Language Helper

Multi-language app tool with an efficient generator and a custom GPT-4 translator for easy localization.

Features

• Easy to control the language translations in your application. Automatically uses the current device locale upon first open, if possible.

• You can completely control the translated text with LanguageConditions.

• Supports analyzing which text is missing in a specific language or is in your app but not in your language data, and vice versa.

• Supports extracting the needed text for translation from all .dart files in your project with a single command (Not using build_runner nor custom parser so it very fast and reliable).

• A Language Helper Translator on Chat GPT-4 that make it easier to translate the language data to a destination language.

Contents

• Set Up: Only this step is required while developing
  • Add The language_helper To The Project
  • Add An Empty LanguageHelper While Developing
  • Add .tr Or .trP To All Needed Strings
• Generator Flow Usage
  • Dart Map
  • JSON
• Manual Flow Usage
  • Create The Translations
  • Add To The Project
• Using LanguageBuilder To Update The Strings
• Control The Translation
  • Change The Language
  • Add A New Language Data
  • Get The List Of Supported Language Code
  • Listen To The Language Changing State
• Advanced Language Helper Generator
  • Modify The Input Path
  • Modify The Output Path
  • Convert From LanguageData to JSON
• Language Data Serialization
• Language Helper Translator (A Custom Chat GPT-4)
• Additional Information
• Contributions

Set Up

While developing, we just need to finish the Set Up steps. All other steps can be done when the app is ready to implement the localizations.

Add The language_helper To The Project

flutter pub add language_helper

Add An Empty LanguageHelper While Developing

final languageHelper = LanguageHelper.instance;

main() async {
  WidgetsFlutterBinding.ensureInitialized();
  await languageHelper.initial(data: []);
  runApp(const MyApp());
}

Add .tr Or .trP To All Needed Strings

Normal translation

Text('Translate this text'.tr)

Translate with parameters

Text('Hello @{name}'.trP({'name': name}))

Plural (Read Manual Flow Usage to know how to use it)

Text('We have @{number} dollar'.trP({'number': number}))

iOS configuration

Add the supported localizations to the Infi.plist. For instance with en and vi supported:

<key>CFBundleLocalizations</key>
<array>
   <string>en</string>
   <string>vi</string>
</array>

Generator Flow Usage

Dart Map

Generate:

dart run language_helper:generate

The data will be generated in this path by default:

|-- .lib
|   |--- resources
|   |    |--- language_helper
|   |    |    |--- language_data.dart
|   |    |    |--- languages
|   |    |    |    |--- _generated.dart   ; This file will be overwritten when re-generating

Implement to your project:

final languageHelper = LanguageHelper.instance;

final languageDataProvider = LanguageDataProvider.data(languageData);

main() async {
  await languageHelper.initial(
      data: [languageDataProvider],
  );

  runApp(const MyApp());
}

JSON

When using JSON, you can store your translation data in assets or on network

dart run language_helper:generate --json

The data will be generated in this path by default:

|-- assets
|   |--- resources
|   |    |--- language_helper
|   |    |    |--- codes.json
|   |    |    |--- languages
|   |    |    |    |--- _generated.json ; This file will be overwritten when re-generating

Implement to your project

Define the language data:

final languageHelper = LanguageHelper.instance;

// Network
final languageDataProvider = LanguageDataProvider.network('https://example.com/resources');

// Assets
final languageDataProvider = LanguageDataProvider.asset('assets/resources');

Add to the LanguageHelper instance:

final languageHelper = LanguageHelper.instance;

main() async {
  await languageHelper.initial(
      data: [languageDataProvider],
  );

  runApp(const MyApp());
}

Combine all of them to improve the translation:

main() async {
  await languageHelper.initial(
      data: [
        LanguageDataProvider.network('https://example.com/resources'),
        LanguageDataProvider.asset('assets/resources'),
        LanguageDataProvider.data(languageData),
      ],
  );

  runApp(const MyApp());
}

The package will get the translation data in order from top to bottom.

Manual Flow Usage

Create The Translations

Dart Map:

final en = {
  'Translate this text': 'Translate this text',
  'Hello @{name}': 'Hello @{name}',
  'We have @{number} dollar': LanguageConditions(
    param: 'number',
    conditions: {
      '0': 'We have zero dollar',
      '1': 'We have one dollar',

      // Default value.
      '_': 'We have @{number} dollars',
    }
  ),
};

const vi = {
  'Translate this text': 'Dịch chữ này',
  'Hello @{name}': 'Xin chào @{name}',
  'We have @{number} dollar': 'Chúng ta có @{number} đô-la', 
};

LanguageData languageData = {
  LanguageCodes.en: en,
  LanguageCodes.vi: vi,
};

final languageDataProvider = LanguageDataProvider.data(languageData);

With LanguageConditions, you can completely control which text is returned according to the parameters' conditions. You can use 'default' or '_' to set the default value for the condition.

