[Backport 2.x] Add support for scored named queries (#11626) (#12427)

* Add support for scored named queries (#11626) Opensearch already support labelling the queries, that returns as a list in the returned results, of which query it matched. However one of the use case while doing hybrid search with query text and dense vector is to determine individual scores for each query type. This is very useful in further analysis and building offline model to generate better weights for ranking score. Hence adding this feature that sends the client to add the score for each matched query. --------- Signed-off-by: Dharin Shah <[email protected]> Signed-off-by: Dharin Shah <[email protected]> Co-authored-by: Dharin Shah <[email protected]> (cherry picked from commit 52b27f4) Signed-off-by: Andrew Ross <[email protected]> * Update version checks Signed-off-by: Andrew Ross <[email protected]> * Update test version guard Co-authored-by: Andriy Redko <[email protected]> Signed-off-by: Andrew Ross <[email protected]> --------- Signed-off-by: Dharin Shah <[email protected]> Signed-off-by: Dharin Shah <[email protected]> Signed-off-by: Andrew Ross <[email protected]> Co-authored-by: Dharin Shah <[email protected]> Co-authored-by: Dharin Shah <[email protected]> Co-authored-by: Andriy Redko <[email protected]>
opensearch-project · Feb 23, 2024 · 4e03b28 · 4e03b28
1 parent 44f1417
commit 4e03b28
Show file tree

Hide file tree

Showing 23 changed files with 672 additions and 105 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -10,6 +10,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - Introduce query level setting `index.query.max_nested_depth` limiting nested queries ([#3268](https://github.com/opensearch-project/OpenSearch/issues/3268)
 - Add toString methods to MultiSearchRequest, MultiGetRequest and CreateIndexRequest ([#12163](https://github.com/opensearch-project/OpenSearch/pull/12163))
 - Fix error in RemoteSegmentStoreDirectory when debug logging is enabled ([#12328](https://github.com/opensearch-project/OpenSearch/pull/12328))
+- Support for returning scores in matched queries ([#11626](https://github.com/opensearch-project/OpenSearch/pull/11626))
 
 ### Dependencies
 - Bump `com.squareup.okio:okio` from 3.7.0 to 3.8.0 ([#12290](https://github.com/opensearch-project/OpenSearch/pull/12290))

diff --git a/release-notes/opensearch.release-notes-2.12.0.md b/release-notes/opensearch.release-notes-2.12.0.md
@@ -52,7 +52,6 @@
 - Remove concurrent segment search feature flag for GA launch ([#12074](https://github.com/opensearch-project/OpenSearch/pull/12074))
 - Enable Fuzzy codec for doc id fields using a bloom filter ([#11022](https://github.com/opensearch-project/OpenSearch/pull/11022))
 - [Metrics Framework] Adds support for Histogram metric ([#12062](https://github.com/opensearch-project/OpenSearch/pull/12062))
-- Support for returning scores in matched queries ([#11626](https://github.com/opensearch-project/OpenSearch/pull/11626))
 
 ### Dependencies
 - Bumps jetty version to 9.4.52.v20230823 to fix GMS-2023-1857 ([#9822](https://github.com/opensearch-project/OpenSearch/pull/9822))

diff --git a/rest-api-spec/src/main/resources/rest-api-spec/api/search.json b/rest-api-spec/src/main/resources/rest-api-spec/api/search.json
@@ -229,6 +229,11 @@
       "search_pipeline": {
         "type": "string",
         "description": "The search pipeline to use to execute this request"
+      },
+      "include_named_queries_score":{
+        "type": "boolean",
+        "description":"Indicates whether hit.matched_queries should be rendered as a map that includes the name of the matched query associated with its score (true) or as an array containing the name of the matched queries (false)",
+        "default":false
       }
     },
     "body":{

