|
| 1 | +/* |
| 2 | + * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one |
| 3 | + * or more contributor license agreements. Licensed under the Elastic License |
| 4 | + * 2.0; you may not use this file except in compliance with the Elastic License |
| 5 | + * 2.0. |
| 6 | + */ |
| 7 | + |
| 8 | +package org.elasticsearch.xpack.logsdb.patternedtext; |
| 9 | + |
| 10 | +import org.apache.lucene.index.SortedSetDocValues; |
| 11 | +import org.apache.lucene.util.BytesRef; |
| 12 | +import org.elasticsearch.test.ESTestCase; |
| 13 | + |
| 14 | +import java.io.IOException; |
| 15 | +import java.util.ArrayList; |
| 16 | +import java.util.Arrays; |
| 17 | +import java.util.HashSet; |
| 18 | +import java.util.List; |
| 19 | +import java.util.Objects; |
| 20 | +import java.util.stream.Collectors; |
| 21 | + |
| 22 | +import static org.apache.lucene.search.DocIdSetIterator.NO_MORE_DOCS; |
| 23 | + |
| 24 | +public class PatternTextDocValuesTests extends ESTestCase { |
| 25 | + |
| 26 | + private static PatternedTextDocValues makeDocValueSparseArgs() { |
| 27 | + var template = new SimpleSortedSetDocValues("%W dog", "cat", "%W mouse %W", "hat %W"); |
| 28 | + var args = new SimpleSortedSetDocValues("1", null, "2 3", "4"); |
| 29 | + return new PatternedTextDocValues(template, args); |
| 30 | + } |
| 31 | + |
| 32 | + private static PatternedTextDocValues makeDocValuesDenseArgs() { |
| 33 | + var template = new SimpleSortedSetDocValues("%W moose", "%W goose %W", "%W mouse %W", "%W house"); |
| 34 | + var args = new SimpleSortedSetDocValues("1", "4 5", "2 3", "7"); |
| 35 | + return new PatternedTextDocValues(template, args); |
| 36 | + } |
| 37 | + |
| 38 | + private static PatternedTextDocValues makeDocValueMissingValues() { |
| 39 | + var template = new SimpleSortedSetDocValues("%W cheddar", "cat", null, "%W cheese"); |
| 40 | + var args = new SimpleSortedSetDocValues("1", null, null, "4"); |
| 41 | + return new PatternedTextDocValues(template, args); |
| 42 | + } |
| 43 | + |
| 44 | + public void testNextDoc() throws IOException { |
| 45 | + var docValues = randomBoolean() ? makeDocValueSparseArgs() : makeDocValuesDenseArgs(); |
| 46 | + assertEquals(-1, docValues.docID()); |
| 47 | + assertEquals(0, docValues.nextDoc()); |
| 48 | + assertEquals(1, docValues.nextDoc()); |
| 49 | + assertEquals(2, docValues.nextDoc()); |
| 50 | + assertEquals(3, docValues.nextDoc()); |
| 51 | + assertEquals(NO_MORE_DOCS, docValues.nextDoc()); |
| 52 | + } |
| 53 | + |
| 54 | + public void testNextDocMissing() throws IOException { |
| 55 | + var docValues = makeDocValueMissingValues(); |
| 56 | + assertEquals(-1, docValues.docID()); |
| 57 | + assertEquals(0, docValues.nextDoc()); |
| 58 | + assertEquals(1, docValues.nextDoc()); |
| 59 | + assertEquals(3, docValues.nextDoc()); |
| 60 | + assertEquals(NO_MORE_DOCS, docValues.nextDoc()); |
| 61 | + } |
| 62 | + |
| 63 | + public void testAdvance1() throws IOException { |
| 64 | + var docValues = randomBoolean() ? makeDocValueSparseArgs() : makeDocValuesDenseArgs(); |
| 65 | + assertEquals(-1, docValues.docID()); |
| 66 | + assertEquals(0, docValues.nextDoc()); |
| 67 | + assertEquals(1, docValues.advance(1)); |
| 68 | + assertEquals(2, docValues.advance(2)); |
| 69 | + assertEquals(3, docValues.advance(3)); |
| 70 | + assertEquals(NO_MORE_DOCS, docValues.advance(4)); |
| 71 | + } |
| 72 | + |
| 73 | + public void testAdvanceFarther() throws IOException { |
| 74 | + var docValues = randomBoolean() ? makeDocValueSparseArgs() : makeDocValuesDenseArgs(); |
| 75 | + assertEquals(2, docValues.advance(2)); |
| 76 | + // repeats says on value |
| 77 | + assertEquals(2, docValues.advance(2)); |
| 78 | + } |
| 79 | + |
