You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@lucene.apache.org by GitBox <gi...@apache.org> on 2020/09/22 21:36:02 UTC
[GitHub] [lucene-solr] mikemccand commented on a change in pull request #1893: LUCENE-9444 Utility class to get facet labels from taxonomy for a fac…
mikemccand commented on a change in pull request #1893:
URL: https://github.com/apache/lucene-solr/pull/1893#discussion_r493041886
##########
File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/TaxonomyFacetLabels.java
##########
@@ -0,0 +1,184 @@
+/*
+ * 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.
+ */
+package org.apache.lucene.facet.taxonomy;
+
+import org.apache.lucene.facet.FacetsConfig;
+import org.apache.lucene.index.LeafReaderContext;
+import org.apache.lucene.util.IntsRef;
+
+import java.io.IOException;
+
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.INVALID_ORDINAL;
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.ROOT_ORDINAL;
+
+/**
+ * Utility class to easily retrieve previously indexed facet labels, allowing you to skip also adding stored fields for these values,
+ * reducing your index size.
+ *
+ * @lucene.experimental
+ **/
+public class TaxonomyFacetLabels {
+
+ /**
+ * Index field name provided to the constructor
+ */
+ private final String indexFieldName;
+
+ /**
+ * {@code TaxonomyReader} provided to the constructor
+ */
+ private final TaxonomyReader taxoReader;
+
+ /**
+ * {@code FacetsConfig} provided to the constructor
+ */
+ private final FacetsConfig config;
+
+ /**
+ * {@code OrdinalsReader} to decode ordinals previously indexed into the {@code BinaryDocValues} facet field
+ */
+ private final OrdinalsReader ordsReader;
+
+ /**
+ * Sole constructor. Do not close the provided {@link TaxonomyReader} while still using this instance!
+ */
+ public TaxonomyFacetLabels(TaxonomyReader taxoReader, FacetsConfig config, String indexFieldName) throws IOException {
+ this.taxoReader = taxoReader;
+ this.config = config;
+ this.indexFieldName = indexFieldName;
+ this.ordsReader = new DocValuesOrdinalsReader(indexFieldName);
+ }
+
+ /**
+ * Create and return an instance of {@link FacetLabelReader} to retrieve facet labels for
+ * multiple documents and (optionally) for a specific dimension. You must create this per-segment,
+ * and then step through all hits, in order, for that segment.
+ *
+ * <p><b>NOTE</b>: This class is not thread-safe, so you must use a new instance of this
+ * class for each thread.</p>
+ *
+ * @param readerContext LeafReaderContext used to access the {@code BinaryDocValues} facet field
+ * @return an instance of {@link FacetLabelReader}
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabelReader getFacetLabelReader(LeafReaderContext readerContext) throws IOException {
+ return new FacetLabelReader(ordsReader, readerContext);
+ }
+
+ /**
+ * Utility class to retrieve facet labels for multiple documents.
+ *
+ * @lucene.experimental
+ */
+ public class FacetLabelReader {
+ private final OrdinalsReader.OrdinalsSegmentReader ordinalsSegmentReader;
+ private final IntsRef decodedOrds = new IntsRef();
+ private int currentDocId = -1;
+ private int currentPos = -1;
+
+ // Lazily set when nextFacetLabel(int docId, String facetDimension) is first called
+ private int[] parents;
+
+ /**
+ * Sole constructor.
+ */
+ public FacetLabelReader(OrdinalsReader ordsReader, LeafReaderContext readerContext) throws IOException {
+ ordinalsSegmentReader = ordsReader.getReader(readerContext);
+ }
+
+ /**
+ * Retrieves the next {@link FacetLabel} for the specified {@code docId}, or {@code null} if there are no more.
+ * This method has state: if the provided {@code docId} is the same as the previous invocation, it returns the
+ * next {@link FacetLabel} for that document. Otherwise, it advances to the new {@code docId} and provides the
+ * first {@link FacetLabel} for that document, or {@code null} if that document has no indexed facets. Each
+ * new {@code docId} must be in strictly monotonic (increasing) order.
+ *
+ * @param docId input docId provided in monotonic (non-decreasing) order
+ * @return the first or next {@link FacetLabel}, or {@code null} if there are no more
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabel nextFacetLabel(int docId) throws IOException {
+ if (currentDocId != docId) {
+ assert docId > currentDocId: "docs out of order! currentDocId=" + currentDocId + " docId=" + docId;
Review comment:
Hmm let's upgrade this to a real `if` not an `assert` (that requires that assertions are enabled), and change to throw `IllegalArgumentException`? In general if it is a problem that a user could legitimately trip up on, it should be a real `if`, as long as performance impact of such micro-policing is acceptable (which should be the case here).
##########
File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/TaxonomyFacetLabels.java
##########
@@ -0,0 +1,184 @@
+/*
+ * 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.
+ */
+package org.apache.lucene.facet.taxonomy;
+
+import org.apache.lucene.facet.FacetsConfig;
+import org.apache.lucene.index.LeafReaderContext;
+import org.apache.lucene.util.IntsRef;
+
+import java.io.IOException;
+
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.INVALID_ORDINAL;
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.ROOT_ORDINAL;
+
+/**
+ * Utility class to easily retrieve previously indexed facet labels, allowing you to skip also adding stored fields for these values,
+ * reducing your index size.
