|
| 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.mapper; |
| 10 | + |
| 11 | +import org.apache.lucene.document.Document; |
| 12 | +import org.apache.lucene.document.Field; |
| 13 | +import org.apache.lucene.document.TextField; |
| 14 | +import org.apache.lucene.index.DirectoryReader; |
| 15 | +import org.apache.lucene.index.IndexReader; |
| 16 | +import org.apache.lucene.index.IndexWriter; |
| 17 | +import org.apache.lucene.index.IndexWriterConfig; |
| 18 | +import org.apache.lucene.search.IndexSearcher; |
| 19 | +import org.apache.lucene.search.Query; |
| 20 | +import org.apache.lucene.search.TopDocs; |
| 21 | +import org.apache.lucene.store.Directory; |
| 22 | +import org.opensearch.common.lucene.Lucene; |
| 23 | +import org.opensearch.core.index.Index; |
| 24 | +import org.opensearch.index.query.QueryBuilders; |
| 25 | +import org.opensearch.index.query.QueryShardContext; |
| 26 | +import org.opensearch.script.DerivedFieldScript; |
| 27 | + |
| 28 | +import java.io.IOException; |
| 29 | +import java.util.ArrayList; |
| 30 | +import java.util.List; |
| 31 | +import java.util.Set; |
| 32 | + |
| 33 | +import org.mockito.Mockito; |
| 34 | + |
| 35 | +import static org.mockito.Mockito.when; |
| 36 | + |
| 37 | +public class DerivedFieldMapperQueryTests extends MapperServiceTestCase { |
| 38 | + |
| 39 | + // First element is the document ingested, other elements corresponds to value returned against a given derived field script |
| 40 | + // Raw Message, Request Succeeded (boolean), Timestamp (long), Client IP, Method, Request Size (double), Duration (long) |
| 41 | + private static final Object[][] raw_requests = new Object[][] { |
| 42 | + { |
| 43 | + "40.135.0.0 GET /images/hm_bg.jpg?size=1.5KB HTTP/1.0 200 2024-03-20T08:30:45 1500", |
| 44 | + true, |
| 45 | + 1710923445000L, |
| 46 | + "40.135.0.0", |
| 47 | + "GET", |
| 48 | + 1.5, |
| 49 | + 1500L }, |
| 50 | + { |
| 51 | + "232.0.0.0 GET /images/hm_bg.jpg?size=2.3KB HTTP/1.0 400 2024-03-20T09:15:20 2300", |
| 52 | + false, |
| 53 | + 1710926120000L, |
| 54 | + "232.0.0.0", |
| 55 | + "GET", |
| 56 | + 2.3, |
| 57 | + 2300L }, |
| 58 | + { |
| 59 | + "26.1.0.0 DELETE /images/hm_bg.jpg?size=3.7KB HTTP/1.0 200 2024-03-20T10:05:55 3700", |
| 60 | + true, |
| 61 | + 1710929155000L, |
| 62 | + "26.1.0.0", |
| 63 | + "DELETE", |
| 64 | + 3.7, |
| 65 | + 3700L }, |
| 66 | + { |
| 67 | + "247.37.0.0 GET /french/splash_inet.html?size=4.1KB HTTP/1.0 400 2024-03-20T11:20:10 4100", |
| 68 | + false, |
| 69 | + 1710933610000L, |
| 70 | + "247.37.0.0", |
| 71 | + "GET", |
| 72 | + 4.1, |
| 73 | + 4100L }, |
| 74 | + { |
| 75 | + "247.37.0.0 DELETE /french/splash_inet.html?size=5.8KB HTTP/1.0 400 2024-03-20T12:45:30 5800", |
| 76 | + false, |
| 77 | + 1710938730000L, |
| 78 | + "247.37.0.0", |
| 79 | + "DELETE", |
| 80 | + 5.8, |
| 81 | + 5800L }, |
| 82 | + { |
| 83 | + "10.20.30.40 GET /path/to/resource?size=6.3KB HTTP/1.0 200 2024-03-20T13:10:15 6300", |
| 84 | + true, |
| 85 | + 1710940215000L, |
| 86 | + "10.20.30.40", |
| 87 | + "GET", |
| 88 | + 6.3, |
| 89 | + 6300L }, |
| 90 | + { |
| 91 | + "50.60.70.80 GET /path/to/resource?size=7.2KB HTTP/1.0 404 2024-03-20T14:20:50 7200", |
| 92 | + false, |
| 93 | + 1710944450000L, |
| 94 | + "50.60.70.80", |
| 95 | + "GET", |
| 96 | + 7.2, |
| 97 | + 7200L }, |
| 98 | + { |
| 99 | + "127.0.0.1 PUT /path/to/resource?size=8.9KB HTTP/1.0 500 2024-03-20T15:30:25 8900", |
