Complete keyword changes for star tree

CHANGELOG.md

@@ -14,6 +14,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - Increase segrep pressure checkpoint default limit to 30 ([#16577](https://github.com/opensearch-project/OpenSearch/pull/16577/files))
 - Add dynamic setting allowing size > 0 requests to be cached in the request cache ([#16483](https://github.com/opensearch-project/OpenSearch/pull/16483))
 - Make IndexStoreListener a pluggable interface ([#16583](https://github.com/opensearch-project/OpenSearch/pull/16583))
+- Support for keyword fields in star-tree index ([#16233](https://github.com/opensearch-project/OpenSearch/pull/16233))
 - Add a flag in QueryShardContext to differentiate inner hit query ([#16600](https://github.com/opensearch-project/OpenSearch/pull/16600))
 
 ### Dependencies

server/src/internalClusterTest/java/org/opensearch/index/mapper/StarTreeMapperIT.java

@@ -56,7 +56,7 @@ public class StarTreeMapperIT extends OpenSearchIntegTestCase {
         .put(IndexSettings.INDEX_TRANSLOG_FLUSH_THRESHOLD_SIZE_SETTING.getKey(), new ByteSizeValue(512, ByteSizeUnit.MB))
         .build();
 
-    private static XContentBuilder createMinimalTestMapping(boolean invalidDim, boolean invalidMetric, boolean keywordDim) {
+    private static XContentBuilder createMinimalTestMapping(boolean invalidDim, boolean invalidMetric, boolean ipdim) {
         try {
             return jsonBuilder().startObject()
                 .startObject("composite")
@@ -68,12 +68,15 @@ private static XContentBuilder createMinimalTestMapping(boolean invalidDim, bool
                 .endObject()
                 .startArray("ordered_dimensions")
                 .startObject()
-                .field("name", getDim(invalidDim, keywordDim))
+                .field("name", getDim(invalidDim, ipdim))
+                .endObject()
+                .startObject()
+                .field("name", "keyword_dv")
                 .endObject()
                 .endArray()
                 .startArray("metrics")
                 .startObject()
-                .field("name", getDim(invalidMetric, false))
+                .field("name", getMetric(invalidMetric, false))
                 .endObject()
                 .endArray()
                 .endObject()
@@ -99,6 +102,10 @@ private static XContentBuilder createMinimalTestMapping(boolean invalidDim, bool
                 .field("type", "keyword")
                 .field("doc_values", false)
                 .endObject()
+                .startObject("ip")
+                .field("type", "ip")
+                .field("doc_values", false)
+                .endObject()
                 .endObject()
                 .endObject();
         } catch (IOException e) {
@@ -356,10 +363,19 @@ private XContentBuilder getMappingWithDuplicateFields(boolean isDuplicateDim, bo
     }
 
     private static String getDim(boolean hasDocValues, boolean isKeyword) {
+        if (hasDocValues) {
+            return random().nextBoolean() ? "numeric" : "keyword";
+        } else if (isKeyword) {
+            return "ip";
+        }
+        return "numeric_dv";
+    }
+
+    private static String getMetric(boolean hasDocValues, boolean isKeyword) {
         if (hasDocValues) {
             return "numeric";
         } else if (isKeyword) {
-            return "keyword";
+            return "ip";
         }
         return "numeric_dv";
     }
@@ -398,6 +414,7 @@ public void testValidCompositeIndex() {
                         assertEquals(expectedTimeUnits.get(i).shortName(), dateDim.getSortedCalendarIntervals().get(i).shortName());
                     }
                     assertEquals("numeric_dv", starTreeFieldType.getDimensions().get(1).getField());
+                    assertEquals("keyword_dv", starTreeFieldType.getDimensions().get(2).getField());
                     assertEquals("numeric_dv", starTreeFieldType.getMetrics().get(0).getField());
                     List<MetricStat> expectedMetrics = Arrays.asList(MetricStat.VALUE_COUNT, MetricStat.SUM, MetricStat.AVG);
                     assertEquals(expectedMetrics, starTreeFieldType.getMetrics().get(0).getMetrics());
@@ -665,10 +682,7 @@ public void testInvalidDimCompositeIndex() {
             IllegalArgumentException.class,
             () -> prepareCreate(TEST_INDEX).setSettings(settings).setMapping(createMinimalTestMapping(true, false, false)).get()
         );
-        assertEquals(
-            "Aggregations not supported for the dimension field [numeric] with field type [integer] as part of star tree field",
-            ex.getMessage()
-        );
+        assertTrue(ex.getMessage().startsWith("Aggregations not supported for the dimension field "));
     }
 