+ *
+ * @lucene.experimental
+ **/
+public class TaxonomyFacetLabels {
+
+ /**
+ * Index field name provided to the constructor
+ */
+ private final String indexFieldName;
+
+ /**
+ * {@code TaxonomyReader} provided to the constructor
+ */
+ private final TaxonomyReader taxoReader;
+
+ /**
+ * {@code FacetsConfig} provided to the constructor
+ */
+ private final FacetsConfig config;
+
+ /**
+ * {@code OrdinalsReader} to decode ordinals previously indexed into the {@code BinaryDocValues} facet field
+ */
+ private final OrdinalsReader ordsReader;
+
+ /**
+ * Sole constructor. Do not close the provided {@link TaxonomyReader} while still using this instance!
+ */
+ public TaxonomyFacetLabels(TaxonomyReader taxoReader, FacetsConfig config, String indexFieldName) throws IOException {
+ this.taxoReader = taxoReader;
+ this.config = config;
+ this.indexFieldName = indexFieldName;
+ this.ordsReader = new DocValuesOrdinalsReader(indexFieldName);
+ }
+
+ /**
+ * Create and return an instance of {@link FacetLabelReader} to retrieve facet labels for
+ * multiple documents and (optionally) for a specific dimension. You must create this per-segment,
+ * and then step through all hits, in order, for that segment.
+ *
+ * <p><b>NOTE</b>: This class is not thread-safe, so you must use a new instance of this
+ * class for each thread.</p>
+ *
+ * @param readerContext LeafReaderContext used to access the {@code BinaryDocValues} facet field
+ * @return an instance of {@link FacetLabelReader}
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabelReader getFacetLabelReader(LeafReaderContext readerContext) throws IOException {
+ return new FacetLabelReader(ordsReader, readerContext);
+ }
+
+ /**
+ * Utility class to retrieve facet labels for multiple documents.
+ *
+ * @lucene.experimental
+ */
+ public class FacetLabelReader {
+ private final OrdinalsReader.OrdinalsSegmentReader ordinalsSegmentReader;
+ private final IntsRef decodedOrds = new IntsRef();
+ private int currentDocId = -1;
+ private int currentPos = -1;
+
+ // Lazily set when nextFacetLabel(int docId, String facetDimension) is first called
+ private int[] parents;
+
+ /**
+ * Sole constructor.
+ */
+ public FacetLabelReader(OrdinalsReader ordsReader, LeafReaderContext readerContext) throws IOException {
+ ordinalsSegmentReader = ordsReader.getReader(readerContext);
+ }
+
+ /**
+ * Retrieves the next {@link FacetLabel} for the specified {@code docId}, or {@code null} if there are no more.
+ * This method has state: if the provided {@code docId} is the same as the previous invocation, it returns the
+ * next {@link FacetLabel} for that document. Otherwise, it advances to the new {@code docId} and provides the
+ * first {@link FacetLabel} for that document, or {@code null} if that document has no indexed facets. Each
+ * new {@code docId} must be in strictly monotonic (increasing) order.
+ *
+ * @param docId input docId provided in monotonic (non-decreasing) order
+ * @return the first or next {@link FacetLabel}, or {@code null} if there are no more
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabel nextFacetLabel(int docId) throws IOException {
+ if (currentDocId != docId) {
+ assert docId > currentDocId: "docs out of order! currentDocId=" + currentDocId + " docId=" + docId;
+ ordinalsSegmentReader.get(docId, decodedOrds);
+ currentDocId = docId;
+ currentPos = decodedOrds.offset;
+ }
+
+ int endPos = decodedOrds.offset + decodedOrds.length;
+ assert currentPos <= endPos;
Review comment:
This one is a good usage of `assert` because this should only be violated if there is a bug in this code or our default `Codec`, etc. No accidental abuse by users could cause this.
##########
File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/TaxonomyFacetLabels.java
##########
@@ -0,0 +1,184 @@
+/*
+ * 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.
+ */
+package org.apache.lucene.facet.taxonomy;
+
+import org.apache.lucene.facet.FacetsConfig;
+import org.apache.lucene.index.LeafReaderContext;
+import org.apache.lucene.util.IntsRef;
+
+import java.io.IOException;
+
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.INVALID_ORDINAL;
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.ROOT_ORDINAL;
+
+/**
+ * Utility class to easily retrieve previously indexed facet labels, allowing you to skip also adding stored fields for these values,
+ * reducing your index size.
+ *
+ * @lucene.experimental
+ **/
+public class TaxonomyFacetLabels {
+
+ /**
+ * Index field name provided to the constructor
+ */
+ private final String indexFieldName;
+
+ /**
+ * {@code TaxonomyReader} provided to the constructor
+ */
+ private final TaxonomyReader taxoReader;
+
+ /**
+ * {@code FacetsConfig} provided to the constructor
+ */
+ private final FacetsConfig config;
+
+ /**
+ * {@code OrdinalsReader} to decode ordinals previously indexed into the {@code BinaryDocValues} facet field
+ */
+ private final OrdinalsReader ordsReader;
+
+ /**
+ * Sole constructor. Do not close the provided {@link TaxonomyReader} while still using this instance!
+ */
+ public TaxonomyFacetLabels(TaxonomyReader taxoReader, FacetsConfig config, String indexFieldName) throws IOException {
+ this.taxoReader = taxoReader;
+ this.config = config;
+ this.indexFieldName = indexFieldName;
+ this.ordsReader = new DocValuesOrdinalsReader(indexFieldName);
+ }
+
+ /**
+ * Create and return an instance of {@link FacetLabelReader} to retrieve facet labels for
+ * multiple documents and (optionally) for a specific dimension. You must create this per-segment,
+ * and then step through all hits, in order, for that segment.
