利用依赖倒置和依赖注入实现应用程序解耦

首先，我们虚拟一个系统环境(e-Shop)，即网上商店的应用程序，其中有一个应用: 获取指定分类下的所有产品信息。我们按照传统的思路来实现。图1展示了系统的设计图。

新建一个类库工程EShop.Service. 然后添加相应的类到工程中。

namespace EShop.Service{    public class Product    {    }

    public class ProductRepository    {        public IList<Product> GetProductsByCategory(int categoryId)        {            IList<Product> products = new List<Product>();            //进行数据库操作            return products;        }    }

    public class ProductService    {        private ProductRepository productRepository;

        public ProductService()        {            productRepository = new ProductRepository();        }

        public IList<Product> GetProductsByCategory(int categoryId)        {            IList<Product> products;            string storageKey = String.Format("products_in_category_id_{0}", categoryId);            products = (List<Product>)HttpContext.Current.Cache.Get(storageKey);            if (products == null)            {                products = productRepository.GetProductsByCategory(categoryId);                HttpContext.Current.Cache.Insert(storageKey, products);            }            return products;        }    }}

从以上的程序段是否能发现不合理之处呢？我归纳了一下，大致有以下几点：

1. ProductService依赖于ProductRepository，一旦后者的API发生变化，则前者必然跟着发生变化。

2. 难以测试ProductService的方法，因为ProductRepository并未真正连接到数据库，另外还依赖于HttpContext，两者之间紧耦合。

3. 目前采用HttpContext来进行缓存，如果要更换缓存机制(如：Velocity或Memcached)，ProductService将进行更改。

基于以上几点不合理之处，我们将一一进行重构和优化。

针对1：我们采用依赖倒置原则(Dependency Inversion Principle)——依赖于抽象而不是具体实现来解决。我们加入了接口IProductRepository。

public interface IProductRepository{    IList<Product> GetProductsByCategory(int categoryId);}public class ProductRepository : IProductRepository{    //...}public class ProductService{    private IProductRepository productRepository;

    public ProductService()    {        productRepository = new ProductRepository();    }    //...}

针对2：我们采用依赖注入原则(Dependency Injection Principle)——通过将抽象注入到构造函数、方法或属性来解决。我们修改ProductService构造函数。

public class ProductService{    private IProductRepository productRepository;

    public ProductService(IProductRepository productRepository)    {        this.productRepository = productRepository;    }    //...}

针对3：我们采用适配器模式(Adapter Pattern)——转换已有接口与客户期望的目标接口使之兼容来解决。我们添加了接口ICacheStorage和类HttpContextCacheAdapter。

public interface ICacheStorage{    void Remove(string key);    void Store(string key, object data);    T Retrieve<T>(string key);}

public class HttpContextCacheAdapter : ICacheStoragepublic class HttpContextCacheAdapter : ICacheStorage{    public void Remove(string key)    {        HttpContext.Current.Cache.Remove(key);    }

    public void Store(string key, object data)    {        HttpContext.Current.Cache.Insert(key, data);    }

    public T Retrieve<T>(string key)    {        T item = (T)HttpContext.Current.Cache.Get(key);        if (item == null)        {            item = default(T);        }        return item;    }}

public class ProductService{    private IProductRepository productRepository;    private ICacheStorage cacheStorage;

    public ProductService(IProductRepository productRepository, ICacheStorage cacheStorage)    {        this.productRepository = productRepository;        this.cacheStorage = cacheStorage;    }

    public IList<Product> GetProductsByCategory(int categoryId)    {        IList<Product> products;        string storageKey = String.Format("products_in_category_id_{0}", categoryId);        products = cacheStorage.Retrieve<List<Product>>(storageKey);        if (products == null)        {            products = productRepository.GetProductsByCategory(categoryId);            cacheStorage.Store(storageKey, products);        }        return products;    }}

综合以上的解决方案，这里贴出重构后完整的代码：

namespace EShop.Service{    public class Product    {    }

    public interface IProductRepository    {        IList<Product> GetProductsByCategory(int categoryId);    }

    public class ProductRepository : IProductRepository    {        public IList<Product> GetProductsByCategory(int categoryId)        {            IList<Product> products = new List<Product>();            //进行数据库操作             return products;        }    }

    public interface ICacheStorage    {        void Remove(string key);        void Store(string key, object data);        T Retrieve<T>(string key);    }

    public class HttpContextCacheAdapter : ICacheStorage    {        public void Remove(string key)        {            HttpContext.Current.Cache.Remove(key);        }

        public void Store(string key, object data)        {            HttpContext.Current.Cache.Insert(key, data);        }

        public T Retrieve<T>(string key)        {            T item = (T)HttpContext.Current.Cache.Get(key);            if (item == null)            {                item = default(T);            }            return item;        }    }

    //有时为方便起见，我们定义空对象(空缓存适配器)    public class NullCacheAdapter : ICacheStorage    {        public void Remove(string key)        {        }

        public void Store(string key, object data)        {        }

        public T Retrieve<T>(string key)        {            return default(T);        }    }

    public class ProductService    {        private IProductRepository productRepository;        private ICacheStorage cacheStorage;

        public ProductService(IProductRepository productRepository, ICacheStorage cacheStorage)        {            this.productRepository = productRepository;            this.cacheStorage = cacheStorage;        }

        public IList<Product> GetProductsByCategory(int categoryId)        {            IList<Product> products;            string storageKey = String.Format("products_in_category_id_{0}", categoryId);            products = cacheStorage.Retrieve<List<Product>>(storageKey);            if (products == null)            {                products = productRepository.GetProductsByCategory(categoryId);                cacheStorage.Store(storageKey, products);            }            return products;        }    }}

附上重构后的设计图：