diff --git a/rest-api-spec/src/main/resources/rest-api-spec/test/search/350_matched_queries.yml b/rest-api-spec/src/main/resources/rest-api-spec/test/search/350_matched_queries.yml
@@ -0,0 +1,103 @@
+setup:
+  - skip:
+      version: " - 2.12.99"
+      reason: "implemented for versions post 2.12.99"
+
+---
+"matched queries":
+  - do:
+      indices.create:
+        index: test
+
+  - do:
+      bulk:
+        refresh: true
+        body:
+          - '{ "index" : { "_index" : "test_1", "_id" : "1" } }'
+          - '{"field" : 1 }'
+          - '{ "index" : { "_index" : "test_1", "_id" : "2" } }'
+          - '{"field" : [1, 2] }'
+
+  - do:
+      search:
+        index: test_1
+        body:
+          query:
+            bool: {
+              should: [
+                {
+                  match: {
+                    field: {
+                      query: 1,
+                      _name: match_field_1
+                    }
+                  }
+                },
+                {
+                  match: {
+                    field: {
+                      query: 2,
+                      _name: match_field_2,
+                      boost: 10
+                    }
+                  }
+                }
+              ]
+            }
+
+  - match: {hits.total.value: 2}
+  - length: {hits.hits.0.matched_queries: 2}
+  - match: {hits.hits.0.matched_queries: [ "match_field_1", "match_field_2" ]}
+  - length: {hits.hits.1.matched_queries: 1}
+  - match: {hits.hits.1.matched_queries: [ "match_field_1" ]}
+
+---
+
+"matched queries with scores":
+  - do:
+      indices.create:
+        index: test
+
+  - do:
+      bulk:
+        refresh: true
+        body:
+          - '{ "index" : { "_index" : "test_1", "_id" : "1" } }'
+          - '{"field" : 1 }'
+          - '{ "index" : { "_index" : "test_1", "_id" : "2" } }'
+          - '{"field" : [1, 2] }'
+
+  - do:
+      search:
+        include_named_queries_score: true
+        index: test_1
+        body:
+          query:
+            bool: {
+              should: [
+                {
+                  match: {
+                    field: {
+                      query: 1,
+                      _name: match_field_1
+                    }
+                  }
+                },
+                {
+                  match: {
+                    field: {
+                      query: 2,
+                      _name: match_field_2,
+                      boost: 10
+                    }
+                  }
+                }
+              ]
+            }
+
+  - match: { hits.total.value: 2 }
+  - length: { hits.hits.0.matched_queries: 2 }
+  - match: { hits.hits.0.matched_queries.match_field_1: 1 }
+  - match: { hits.hits.0.matched_queries.match_field_2: 10 }
+  - length: { hits.hits.1.matched_queries: 1 }
+  - match: { hits.hits.1.matched_queries.match_field_1: 1 }
diff --git a/...rc/internalClusterTest/java/org/opensearch/indices/recovery/IndexPrimaryRelocationIT.java b/...rc/internalClusterTest/java/org/opensearch/indices/recovery/IndexPrimaryRelocationIT.java
@@ -66,19 +66,16 @@ public void testPrimaryRelocationWhileIndexing() throws Exception {
         ensureGreen("test");
         AtomicInteger numAutoGenDocs = new AtomicInteger();
         final AtomicBoolean finished = new AtomicBoolean(false);
-        Thread indexingThread = new Thread() {
-            @Override
-            public void run() {
-                while (finished.get() == false && numAutoGenDocs.get() < 10_000) {
-                    IndexResponse indexResponse = client().prepareIndex("test").setId("id").setSource("field", "value").get();
-                    assertEquals(DocWriteResponse.Result.CREATED, indexResponse.getResult());
-                    DeleteResponse deleteResponse = client().prepareDelete("test", "id").get();
-                    assertEquals(DocWriteResponse.Result.DELETED, deleteResponse.getResult());
-                    client().prepareIndex("test").setSource("auto", true).get();
-                    numAutoGenDocs.incrementAndGet();
-                }
+        Thread indexingThread = new Thread(() -> {
+            while (finished.get() == false && numAutoGenDocs.get() < 10_000) {
+                IndexResponse indexResponse = client().prepareIndex("test").setId("id").setSource("field", "value").get();
+                assertEquals(DocWriteResponse.Result.CREATED, indexResponse.getResult());
+                DeleteResponse deleteResponse = client().prepareDelete("test", "id").get();
+                assertEquals(DocWriteResponse.Result.DELETED, deleteResponse.getResult());
+                client().prepareIndex("test").setSource("auto", true).get();
+                numAutoGenDocs.incrementAndGet();
             }
-        };
+        });
         indexingThread.start();
 
         ClusterState initialState = client().admin().cluster().prepareState().get().getState();

diff --git a/...r/src/internalClusterTest/java/org/opensearch/search/fetch/subphase/MatchedQueriesIT.java b/...r/src/internalClusterTest/java/org/opensearch/search/fetch/subphase/MatchedQueriesIT.java
@@ -61,7 +61,9 @@
 import static org.opensearch.search.SearchService.CLUSTER_CONCURRENT_SEGMENT_SEARCH_SETTING;
 import static org.opensearch.test.hamcrest.OpenSearchAssertions.assertHitCount;
 import static org.hamcrest.Matchers.equalTo;
+import static org.hamcrest.Matchers.greaterThan;
 import static org.hamcrest.Matchers.hasItemInArray;
+import static org.hamcrest.Matchers.hasKey;
 
 public class MatchedQueriesIT extends ParameterizedStaticSettingsOpenSearchIntegTestCase {
 
