Add field type for version strings (#59773)

This PR adds a new 'version' field type that allows indexing string values
representing software versions similar to the ones defined in the Semantic
Versioning definition (semver.org). The field behaves very similar to a
'keyword' field but allows efficient sorting and range queries that take into
accound the special ordering needed for version strings. For example, the main
version parts are sorted numerically (ie 2.0.0 < 11.0.0) whereas this wouldn't
be possible with 'keyword' fields today.

Valid version values are similar to the Semantic Versioning definition, with the
notable exception that in addition to the "main" version consiting of
major.minor.patch, we allow less or more than three numeric identifiers, i.e.
"1.2" or "1.4.6.123.12" are treated as valid too.

Relates to #48878

Author: Christoph Büscher

docs/reference/mapping/types.asciidoc

@@ -51,6 +51,8 @@ Dates::                 Date types, including <<date,`date`>> and
 <<range,Range>>::   Range types, such as `long_range`, `double_range`,
                     `date_range`, and `ip_range`.
 <<ip,`ip`>>::       IPv4 and IPv6 addresses.
+<<version,Version>>::  Software versions. Supports https://semver.org/[Semantic Versioning]
+precedence rules.
 {plugins}/mapper-murmur3.html[`murmur3`]:: Compute and stores hashes of
 values.
 
@@ -148,6 +150,8 @@ include::types/geo-shape.asciidoc[]
 
 include::types/ip.asciidoc[]
 
+include::types/version.asciidoc[]
+
 include::types/parent-join.asciidoc[]
 
 include::types/keyword.asciidoc[]

docs/reference/mapping/types/keyword.asciidoc

@@ -129,3 +129,4 @@ The following parameters are accepted by `keyword` fields:
 include::constant-keyword.asciidoc[]
 
 include::wildcard.asciidoc[]
+

docs/reference/mapping/types/version.asciidoc

@@ -0,0 +1,70 @@
+[role="xpack"]
+[testenv="basic"]
+[[version]]
+=== Version field type
+++++
+<titleabbrev>Version</titleabbrev>
+++++
+
+The `version` field type is a specialization of the `keyword` field for
+handling software version values and to support specialized precedence
+rules for them. Precedence is defined following the rules outlined by
+https://semver.org/[Semantic Versioning], which for example means that
+major, minor and patch version parts are sorted numerically (i.e. 
+"2.1.0" < "2.4.1" < "2.11.2") and pre-release versions are sorted before
+release versions (i.e. "1.0.0-alpha" < "1.0.0").
+
+You index a `version` field as follows
+
+[source,console]
+--------------------------------------------------
+PUT my-index-000001
+{
+  "mappings": {
+    "properties": {
+      "my_version": {
+        "type": "version"
+      }
+    }
+  }
+}
+
+--------------------------------------------------
+
+The field offers the same search capabilities as a regular keyword field. It 
+can e.g. be searched for exact matches using `match` or `term` queries and
+supports prefix and wildcard searches. The main benefit is that `range` queries
+will honor Semver ordering, so a `range` query between "1.0.0" and "1.5.0"
+will include versions of "1.2.3" but not "1.11.2" for example. Note that this
+would be different when using a regular `keyword` field for indexing where ordering
+is alphabetical.
+
+Software versions are expected to follow the
+https://semver.org/[Semantic Versioning rules] schema and precedence rules with
+the notable exception that more or less than three main version identifiers are
+allowed (i.e. "1.2" or "1.2.3.4" qualify as valid versions while they wouldn't under
+strict Semver rules). Version strings that are not valid under the Semver definition
+(e.g. "1.2.alpha.4") can still be indexed and retrieved as exact matches, however they
+will all appear _after_ any valid version with regular alphabetical ordering. The empty
+String "" is considered invalid and sorted after all valid versions, but before other
+invalid ones.
+
+[discrete]
+[[version-params]]
+==== Parameters for version fields
+
+The following parameters are accepted by `version` fields:
+
+[horizontal]
+
+<<mapping-field-meta,`meta`>>::
+
+    Metadata about the field.
+
+[discrete]
+==== Limitations
+
+This field type isn't optimized for heavy wildcard, regex or fuzzy searches. While those
+type of queries work in this field, you should consider using a regular `keyword` field if
+you strongly rely on these kind of queries.
+

server/src/main/java/org/elasticsearch/index/mapper/TermBasedFieldType.java

@@ -34,9 +34,10 @@
 
 /** Base {@link MappedFieldType} implementation for a field that is indexed
  *  with the inverted index. */
-abstract class TermBasedFieldType extends SimpleMappedFieldType {
+public abstract class TermBasedFieldType extends SimpleMappedFieldType {
 
-    TermBasedFieldType(String name, boolean isSearchable, boolean hasDocValues, TextSearchInfo textSearchInfo, Map<String, String> meta) {
+    public TermBasedFieldType(String name, boolean isSearchable, boolean hasDocValues, TextSearchInfo textSearchInfo,
+        Map<String, String> meta) {
         super(name, isSearchable, hasDocValues, textSearchInfo, meta);
     }
 

x-pack/plugin/analytics/src/main/java/org/elasticsearch/xpack/analytics/stringstats/StringStatsAggregator.java