     public void testMaxDimsCompositeIndex() {
@@ -734,7 +748,7 @@ public void testUnsupportedDim() {
             () -> prepareCreate(TEST_INDEX).setSettings(settings).setMapping(createMinimalTestMapping(false, false, true)).get()
         );
         assertEquals(
-            "Failed to parse mapping [_doc]: unsupported field type associated with dimension [keyword] as part of star tree field [startree-1]",
+            "Failed to parse mapping [_doc]: unsupported field type associated with dimension [ip] as part of star tree field [startree-1]",
             ex.getMessage()
         );
     }

server/src/main/java/org/apache/lucene/index/DocValuesWriterWrapper.java

@@ -0,0 +1,18 @@
+/*
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * The OpenSearch Contributors require contributions made to
+ * this file be licensed under the Apache-2.0 license or a
+ * compatible open source license.
+ */
+
+package org.apache.lucene.index;
+
+import org.apache.lucene.search.DocIdSetIterator;
+
+/**
+ * Base wrapper class for DocValuesWriter.
+ */
+public interface DocValuesWriterWrapper<T extends DocIdSetIterator> {
+    T getDocValues();
+}

server/src/main/java/org/apache/lucene/index/SortedNumericDocValuesWriterWrapper.java

@@ -18,9 +18,9 @@
  *
  * @opensearch.experimental
  */
-public class SortedNumericDocValuesWriterWrapper {
+public class SortedNumericDocValuesWriterWrapper implements DocValuesWriterWrapper<SortedNumericDocValues> {
 
-    private final SortedNumericDocValuesWriter sortedNumericDocValuesWriter;
+    private final SortedNumericDocValuesWriter sortedNumericDocValuesWriterDelegate;
 
     /**
      * Sole constructor. Constructs a new {@link SortedNumericDocValuesWriterWrapper} instance.
@@ -29,7 +29,7 @@ public class SortedNumericDocValuesWriterWrapper {
      * @param counter a counter for tracking memory usage
      */
     public SortedNumericDocValuesWriterWrapper(FieldInfo fieldInfo, Counter counter) {
-        sortedNumericDocValuesWriter = new SortedNumericDocValuesWriter(fieldInfo, counter);
+        sortedNumericDocValuesWriterDelegate = new SortedNumericDocValuesWriter(fieldInfo, counter);
     }
 
     /**
@@ -39,15 +39,16 @@ public SortedNumericDocValuesWriterWrapper(FieldInfo fieldInfo, Counter counter)
      * @param value the value to add
      */
     public void addValue(int docID, long value) {
-        sortedNumericDocValuesWriter.addValue(docID, value);
+        sortedNumericDocValuesWriterDelegate.addValue(docID, value);
     }
 
     /**
      * Returns the {@link SortedNumericDocValues} instance containing the sorted numeric doc values
      *
      * @return the {@link SortedNumericDocValues} instance
      */
+    @Override
     public SortedNumericDocValues getDocValues() {
-        return sortedNumericDocValuesWriter.getDocValues();
+        return sortedNumericDocValuesWriterDelegate.getDocValues();
     }
 }

server/src/main/java/org/apache/lucene/index/SortedSetDocValuesWriterWrapper.java

@@ -0,0 +1,58 @@
+/*
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * The OpenSearch Contributors require contributions made to
+ * this file be licensed under the Apache-2.0 license or a
+ * compatible open source license.
+ */
+
+package org.apache.lucene.index;
+
+import org.apache.lucene.util.ByteBlockPool;
+import org.apache.lucene.util.BytesRef;
+import org.apache.lucene.util.Counter;
+
+/**
+ * A wrapper class for writing sorted set doc values.
+ * <p>
+ * This class provides a convenient way to add sorted set doc values to a field
+ * and retrieve the corresponding {@link SortedSetDocValues} instance.
+ *
+ * @opensearch.experimental
+ */
+public class SortedSetDocValuesWriterWrapper implements DocValuesWriterWrapper<SortedSetDocValues> {
+
+    private final SortedSetDocValuesWriter sortedSetDocValuesWriterDelegate;
+
+    /**
+     * Sole constructor. Constructs a new {@link SortedSetDocValuesWriterWrapper} instance.
+     *
+     * @param fieldInfo the field information for the field being written
+     * @param counter a counter for tracking memory usage
+     * @param byteBlockPool a byte block pool for allocating byte blocks
+     * @see SortedSetDocValuesWriter
+     */
+    public SortedSetDocValuesWriterWrapper(FieldInfo fieldInfo, Counter counter, ByteBlockPool byteBlockPool) {
+        sortedSetDocValuesWriterDelegate = new SortedSetDocValuesWriter(fieldInfo, counter, byteBlockPool);
+    }
+
+    /**
+     * Adds a bytes ref value to the sorted set doc values for the specified document.
+     *
+     * @param docID the document ID
+     * @param value the value to add
+     */
+    public void addValue(int docID, BytesRef value) {
+        sortedSetDocValuesWriterDelegate.addValue(docID, value);
+    }
+
+    /**
+     * Returns the {@link SortedSetDocValues} instance containing the sorted numeric doc values
+     *
+     * @return the {@link SortedSetDocValues} instance
+     */
+    @Override
+    public SortedSetDocValues getDocValues() {
+        return sortedSetDocValuesWriterDelegate.getDocValues();
+    }
+}