| 100 | + false, |
| 101 | + 1710948625000L, |
| 102 | + "127.0.0.1", |
| 103 | + "PUT", |
| 104 | + 8.9, |
| 105 | + 8900L }, |
| 106 | + { |
| 107 | + "127.0.0.1 GET /path/to/resource?size=9.4KB HTTP/1.0 200 2024-03-20T16:40:15 9400", |
| 108 | + true, |
| 109 | + 1710952815000L, |
| 110 | + "127.0.0.1", |
| 111 | + "GET", |
| 112 | + 9.4, |
| 113 | + 9400L }, |
| 114 | + { |
| 115 | + "192.168.1.1 GET /path/to/resource?size=10.7KB HTTP/1.0 400 2024-03-20T17:50:40 10700", |
| 116 | + false, |
| 117 | + 1710957040000L, |
| 118 | + "192.168.1.1", |
| 119 | + "GET", |
| 120 | + 10.7, |
| 121 | + 10700L } }; |
| 122 | + |
| 123 | + public void testAllPossibleQueriesOnDerivedFields() throws IOException { |
| 124 | + MapperService mapperService = createMapperService(topMapping(b -> { |
| 125 | + b.startObject("properties"); |
| 126 | + { |
| 127 | + b.startObject("raw_message"); |
| 128 | + { |
| 129 | + b.field("type", "text"); |
| 130 | + } |
| 131 | + b.endObject(); |
| 132 | + } |
| 133 | + b.endObject(); |
| 134 | + b.startObject("derived"); |
| 135 | + { |
| 136 | + b.startObject("request_succeeded"); |
| 137 | + { |
| 138 | + b.field("type", "boolean"); |
| 139 | + b.field("script", ""); |
| 140 | + } |
| 141 | + b.endObject(); |
| 142 | + b.startObject("@timestamp"); |
| 143 | + { |
| 144 | + b.field("type", "date"); |
| 145 | + b.field("script", ""); |
| 146 | + } |
| 147 | + b.endObject(); |
| 148 | + b.startObject("client_ip"); |
| 149 | + { |
| 150 | + b.field("type", "ip"); |
| 151 | + b.field("script", ""); |
| 152 | + } |
| 153 | + b.endObject(); |
| 154 | + b.startObject("method"); |
| 155 | + { |
| 156 | + b.field("type", "keyword"); |
| 157 | + b.field("script", ""); |
| 158 | + } |
| 159 | + b.endObject(); |
| 160 | + b.startObject("request_size"); |
| 161 | + { |
| 162 | + b.field("type", "double"); |
| 163 | + b.field("script", ""); |
| 164 | + } |
| 165 | + b.endObject(); |
| 166 | + b.startObject("duration"); |
| 167 | + { |
| 168 | + b.field("type", "long"); |
| 169 | + b.field("script", ""); |
| 170 | + } |
| 171 | + b.endObject(); |
| 172 | + } |
| 173 | + b.endObject(); |
| 174 | + })); |
| 175 | + |
| 176 | + List<Document> docs = new ArrayList<>(); |
| 177 | + for (Object[] request : raw_requests) { |
| 178 | + Document document = new Document(); |
| 179 | + document.add(new TextField("raw_message", (String) request[0], Field.Store.YES)); |
| 180 | + docs.add(document); |
| 181 | + } |
| 182 | + |
| 183 | + int[] scriptIndex = { 1 }; |
| 184 | + |
| 185 | + // Mock DerivedFieldScript.Factory |
| 186 | + DerivedFieldScript.Factory factory = (params, lookup) -> (DerivedFieldScript.LeafFactory) ctx -> new DerivedFieldScript( |
| 187 | + params, |
| 188 | + lookup, |
| 189 | + ctx |
| 190 | + ) { |
| 191 | + int docId = 0; |
| 192 | + |
| 193 | + @Override |
| 194 | + public void setDocument(int docId) { |
| 195 | + super.setDocument(docId); |
| 196 | + this.docId = docId; |
| 197 | + } |
| 198 | + |
| 199 | + @Override |
| 200 | + public Object execute() { |
| 201 | + return raw_requests[docId][scriptIndex[0]]; |
| 202 | + } |
| 203 | + }; |
| 204 | + |
| 205 | + QueryShardContext queryShardContext = createQueryShardContext(mapperService); |
| 206 | + when(queryShardContext.compile(Mockito.any(), Mockito.any())).thenReturn(factory); |
| 207 | + when(queryShardContext.sourcePath("raw_message")).thenReturn(Set.of("raw_message")); |
