|
| 1 | +/* |
| 2 | + * SPDX-License-Identifier: Apache-2.0 |
| 3 | + * |
| 4 | + * The OpenSearch Contributors require contributions made to |
| 5 | + * this file be licensed under the Apache-2.0 license or a |
| 6 | + * compatible open source license. |
| 7 | + */ |
| 8 | + |
| 9 | +package org.opensearch.index.codec.customcodecs; |
| 10 | + |
| 11 | +import org.apache.lucene.codecs.StoredFieldsFormat; |
| 12 | +import org.apache.lucene.codecs.StoredFieldsReader; |
| 13 | +import org.apache.lucene.codecs.StoredFieldsWriter; |
| 14 | +import org.apache.lucene.codecs.compressing.CompressionMode; |
| 15 | +import org.apache.lucene.codecs.lucene90.compressing.Lucene90CompressingStoredFieldsFormat; |
| 16 | +import org.apache.lucene.index.FieldInfos; |
| 17 | +import org.apache.lucene.index.SegmentInfo; |
| 18 | +import org.apache.lucene.store.Directory; |
| 19 | +import org.apache.lucene.store.IOContext; |
| 20 | + |
| 21 | +import java.io.IOException; |
| 22 | +import java.util.Objects; |
| 23 | + |
| 24 | +import static org.opensearch.index.codec.customcodecs.backward_codecs.lucene99.Lucene99CustomCodec.DEFAULT_COMPRESSION_LEVEL; |
| 25 | + |
| 26 | +/** Stored field format used by pluggable codec */ |
| 27 | +public class Lucene101CustomStoredFieldsFormat extends StoredFieldsFormat { |
| 28 | + |
| 29 | + /** A key that we use to map to a mode */ |
| 30 | + public static final String MODE_KEY = Lucene101CustomStoredFieldsFormat.class.getSimpleName() + ".mode"; |
| 31 | + |
| 32 | + protected static final int ZSTD_BLOCK_LENGTH = 10 * 48 * 1024; |
| 33 | + protected static final int ZSTD_MAX_DOCS_PER_BLOCK = 4096; |
| 34 | + protected static final int ZSTD_BLOCK_SHIFT = 10; |
| 35 | + |
| 36 | + private final CompressionMode zstdCompressionMode; |
| 37 | + private final CompressionMode zstdNoDictCompressionMode; |
| 38 | + |
| 39 | + private final Lucene101CustomCodec.Mode mode; |
| 40 | + private final int compressionLevel; |
| 41 | + |
| 42 | + /** default constructor */ |
| 43 | + public Lucene101CustomStoredFieldsFormat() { |
| 44 | + this(Lucene101CustomCodec.Mode.ZSTD, DEFAULT_COMPRESSION_LEVEL); |
| 45 | + } |
| 46 | + |
| 47 | + /** |
| 48 | + * Creates a new instance. |
| 49 | + * |
| 50 | + * @param mode The mode represents ZSTD or ZSTDNODICT |
| 51 | + */ |
| 52 | + public Lucene101CustomStoredFieldsFormat(Lucene101CustomCodec.Mode mode) { |
| 53 | + this(mode, DEFAULT_COMPRESSION_LEVEL); |
| 54 | + } |
| 55 | + |
| 56 | + /** |
| 57 | + * Creates a new instance with the specified mode and compression level. |
| 58 | + * |
| 59 | + * @param mode The mode represents ZSTD or ZSTDNODICT |
| 60 | + * @param compressionLevel The compression level for the mode. |
| 61 | + */ |
| 62 | + public Lucene101CustomStoredFieldsFormat(Lucene101CustomCodec.Mode mode, int compressionLevel) { |
| 63 | + this.mode = Objects.requireNonNull(mode); |
| 64 | + this.compressionLevel = compressionLevel; |
| 65 | + zstdCompressionMode = new ZstdCompressionMode(compressionLevel); |
| 66 | + zstdNoDictCompressionMode = new ZstdNoDictCompressionMode(compressionLevel); |
| 67 | + } |
| 68 | + |
| 69 | + /** |
| 70 | + * Returns a {@link StoredFieldsReader} to load stored fields. |
