|
| 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 | +package org.elasticsearch.xpack.textstructure.rest; |
| 8 | + |
| 9 | +import com.carrotsearch.randomizedtesting.annotations.Name; |
| 10 | +import com.carrotsearch.randomizedtesting.annotations.ParametersFactory; |
| 11 | + |
| 12 | +import org.apache.http.entity.ContentType; |
| 13 | +import org.apache.http.entity.StringEntity; |
| 14 | +import org.elasticsearch.client.Request; |
| 15 | +import org.elasticsearch.client.Response; |
| 16 | +import org.elasticsearch.test.cluster.ElasticsearchCluster; |
| 17 | +import org.elasticsearch.test.cluster.local.distribution.DistributionType; |
| 18 | +import org.elasticsearch.test.rest.ESRestTestCase; |
| 19 | +import org.junit.ClassRule; |
| 20 | + |
| 21 | +import java.io.IOException; |
| 22 | +import java.util.Arrays; |
| 23 | +import java.util.List; |
| 24 | +import java.util.Map; |
| 25 | + |
| 26 | +import static org.hamcrest.Matchers.containsInAnyOrder; |
| 27 | +import static org.hamcrest.Matchers.containsString; |
| 28 | +import static org.hamcrest.Matchers.equalTo; |
| 29 | +import static org.hamcrest.Matchers.hasKey; |
| 30 | + |
| 31 | +public class TextStructureTimestampFormatsIT extends ESRestTestCase { |
| 32 | + |
| 33 | + public static final String[] ISO_08601_JAVA_FORMATS = new String[] { "yyyy-MM-dd HH:mm:ss" }; |
| 34 | + public static final String ISO_08601_TIMESTAMP_GROK_PATTERN = "%{TIMESTAMP_ISO8601:timestamp}"; |
| 35 | + |
| 36 | + public static final String[] TIMESTAMP_YMD_JAVA_FORMATS = new String[] { |
| 37 | + "yyyy/MM/dd HH:mm:ss", |
| 38 | + "yyyy.MM.dd HH:mm:ss", |
| 39 | + "yyyy-MM-dd HH:mm:ss" }; |
| 40 | + public static final String TIMESTAMP_YMD_TIMESTAMP_GROK_PATTERN = "%{TIMESTAMP_YMD:timestamp}"; |
| 41 | + |
| 42 | + public static final String[] MONTH_EXPLICIT_NAME_JAVA_FORMATS = new String[] { "MMM d, yyyy" }; |
| 43 | + |
| 44 | + private final String ecsCompatibility; |
| 45 | + |
| 46 | + @ClassRule |
| 47 | + public static ElasticsearchCluster cluster = ElasticsearchCluster.local() |
| 48 | + .distribution(DistributionType.DEFAULT) |
| 49 | + .module("x-pack-text-structure") |
| 50 | + .setting("xpack.security.enabled", "false") |
| 51 | + .build(); |
| 52 | + |
| 53 | + public TextStructureTimestampFormatsIT(@Name("ecs_compatibility") String ecsCompatibility) { |
| 54 | + this.ecsCompatibility = ecsCompatibility; |
| 55 | + } |
| 56 | + |
| 57 | + @Override |
| 58 | + protected String getTestRestCluster() { |
| 59 | + return cluster.getHttpAddresses(); |
| 60 | + } |
| 61 | + |
| 62 | + @ParametersFactory |
| 63 | + public static Iterable<Object[]> parameters() { |
| 64 | + return Arrays.asList(new Object[] { "v1" }, new Object[] { "disabled" }); |
| 65 | + } |
| 66 | + |
| 67 | + public void testTimestampYearYmdSlashFormat() throws IOException { |
| 68 | + // use a multi-line sample to ensure we are detecting ndjson format |
| 69 | + Map<String, Object> responseMap = executeAndVerifyRequest(""" |
| 70 | + "2025/07/10 10:30:35" |
| 71 | + "2025/07/10 10:31:42" |
| 72 | + "2025/07/10 10:32:15" |
| 73 | + """, ecsCompatibility); |
| 74 | + verifyTimestampDetected(responseMap, "date"); |
| 75 | + verifyTimestampFormat(responseMap, TIMESTAMP_YMD_TIMESTAMP_GROK_PATTERN, TIMESTAMP_YMD_JAVA_FORMATS); |
| 76 | + } |
| 77 | + |
| 78 | + public void testTimestampYearYmdSlashFormat_WithDotAndMillis() throws IOException { |
| 79 | + // use a multi-line sample to ensure we are detecting ndjson format |
| 80 | + Map<String, Object> responseMap = executeAndVerifyRequest(""" |
| 81 | + "2025/07/10 10:30:35.123" |
| 82 | + "2025/07/10 10:31:42.123" |
| 83 | + "2025/07/10 10:32:15.123" |
| 84 | + """, ecsCompatibility); |
| 85 | + verifyTimestampDetected(responseMap, "date"); |
| 86 | + verifyTimestampFormat( |
| 87 | + responseMap, |
| 88 | + TIMESTAMP_YMD_TIMESTAMP_GROK_PATTERN, |
| 89 | + "yyyy/MM/dd HH:mm:ss.SSS", |
| 90 | + "yyyy.MM.dd HH:mm:ss.SSS", |
| 91 | + "yyyy-MM-dd HH:mm:ss.SSS" |
| 92 | + ); |
| 93 | + } |
| 94 | + |
| 95 | + public void testTimestampYearYmdSlashFormat_WithSlashAndNanos() throws IOException { |
| 96 | + // use a multi-line sample to ensure we are detecting ndjson format |
