Adding User Behavior Insights functionality.

CHANGELOG.md

@@ -39,6 +39,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - [Search Pipeline] Handle default pipeline for multiple indices ([#13276](https://github.com/opensearch-project/OpenSearch/pull/13276))
 - Add support for deep copying SearchRequest ([#12295](https://github.com/opensearch-project/OpenSearch/pull/12295))
 - Support multi ranges traversal when doing date histogram rewrite optimization. ([#13317](https://github.com/opensearch-project/OpenSearch/pull/13317))
+- Add User Behavior Insights. ([#13545](https://github.com/opensearch-project/OpenSearch/issues/13545))
 
 ### Dependencies
 - Bump `org.apache.commons:commons-configuration2` from 2.10.0 to 2.10.1 ([#12896](https://github.com/opensearch-project/OpenSearch/pull/12896))

gradle/missing-javadoc.gradle

@@ -124,6 +124,7 @@ configure([
   project(":modules:repository-url"),
   project(":modules:systemd"),
   project(":modules:transport-netty4"),
+  project(":modules:ubi"),
   project(":plugins:analysis-icu"),
   project(":plugins:analysis-kuromoji"),
   project(":plugins:analysis-nori"),

modules/ubi/README.md

@@ -0,0 +1,59 @@
+# User Behavior Insights (UBI)
+
+UBI facilitates storing queries and events for the purposes of improving search relevance as descrbed by [[RFC] User Behavior Insights](https://github.com/opensearch-project/OpenSearch/issues/12084).
+
+## Indexing Queries
+
+For UBI to index a query, add a `ubi` block to the `ext` in the search request containing a `query_id`:
+
+```
+curl -s http://localhost:9200/ecommerce/_search -H "Content-type: application/json" -d'
+{
+  "query": {
+    "match": {
+      "title": "toner OR ink"
+    }
+  },
+  "ext": {
+    "ubi": {
+      "query_id": "1234512345"
+    }
+  }
+}
+```
+
+There are optional values that can be included in the `ubi` block along with the `query_id`. Those values are:
+* `client_id` - A unique identifier for the source of the query. This may represent a user or some other mechanism.
+* `user_query` - The user-entered query for this search. For example, in the search request above, the `user_query` may have been `toner ink`.
+
+With these optional values, a sample query would look like:
+
+```
+curl -s http://localhost:9200/ecommerce/_search -H "Content-type: application/json" -d'
+{
+  "query": {
+    "match": {
+      "title": "toner OR ink"
+    }
+  },
+  "ext": {
+    "ubi": {
+      "query_id": "1234512345",
+      "client_id": "abcdefg",
+      "user_query": "toner ink"
+    }
+  }
+}
+```
+
+If a search request does not contain a `ubi` block in `ext`, the query will *not* be indexed.
+
+Queries are indexed into an index called `ubi_queries`.
+
+## Indexing Events
+
+UBI facilitates indexing both queries and client-side events. These client-side events may be product clicks, scroll-depth,
+adding a product to a cart, or other actions. UBI indexes these events in an index called `ubi_events`. This index is
+automatically created the first time a query containing a `ubi` section in `ext` (example above).
+
+Client-side events can be indexed into the `ubi_events` index by your method of choice.

modules/ubi/build.gradle

@@ -0,0 +1,12 @@
+apply plugin: 'opensearch.yaml-rest-test'
+
+opensearchplugin {
+  description 'Integrates OpenSearch with systemd'
+  classname 'org.opensearch.ubi.UbiModulePlugin'
+}
+
+dependencies {
+  // required for the yaml test to run
+  yamlRestTestImplementation "org.apache.logging.log4j:log4j-core:${versions.log4j}"
+  runtimeOnly "org.apache.logging.log4j:log4j-core:${versions.log4j}"
+}

modules/ubi/src/main/java/org/opensearch/ubi/QueryRequest.java

@@ -0,0 +1,77 @@
+/*
+ * 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.opensearch.ubi;
+
+/**
+ * A query received by OpenSearch.
+ */
+public class QueryRequest {
+
+    private final long timestamp;
+    private final String queryId;
+    private final String userId;
+    private final String userQuery;
+    private final QueryResponse queryResponse;
+
+    /**
+     * Creates a query request.
+     * @param queryId The ID of the query.
+     * @param userQuery The user-entered query.
+     * @param userId The ID of the user that initiated the query.
+     * @param queryResponse The {@link QueryResponse} for this query request.
+     */
+    public QueryRequest(final String queryId, final String userQuery, final String userId, final QueryResponse queryResponse) {
+        this.timestamp = System.currentTimeMillis();
+        this.queryId = queryId;
+        this.userId = userId;
+        this.userQuery = userQuery;
+        this.queryResponse = queryResponse;
+    }
+
+    /**
+     * Gets the timestamp.
+     * @return The timestamp.
+     */
+    public long getTimestamp() {
+        return timestamp;
+    }
+
+    /**
+     * Gets the query ID.
+     * @return The query ID.
+     */
+    public String getQueryId() {
+        return queryId;
+    }
+
+    /**
+     * Gets the user query.
+     * @return The user query.
+     */
+    public String getUserQuery() {
+        return userQuery;
+    }
+
+    /**
+     * Gets the user ID.
+     * @return The user ID.
+     */
+    public String getUserId() {
+        return userId;
+    }
+
+    /**
+     * Gets the query response for this query request.
+     * @return The {@link QueryResponse} for this query request.
+     */
+    public QueryResponse getQueryResponse() {
+        return queryResponse;
+    }
+
+}

modules/ubi/src/main/java/org/opensearch/ubi/QueryResponse.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.opensearch.ubi;
+
+import java.util.List;
+
+/**
+ * A query response.
+ */
+public class QueryResponse {
+
+    private final String queryId;
+    private final String queryResponseId;
+    private final List<String> queryResponseObjectIds;
+
+    /**
+     * Creates a query response.
+     * @param queryId The ID of the query.
+     * @param queryResponseId The ID of the query response.
+     * @param queryResponseObjectIds A list of IDs for the hits in the query.
+     */
+    public QueryResponse(final String queryId, final String queryResponseId, final List<String> queryResponseObjectIds) {
+        this.queryId = queryId;
+        this.queryResponseId = queryResponseId;
+        this.queryResponseObjectIds = queryResponseObjectIds;
+    }
+
+    /**
+     * Gets the query ID.
+     * @return The query ID.
+     */
+    public String getQueryId() {
+        return queryId;
+    }
+
+    /**
+     * Gets the query response ID.
+     * @return The query response ID.
+     */
+    public String getQueryResponseId() {
+        return queryResponseId;
+    }
+
+    /**
+     * Gets the list of query response hit IDs.
+     * @return A list of query response hit IDs.
+     */
+    public List<String> getQueryResponseObjectIds() {
+        return queryResponseObjectIds;
+    }
+
+}