+ *
+ * <p><b>NOTE</b>: This class is not thread-safe, so you must use a new instance of this
+ * class for each thread.</p>
+ *
+ * @param readerContext LeafReaderContext used to access the {@code BinaryDocValues} facet field
+ * @return an instance of {@link FacetLabelReader}
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabelReader getFacetLabelReader(LeafReaderContext readerContext) throws IOException {
+ return new FacetLabelReader(ordsReader, readerContext);
+ }
+
+ /**
+ * Utility class to retrieve facet labels for multiple documents.
+ *
+ * @lucene.experimental
+ */
+ public class FacetLabelReader {
+ private final OrdinalsReader.OrdinalsSegmentReader ordinalsSegmentReader;
+ private final IntsRef decodedOrds = new IntsRef();
+ private int currentDocId = -1;
+ private int currentPos = -1;
+
+ // Lazily set when nextFacetLabel(int docId, String facetDimension) is first called
+ private int[] parents;
+
+ /**
+ * Sole constructor.
+ */
+ public FacetLabelReader(OrdinalsReader ordsReader, LeafReaderContext readerContext) throws IOException {
+ ordinalsSegmentReader = ordsReader.getReader(readerContext);
+ }
+
+ /**
+ * Retrieves the next {@link FacetLabel} for the specified {@code docId}, or {@code null} if there are no more.
+ * This method has state: if the provided {@code docId} is the same as the previous invocation, it returns the
+ * next {@link FacetLabel} for that document. Otherwise, it advances to the new {@code docId} and provides the
+ * first {@link FacetLabel} for that document, or {@code null} if that document has no indexed facets. Each
+ * new {@code docId} must be in strictly monotonic (increasing) order.
+ *
+ * @param docId input docId provided in monotonic (non-decreasing) order
Review comment:
Can we add a **Note** here, that the returned `FacetLabel` are not necessarily in the same order in which they were indexed?
##########
File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/TaxonomyFacetLabels.java
##########
@@ -0,0 +1,184 @@
+/*
+ * 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.
+ */
+package org.apache.lucene.facet.taxonomy;
+
+import org.apache.lucene.facet.FacetsConfig;
+import org.apache.lucene.index.LeafReaderContext;
+import org.apache.lucene.util.IntsRef;
+
+import java.io.IOException;
+
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.INVALID_ORDINAL;
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.ROOT_ORDINAL;
+
+/**
+ * Utility class to easily retrieve previously indexed facet labels, allowing you to skip also adding stored fields for these values,
+ * reducing your index size.
+ *
+ * @lucene.experimental
+ **/
+public class TaxonomyFacetLabels {
+
+ /**
+ * Index field name provided to the constructor
+ */
+ private final String indexFieldName;
+
+ /**
+ * {@code TaxonomyReader} provided to the constructor
+ */
+ private final TaxonomyReader taxoReader;
+
+ /**
+ * {@code FacetsConfig} provided to the constructor
+ */
+ private final FacetsConfig config;
+
+ /**
+ * {@code OrdinalsReader} to decode ordinals previously indexed into the {@code BinaryDocValues} facet field
+ */
+ private final OrdinalsReader ordsReader;
+
+ /**
+ * Sole constructor. Do not close the provided {@link TaxonomyReader} while still using this instance!
+ */
+ public TaxonomyFacetLabels(TaxonomyReader taxoReader, FacetsConfig config, String indexFieldName) throws IOException {
+ this.taxoReader = taxoReader;
+ this.config = config;
+ this.indexFieldName = indexFieldName;
+ this.ordsReader = new DocValuesOrdinalsReader(indexFieldName);
+ }
+
+ /**
+ * Create and return an instance of {@link FacetLabelReader} to retrieve facet labels for
+ * multiple documents and (optionally) for a specific dimension. You must create this per-segment,
+ * and then step through all hits, in order, for that segment.
+ *
+ * <p><b>NOTE</b>: This class is not thread-safe, so you must use a new instance of this
+ * class for each thread.</p>
+ *
+ * @param readerContext LeafReaderContext used to access the {@code BinaryDocValues} facet field
+ * @return an instance of {@link FacetLabelReader}
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabelReader getFacetLabelReader(LeafReaderContext readerContext) throws IOException {
+ return new FacetLabelReader(ordsReader, readerContext);
+ }
+
+ /**
+ * Utility class to retrieve facet labels for multiple documents.
+ *
+ * @lucene.experimental
+ */
+ public class FacetLabelReader {
+ private final OrdinalsReader.OrdinalsSegmentReader ordinalsSegmentReader;
+ private final IntsRef decodedOrds = new IntsRef();
+ private int currentDocId = -1;
+ private int currentPos = -1;
+
+ // Lazily set when nextFacetLabel(int docId, String facetDimension) is first called
+ private int[] parents;
+
+ /**
+ * Sole constructor.
+ */
+ public FacetLabelReader(OrdinalsReader ordsReader, LeafReaderContext readerContext) throws IOException {
+ ordinalsSegmentReader = ordsReader.getReader(readerContext);
+ }
+
+ /**
+ * Retrieves the next {@link FacetLabel} for the specified {@code docId}, or {@code null} if there are no more.