@@ -95,15 +97,18 @@ public void testSimpleMatchedQueryFromFilteredQuery() throws Exception {
                             .should(rangeQuery("number").gte(2).queryName("test2"))
                     )
             )
+            .setIncludeNamedQueriesScore(true)
             .get();
         assertHitCount(searchResponse, 3L);
         for (SearchHit hit : searchResponse.getHits()) {
             if (hit.getId().equals("3") || hit.getId().equals("2")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("test2"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("test2"));
+                assertThat(hit.getMatchedQueryScore("test2"), equalTo(1f));
             } else if (hit.getId().equals("1")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("test1"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("test1"));
+                assertThat(hit.getMatchedQueryScore("test1"), equalTo(1f));
             } else {
                 fail("Unexpected document returned with id " + hit.getId());
             }
@@ -113,15 +118,18 @@ public void testSimpleMatchedQueryFromFilteredQuery() throws Exception {
             .setQuery(
                 boolQuery().should(rangeQuery("number").lte(2).queryName("test1")).should(rangeQuery("number").gt(2).queryName("test2"))
             )
+            .setIncludeNamedQueriesScore(true)
             .get();
         assertHitCount(searchResponse, 3L);
         for (SearchHit hit : searchResponse.getHits()) {
             if (hit.getId().equals("1") || hit.getId().equals("2")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("test1"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("test1"));
+                assertThat(hit.getMatchedQueryScore("test1"), equalTo(1f));
             } else if (hit.getId().equals("3")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("test2"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("test2"));
+                assertThat(hit.getMatchedQueryScore("test2"), equalTo(1f));
             } else {
                 fail("Unexpected document returned with id " + hit.getId());
             }
@@ -147,12 +155,15 @@ public void testSimpleMatchedQueryFromTopLevelFilter() throws Exception {
         assertHitCount(searchResponse, 3L);
         for (SearchHit hit : searchResponse.getHits()) {
             if (hit.getId().equals("1")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(2));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("name"));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("title"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(2));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("name"));
+                assertThat(hit.getMatchedQueryScore("name"), greaterThan(0f));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("title"));
+                assertThat(hit.getMatchedQueryScore("title"), greaterThan(0f));
             } else if (hit.getId().equals("2") || hit.getId().equals("3")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("name"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("name"));
+                assertThat(hit.getMatchedQueryScore("name"), greaterThan(0f));
             } else {
                 fail("Unexpected document returned with id " + hit.getId());
             }
@@ -168,12 +179,15 @@ public void testSimpleMatchedQueryFromTopLevelFilter() throws Exception {
         assertHitCount(searchResponse, 3L);
         for (SearchHit hit : searchResponse.getHits()) {
             if (hit.getId().equals("1")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(2));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("name"));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("title"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(2));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("name"));
+                assertThat(hit.getMatchedQueryScore("name"), greaterThan(0f));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("title"));
+                assertThat(hit.getMatchedQueryScore("title"), greaterThan(0f));
             } else if (hit.getId().equals("2") || hit.getId().equals("3")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("name"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("name"));
+                assertThat(hit.getMatchedQueryScore("name"), greaterThan(0f));
             } else {
                 fail("Unexpected document returned with id " + hit.getId());
             }
@@ -197,9 +211,11 @@ public void testSimpleMatchedQueryFromTopLevelFilterAndFilteredQuery() throws Ex
         assertHitCount(searchResponse, 3L);
         for (SearchHit hit : searchResponse.getHits()) {
             if (hit.getId().equals("1") || hit.getId().equals("2") || hit.getId().equals("3")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(2));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("name"));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("title"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(2));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("name"));
+                assertThat(hit.getMatchedQueryScore("name"), greaterThan(0f));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("title"));
+                assertThat(hit.getMatchedQueryScore("title"), greaterThan(0f));
             } else {
                 fail("Unexpected document returned with id " + hit.getId());
             }
@@ -231,13 +247,15 @@ public void testRegExpQuerySupportsName() throws InterruptedException {
 
         SearchResponse searchResponse = client().prepareSearch()
             .setQuery(QueryBuilders.regexpQuery("title", "title1").queryName("regex"))
+            .setIncludeNamedQueriesScore(true)
             .get();
         assertHitCount(searchResponse, 1L);
 
         for (SearchHit hit : searchResponse.getHits()) {
             if (hit.getId().equals("1")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("regex"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("regex"));
+                assertThat(hit.getMatchedQueryScore("regex"), equalTo(1f));
             } else {
                 fail("Unexpected document returned with id " + hit.getId());
             }
@@ -252,15 +270,17 @@ public void testPrefixQuerySupportsName() throws InterruptedException {
         refresh();
         indexRandomForConcurrentSearch("test1");
 
