Google ML Kit for Flutter

This project acts as a bridge between Flutter and Google's native ML Kit APIs. It uses Flutter Platform Channels to communicate between the Dart code and the native platform (Android/iOS).

Key Concepts:

• No Dart-side Processing: No Machine Learning processing is performed in Flutter/Dart. All calls are passed to the native platform via MethodChannel (Android) or FlutterMethodChannel (iOS) and executed using Google's native APIs.
• Asynchronous Communication: Messages and responses are passed asynchronously to ensure the user interface remains responsive.
• Platform Support: Google's ML Kit APIs are developed natively only for iOS and Android. This plugin does not support Web or other platforms.

Debugging Guidance:

Because the processing happens on the native side, if you encounter issues with an ML model, you should first determine if the issue is with the plugin or the native API:

1. Run the official native example apps by Google for iOS and Android.
2. If the issue is reproducible in the native Google apps, report the issue to Google.
3. If the native apps work correctly but the Flutter plugin fails, report the issue to the plugin authors.

To use the translation plugin on iOS, you must meet the following requirements:

• Minimum iOS Deployment Target: 15.5
• Xcode: 15.3.0 or newer
• Swift: 5
• Architecture: ML Kit does not support 32-bit architectures (i386 and armv7). You must exclude armv7 in your Xcode build settings.

Steps to exclude armv7:

1. In Xcode, go to Project > Runner > Building Settings.
2. Locate Excluded Architectures > Any SDK and add armv7.

Podfile Configuration: Your Podfile must be configured to set the deployment target and exclude the unsupported architecture.

platform :ios, '15.5'  # or newer version

...

# add this line:
$iOSVersion = '15.5'  # or newer version

post_install do |installer|
  # add these lines:
  installer.pods_project.build_configurations.each do |config|
    config.build_settings["EXCLUDED_ARCHS[sdk=*]"] = "armv7"
    config.build_settings['IPHONEOS_DEPLOYMENT_TARGET'] = $iOSVersion
  end

  installer.pods_project.targets.each do |target|
    flutter_additional_ios_build_settings(target)

    # add these lines:
    target.build_configurations.each do |config|
      if Gem::Version.new($iOSVersion) > Gem::Version.new(config.build_settings['IPHONEOS_DEPLOYMENT_TARGET'])
        config.build_settings['IPHONEOS_DEPLOYMENT_TARGET'] = $iOSVersion
      end
    end
  end
end

The Prompt class is the primary interface for interacting with Google's ML Kit GenAI Prompt API. You can use it to generate text from simple text prompts or multimodal prompts (text combined with an image).

Workflow:

1. Initialize: Create an instance of Prompt.
2. Check Status: Use checkFeatureStatus() to see if the feature is available or downloadable. If it is downloadable, you must call downloadFeature() before running inference.
3. Run Inference: Use runInference() with your text prompt.
4. Cleanup: Call close() to release resources.

final prompt = Prompt();

// 1. Check status and handle downloads if necessary
final status = await prompt.checkFeatureStatus();
if (status == FeatureStatus.downloadable) {
  await prompt.downloadFeature(
    onDownloadCompleted: () {
      // Start prompt generation
    },
  );
} else if (status == FeatureStatus.available) {
  // Start prompt generation
}

// 2. Generate text
final text = "Write a 3 sentence story about a magical dog.";
final response = await prompt.runInference(text);
print('Response: $response');

// 3. Release resources
prompt.close();

To use google_mlkit_barcode_scanning on iOS, you must meet the following requirements:

• Minimum iOS Deployment Target: 15.5
• Xcode: 15.3.0 or newer
• Swift: 5
• Architecture: ML Kit does not support 32-bit architectures (i386 and armv7). You must exclude armv7 in your Xcode build settings.

Steps to exclude armv7: Go to Project > Runner > Building Settings > Excluded Architectures > Any SDK > armv7.

Podfile Configuration: Ensure your Podfile sets the platform to 15.5 or newer and includes the necessary post_install hooks to exclude armv7 and enforce the deployment target across all pods.

platform :ios, '15.5'  # or newer version

...

# add this line:
$iOSVersion = '15.5'  # or newer version

post_install do |installer|
  # add these lines:
  installer.pods_project.build_configurations.each do |config|
    config.build_settings["EXCLUDED_ARCHS[sdk=*]"] = "armv7"
    config.build_settings['IPHONEOS_DEPLOYMENT_TARGET'] = $iOSVersion
  end

  installer.pods_project.targets.each do |target|
    flutter_additional_ios_build_settings(target)

    # add these lines:
    target.build_configurations.each do |config|
      if Gem::Version.new($iOSVersion) > Gem::Version.new(config.build_settings['IPHONEOS_DEPLOYMENT_TARGET'])
        config.build_settings['IPHONEOS_DEPLOYMENT_TARGET'] = $iOSVersion
      end
    end
  end
end

You can use a custom TensorFlow Lite model bundled with your app assets.

1. Add the .tflite model to your pubspec.yaml assets.
2. Android Setup: Add the following to your app/build.gradle to prevent Gradle from compressing the model file:

   android {
       aaptOptions {
           noCompress "tflite"
       }
   }

3. Copy the model from assets to the application support directory using a helper function.
4. Initialize ImageLabeler using LocalLabelerOptions with the model's file path.

// Helper to get model path
Future<String> getModelPath(String asset) async {
  final path = '${(await getApplicationSupportDirectory()).path}/$asset';
  await Directory(dirname(path)).create(recursive: true);
  final file = File(path);
  if (!await file.exists()) {
    final byteData = await rootBundle.load(asset);
    await file.writeAsBytes(byteData.buffer
            .asUint8List(byteData.offsetInBytes, byteData.lengthInBytes));
  }
  return file.path;
}

// Usage
final modelPath = await getModelPath('assets/ml/object_labeler.tflite');
final options = LocalLabelerOptions(
  confidenceThreshold: confidenceThreshold,
  modelPath: modelPath,
);
final imageLabeler = ImageLabeler(options: options);