+ * This method has state: if the provided {@code docId} is the same as the previous invocation, it returns the
+ * next {@link FacetLabel} for that document. Otherwise, it advances to the new {@code docId} and provides the
+ * first {@link FacetLabel} for that document, or {@code null} if that document has no indexed facets. Each
+ * new {@code docId} must be in strictly monotonic (increasing) order.
+ *
+ * @param docId input docId provided in monotonic (non-decreasing) order
+ * @return the first or next {@link FacetLabel}, or {@code null} if there are no more
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabel nextFacetLabel(int docId) throws IOException {
+ if (currentDocId != docId) {
+ assert docId > currentDocId: "docs out of order! currentDocId=" + currentDocId + " docId=" + docId;
+ ordinalsSegmentReader.get(docId, decodedOrds);
+ currentDocId = docId;
+ currentPos = decodedOrds.offset;
+ }
+
+ int endPos = decodedOrds.offset + decodedOrds.length;
+ assert currentPos <= endPos;
+
+ if (currentPos == endPos) {
+ // no more FacetLabels
+ return null;
+ }
+
+ int ord = decodedOrds.ints[currentPos++];
+ return taxoReader.getPath(ord);
+ }
+
+ private boolean isDescendant(int ord, int ancestorOrd) {
+ while (ord != INVALID_ORDINAL && ord != ROOT_ORDINAL) {
+ if (parents[ord] == ancestorOrd) {
+ return true;
+ }
+ ord = parents[ord];
+ }
+ return false;
+ }
+
+ /**
+ * Retrieves the next {@link FacetLabel} for the specified {@code docId} under the requested {@code facetDimension},
+ * or {@code null} if there are no more. This method has state: if the provided {@code docId} is the same as the
+ * previous invocation, it returns the next {@link FacetLabel} for that document. Otherwise, it advances to
+ * the new {@code docId} and provides the first {@link FacetLabel} for that document, or {@code null} if that document
+ * has no indexed facets. Each new {@code docId} must be in strictly monotonic (increasing) order.
+ *
+ * <p><b>NOTE</b>: this method loads the {@code int[] parents} array from the taxonomy index</p>
+ *
+ * @param docId input docId provided in non-decreasing order
+ * @return the first or next {@link FacetLabel}, or {@code null} if there are no more
+ * @throws IOException if {@link TaxonomyReader} has problems getting path for an ordinal
+ */
+ public FacetLabel nextFacetLabel(int docId, String facetDimension) throws IOException {
+ final int parentOrd = taxoReader.getOrdinal(new FacetLabel(facetDimension));
+ assert parentOrd != INVALID_ORDINAL : "Category ordinal not found for facet dimension: " + facetDimension;
Review comment:
Can we upgrade this to a real `if`? Caller might legitimately pass in a non-existing `facetDimension`?
##########
File path: lucene/facet/src/test/org/apache/lucene/facet/taxonomy/TestTaxonomyLabels.java
##########
@@ -0,0 +1,192 @@
+/*
+ * 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.
+ */
+package org.apache.lucene.facet.taxonomy;
+
+import org.apache.lucene.document.Document;
+import org.apache.lucene.facet.FacetField;
+import org.apache.lucene.facet.FacetTestCase;
+import org.apache.lucene.facet.FacetsCollector;
+import org.apache.lucene.facet.FacetsCollector.MatchingDocs;
+import org.apache.lucene.facet.FacetsConfig;
+import org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader;
+import org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyWriter;
+import org.apache.lucene.index.IndexWriterConfig;
+import org.apache.lucene.index.RandomIndexWriter;
+import org.apache.lucene.search.DocIdSetIterator;
+import org.apache.lucene.search.IndexSearcher;
+import org.apache.lucene.search.MatchAllDocsQuery;
+import org.apache.lucene.store.Directory;
+import org.apache.lucene.util.IOUtils;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.List;
+import java.util.Set;
+import java.util.stream.Collectors;
+
+public class TestTaxonomyLabels extends FacetTestCase {
+
+ private List<Document> prepareDocuments() {
+ List<Document> docs = new ArrayList<>();
+
+ Document doc = new Document();
+ doc.add(new FacetField("Author", "Bob"));
+ doc.add(new FacetField("Publish Date", "2010", "10", "15"));
+ docs.add(doc);
+
+ doc = new Document();
+ doc.add(new FacetField("Author", "Lisa"));
+ doc.add(new FacetField("Publish Date", "2010", "10", "20"));
+ docs.add(doc);
+
+ doc = new Document();
+ doc.add(new FacetField("Author", "Tom"));
+ doc.add(new FacetField("Publish Date", "2012", "1", "1"));
+ docs.add(doc);
+
+ doc = new Document();
+ doc.add(new FacetField("Author", "Susan"));
+ doc.add(new FacetField("Publish Date", "2012", "1", "7"));