| 208 | + when(queryShardContext.index()).thenReturn(new Index("test_index", "uuid")); |
| 209 | + |
| 210 | + // Index and Search |
| 211 | + try (Directory dir = newDirectory()) { |
| 212 | + IndexWriter iw = new IndexWriter(dir, new IndexWriterConfig(Lucene.STANDARD_ANALYZER)); |
| 213 | + for (Document d : docs) { |
| 214 | + iw.addDocument(d); |
| 215 | + } |
| 216 | + try (IndexReader reader = DirectoryReader.open(iw)) { |
| 217 | + iw.close(); |
| 218 | + |
| 219 | + IndexSearcher searcher = new IndexSearcher(reader); |
| 220 | + Query query = QueryBuilders.termQuery("request_succeeded", "true").toQuery(queryShardContext); |
| 221 | + TopDocs topDocs = searcher.search(query, 10); |
| 222 | + assertEquals(4, topDocs.totalHits.value); |
| 223 | + |
| 224 | + // IP Field Term Query |
| 225 | + scriptIndex[0] = 3; |
| 226 | + query = QueryBuilders.termQuery("client_ip", "192.168.0.0/16").toQuery(queryShardContext); |
| 227 | + topDocs = searcher.search(query, 10); |
| 228 | + assertEquals(1, topDocs.totalHits.value); |
| 229 | + |
| 230 | + scriptIndex[0] = 4; |
| 231 | + query = QueryBuilders.termsQuery("method", "DELETE", "PUT").toQuery(queryShardContext); |
| 232 | + topDocs = searcher.search(query, 10); |
| 233 | + assertEquals(3, topDocs.totalHits.value); |
| 234 | + |
| 235 | + query = QueryBuilders.termsQuery("method", "delete").toQuery(queryShardContext); |
| 236 | + topDocs = searcher.search(query, 10); |
| 237 | + assertEquals(0, topDocs.totalHits.value); |
| 238 | + |
| 239 | + query = QueryBuilders.termQuery("method", "delete").caseInsensitive(true).toQuery(queryShardContext); |
| 240 | + topDocs = searcher.search(query, 10); |
| 241 | + assertEquals(2, topDocs.totalHits.value); |
| 242 | + |
| 243 | + // Range queries of types - date, long and double |
| 244 | + scriptIndex[0] = 2; |
| 245 | + query = QueryBuilders.rangeQuery("@timestamp").from("2024-03-20T14:20:50").toQuery(queryShardContext); |
| 246 | + topDocs = searcher.search(query, 10); |
| 247 | + assertEquals(4, topDocs.totalHits.value); |
| 248 | + |
| 249 | + scriptIndex[0] = 5; |
| 250 | + query = QueryBuilders.rangeQuery("request_size").from("4.1").toQuery(queryShardContext); |
| 251 | + topDocs = searcher.search(query, 10); |
| 252 | + assertEquals(7, topDocs.totalHits.value); |
| 253 | + |
| 254 | + scriptIndex[0] = 6; |
| 255 | + query = QueryBuilders.rangeQuery("duration").from("5800").toQuery(queryShardContext); |
| 256 | + topDocs = searcher.search(query, 10); |
| 257 | + assertEquals(6, topDocs.totalHits.value); |
| 258 | + |
| 259 | + scriptIndex[0] = 4; |
| 260 | + |
| 261 | + // Prefix Query |
| 262 | + query = QueryBuilders.prefixQuery("method", "DE").toQuery(queryShardContext); |
| 263 | + topDocs = searcher.search(query, 10); |
| 264 | + assertEquals(2, topDocs.totalHits.value); |
| 265 | + |
| 266 | + scriptIndex[0] = 4; |
| 267 | + query = QueryBuilders.wildcardQuery("method", "G*").toQuery(queryShardContext); |
| 268 | + topDocs = searcher.search(query, 10); |
| 269 | + assertEquals(7, topDocs.totalHits.value); |
| 270 | + |
| 271 | + // Regexp Query |
| 272 | + scriptIndex[0] = 4; |
| 273 | + query = QueryBuilders.regexpQuery("method", ".*LET.*").toQuery(queryShardContext); |
| 274 | + topDocs = searcher.search(query, 10); |
| 275 | + assertEquals(2, topDocs.totalHits.value); |
| 276 | + } |
| 277 | + } |
| 278 | + } |
| 279 | +} |
0 commit comments