| 80 | + public void testAdvanceSkipsValuesIfMissing() throws IOException { |
| 81 | + var docValues = makeDocValueMissingValues(); |
| 82 | + assertEquals(3, docValues.advance(2)); |
| 83 | + } |
| 84 | + |
| 85 | + public void testAdvanceExactMissing() throws IOException { |
| 86 | + var docValues = makeDocValueMissingValues(); |
| 87 | + assertTrue(docValues.advanceExact(1)); |
| 88 | + assertFalse(docValues.advanceExact(2)); |
| 89 | + assertEquals(3, docValues.docID()); |
| 90 | + } |
| 91 | + |
| 92 | + public void testValueAll() throws IOException { |
| 93 | + var docValues = makeDocValuesDenseArgs(); |
| 94 | + assertEquals(0, docValues.nextDoc()); |
| 95 | + assertEquals("1 moose", docValues.binaryValue().utf8ToString()); |
| 96 | + assertEquals(1, docValues.nextDoc()); |
| 97 | + assertEquals("4 goose 5", docValues.binaryValue().utf8ToString()); |
| 98 | + assertEquals(2, docValues.nextDoc()); |
| 99 | + assertEquals("2 mouse 3", docValues.binaryValue().utf8ToString()); |
| 100 | + assertEquals(3, docValues.nextDoc()); |
| 101 | + assertEquals("7 house", docValues.binaryValue().utf8ToString()); |
| 102 | + } |
| 103 | + |
| 104 | + public void testValueMissing() throws IOException { |
| 105 | + var docValues = makeDocValueMissingValues(); |
| 106 | + assertEquals(0, docValues.nextDoc()); |
| 107 | + assertEquals("1 cheddar", docValues.binaryValue().utf8ToString()); |
| 108 | + assertEquals(1, docValues.nextDoc()); |
| 109 | + assertEquals("cat", docValues.binaryValue().utf8ToString()); |
| 110 | + assertEquals(3, docValues.nextDoc()); |
| 111 | + assertEquals("4 cheese", docValues.binaryValue().utf8ToString()); |
| 112 | + } |
| 113 | + |
| 114 | + static class SimpleSortedSetDocValues extends SortedSetDocValues { |
| 115 | + |
| 116 | + private final List<String> ordToValues; |
| 117 | + private final List<Integer> docToOrds; |
| 118 | + private int currDoc = -1; |
| 119 | + |
| 120 | + SimpleSortedSetDocValues(String... docIdToValue) { |
| 121 | + ordToValues = Arrays.stream(docIdToValue).filter(Objects::nonNull) |
| 122 | + .collect(Collectors.toSet()) |
| 123 | + .stream().sorted().toList(); |
| 124 | + docToOrds = Arrays.stream(docIdToValue).map(v -> v == null ? null : ordToValues.indexOf(v)).toList(); |
| 125 | + } |
| 126 | + |
| 127 | + @Override |
| 128 | + public long nextOrd() { |
| 129 | + return docToOrds.get(currDoc); |
| 130 | + } |
| 131 | + |
| 132 | + @Override |
| 133 | + public int docValueCount() { |
| 134 | + return 1; |
| 135 | + } |
| 136 | + |
| 137 | + @Override |
| 138 | + public BytesRef lookupOrd(long ord) { |
| 139 | + return new BytesRef(ordToValues.get((int) ord)); |
| 140 | + } |
| 141 | + |
| 142 | + @Override |
| 143 | + public long getValueCount() { |
| 144 | + return ordToValues.size(); |
| 145 | + } |
| 146 | + |
| 147 | + @Override |
| 148 | + public boolean advanceExact(int target) { |
| 149 | + return advance(target) == target; |
| 150 | + } |
| 151 | + |
| 152 | + @Override |
| 153 | + public int docID() { |
| 154 | + return currDoc >= docToOrds.size() ? NO_MORE_DOCS : currDoc; |
| 155 | + } |
| 156 | + |
| 157 | + @Override |
| 158 | + public int nextDoc() throws IOException { |
| 159 | + return advance(currDoc + 1); |
| 160 | + } |
| 161 | + |
| 162 | + @Override |
| 163 | + public int advance(int target) { |
| 164 | + for (currDoc = target; currDoc < docToOrds.size(); currDoc++) { |
| 165 | + if (docToOrds.get(currDoc) != null) { |
| 166 | + return currDoc; |
| 167 | + } |
| 168 | + } |
| 169 | + return NO_MORE_DOCS; |
| 170 | + } |
| 171 | + |
| 172 | + @Override |
| 173 | + public long cost() { |
| 174 | + return 1; |
| 175 | + } |
| 176 | + } |
| 177 | +} |
0 commit comments