Add support for overriding API version (#114)

This API will be required for authors to use to try function calling
while the backend support is still in version `v1beta`.

Add a RequestOptions class. For now this supports only the `apiVersion`
configuration. In the future this may add a `timeout` (#44) if implement
a deeper HTTP timeout functionality than `Future.timeout`.

pkgs/google_generative_ai/CHANGELOG.md

@@ -1,3 +1,8 @@
+## 0.2.4-wip
+
+- Allow specifying an API version in a `requestOptions` argument when
+  constructing a model.
+
 ## 0.2.3
 
 - Update the package version that is sent with the HTTP client name.

pkgs/google_generative_ai/lib/google_generative_ai.dart

@@ -66,4 +66,4 @@ export 'src/error.dart'
         InvalidApiKey,
         ServerException,
         UnsupportedUserLocation;
-export 'src/model.dart' show GenerativeModel;
+export 'src/model.dart' show GenerativeModel, RequestOptions;

pkgs/google_generative_ai/lib/src/model.dart

@@ -21,8 +21,8 @@ import 'client.dart';
 import 'content.dart';
 
 const _apiVersion = 'v1';
-final _googleAIBaseUri =
-    Uri.https('generativelanguage.googleapis.com', _apiVersion);
+Uri _googleAIBaseUri(RequestOptions? options) => Uri.https(
+    'generativelanguage.googleapis.com', options?.apiVersion ?? _apiVersion);
 
 enum Task {
   generateContent,
@@ -32,6 +32,19 @@ enum Task {
   batchEmbedContents;
 }
 
+/// Configuration for how a [GenerativeModel] makes requests.
+///
+/// This allows overriding the API version in use which may be required to use
+/// some beta features.
+final class RequestOptions {
+  /// The API version used to make requests.
+  ///
+  /// By default the version is `v1`. This may be specified as `v1beta` to use
+  /// beta features.
+  final String? apiVersion;
+  const RequestOptions({this.apiVersion});
+}
+
 /// A multimodel generative model (like Gemini).
 ///
 /// Allows generating content, creating embeddings, and counting the number of
@@ -73,13 +86,14 @@ final class GenerativeModel {
     List<SafetySetting> safetySettings = const [],
     GenerationConfig? generationConfig,
     http.Client? httpClient,
+    RequestOptions? requestOptions,
   }) =>
       GenerativeModel._withClient(
         client: HttpApiClient(apiKey: apiKey, httpClient: httpClient),
         model: model,
         safetySettings: safetySettings,
         generationConfig: generationConfig,
-        baseUri: _googleAIBaseUri,
+        baseUri: _googleAIBaseUri(requestOptions),
       );
 
   GenerativeModel._withClient({
@@ -243,17 +257,20 @@ final class GenerativeModel {
 /// Creates a model with an overridden [ApiClient] for testing.
 ///
 /// Package private test-only method.
-GenerativeModel createModelWithClient(
-        {required String model,
-        required ApiClient client,
-        List<SafetySetting> safetySettings = const [],
-        GenerationConfig? generationConfig}) =>
+GenerativeModel createModelWithClient({
+  required String model,
+  required ApiClient client,
+  List<SafetySetting> safetySettings = const [],
+  GenerationConfig? generationConfig,
+  RequestOptions? requestOptions,
+}) =>
     GenerativeModel._withClient(
-        client: client,
-        model: model,
-        safetySettings: safetySettings,
-        generationConfig: generationConfig,
-        baseUri: _googleAIBaseUri);
+      client: client,
+      model: model,
+      safetySettings: safetySettings,
+      generationConfig: generationConfig,
+      baseUri: _googleAIBaseUri(requestOptions),
+    );
 
 /// Creates a model with an overridden base URL to communicate with a different
 /// backend.

pkgs/google_generative_ai/lib/src/version.dart

@@ -12,4 +12,4 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-const packageVersion = '0.2.3';
+const packageVersion = '0.2.4-wip';

pkgs/google_generative_ai/pubspec.yaml

@@ -1,5 +1,5 @@
 name: google_generative_ai
-version: 0.2.3
+version: 0.2.4-wip
 description: >-
   The Google AI Dart SDK enables developers to use Google's state-of-the-art
   generative AI models (like Gemini).

pkgs/google_generative_ai/test/generative_model_test.dart

@@ -24,9 +24,10 @@ void main() {
     const defaultModelName = 'some-model';
 
     (StubClient, GenerativeModel) createModel(
-        [String modelName = defaultModelName]) {
+        [String modelName = defaultModelName, RequestOptions? requestOptions]) {
       final client = StubClient();
-      final model = createModelWithClient(model: modelName, client: client);
+      final model = createModelWithClient(
+          model: modelName, client: client, requestOptions: requestOptions);
       return (client, model);
     }
 
@@ -108,6 +109,46 @@ void main() {
           ], null)));
     });
 
+    test('allows specifying an API version', () async {
+      final (client, model) = createModel(
+          defaultModelName, RequestOptions(apiVersion: 'override_version'));
+      final prompt = 'Some prompt';
+      final result = 'Some response';
+      client.stub(
+        Uri.parse('https://generativelanguage.googleapis.com/override_version/'
+            'models/some-model:generateContent'),
+        {
+          'contents': [
+            {
+              'role': 'user',
+              'parts': [
+                {'text': prompt}
+              ]
+            }
+          ]
+        },
+        {
+          'candidates': [
+            {
+              'content': {
+                'role': 'model',
+                'parts': [
+                  {'text': result}
+                ]
+              }
+            }
+          ]
+        },
+      );
+      final response = await model.generateContent([Content.text(prompt)]);
+      expect(
+          response,
+          matchesGenerateContentResponse(GenerateContentResponse([
+            Candidate(
+                Content('model', [TextPart(result)]), null, null, null, null),
+          ], null)));
+    });
+
     group('generate unary content', () {
       test('can make successful request', () async {
         final (client, model) = createModel();