Files
cherrypick/cherrypick_generator/README.md
Sergey Penkovsky a9101513e1 Add CI/CD badges to package README files
- Added Melos + FVM CI badges to all package README.md files
- Standardized badge formatting across all packages
- Improved project visibility with build status indicators
2025-09-08 17:16:51 +03:00

208 lines
6.0 KiB
Markdown

[![Melos + FVM CI](https://github.com/pese-git/cherrypick/actions/workflows/pipeline.yml/badge.svg)](https://github.com/pese-git/cherrypick/actions/workflows/pipeline.yml)
---
# Cherrypick Generator
**Cherrypick Generator** is a Dart code generation library for automating dependency injection (DI) boilerplate. It processes classes and fields annotated with [cherrypick_annotations](https://pub.dev/packages/cherrypick_annotations) and generates registration code for services, modules, and field injection for classes marked as `@injectable`. It supports advanced DI features such as scopes, named bindings, parameters, and asynchronous dependencies.
---
## Features
- **Automatic Field Injection:**
Detects classes annotated with `@injectable()`, and generates mixins to inject all fields annotated with `@inject()`, supporting scope and named qualifiers.
- **Module and Service Registration:**
For classes annotated with `@module()`, generates service registration code for methods using annotations such as `@provide`, `@instance`, `@singleton`, `@named`, and `@params`.
- **Scope & Named Qualifier Support:**
Supports advanced DI features:
  • Field-level scoping with `@scope('scopename')`
  • Named dependencies via `@named('value')`
- **Synchronous & Asynchronous Support:**
Handles both synchronous and asynchronous services (including `Future<T>`) for both field injection and module registration.
- **Parameters and Runtime Arguments:**
Recognizes and wires both injected dependencies and runtime parameters using `@params`.
- **Error Handling:**
Validates annotations at generation time. Provides helpful errors for incorrect usage (e.g., using `@injectable` on non-class elements).
---
## How It Works
### 1. Annotate your code
Use annotations from [cherrypick_annotations](https://pub.dev/packages/cherrypick_annotations):
- `@injectable()` — on classes to enable field injection
- `@inject()` — on fields to specify they should be injected
- `@scope()`, `@named()` — on fields or parameters for advanced wiring
- `@module()` — on classes to mark as DI modules
- `@provide`, `@instance`, `@singleton`, `@params` — on methods and parameters for module-based DI
### 2. Run the generator
Use `build_runner` to process your code and generate `.module.cherrypick.g.dart` and `.inject.cherrypick.g.dart` files.
### 3. Use the output in your application
- For modules: Register DI providers using the generated `$YourModule` class.
- For services: Enable field injection on classes using the generated mixin.
---
## Field Injection Example
Given the following:
```dart
import 'package:cherrypick_annotations/cherrypick_annotations.dart';
@injectable()
class MyWidget with _$MyWidget {
@inject()
late final AuthService auth;
@inject()
@scope('profile')
late final ProfileManager manager;
@inject()
@named('special')
late final ApiClient specialApi;
}
```
**The generator will output (simplified):**
```dart
mixin _$MyWidget {
void _inject(MyWidget instance) {
instance.auth = CherryPick.openRootScope().resolve<AuthService>();
instance.manager = CherryPick.openScope(scopeName: 'profile').resolve<ProfileManager>();
instance.specialApi = CherryPick.openRootScope().resolve<ApiClient>(named: 'special');
}
}
```
You can then mix this into your widget to enable automatic DI at runtime.
---
## Module Registration Example
Given:
```dart
import 'package:cherrypick_annotations/cherrypick_annotations.dart';
@module()
class MyModule {
@singleton
@instance
AuthService provideAuth(Api api);
@provide
@named('logging')
Future<Logger> provideLogger(@params Map<String, dynamic> args);
}
```
**The generator will output (simplified):**
```dart
final class $MyModule extends MyModule {
@override
void builder(Scope currentScope) {
bind<AuthService>()
.toInstance(provideAuth(currentScope.resolve<Api>()))
.singleton();
bind<Logger>()
.toProvideAsyncWithParams((args) => provideLogger(args))
.withName('logging');
}
}
```
---
## Key Points
- **Rich Annotation Support:**
Mix and match field, parameter, and method annotations for maximum flexibility.
- **Scope and Named Resolution:**
Use `@scope('...')` and `@named('...')` to precisely control where and how dependencies are wired.
- **Async/Synchronous:**
The generator distinguishes between sync (`resolve<T>`) and async (`resolveAsync<T>`) dependencies.
- **Automatic Mixins:**
For classes with `@injectable()`, a mixin is generated that injects all relevant fields (using constructor or setter).
- **Comprehensive Error Checking:**
Misapplied annotations (e.g., `@injectable()` on non-class) produce clear build-time errors.
---
## Usage
1. **Add dependencies**
```yaml
dependencies:
cherrypick_annotations: ^latest
dev_dependencies:
cherrypick_generator: ^latest
build_runner: ^2.1.0
```
2. **Annotate your classes and modules as above**
3. **Run the generator**
```shell
dart run build_runner build
# or, if using Flutter:
flutter pub run build_runner build
```
4. **Use generated code**
- Import the generated `.inject.cherrypick.g.dart` or `.cherrypick.g.dart` files where needed
---
## Advanced Usage
- **Combining Modules and Field Injection:**
It's possible to mix both style of DI — modules for binding, and field injection for consuming services.
- **Parameter and Named Injection:**
Use `@named` on both provider and parameter for named registration and lookup; use `@params` to pass runtime arguments.
- **Async Factories:**
Methods returning Future<T> generate async bindings and async field resolution logic.
---
## Developer Notes
- The generator relies on the Dart analyzer, `source_gen`, and `build` packages.
- All classes and methods are parsed for annotations.
- Improper annotation usage will result in generator errors.
---
## License
```
Licensed under the Apache License, Version 2.0
```
---
## Contribution
Pull requests and issues are welcome! Please open GitHub issues or submit improvements.
---