+ docs.add(doc);
+
+ doc = new Document();
+ doc.add(new FacetField("Author", "Frank"));
+ doc.add(new FacetField("Publish Date", "1999", "5", "5"));
+ docs.add(doc);
+
+ return docs;
+ }
+
+ private List<Integer> allDocIds(MatchingDocs m, boolean decreasingDocIds) throws IOException {
+ DocIdSetIterator disi = m.bits.iterator();
+ List<Integer> docIds = new ArrayList<>();
+ while (disi.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
+ docIds.add(disi.docID());
+ }
+
+ if (decreasingDocIds == true) {
+ Collections.reverse(docIds);
+ }
+ return docIds;
+ }
+
+ private List<FacetLabel> lookupFacetLabels(TaxonomyFacetLabels taxoLabels,
+ List<MatchingDocs> matchingDocs) throws IOException {
+ return lookupFacetLabels(taxoLabels, matchingDocs, null, false);
+ }
+
+ private List<FacetLabel> lookupFacetLabels(TaxonomyFacetLabels taxoLabels,
+ List<MatchingDocs> matchingDocs,
+ String dimension) throws IOException {
+ return lookupFacetLabels(taxoLabels, matchingDocs, dimension, false);
+ }
+
+ private List<FacetLabel> lookupFacetLabels(TaxonomyFacetLabels taxoLabels, List<MatchingDocs> matchingDocs, String dimension,
+ boolean decreasingDocIds) throws IOException {
+ List<FacetLabel> facetLabels = new ArrayList<>();
+
+ for (MatchingDocs m : matchingDocs) {
+ TaxonomyFacetLabels.FacetLabelReader facetLabelReader = taxoLabels.getFacetLabelReader(m.context);
+ List<Integer> docIds = allDocIds(m, decreasingDocIds);
+ FacetLabel facetLabel;
+ for (Integer docId : docIds) {
+ while (true) {
+ if (dimension != null) {
+ facetLabel = facetLabelReader.nextFacetLabel(docId, dimension);
+ } else {
+ facetLabel = facetLabelReader.nextFacetLabel(docId);
+ }
+
+ if (facetLabel == null) {
+ break;
+ }
+ facetLabels.add(facetLabel);
+ }
+ }
+ }
+
+ return facetLabels;
+ }
+
+
+ public void testBasic() throws Exception {
+ Directory dir = newDirectory();
+ Directory taxoDir = newDirectory();
+
+ // Writes facet ords to a separate directory from the main index:
+ DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir, IndexWriterConfig.OpenMode.CREATE);
+ RandomIndexWriter writer = new RandomIndexWriter(random(), dir);
+
+ FacetsConfig config = new FacetsConfig();
+ config.setHierarchical("Publish Date", true);
+
+ for (Document doc : prepareDocuments()) {
+ writer.addDocument(config.build(taxoWriter, doc));
+ }
+
+ // NRT open
+ IndexSearcher searcher = newSearcher(writer.getReader());
+ // NRT open
+ TaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoWriter);
+
+ FacetsCollector fc = new FacetsCollector();
+ searcher.search(new MatchAllDocsQuery(), fc);
+
+ TaxonomyFacetLabels taxoLabels = new TaxonomyFacetLabels(taxoReader, config, FacetsConfig.DEFAULT_INDEX_FIELD_NAME);
+
+ // Check labels for all dimensions
+ List<FacetLabel> facetLabels = lookupFacetLabels(taxoLabels, fc.getMatchingDocs());
+ assertTrue(facetLabels.size() == 10);
+ assertTrue(facetLabels.stream()
+ .filter(l -> "Author".equals(l.components[0]))
+ .map(l -> l.components[1]).collect(Collectors.toSet())
+ .equals(Set.of("Bob", "Lisa", "Susan", "Frank", "Tom")));
+
+ assertTrue(facetLabels.stream()
+ .filter(l -> "Publish Date".equals(l.components[0]))
+ .map(l -> String.join("/", l.components[1], l.components[2], l.components[3]))
+ .collect(Collectors.toSet())
+ .equals(Set.of("2010/10/15", "2010/10/20", "2012/1/1", "2012/1/7", "1999/5/5")));
+
+ // Check labels for specific dimension
+ facetLabels = lookupFacetLabels(taxoLabels, fc.getMatchingDocs(), "Publish Date");
+ assertTrue(facetLabels.size() == 5);
+ assertTrue(facetLabels.stream()
+ .filter(l -> "Publish Date".equals(l.components[0]))
+ .map(l -> String.join("/", l.components[1], l.components[2], l.components[3]))
+ .collect(Collectors.toSet())
+ .equals(Set.of("2010/10/15", "2010/10/20", "2012/1/1", "2012/1/7", "1999/5/5")));
+
+ try {
+ facetLabels = lookupFacetLabels(taxoLabels, fc.getMatchingDocs(), null, true);
+ fail("Assertion error was not thrown for using docIds supplied in decreasing order");
+ } catch (AssertionError ae) {
Review comment:
Ahh, be careful here! Lucene's `test-framework` sometimes randomly runs without assertions enabled, to help us catch bugs where we accidentally create code relying on assertions (it has happened, do not ask who).
So when tests run with assertions disabled then this test would false-fail? But, if we upgrade to real `if` that would also fix this.
##########
File path: lucene/facet/src/test/org/apache/lucene/facet/taxonomy/TestTaxonomyLabels.java
##########
@@ -0,0 +1,192 @@
+/*
+ * 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.