JSON:

assets/resources/language_helper/codes.json: Contains all language codes

["en", "vi"]

assets/resources/language_helper/languages/en.json:

{
  "Translate this text": "Translate this text",
  "Hello @{name}": "Hello @{name}",
  "We have @{number} dollar": {
    "param": "number",
    "conditions": {
      "0": "We have zero dollar",
      "1": "We have one dollar",

      // Default value.
      "_": "We have @{number} dollars",
    }
  }
}

assets/resources/language_helper/languages/vi.json:

{
  "Translate this text": "Dịch chữ này",
  "Hello @{name}": "Xin chào @{name}",
  "We have @{number} dollar": "Chúng ta có @{number} đô-la", 
}

Remember to add those files to the pubspec.yaml:

flutter:
  assets:
    - assets/resources/language_helper/codes.json
    - assets/resources/language_helper/languages/

final languageDataProvider = LanguageDataProvider.asset('assets/resources');

Add To The Project

final languageHelper = LanguageHelper.instance;

main() async {
  await languageHelper.initial(
      data: [languageDataProvider],
  );

  runApp(const MyApp());
}

Using LanguageBuilder To Update The Strings

In the MaterialApp

class App extends StatelessWidget {
  const App({super.key});

  @override
  Widget build(BuildContext context) {
    return LanguageBuilder(
      builder: (context) {
        return MaterialApp(
          localizationsDelegates: languageHelper.delegates,
          supportedLocales: languageHelper.locales,
          locale: languageHelper.locale,
          home: const HomePage(),
        );
      }
    );
  }
}

In your Widgets

Using LanguageBuilder:

LanguageBuilder(
    builder: (context) {
        return Scaffold(
          body: Column(
            children: [
              Text('Hello @{name}'.tr),
              Text('We have @{number} dollar'.tr),
            ],
          ),
        );
    },
),

Using Tr (A short version of LanguageBuilder):

Tr((_) => Text('Hello @{name}'.tr)),

Control The Translation

Change The Language

languageHelper.change(LanguageCodes.vi);

Add A New Language Data

languageHelper.addData(LanguageDataProvider.data(newLanguageData));
languageHelper.addDataOverrides(LanguageDataProvider.data(newLanguageDataOverrides));

The addData and addDataOverrides have activate parameter which automaticaly rebuild all needed LanguageBuilder, so notice that you may get the setState issue because of the rebuilding of the LanguageBuilder when it's still building. If the error occurs, you may need to set it to false and activate the new data yourself by using reload method.

Get The List Of Supported Language Code

// List of [LanguageCodes] from both of the [data] and [dataOverrides] without duplicated
final codes = languageHelper.codes;

// List of [LanguageCodes] from the [dataOverrides]
final codesOverrides = languageHelper.codesOverrides;

Listen To The Language Changing State

Beside the onChanged callback, you can listen to the language changed events by using stream:

final sub = languageHelper.stream.listen((code) => print(code));

Note: Remember to sub.cancel() when it's not in use to avoid memory leaks.

Analyze The Translation

Currently works properly with LanguageDataProvider.data method

languageHelper.analyze();

This function will automatically be called in initial when isDebug is true.

Advanced Language Helper Generator

Modify The Input Path

Add --path option to your command:

dart run language_helper:generate --path=./lib

Modify The Output Path

Add --output option to your command:

dart run language_helper:generate --output=./lib/resources

Convert From LanguageData to JSON

• Create a bin folder in the same level with your lib.
• Create a export_json.dart file in your bin.
• Add this code to your export_json.dart:

void main() {
  test('', () {
    languageData.exportJson('./assets/resources');
  });
}

• Add the missing imports.
• Run flutter test ./bin/export_json.dart.
• The JSON will be generated in this path:

assets
|  |- language_helper
|  |  |- codes.json
|  |  |  |- languages
|  |  |  |  |- en.json
|  |  |  |  |- vi.json
|  |  |  |  |- ...

Language Data Serialization

Convert LanguageData to JSON:

final json = data.toJson();

Convert JSON to LanguageData:

final data = LanguageDataSerializer.fromJson(json);

Language Helper Translator (A Custom Chat GPT-4)

• Assume that here is our language data:

final en = {
  'Hello @{name}': 'Hello @{name}',
  'We have @{number} dollar': LanguageConditions(
    param: 'number',
    conditions: {
      '0': 'We have zero dollar',
      '1': 'We have one dollar',

      // Default value.
      '_': 'We have @{number} dollars',
    }
  ),
};

• Go to Language Helper Translator. You should open a New Chat a few times to let the AI read the instructions carefully to improve the translation (just my personal experience).
• Use this template to translate the data. Be sure to replace [] with the appropriate infomation:

This is the translation of the [app/game] that [purpose of the app/game to help the AI understand the context]. Translate it into [destination language]:

final en = {
  'Hello @{name}': 'Hello @{name}',
  'We have @{number} dollar': LanguageConditions(
    param: 'number',
    conditions: {
      '0': 'We have zero dollar',
      '1': 'We have one dollar',

      // Default value.
      '_': 'We have @{number} dollars',
    }
  ),
};

• The GPT will keeps all keys and comments in their original text, positions them exactly as they appear in the source, keeps the @{param} and @param in their appropriate places during the translation.

Additional Information

• The app will try to use the Devicelocale to set the initialCode if it is not set. If the Devicelocale is unavailable, it will use the first language in data instead.

• No matter how many LanguageBuilder that you use, the plugin only rebuilds the outest (the root) widget of LanguageBuilder, so it significantly improves performance. If you want to force rebuild some Widget, you can set the forceRebuild parameter in the LanguageBuilder to true.

• The LanguageCodes contains all the common languages with additional information like name in English (englishName) and name in native language (nativeName).

• The @{param} works in all cases (We should use this way to avoid issues when translating with Language Helper Translator). The @param only work if the text ends with a white space, end of line, or end with a new line.

• The addData and addDataOverrides have activate parameter which automaticaly rebuild all needed LanguageBuilder, so notice that you may get the setState issue because of the rebuilding of the LanguageBuilder when it's still building. If the error occurs, you may need to set it to false and activate the new data yourself by using reload method.

• The assets data is preferred between assets and network because we still haven't a way to cache it.

• Use the isInitialized (bool) and ensureInitialized (Future) to check whether the initial is run.

Contributions

As the project is currently in its early stages, it may contain bugs or other issues. Should you experience any problems, we kindly ask that you file an issue to let us know. Additionally, we welcome contributions in the form of pull requests (PRs) to help enhance the project.