@@ -101,7 +101,7 @@ public void collect(int doc, long bucket) throws IOException {
                     for (int i = 0; i < valuesCount; i++) {
                         BytesRef value = values.nextValue();
                         if (value.length > 0) {
-                            String valueStr = value.utf8ToString();
+                            String valueStr = (String) format.format(value);
                             int length = valueStr.length();
                             totalLength.increment(bucket, length);
 

x-pack/plugin/src/test/resources/rest-api-spec/test/versionfield/10_basic.yml

@@ -0,0 +1,107 @@
+# Integration tests for the version field
+#
+---
+setup:
+
+  - skip:
+      features: headers
+      version: " - 7.99.99"
+      reason: "version field is added to 8.0 first"
+
+  - do:
+        indices.create:
+          index: test_index
+          body:
+              mappings:
+                properties:
+                  version:
+                    type: version
+
+  - do:
+        bulk:
+          refresh: true
+          body:
+            - '{ "index" : { "_index" : "test_index", "_id" : "1" } }'
+            - '{"version": "1.1.0" }'
+            - '{ "index" : { "_index" : "test_index", "_id" : "2" } }'
+            - '{"version": "2.0.0-beta" }'
+            - '{ "index" : { "_index" : "test_index", "_id" : "3" } }'
+            - '{"version": "3.1.0" }'
+
+---
+"Store malformed":
+  - do:
+        indices.create:
+          index: test_malformed
+          body:
+              mappings:
+                properties:
+                  version:
+                    type: version
+
+  - do:
+        bulk:
+          refresh: true
+          body:
+            - '{ "index" : { "_index" : "test_malformed", "_id" : "1" } }'
+            - '{"version": "1.1.0" }'
+            - '{ "index" : { "_index" : "test_malformed", "_id" : "2" } }'
+            - '{"version": "2.0.0-beta" }'
+            - '{ "index" : { "_index" : "test_malformed", "_id" : "3" } }'
+            - '{"version": "v3.1.0" }'
+            - '{ "index" : { "_index" : "test_malformed", "_id" : "4" } }'
+            - '{"version": "1.el6" }'
+            
+  - do:
+        search:
+          index: test_malformed
+          body:
+            query: { "match" : { "version" : "1.el6" } }
+
+  - do:
+        search:
+          index: test_malformed
+          body:
+            query: { "match_all" : { } }
+            sort:
+              version: asc
+
+  - match: { hits.total.value: 4 }
+  - match: { hits.hits.0._source.version: "1.1.0" }
+  - match: { hits.hits.1._source.version: "2.0.0-beta" }
+  - match: { hits.hits.2._source.version: "1.el6" }
+  - match: { hits.hits.3._source.version: "v3.1.0" }
+
+---
+"Basic ranges":
+  - do:
+        search:
+          index: test_index
+          body:
+            query: { "range" : { "version" : { "gt" : "1.1.0", "lt" : "9999" } } }
+
+  - match: { hits.total.value: 2 }
+
+  - do:
+        search:
+          index: test_index
+          body:
+            query: { "range" : { "version" : { "gte" : "1.1.0", "lt" : "9999" } } }
+
+  - match: { hits.total.value: 3 }
+  
+  - do:
+        search:
+          index: test_index
+          body:
+            query: { "range" : { "version" : { "gte" : "2.0.0", "lt" : "9999" } } }
+
+  - match: { hits.total.value: 1 }
+  
+  - do:
+        search:
+          index: test_index
+          body:
+            query: { "range" : { "version" : { "gte" : "2.0.0-alpha", "lt" : "9999" } } }
+
+  - match: { hits.total.value: 2 }

x-pack/plugin/src/test/resources/rest-api-spec/test/versionfield/20_scripts.yml

@@ -0,0 +1,54 @@
+# Integration tests for the version field
+#
+---
+setup:
+
+  - skip:
+      features: headers
+      version: " - 7.99.99"
+      reason: "version field is added to 8.0 first"
+
+  - do:
+        indices.create:
+          index: test_index
+          body:
+              mappings:
+                properties:
+                  version:
+                    type: version
+
+  - do:
+        bulk:
+          refresh: true
+          body:
+            - '{ "index" : { "_index" : "test_index", "_id" : "1" } }'
+            - '{"version": "1.1.12" }'
+            - '{ "index" : { "_index" : "test_index", "_id" : "2" } }'
+            - '{"version": "2.0.0-beta" }'
+            - '{ "index" : { "_index" : "test_index", "_id" : "3" } }'
+            - '{"version": "3.1.0" }'
+
+---
+"Filter script":
+  - do:
+        search:
+          index: test_index
+          body:
+            query: { "script" : { "script" : { "source": "doc['version'].value.length() > 5"} } }
+
+  - match: { hits.total.value: 2 }
+  - match: { hits.hits.0._source.version: "1.1.12" }
+  - match: { hits.hits.1._source.version: "2.0.0-beta" }
+
+---
+"Sort script":
+  - do:
+        search:
+          index: test_index
+          body:
+            sort: { "_script" : { "type" : "number", "script" : { "source": "doc['version'].value.length()" } } }
+
+  - match: { hits.total.value: 3 }
+  - match: { hits.hits.0._source.version: "3.1.0" }
+  - match: { hits.hits.1._source.version: "1.1.12" }
+  - match: { hits.hits.2._source.version: "2.0.0-beta" }