-        SearchResponse searchResponse = client().prepareSearch()
+        var query = client().prepareSearch()
             .setQuery(QueryBuilders.prefixQuery("title", "title").queryName("prefix"))
-            .get();
+            .setIncludeNamedQueriesScore(true);
+        var searchResponse = query.get();
         assertHitCount(searchResponse, 1L);
 
         for (SearchHit hit : searchResponse.getHits()) {
             if (hit.getId().equals("1")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("prefix"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("prefix"));
+                assertThat(hit.getMatchedQueryScore("prefix"), equalTo(1f));
             } else {
                 fail("Unexpected document returned with id " + hit.getId());
             }
@@ -282,8 +302,9 @@ public void testFuzzyQuerySupportsName() throws InterruptedException {
 
         for (SearchHit hit : searchResponse.getHits()) {
             if (hit.getId().equals("1")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("fuzzy"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("fuzzy"));
+                assertThat(hit.getMatchedQueryScore("fuzzy"), greaterThan(0f));
             } else {
                 fail("Unexpected document returned with id " + hit.getId());
             }
@@ -300,13 +321,15 @@ public void testWildcardQuerySupportsName() throws InterruptedException {
 
         SearchResponse searchResponse = client().prepareSearch()
             .setQuery(QueryBuilders.wildcardQuery("title", "titl*").queryName("wildcard"))
+            .setIncludeNamedQueriesScore(true)
             .get();
         assertHitCount(searchResponse, 1L);
 
         for (SearchHit hit : searchResponse.getHits()) {
             if (hit.getId().equals("1")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("wildcard"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("wildcard"));
+                assertThat(hit.getMatchedQueryScore("wildcard"), equalTo(1f));
             } else {
                 fail("Unexpected document returned with id " + hit.getId());
             }
@@ -328,8 +351,9 @@ public void testSpanFirstQuerySupportsName() throws InterruptedException {
 
         for (SearchHit hit : searchResponse.getHits()) {
             if (hit.getId().equals("1")) {
-                assertThat(hit.getMatchedQueries().length, equalTo(1));
-                assertThat(hit.getMatchedQueries(), hasItemInArray("span"));
+                assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                assertThat(hit.getMatchedQueriesAndScores(), hasKey("span"));
+                assertThat(hit.getMatchedQueryScore("span"), greaterThan(0f));
             } else {
                 fail("Unexpected document returned with id " + hit.getId());
             }
@@ -363,11 +387,13 @@ public void testMatchedWithShould() throws Exception {
             assertHitCount(searchResponse, 2L);
             for (SearchHit hit : searchResponse.getHits()) {
                 if (hit.getId().equals("1")) {
-                    assertThat(hit.getMatchedQueries().length, equalTo(1));
-                    assertThat(hit.getMatchedQueries(), hasItemInArray("dolor"));
+                    assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                    assertThat(hit.getMatchedQueriesAndScores(), hasKey("dolor"));
+                    assertThat(hit.getMatchedQueryScore("dolor"), greaterThan(0f));
                 } else if (hit.getId().equals("2")) {
-                    assertThat(hit.getMatchedQueries().length, equalTo(1));
-                    assertThat(hit.getMatchedQueries(), hasItemInArray("elit"));
+                    assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+                    assertThat(hit.getMatchedQueriesAndScores(), hasKey("elit"));
+                    assertThat(hit.getMatchedQueryScore("elit"), greaterThan(0f));
                 } else {
                     fail("Unexpected document returned with id " + hit.getId());
                 }
@@ -391,7 +417,10 @@ public void testMatchedWithWrapperQuery() throws Exception {
         for (QueryBuilder query : queries) {
             SearchResponse searchResponse = client().prepareSearch().setQuery(query).get();
             assertHitCount(searchResponse, 1L);
-            assertThat(searchResponse.getHits().getAt(0).getMatchedQueries()[0], equalTo("abc"));
+            SearchHit hit = searchResponse.getHits().getAt(0);
+            assertThat(hit.getMatchedQueriesAndScores().size(), equalTo(1));
+            assertThat(hit.getMatchedQueriesAndScores(), hasKey("abc"));
+            assertThat(hit.getMatchedQueryScore("abc"), greaterThan(0f));
         }
     }
 }