+ */
+package org.apache.lucene.facet.taxonomy;
+
+import org.apache.lucene.document.Document;
+import org.apache.lucene.facet.FacetField;
+import org.apache.lucene.facet.FacetTestCase;
+import org.apache.lucene.facet.FacetsCollector;
+import org.apache.lucene.facet.FacetsCollector.MatchingDocs;
+import org.apache.lucene.facet.FacetsConfig;
+import org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader;
+import org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyWriter;
+import org.apache.lucene.index.IndexWriterConfig;
+import org.apache.lucene.index.RandomIndexWriter;
+import org.apache.lucene.search.DocIdSetIterator;
+import org.apache.lucene.search.IndexSearcher;
+import org.apache.lucene.search.MatchAllDocsQuery;
+import org.apache.lucene.store.Directory;
+import org.apache.lucene.util.IOUtils;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.List;
+import java.util.Set;
+import java.util.stream.Collectors;
+
+public class TestTaxonomyLabels extends FacetTestCase {
+
+ private List<Document> prepareDocuments() {
+ List<Document> docs = new ArrayList<>();
+
+ Document doc = new Document();
+ doc.add(new FacetField("Author", "Bob"));
+ doc.add(new FacetField("Publish Date", "2010", "10", "15"));
+ docs.add(doc);
+
+ doc = new Document();
+ doc.add(new FacetField("Author", "Lisa"));
+ doc.add(new FacetField("Publish Date", "2010", "10", "20"));
+ docs.add(doc);
+
+ doc = new Document();
+ doc.add(new FacetField("Author", "Tom"));
+ doc.add(new FacetField("Publish Date", "2012", "1", "1"));
+ docs.add(doc);
+
+ doc = new Document();
+ doc.add(new FacetField("Author", "Susan"));
+ doc.add(new FacetField("Publish Date", "2012", "1", "7"));
+ docs.add(doc);
+
+ doc = new Document();
+ doc.add(new FacetField("Author", "Frank"));
+ doc.add(new FacetField("Publish Date", "1999", "5", "5"));
+ docs.add(doc);
+
+ return docs;
+ }
+
+ private List<Integer> allDocIds(MatchingDocs m, boolean decreasingDocIds) throws IOException {
+ DocIdSetIterator disi = m.bits.iterator();
+ List<Integer> docIds = new ArrayList<>();
+ while (disi.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
+ docIds.add(disi.docID());
+ }
+
+ if (decreasingDocIds == true) {
+ Collections.reverse(docIds);
+ }
+ return docIds;
+ }
+
+ private List<FacetLabel> lookupFacetLabels(TaxonomyFacetLabels taxoLabels,
+ List<MatchingDocs> matchingDocs) throws IOException {
+ return lookupFacetLabels(taxoLabels, matchingDocs, null, false);
+ }
+
+ private List<FacetLabel> lookupFacetLabels(TaxonomyFacetLabels taxoLabels,
+ List<MatchingDocs> matchingDocs,
+ String dimension) throws IOException {
+ return lookupFacetLabels(taxoLabels, matchingDocs, dimension, false);
+ }
+
+ private List<FacetLabel> lookupFacetLabels(TaxonomyFacetLabels taxoLabels, List<MatchingDocs> matchingDocs, String dimension,
+ boolean decreasingDocIds) throws IOException {
+ List<FacetLabel> facetLabels = new ArrayList<>();
+
+ for (MatchingDocs m : matchingDocs) {
+ TaxonomyFacetLabels.FacetLabelReader facetLabelReader = taxoLabels.getFacetLabelReader(m.context);
+ List<Integer> docIds = allDocIds(m, decreasingDocIds);
+ FacetLabel facetLabel;
+ for (Integer docId : docIds) {
+ while (true) {
+ if (dimension != null) {
+ facetLabel = facetLabelReader.nextFacetLabel(docId, dimension);
+ } else {
+ facetLabel = facetLabelReader.nextFacetLabel(docId);
+ }
+
+ if (facetLabel == null) {
+ break;
+ }
+ facetLabels.add(facetLabel);
+ }
+ }
+ }
+
+ return facetLabels;
+ }
+
+
+ public void testBasic() throws Exception {
Review comment:
Could we make a small change to an existing randomized test, maybe `TestTaxonomyFacetCounts.testRandom`, to use this API to randomly retrieve N facet labels? (So we exercise the class beyond the simplish `testBasic` ... maybe we uncover a rare bug, somewhere).
##########
File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/TaxonomyFacetLabels.java
##########
@@ -0,0 +1,184 @@
+/*
+ * 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.
+ */
+package org.apache.lucene.facet.taxonomy;
+
+import org.apache.lucene.facet.FacetsConfig;
+import org.apache.lucene.index.LeafReaderContext;
+import org.apache.lucene.util.IntsRef;
+
+import java.io.IOException;
+
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.INVALID_ORDINAL;
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.ROOT_ORDINAL;
+
+/**
+ * Utility class to easily retrieve previously indexed facet labels, allowing you to skip also adding stored fields for these values,
+ * reducing your index size.
+ *
+ * @lucene.experimental
+ **/
+public class TaxonomyFacetLabels {
+
+ /**
+ * Index field name provided to the constructor
+ */
+ private final String indexFieldName;
+
+ /**
+ * {@code TaxonomyReader} provided to the constructor
+ */
+ private final TaxonomyReader taxoReader;
+
+ /**
+ * {@code FacetsConfig} provided to the constructor
+ */
+ private final FacetsConfig config;
+
+ /**
+ * {@code OrdinalsReader} to decode ordinals previously indexed into the {@code BinaryDocValues} facet field
+ */
+ private final OrdinalsReader ordsReader;
+
+ /**
+ * Sole constructor. Do not close the provided {@link TaxonomyReader} while still using this instance!
+ */
+ public TaxonomyFacetLabels(TaxonomyReader taxoReader, FacetsConfig config, String indexFieldName) throws IOException {
+ this.taxoReader = taxoReader;
+ this.config = config;
+ this.indexFieldName = indexFieldName;
+ this.ordsReader = new DocValuesOrdinalsReader(indexFieldName);
+ }
+
+ /**
+ * Create and return an instance of {@link FacetLabelReader} to retrieve facet labels for
+ * multiple documents and (optionally) for a specific dimension. You must create this per-segment,
+ * and then step through all hits, in order, for that segment.