server/src/main/java/org/opensearch/index/codec/composite/composite912/Composite912DocValuesReader.java

@@ -14,7 +14,7 @@
 import org.apache.lucene.codecs.DocValuesProducer;
 import org.apache.lucene.index.BinaryDocValues;
 import org.apache.lucene.index.CorruptIndexException;
-import org.apache.lucene.index.DocValues;
+import org.apache.lucene.index.DocValuesType;
 import org.apache.lucene.index.FieldInfo;
 import org.apache.lucene.index.FieldInfos;
 import org.apache.lucene.index.IndexFileNames;
@@ -40,6 +40,7 @@
 
 import java.io.IOException;
 import java.util.ArrayList;
+import java.util.HashMap;
 import java.util.LinkedHashMap;
 import java.util.List;
 import java.util.Map;
@@ -111,7 +112,7 @@ public Composite912DocValuesReader(DocValuesProducer producer, SegmentReadState
                     readState.segmentInfo.getId(),
                     readState.segmentSuffix
                 );
-
+                Map<String, DocValuesType> dimensionFieldTypeMap = new HashMap<>();
                 while (true) {
 
                     // validate magic marker
@@ -155,13 +156,16 @@ public Composite912DocValuesReader(DocValuesProducer producer, SegmentReadState
                             compositeIndexInputMap.put(compositeFieldName, starTreeIndexInput);
                             compositeIndexMetadataMap.put(compositeFieldName, starTreeMetadata);
 
-                            List<String> dimensionFields = starTreeMetadata.getDimensionFields();
-
+                            Map<String, DocValuesType> dimensionFieldToDocValuesMap = starTreeMetadata.getDimensionFields();
                             // generating star tree unique fields (fully qualified name for dimension and metrics)
-                            for (String dimensions : dimensionFields) {
-                                fields.add(fullyQualifiedFieldNameForStarTreeDimensionsDocValues(compositeFieldName, dimensions));
+                            for (Map.Entry<String, DocValuesType> dimensionEntry : dimensionFieldToDocValuesMap.entrySet()) {
+                                String dimName = fullyQualifiedFieldNameForStarTreeDimensionsDocValues(
+                                    compositeFieldName,
+                                    dimensionEntry.getKey()
+                                );
+                                fields.add(dimName);
+                                dimensionFieldTypeMap.put(dimName, dimensionEntry.getValue());
                             }
-
                             // adding metric fields
                             for (Metric metric : starTreeMetadata.getMetrics()) {
                                 for (MetricStat metricStat : metric.getBaseMetrics()) {
@@ -184,7 +188,7 @@ public Composite912DocValuesReader(DocValuesProducer producer, SegmentReadState
 
                 // populates the dummy list of field infos to fetch doc id set iterators for respective fields.
                 // the dummy field info is used to fetch the doc id set iterators for respective fields based on field name
-                FieldInfos fieldInfos = new FieldInfos(getFieldInfoList(fields));
+                FieldInfos fieldInfos = new FieldInfos(getFieldInfoList(fields, dimensionFieldTypeMap));
                 this.readState = new SegmentReadState(
                     readState.directory,
                     readState.segmentInfo,
@@ -291,17 +295,4 @@ public CompositeIndexValues getCompositeIndexValues(CompositeIndexFieldInfo comp
 
     }
 
-    /**
-     * Returns the sorted numeric doc values for the given sorted numeric field.
-     * If the sorted numeric field is null, it returns an empty doc id set iterator.
-     * <p>
-     * Sorted numeric field can be null for cases where the segment doesn't hold a particular value.
-     *
-     * @param sortedNumeric the sorted numeric doc values for a field
-     * @return empty sorted numeric values if the field is not present, else sortedNumeric
-     */
-    public static SortedNumericDocValues getSortedNumericDocValues(SortedNumericDocValues sortedNumeric) {
-        return sortedNumeric == null ? DocValues.emptySortedNumeric() : sortedNumeric;
-    }
-
 }