DecoratR

Intuitive .NET library for implementing the Decorator pattern with Microsoft's Dependency Injection container. DecoratR provides a fluent API to chain decorators around your services, enabling cross-cutting concerns like logging, caching, retry logic, and more.

Table of Contents

• Installation
• Quick Start
• Core Concepts
  • Decorator Chain Order
  • Constructor Requirements
• Basic Usage
  • Simple Decorator Chain
  • Conditional Decoration
• Keyed Services
  • Complex Keys
• Custom Factory Methods
  • Advanced Factory Examples
• Generic Decorators
• Lifetime Management
• Best Practices
  • 1. Decorator Ordering
  • 2. Error Handling
  • 3. Performance Considerations
  • 4. Testing
• API Reference
  • Extension Methods
  • IDecorationBuilder Methods
• Requirements
• Contributing
• License

Installation

dotnet add package DecoratR

Quick Start

using DecoratR; using Microsoft.Extensions.DependencyInjection; var services = new ServiceCollection(); // Basic decoration services.Decorate<IUserService>() .With<LoggingDecorator>() .Then<CacheDecorator>() .Then<UserService>() .Apply(); var provider = services.BuildServiceProvider(); var userService = provider.GetService<IUserService>(); // Result: LoggingDecorator -> CacheDecorator -> UserService

Core Concepts

Decorator Chain Order

Decorators are applied in the order they are defined. The last decorator added should be your base implementation:

services.Decorate<IService>() .With<FirstDecorator>() // Outermost decorator .Then<SecondDecorator>() // Middle decorator .Then<BaseService>() // Base implementation (innermost) .Apply();

Constructor Requirements

Decorators (except the base implementation) must have a constructor that accepts the service type as the first parameter:

public class LoggingDecorator(IService inner) : IService { public string Execute() => $"Log({inner.Execute()})"; }

Basic Usage

Simple Decorator Chain

using System.Collections.Concurrent; // Service interface public interface IOrderService { Task<Order> GetOrderAsync(int orderId); } // Logging decorator public class LoggingDecorator(IOrderService inner) : IOrderService { public async Task<Order> GetOrderAsync(int orderId) { Console.WriteLine($"Getting order {orderId}"); return await inner.GetOrderAsync(orderId); } } // Cache decorator public class CacheDecorator(IOrderService inner) : IOrderService { private static readonly ConcurrentDictionary<int, Order> _cache = new(); public async Task<Order> GetOrderAsync(int orderId) { if (_cache.TryGetValue(orderId, out var cachedOrder)) return cachedOrder; var order = await inner.GetOrderAsync(orderId); _cache.TryAdd(orderId, order); return order; } } // Configuration services.Decorate<IOrderService>() .With<LoggingDecorator>() .Then<CacheDecorator>() .Then<OrderService>() .AsScoped() .Apply();

Conditional Decoration

Apply decorators based on runtime conditions:

services.Decorate<IOrderService>() .With<LoggingDecorator>() .ThenIf<RetryDecorator>(env.IsDevelopment()) // Only in development .ThenIf<CacheDecorator>(enableCaching) // Based on configuration .Then<OrderService>() .Apply();

Keyed Services

Note: Keyed services require .NET 8.0 or later. This feature is not available when targeting .NET 6.0 or 7.0.

DecoratR fully supports .NET 8+ keyed services, allowing you to create different decorator chains for the same service type:

// Different configurations for different contexts services.Decorate<IOrderService>("internal") .With<LoggingDecorator>() .Then<OrderService>() .Apply(); services.Decorate<IOrderService>("external") .With<LoggingDecorator>() .Then<RateLimitingDecorator>() .Then<SecurityDecorator>() .Then<OrderService>() .Apply(); services.Decorate<IOrderService>("cached") .With<CacheDecorator>() .Then<OrderService>() .AsSingleton() .Apply(); // Usage var internalService = provider.GetRequiredKeyedService<IOrderService>("internal"); var externalService = provider.GetRequiredKeyedService<IOrderService>("external"); var cachedService = provider.GetRequiredKeyedService<IOrderService>("cached");

Complex Keys

You can use any object as a key, including anonymous objects:

var productionKey = new { Environment = "Production", Version = "v2" }; services.Decorate<IOrderService>(productionKey) .With<AuditDecorator>() .Then<SecurityDecorator>() .Then<OrderService>() .Apply(); // Usage var service = provider.GetRequiredKeyedService<IOrderService>(productionKey);

Custom Factory Methods

For decorators that require complex initialization or dependencies not easily handled by the DI container:

services.AddSingleton<IMetrics, MetricsService>(); services.AddSingleton<IConfiguration, ConfigurationService>(); services.Decorate<IOrderService>() .With((serviceProvider, inner) => new MetricsDecorator( inner, serviceProvider.GetRequiredService<IMetrics>(), serviceProvider.GetRequiredService<IConfiguration>().GetValue<string>("MetricsPrefix"))) .Then<OrderService>() .Apply();