x-pack/plugin/versionfield/build.gradle

@@ -0,0 +1,20 @@
+evaluationDependsOn(xpackModule('core'))
+
+apply plugin: 'elasticsearch.esplugin'
+apply plugin: 'elasticsearch.internal-cluster-test'
+
+esplugin {
+  name 'versionfield'
+  description 'A plugin for a field type to store sofware versions'
+  classname 'org.elasticsearch.xpack.versionfield.VersionFieldPlugin'
+  extendedPlugins = ['x-pack-core', 'lang-painless']
+}
+archivesBaseName = 'x-pack-versionfield'
+
+dependencies {
+  compileOnly project(path: xpackModule('core'), configuration: 'default')
+  compileOnly project(':modules:lang-painless:spi')
+  compileOnly project(':modules:lang-painless')
+  testImplementation project(path: xpackModule('core'), configuration: 'testArtifacts')
+  testImplementation project(path: xpackModule('analytics'), configuration: 'default')
+}

x-pack/plugin/versionfield/src/internalClusterTest/java/org/elasticsearch/xpack/versionfield/VersionFieldIT.java

@@ -0,0 +1,76 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License;
+ * you may not use this file except in compliance with the Elastic License.
+ */
+
+package org.elasticsearch.xpack.versionfield;
+
+import org.elasticsearch.action.search.SearchResponse;
+import org.elasticsearch.common.xcontent.XContentFactory;
+import org.elasticsearch.plugins.Plugin;
+import org.elasticsearch.search.aggregations.AggregationBuilders;
+import org.elasticsearch.search.aggregations.bucket.terms.Terms;
+import org.elasticsearch.search.aggregations.bucket.terms.Terms.Bucket;
+import org.elasticsearch.test.ESIntegTestCase;
+import org.elasticsearch.xpack.core.LocalStateCompositeXPackPlugin;
+
+import java.util.Collection;
+import java.util.List;
+
+import static org.elasticsearch.common.xcontent.XContentFactory.jsonBuilder;
+
+public class VersionFieldIT extends ESIntegTestCase {
+
+    @Override
+    protected Collection<Class<? extends Plugin>> nodePlugins() {
+        return List.of(VersionFieldPlugin.class, LocalStateCompositeXPackPlugin.class);
+    }
+
+    public void testTermsAggregation() throws Exception {
+        String indexName = "test";
+        createIndex(indexName);
+
+        client().admin()
+            .indices()
+            .preparePutMapping(indexName)
+            .setSource(
+                XContentFactory.jsonBuilder()
+                    .startObject()
+                    .startObject("_doc")
+                    .startObject("properties")
+                    .startObject("version")
+                    .field("type", "version")
+                    .endObject()
+                    .endObject()
+                    .endObject()
+                    .endObject()
+            )
+            .get();
+        ensureGreen();
+
+        client().prepareIndex(indexName).setId("1").setSource(jsonBuilder().startObject().field("version", "1.0").endObject()).get();
+        client().prepareIndex(indexName).setId("2").setSource(jsonBuilder().startObject().field("version", "1.3.0").endObject()).get();
+        client().prepareIndex(indexName)
+            .setId("3")
+            .setSource(jsonBuilder().startObject().field("version", "2.1.0-alpha").endObject())
+            .get();
+        client().prepareIndex(indexName).setId("4").setSource(jsonBuilder().startObject().field("version", "2.1.0").endObject()).get();
+        client().prepareIndex(indexName).setId("5").setSource(jsonBuilder().startObject().field("version", "3.11.5").endObject()).get();
+        refresh();
+
+        // terms aggs
+        SearchResponse response = client().prepareSearch(indexName)
+            .addAggregation(AggregationBuilders.terms("myterms").field("version"))
+            .get();
+        Terms terms = response.getAggregations().get("myterms");
+        List<? extends Bucket> buckets = terms.getBuckets();
+
+        assertEquals(5, buckets.size());
+        assertEquals("1.0", buckets.get(0).getKey());
+        assertEquals("1.3.0", buckets.get(1).getKey());
+        assertEquals("2.1.0-alpha", buckets.get(2).getKey());
+        assertEquals("2.1.0", buckets.get(3).getKey());
+        assertEquals("3.11.5", buckets.get(4).getKey());
+    }
+}