+ *
+ * <p><b>NOTE</b>: This class is not thread-safe, so you must use a new instance of this
+ * class for each thread.</p>
+ *
+ * @param readerContext LeafReaderContext used to access the {@code BinaryDocValues} facet field
+ * @return an instance of {@link FacetLabelReader}
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabelReader getFacetLabelReader(LeafReaderContext readerContext) throws IOException {
+ return new FacetLabelReader(ordsReader, readerContext);
+ }
+
+ /**
+ * Utility class to retrieve facet labels for multiple documents.
+ *
+ * @lucene.experimental
+ */
+ public class FacetLabelReader {
+ private final OrdinalsReader.OrdinalsSegmentReader ordinalsSegmentReader;
+ private final IntsRef decodedOrds = new IntsRef();
+ private int currentDocId = -1;
+ private int currentPos = -1;
+
+ // Lazily set when nextFacetLabel(int docId, String facetDimension) is first called
+ private int[] parents;
+
+ /**
+ * Sole constructor.
+ */
+ public FacetLabelReader(OrdinalsReader ordsReader, LeafReaderContext readerContext) throws IOException {
+ ordinalsSegmentReader = ordsReader.getReader(readerContext);
+ }
+
+ /**
+ * Retrieves the next {@link FacetLabel} for the specified {@code docId}, or {@code null} if there are no more.
+ * This method has state: if the provided {@code docId} is the same as the previous invocation, it returns the
+ * next {@link FacetLabel} for that document. Otherwise, it advances to the new {@code docId} and provides the
+ * first {@link FacetLabel} for that document, or {@code null} if that document has no indexed facets. Each
+ * new {@code docId} must be in strictly monotonic (increasing) order.
+ *
+ * @param docId input docId provided in monotonic (non-decreasing) order
+ * @return the first or next {@link FacetLabel}, or {@code null} if there are no more
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabel nextFacetLabel(int docId) throws IOException {
+ if (currentDocId != docId) {
+ assert docId > currentDocId: "docs out of order! currentDocId=" + currentDocId + " docId=" + docId;
+ ordinalsSegmentReader.get(docId, decodedOrds);
+ currentDocId = docId;
+ currentPos = decodedOrds.offset;
+ }
+
+ int endPos = decodedOrds.offset + decodedOrds.length;
+ assert currentPos <= endPos;
+
+ if (currentPos == endPos) {
+ // no more FacetLabels
+ return null;
+ }
+
+ int ord = decodedOrds.ints[currentPos++];
+ return taxoReader.getPath(ord);
+ }
+
+ private boolean isDescendant(int ord, int ancestorOrd) {
+ while (ord != INVALID_ORDINAL && ord != ROOT_ORDINAL) {
+ if (parents[ord] == ancestorOrd) {
+ return true;
+ }
+ ord = parents[ord];
+ }
+ return false;
+ }
+
+ /**
+ * Retrieves the next {@link FacetLabel} for the specified {@code docId} under the requested {@code facetDimension},
+ * or {@code null} if there are no more. This method has state: if the provided {@code docId} is the same as the
+ * previous invocation, it returns the next {@link FacetLabel} for that document. Otherwise, it advances to
+ * the new {@code docId} and provides the first {@link FacetLabel} for that document, or {@code null} if that document
+ * has no indexed facets. Each new {@code docId} must be in strictly monotonic (increasing) order.
+ *
+ * <p><b>NOTE</b>: this method loads the {@code int[] parents} array from the taxonomy index</p>
Review comment:
Add same **Note** here as above comment?
##########
File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/TaxonomyFacetLabels.java
##########
@@ -0,0 +1,184 @@
+/*
+ * 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.
+ */
+package org.apache.lucene.facet.taxonomy;
+
+import org.apache.lucene.facet.FacetsConfig;
+import org.apache.lucene.index.LeafReaderContext;
+import org.apache.lucene.util.IntsRef;
+
+import java.io.IOException;
+
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.INVALID_ORDINAL;
+import static org.apache.lucene.facet.taxonomy.TaxonomyReader.ROOT_ORDINAL;
+
+/**
+ * Utility class to easily retrieve previously indexed facet labels, allowing you to skip also adding stored fields for these values,
+ * reducing your index size.
+ *
+ * @lucene.experimental
+ **/
+public class TaxonomyFacetLabels {
+
+ /**
+ * Index field name provided to the constructor
+ */
+ private final String indexFieldName;
+
+ /**
+ * {@code TaxonomyReader} provided to the constructor
+ */
+ private final TaxonomyReader taxoReader;
+
+ /**
+ * {@code FacetsConfig} provided to the constructor
+ */
+ private final FacetsConfig config;
+
+ /**
+ * {@code OrdinalsReader} to decode ordinals previously indexed into the {@code BinaryDocValues} facet field
+ */
+ private final OrdinalsReader ordsReader;
+
+ /**
+ * Sole constructor. Do not close the provided {@link TaxonomyReader} while still using this instance!