| 97 | + Map<String, Object> responseMap = executeAndVerifyRequest(""" |
| 98 | + "2025/07/10 10:30:35,123456789" |
| 99 | + "2025/07/10 10:31:42,123456789" |
| 100 | + "2025/07/10 10:32:15,123456789" |
| 101 | + """, ecsCompatibility); |
| 102 | + verifyTimestampDetected(responseMap, "date_nanos"); |
| 103 | + verifyTimestampFormat( |
| 104 | + responseMap, |
| 105 | + TIMESTAMP_YMD_TIMESTAMP_GROK_PATTERN, |
| 106 | + "yyyy/MM/dd HH:mm:ss,SSSSSSSSS", |
| 107 | + "yyyy.MM.dd HH:mm:ss,SSSSSSSSS", |
| 108 | + "yyyy-MM-dd HH:mm:ss,SSSSSSSSS" |
| 109 | + ); |
| 110 | + } |
| 111 | + |
| 112 | + public void testTimestampYearYmdDotFormat() throws IOException { |
| 113 | + // use a multi-line sample to ensure we are detecting ndjson format |
| 114 | + Map<String, Object> responseMap = executeAndVerifyRequest(""" |
| 115 | + "2025.07.10 10:30:35" |
| 116 | + "2025.07.10 10:31:42" |
| 117 | + "2025.07.10 10:32:15" |
| 118 | + """, ecsCompatibility); |
| 119 | + verifyTimestampDetected(responseMap, "date"); |
| 120 | + verifyTimestampFormat(responseMap, TIMESTAMP_YMD_TIMESTAMP_GROK_PATTERN, TIMESTAMP_YMD_JAVA_FORMATS); |
| 121 | + } |
| 122 | + |
| 123 | + public void testIso08601TimestampFormat() throws IOException { |
| 124 | + // use a multi-line sample to ensure we are detecting ndjson format |
| 125 | + Map<String, Object> responseMap = executeAndVerifyRequest(""" |
| 126 | + "2025-07-10 10:30:35" |
| 127 | + "2025-07-10 10:31:42" |
| 128 | + "2025-07-10 10:32:15" |
| 129 | + """, ecsCompatibility); |
| 130 | + verifyTimestampDetected(responseMap, "date"); |
| 131 | + // ISO_8601 should have higher priority than TIMESTAMP_YMD |
| 132 | + verifyTimestampFormat(responseMap, ISO_08601_TIMESTAMP_GROK_PATTERN, ISO_08601_JAVA_FORMATS); |
| 133 | + } |
| 134 | + |
| 135 | + public void testMonthExplicitNameFormat() throws IOException { |
| 136 | + // use a multi-line sample to ensure we are detecting ndjson format |
| 137 | + Map<String, Object> responseMap = executeAndVerifyRequest(""" |
| 138 | + "Aug 9, 2025" |
| 139 | + "Aug 10, 2025" |
| 140 | + "Aug 11, 2025" |
| 141 | + """, ecsCompatibility); |
| 142 | + verifyTimestampDetected(responseMap, "date"); |
| 143 | + verifyTimestampFormat(responseMap, "CUSTOM_TIMESTAMP", MONTH_EXPLICIT_NAME_JAVA_FORMATS); |
| 144 | + } |
| 145 | + |
| 146 | + private static Map<String, Object> executeAndVerifyRequest(String sample, String ecsCompatibility) throws IOException { |
| 147 | + Request request = new Request("POST", "/_text_structure/find_structure"); |
| 148 | + request.addParameter("ecs_compatibility", ecsCompatibility); |
| 149 | + request.setEntity(new StringEntity(sample, ContentType.APPLICATION_JSON)); |
| 150 | + Response response = client().performRequest(request); |
| 151 | + assertOK(response); |
| 152 | + return entityAsMap(response); |
| 153 | + } |
| 154 | + |
| 155 | + private static void verifyTimestampDetected(Map<String, Object> responseMap, String expectedType) { |
| 156 | + @SuppressWarnings("unchecked") |
| 157 | + Map<String, Object> mappings = (Map<String, Object>) responseMap.get("mappings"); |
| 158 | + assertThat(mappings, hasKey("properties")); |
| 159 | + @SuppressWarnings("unchecked") |
| 160 | + Map<String, Object> properties = (Map<String, Object>) mappings.get("properties"); |
| 161 | + assertThat(properties, hasKey("@timestamp")); |
| 162 | + @SuppressWarnings("unchecked") |
| 163 | + Map<String, Object> timestamp = (Map<String, Object>) properties.get("@timestamp"); |
| 164 | + assertThat(timestamp.get("type"), equalTo(expectedType)); |
| 165 | + } |
| 166 | + |
| 167 | + private static void verifyTimestampFormat(Map<String, Object> responseMap, String expectedGrokPattern, String... expectedJavaFormats) { |
| 168 | + assertThat(responseMap, hasKey("java_timestamp_formats")); |
| 169 | + @SuppressWarnings("unchecked") |
| 170 | + List<String> javaTimestampFormats = (List<String>) responseMap.get("java_timestamp_formats"); |
| 171 | + assertThat(javaTimestampFormats, containsInAnyOrder(expectedJavaFormats)); |
| 172 | + String grokPattern = (String) responseMap.get("grok_pattern"); |
| 173 | + assertThat(grokPattern, containsString(expectedGrokPattern)); |
| 174 | + } |
| 175 | +} |
0 commit comments