
    assertNotNull(newTaxoReader);
    trBase.close();
    trBase = newTaxoReader;
    
    final ParallelTaxonomyArrays ca2 = trBase.getParallelTaxonomyArrays();
    final int abYoungChildBase2 = ca2.children()[abOrd];
    
    int numRetries = atLeast(50);
    for (int retry = 0; retry < numRetries; retry++) {
      assertConsistentYoungestChild(abPath, abOrd, abYoungChildBase1, abYoungChildBase2, retry, numCategories);
    }

View Full Code Here

    r.close();
    dir.close();
  }


  public static void printStats(TaxonomyReader r, PrintStream out, boolean printTree) throws IOException {
    ParallelTaxonomyArrays arrays = r.getParallelTaxonomyArrays();
    //int[] parents = arrays.parents();
    int[] children = arrays.children();
    int[] siblings = arrays.siblings();
    out.println(r.getSize() + " total categories.");


    int childOrd = children[TaxonomyReader.ROOT_ORDINAL];
    while(childOrd != -1) {
      CategoryPath cp = r.getPath(childOrd);

View Full Code Here

    if (depth > Short.MAX_VALUE - 3) {
      depth = Short.MAX_VALUE -3;
    }


    int endOffset = offset + partitionSize; // one past the largest ordinal in the partition
    ParallelTaxonomyArrays childrenArray = taxonomyReader.getParallelTaxonomyArrays();
    int[] children = childrenArray.children();
    int[] siblings = childrenArray.siblings();
    int totalNumOfDescendantsConsidered = 0; // total number of facets with value != 0, 
    // in the tree. These include those selected as top K in each node, and all the others that
    // were not. Not including rootNode


    // the following priority queue will be used again and again for each node recursed into

View Full Code Here

   */
  private int heapDescendants(int ordinal, Heap<FacetResultNode> pq, FacetResultNode parentResultNode, 
      int offset) throws IOException {
    int partitionSize = facetArrays.arrayLength;
    int endOffset = offset + partitionSize;
    ParallelTaxonomyArrays childrenArray = taxonomyReader.getParallelTaxonomyArrays();
    int[] children = childrenArray.children();
    int[] siblings = childrenArray.siblings();
    FacetResultNode reusable = null;
    int localDepth = 0;
    int depth = facetRequest.getDepth();
    int[] ordinalStack = new int[2+Math.min(Short.MAX_VALUE, depth)];
    int childrenCounter = 0;

View Full Code Here

      for (MatchingDocs md : matchingDocs) {
        aggregator.aggregate(md, clp, facetArrays);
      }
    }
    
    ParallelTaxonomyArrays arrays = taxonomyReader.getParallelTaxonomyArrays();
    
    // compute top-K
    final int[] children = arrays.children();
    final int[] siblings = arrays.siblings();
    List<FacetResult> res = new ArrayList<FacetResult>();
    for (FacetRequest fr : searchParams.facetRequests) {
      int rootOrd = taxonomyReader.getOrdinal(fr.categoryPath);
      if (rootOrd == TaxonomyReader.INVALID_ORDINAL) { // category does not exist
        // Add empty FacetResult:

View Full Code Here

   */
  protected abstract int addSiblings(int ordinal, int[] siblings, PriorityQueue<FacetResultNode> pq);
  
  @Override
  public final FacetResult compute() throws IOException {
    ParallelTaxonomyArrays arrays = taxonomyReader.getParallelTaxonomyArrays();
    final int[] children = arrays.children();
    final int[] siblings = arrays.siblings();
    
    int rootOrd = taxonomyReader.getOrdinal(facetRequest.categoryPath);
        
    FacetResultNode root = new FacetResultNode();
    root.ordinal = rootOrd;

View Full Code Here

TOP

Related Classes of org.apache.lucene.facet.taxonomy.directory.ParallelTaxonomyArrays

org.apache.lucene.facet.search.DepthOneFacetResultsHandler

org.apache.lucene.facet.search.FacetsAccumulator

org.apache.lucene.facet.search.TopKFacetResultsHandler

org.apache.lucene.facet.search.TopKInEachNodeHandler

org.apache.lucene.facet.taxonomy.TestTaxonomyCombined

org.apache.lucene.facet.util.PrintTaxonomyStats

org.apache.lucene.index.DocsAndPositionsEnum

org.apache.lucene.index.CorruptIndexException

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.