Stabilize vector store usage

.github/workflows/pre-release-tests.yml

@@ -17,7 +17,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        version: ["3.0.0"]
+        version: ['latest']
     name: integration-tests-against-rc (dart ${{ matrix.version }})
     steps:
       - name: Checkout

lib/src/results/experimental_features.dart

@@ -11,12 +11,7 @@ part 'experimental_features.g.dart';
   createToJson: false,
 )
 class ExperimentalFeatures {
-  @JsonKey(name: 'vectorStore')
-  final bool vectorStore;
-
-  const ExperimentalFeatures({
-    required this.vectorStore,
-  });
+  const ExperimentalFeatures();
 
   factory ExperimentalFeatures.fromJson(Map<String, dynamic> src) {
     return _$ExperimentalFeaturesFromJson(src);
@@ -29,12 +24,7 @@ class ExperimentalFeatures {
   createFactory: false,
 )
 class UpdateExperimentalFeatures {
-  @JsonKey(name: 'vectorStore')
-  final bool? vectorStore;
-
-  const UpdateExperimentalFeatures({
-    this.vectorStore,
-  });
+  const UpdateExperimentalFeatures();
 
   Map<String, dynamic> toJson() => _$UpdateExperimentalFeaturesToJson(this);
 }

test/search_test.dart

@@ -1,5 +1,4 @@
 import 'package:meilisearch/meilisearch.dart';
-import 'package:meilisearch/src/results/experimental_features.dart';
 import 'package:test/test.dart';
 
 import 'utils/books.dart';
@@ -8,6 +7,8 @@ import 'utils/client.dart';
 import 'utils/wait_for.dart';
 
 void main() {
+  final openAiKeyValue = openAiKey;
+
   group('Search', () {
     setUpClient();
     late String uid;
@@ -542,55 +543,62 @@ void main() {
   });
 
   // Commented because of https://github.com/meilisearch/meilisearch-dart/issues/369
-  group('Experimental', () {
+  group('Vector search', () {
     setUpClient();
     late String uid;
     late MeiliSearchIndex index;
-    late ExperimentalFeatures features;
-    setUp(() async {
-      features = await client.http.updateExperimentalFeatures(
-        UpdateExperimentalFeatures(
-          vectorStore: true,
+    late IndexSettings settings;
+
+    setUpAll(() {
+      settings = IndexSettings(embedders: {
+        'default': OpenAiEmbedder(
+          model: 'text-embedding-3-small',
+          apiKey: openAiKeyValue,
+          documentTemplate: "a book titled '{{ doc.title }}'",
         ),
-      );
-      expect(features.vectorStore, true);
+      });
+    });
 
+    setUp(() async {
       uid = randomUid();
       index = await createIndexWithData(uid: uid, data: vectorBooks);
+      // Configure embedder before running vector search
+      await index.updateSettings(settings).waitFor(client: client);
     });
 
     test('vector search', () async {
-      final vector = [0, 1, 2];
+      // Create a vector with 1536 dimensions (filled with zeros for test purposes)
+      final vector = List.filled(1536, 0.0);
       final res = await index
           .search(
             null,
             SearchQuery(
               vector: vector,
+              hybrid: HybridSearch(
+                embedder: 'default',
+                semanticRatio: 1.0,
+              ),
             ),
           )
           .asSearchResult()
           .mapToContainer();
 
-      expect(res.vector, vector);
       expect(
         res.hits,
         everyElement(
-          isA<MeiliDocumentContainer<Map<String, dynamic>>>()
-              .having(
-                (p0) => p0.vectors,
-                'vectors',
-                isNotNull,
-              )
-              .having(
-                (p0) => p0.semanticScore,
-                'semanticScore',
-                isNotNull,
-              ),
+          isA<MeiliDocumentContainer<Map<String, dynamic>>>().having(
+            (p0) => p0.parsed,
+            'parsed',
+            isNotNull,
+          ),
         ),
       );
     });
-  }, skip: "Requires Experimental API");
-  final openAiKeyValue = openAiKey;
+  },
+      skip: openAiKeyValue == null || openAiKeyValue.isEmpty
+          ? "Requires OPEN_AI_API_KEY environment variable"
+          : null);
+
   group('Embedders', () {
     group(
       'Unit test',
@@ -665,9 +673,6 @@ void main() {
         });
 
         setUp(() async {
-          final features = await client.http.updateExperimentalFeatures(
-              UpdateExperimentalFeatures(vectorStore: true));
-          expect(features.vectorStore, true);
           uid = randomUid();
           index = await createBooksIndex(uid: uid);
         });

test/utils/books_data.dart

@@ -60,23 +60,43 @@ final vectorBooks = [
   {
     "id": 0,
     "title": "Across The Universe",
-    "_vectors": [0, 0.8, -0.2],
+    "_vectors": {
+      "default": {
+        "embeddings": [0, 0.8, -0.2],
+        "regenerate": false,
+      }
+    }
   },
   {
     "id": 1,
     "title": "All Things Must Pass",
-    "_vectors": [1, -0.2, 0],
+    "_vectors": {
+      "default": {
+        "embeddings": [1, -0.2, 0],
+        "regenerate": false,
+      }
+    }
   },
   {
     "id": 2,
     "title": "And Your Bird Can Sing",
-    "_vectors": [-0.2, 4, 6],
+    "_vectors": {
+      "default": {
+        "embeddings": [-0.2, 4, 6],
+        "regenerate": false,
+      }
+    }
   },
   {
     "id": 3,
     "title": "The Matrix",
-    "_vectors": [5, -0.5, 0.3],
-  },
+    "_vectors": {
+      "default": {
+        "embeddings": [5, -0.5, 0.3],
+        "regenerate": false,
+      }
+    },
+  }
 ];
 
 enum CSVHeaderTypes {