feat(genkit_vertexai): support Gemini and multimodal embedders

packages/genkit_google_genai/lib/common.dart

@@ -17,4 +17,11 @@ library;
 
 export 'src/api_client.dart';
 export 'src/common_plugin.dart';
+export 'src/generated/generativelanguage.dart'
+    show
+        BatchEmbedContentsRequest,
+        Content,
+        ContentEmbedding,
+        EmbedContentRequest,
+        Part;
 export 'src/model.dart';

packages/genkit_google_genai/lib/src/api_client.dart

@@ -40,6 +40,15 @@ class GenerativeLanguageBaseClient {
     return EmbedContentResponse.fromJson(res);
   }
 
+  Future<BatchEmbedContentsResponse> batchEmbedContents(
+    BatchEmbedContentsRequest request, {
+    required String model,
+  }) async {
+    final url = '$apiUrlPrefix$model:batchEmbedContents';
+    final res = await _call('POST', url, request.toJson());
+    return BatchEmbedContentsResponse.fromJson(res);
+  }
+
   Future<GenerateContentResponse> generateContent(
     GenerateContentRequest request, {
     required String model,

packages/genkit_google_genai/lib/src/generated/generativelanguage.dart

@@ -1769,7 +1769,7 @@ extension type ContentEmbedding._(Map<String, Object?> _data) {
 
   set values(List<double>? value) => _data['values'] = value;
 
-  /// This field stores the soft tokens tensor frame shape (e.g. [1, 1, 256, 2048]).
+  /// This field stores the soft tokens tensor frame shape (e.g. `[1, 1, 256, 2048]`).
   List<int>? get shape {
     final v = _data['shape'];
     if (v == null) return null;