+ */
+ public TaxonomyFacetLabels(TaxonomyReader taxoReader, FacetsConfig config, String indexFieldName) throws IOException {
+ this.taxoReader = taxoReader;
+ this.config = config;
+ this.indexFieldName = indexFieldName;
+ this.ordsReader = new DocValuesOrdinalsReader(indexFieldName);
+ }
+
+ /**
+ * Create and return an instance of {@link FacetLabelReader} to retrieve facet labels for
+ * multiple documents and (optionally) for a specific dimension. You must create this per-segment,
+ * and then step through all hits, in order, for that segment.
+ *
+ * <p><b>NOTE</b>: This class is not thread-safe, so you must use a new instance of this
+ * class for each thread.</p>
+ *
+ * @param readerContext LeafReaderContext used to access the {@code BinaryDocValues} facet field
+ * @return an instance of {@link FacetLabelReader}
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabelReader getFacetLabelReader(LeafReaderContext readerContext) throws IOException {
+ return new FacetLabelReader(ordsReader, readerContext);
+ }
+
+ /**
+ * Utility class to retrieve facet labels for multiple documents.
+ *
+ * @lucene.experimental
+ */
+ public class FacetLabelReader {
+ private final OrdinalsReader.OrdinalsSegmentReader ordinalsSegmentReader;
+ private final IntsRef decodedOrds = new IntsRef();
+ private int currentDocId = -1;
+ private int currentPos = -1;
+
+ // Lazily set when nextFacetLabel(int docId, String facetDimension) is first called
+ private int[] parents;
+
+ /**
+ * Sole constructor.
+ */
+ public FacetLabelReader(OrdinalsReader ordsReader, LeafReaderContext readerContext) throws IOException {
+ ordinalsSegmentReader = ordsReader.getReader(readerContext);
+ }
+
+ /**
+ * Retrieves the next {@link FacetLabel} for the specified {@code docId}, or {@code null} if there are no more.
+ * This method has state: if the provided {@code docId} is the same as the previous invocation, it returns the
+ * next {@link FacetLabel} for that document. Otherwise, it advances to the new {@code docId} and provides the
+ * first {@link FacetLabel} for that document, or {@code null} if that document has no indexed facets. Each
+ * new {@code docId} must be in strictly monotonic (increasing) order.
+ *
+ * @param docId input docId provided in monotonic (non-decreasing) order
+ * @return the first or next {@link FacetLabel}, or {@code null} if there are no more
+ * @throws IOException when a low-level IO issue occurs
+ */
+ public FacetLabel nextFacetLabel(int docId) throws IOException {
+ if (currentDocId != docId) {
+ assert docId > currentDocId: "docs out of order! currentDocId=" + currentDocId + " docId=" + docId;
+ ordinalsSegmentReader.get(docId, decodedOrds);
+ currentDocId = docId;
+ currentPos = decodedOrds.offset;
+ }
+
+ int endPos = decodedOrds.offset + decodedOrds.length;
+ assert currentPos <= endPos;
+
+ if (currentPos == endPos) {
+ // no more FacetLabels
+ return null;
+ }
+
+ int ord = decodedOrds.ints[currentPos++];
+ return taxoReader.getPath(ord);
+ }
+
+ private boolean isDescendant(int ord, int ancestorOrd) {
+ while (ord != INVALID_ORDINAL && ord != ROOT_ORDINAL) {
+ if (parents[ord] == ancestorOrd) {
+ return true;
+ }
+ ord = parents[ord];
+ }
+ return false;
+ }
+
+ /**
+ * Retrieves the next {@link FacetLabel} for the specified {@code docId} under the requested {@code facetDimension},
+ * or {@code null} if there are no more. This method has state: if the provided {@code docId} is the same as the
+ * previous invocation, it returns the next {@link FacetLabel} for that document. Otherwise, it advances to
+ * the new {@code docId} and provides the first {@link FacetLabel} for that document, or {@code null} if that document
+ * has no indexed facets. Each new {@code docId} must be in strictly monotonic (increasing) order.
+ *
+ * <p><b>NOTE</b>: this method loads the {@code int[] parents} array from the taxonomy index</p>
+ *
+ * @param docId input docId provided in non-decreasing order
+ * @return the first or next {@link FacetLabel}, or {@code null} if there are no more
+ * @throws IOException if {@link TaxonomyReader} has problems getting path for an ordinal
+ */
+ public FacetLabel nextFacetLabel(int docId, String facetDimension) throws IOException {
+ final int parentOrd = taxoReader.getOrdinal(new FacetLabel(facetDimension));
+ assert parentOrd != INVALID_ORDINAL : "Category ordinal not found for facet dimension: " + facetDimension;
+
+ if (currentDocId != docId) {
+ assert docId > currentDocId: "docs out of order! currentDocId=" + currentDocId + " docId=" + docId;
+ ordinalsSegmentReader.get(docId, decodedOrds);
+ currentPos = decodedOrds.offset;
+ currentDocId = docId;
+ }
+
+ if (parents == null) {
+ parents = taxoReader.getParallelTaxonomyArrays().parents();
+ }
+
+ int endPos = decodedOrds.offset + decodedOrds.length;
+ assert currentPos <= endPos;
+
+ for (; currentPos < endPos; ) {
+ int ord = decodedOrds.ints[currentPos++];
+ if (isDescendant(ord, parentOrd) == true) {
+ return taxoReader.getPath(ord);
Review comment:
Hmm when we switch to `BinaryDocValues` for this lookup, we could fix this class to persistently hold a single `BinaryDocValues`, and improve performance for caller that need to look up multiple ordinals, which is the common case.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org
For additional commands, e-mail: issues-help@lucene.apache.org