| 71 | + * @param directory The index directory. |
| 72 | + * @param si The SegmentInfo that stores segment information. |
| 73 | + * @param fn The fieldInfos. |
| 74 | + * @param context The IOContext that holds additional details on the merge/search context. |
| 75 | + */ |
| 76 | + @Override |
| 77 | + public StoredFieldsReader fieldsReader(Directory directory, SegmentInfo si, FieldInfos fn, IOContext context) throws IOException { |
| 78 | + if (si.getAttribute(MODE_KEY) != null) { |
| 79 | + String value = si.getAttribute(MODE_KEY); |
| 80 | + Lucene101CustomCodec.Mode mode = Lucene101CustomCodec.Mode.valueOf(value); |
| 81 | + return impl(mode).fieldsReader(directory, si, fn, context); |
| 82 | + } else { |
| 83 | + throw new IllegalStateException("missing value for " + MODE_KEY + " for segment: " + si.name); |
| 84 | + } |
| 85 | + } |
| 86 | + |
| 87 | + /** |
| 88 | + * Returns a {@link StoredFieldsReader} to write stored fields. |
| 89 | + * @param directory The index directory. |
| 90 | + * @param si The SegmentInfo that stores segment information. |
| 91 | + * @param context The IOContext that holds additional details on the merge/search context. |
| 92 | + */ |
| 93 | + @Override |
| 94 | + public StoredFieldsWriter fieldsWriter(Directory directory, SegmentInfo si, IOContext context) throws IOException { |
| 95 | + String previous = si.putAttribute(MODE_KEY, mode.name()); |
| 96 | + if (previous != null && previous.equals(mode.name()) == false) { |
| 97 | + throw new IllegalStateException( |
| 98 | + "found existing value for " + MODE_KEY + " for segment: " + si.name + " old = " + previous + ", new = " + mode.name() |
| 99 | + ); |
| 100 | + } |
| 101 | + return impl(mode).fieldsWriter(directory, si, context); |
| 102 | + } |
| 103 | + |
| 104 | + StoredFieldsFormat impl(Lucene101CustomCodec.Mode mode) { |
| 105 | + switch (mode) { |
| 106 | + case ZSTD: |
| 107 | + return getCustomCompressingStoredFieldsFormat("CustomStoredFieldsZstd", this.zstdCompressionMode); |
| 108 | + case ZSTD_NO_DICT: |
| 109 | + return getCustomCompressingStoredFieldsFormat("CustomStoredFieldsZstdNoDict", this.zstdNoDictCompressionMode); |
| 110 | + default: |
| 111 | + throw new IllegalStateException("Unsupported compression mode: " + mode); |
| 112 | + } |
| 113 | + } |
| 114 | + |
| 115 | + private StoredFieldsFormat getCustomCompressingStoredFieldsFormat(String formatName, CompressionMode compressionMode) { |
| 116 | + return new Lucene90CompressingStoredFieldsFormat( |
| 117 | + formatName, |
| 118 | + compressionMode, |
| 119 | + ZSTD_BLOCK_LENGTH, |
| 120 | + ZSTD_MAX_DOCS_PER_BLOCK, |
| 121 | + ZSTD_BLOCK_SHIFT |
| 122 | + ); |
| 123 | + } |
| 124 | + |
| 125 | + public Lucene101CustomCodec.Mode getMode() { |
| 126 | + return mode; |
| 127 | + } |
| 128 | + |
| 129 | + /** |
| 130 | + * Returns the compression level. |
| 131 | + */ |
| 132 | + public int getCompressionLevel() { |
| 133 | + return compressionLevel; |
| 134 | + } |
| 135 | + |
| 136 | + public CompressionMode getCompressionMode() { |
| 137 | + return mode == Lucene101CustomCodec.Mode.ZSTD_NO_DICT ? zstdNoDictCompressionMode : zstdCompressionMode; |
| 138 | + } |
| 139 | + |
| 140 | +} |
0 commit comments