Advanced Factory Examples

// Conditional factory with complex logic services.Decorate<IOrderService>() .WithIf(enableAdvancedMetrics, (sp, inner) => { var config = sp.GetRequiredService<IConfiguration>(); var metrics = sp.GetRequiredService<IMetrics>(); var logger = sp.GetRequiredService<ILogger<AdvancedMetricsDecorator>>(); return new AdvancedMetricsDecorator(inner, metrics, config, logger); }) .Then<OrderService>() .Apply(); // Factory for base implementation services.Decorate<IOrderService>() .With<LoggingDecorator>() .Then((serviceProvider, _) => { var connectionString = serviceProvider.GetRequiredService<IConfiguration>() .GetConnectionString("DefaultConnection"); return new DatabaseOrderService(connectionString); }) .Apply();

Generic Decorators

DecoratR supports generic decorators with multiple type parameters:

// Generic decorator with one type parameter public class CacheDecorator<T>(IService<T> inner) : IService<T> { private static readonly Dictionary<string, T> _cache = new(); public async Task<T> GetAsync(string key) { if (_cache.TryGetValue(key, out var cachedValue)) return cachedValue; var value = await inner.GetAsync(key); _cache[key] = value; return value; } } // Multiple generic parameters public class TransformDecorator<TInput, TOutput>(ITransformService<TInput, TOutput> inner) : ITransformService<TInput, TOutput> { public async Task<TOutput> TransformAsync(TInput input) { Console.WriteLine($"Transforming {typeof(TInput).Name} to {typeof(TOutput).Name}"); return await inner.TransformAsync(input); } } // Configuration services.Decorate<IService<string>>() .With<CacheDecorator<string>>() .Then<StringService>() .Apply(); services.Decorate<ITransformService<User, UserDto>>() .With<TransformDecorator<User, UserDto>>() .Then<UserTransformService>() .Apply();

Lifetime Management

Control the lifetime of your decorated services:

// Singleton services.Decorate<IOrderService>() .With<CacheDecorator>() .Then<OrderService>() .AsSingleton() .Apply(); // Scoped services.Decorate<IOrderService>() .With<LoggingDecorator>() .Then<OrderService>() .AsScoped() .Apply(); // Transient (default) services.Decorate<IOrderService>() .With<RetryDecorator>() .Then<OrderService>() .AsTransient() // Optional, as it's the default .Apply(); // Custom lifetime services.Decorate<IOrderService>() .With<MetricsDecorator>() .Then<OrderService>() .WithLifetime(ServiceLifetime.Scoped) .Apply();

Best Practices

1. Decorator Ordering

Consider the logical flow of your decorators:

// Good: Logical order from outside to inside services.Decorate<IService>() .With<SecurityDecorator>() // Authentication/Authorization first .Then<RateLimitingDecorator>() // Rate limiting after security .Then<LoggingDecorator>() // Logging after rate limiting .Then<MetricsDecorator>() // Metrics collection .Then<CacheDecorator>() // Cache closest to data source .Then<BaseService>() // Base implementation .Apply();

2. Error Handling

Implement proper error handling in decorators:

public class ErrorHandlingDecorator(IService inner, ILogger<ErrorHandlingDecorator> logger) : IService { public async Task<Result> ExecuteAsync(Request request) { try { return await inner.ExecuteAsync(request); } catch (Exception ex) { logger.LogError(ex, "Error executing request {RequestId}", request.Id); throw; } } }

3. Performance Considerations

Be mindful of decorator overhead:

// Use conditional decoration for expensive operations var enableDetailedLogging = config.GetValue<bool>("Logging:Detailed"); services.Decorate<IService>() .WithIf<DetailedLoggingDecorator>(enableDetailedLogging) .Then<BaseService>() .Apply();

4. Testing

Decorators are easy to test in isolation:

[Test] public void LoggingDecorator_ShouldLogExecution() { // Arrange var mockInner = new Mock<IService>(); var mockLogger = new Mock<ILogger<LoggingDecorator>>(); var decorator = new LoggingDecorator(mockInner.Object, mockLogger.Object); // Act decorator.Execute(); // Assert mockLogger.Verify(x => x.LogInformation(It.IsAny<string>()), Times.Once); mockInner.Verify(x => x.Execute(), Times.Once); }

API Reference

Extension Methods

• Decorate<TService>() - Begin decoration for a service type
• Decorate<TService>(object serviceKey) - Begin decoration for a keyed service

IDecorationBuilder Methods

• Then<TDecorator>() - Add a decorator to the chain
• With<TDecorator>() - Alias for Then<TDecorator>()
• Then(Func<IServiceProvider, TService, TService> factory) - Add decorator via factory
• ThenIf<TDecorator>(bool condition) - Conditionally add decorator
• WithIf<TDecorator>(bool condition) - Alias for ThenIf<TDecorator>(bool condition)
• ThenIf(bool condition, Func<IServiceProvider, TService, TService> factory) - Conditionally add decorator via factory
• WithLifetime(ServiceLifetime lifetime) - Set service lifetime
• AsSingleton() - Set lifetime to Singleton
• AsScoped() - Set lifetime to Scoped
• AsTransient() - Set lifetime to Transient
• Apply() - Apply the decoration configuration

Requirements

• .NET 6.0 or later
• Microsoft.Extensions.DependencyInjection 6.0.0 or later

Note: Keyed services are only available in .NET 8.0 or later. If you're using .NET 6.0 or 7.0, you can only use the regular (non-keyed) decoration features.

Contributing

Contributions are welcome! Please feel free to submit issues and pull requests on the GitHub repository.

License

This project is licensed under the MIT License. See the LICENSE file for details.