CodingCat.Cache 0.1.1

CodingCat.Cache

Provides handy API to manage caches across different sources.

Using the IKeyBuilder

When doing the caches, sometimes the key to the cache is quite troubling us when a programmer uses contact-email to represent the company's contact email, but the other programmer uses the same key for the CS team email.

We then figured out the better solution maybe enforce all keys require a prefix such as var key = "PageAboutUs-contact-email"; or var key = "PageContactUs-contact-email";

Moreover, when a project needs a distributed cache server (e.g. Redis), but its budget cannot affort both the prod & staging, we would also like to prepend the environment to the keys.

With the IKeyBuilder, we could do this almost without the keyboard.

var environment = Environment.Current.ToString(); // -- PRODUCTION or STAGING or DEBUG
var usingKey = new KeyBuilder<Product>(environment) // -- or new KeyBuilder(this.GetType(), environment);
  .UseKey("mobile")
  .AddSegments("Apple", "iPhone XR"); // -- CodingCat.Demo.Models.Product-DEBUG-Apple-iPhone XR

or

var usingKey = new KeyBuilder<Product>("DEV")
  .UseKey("GetProductById")
  .AddSegment(productId);

or

var usingKey = new KeyBuilder<Product>(environment)
  .UseKey(nameof(GetAllFeaturedProducts)); // -- only the UseKey is required

Using the IStorage

To allow easier dependency injection, we would like to wrap the caching classes into an interface thus we could replace the dependencies without any worries. Currently we support MemoryCache and StackExchange.Redis

var storage = new Storage(TimeSpan.FromSeconds(60)); // -- Memory.Storage
storage.Add(usingKey, JsonConvert.SerializeObject(product));

var product = JsonConvert.DeserializeObject<Product>(
  storage.Get(usingKey)
);

storage.Delete(usingKey);

/// ---- ////
var usingKey = this.keyBuilder
  .UseKey(nameof(GetProductByProductId))
  .AddSegment(productId);
var product = JsonConvert.DeserializedObject<Product>(
  storage.Get(
    usingKey,
    () => JsonConvert.SerializeObject(
      this.Database.Products.GetById(productId)
    ) // -- callback if not found
  )
);

What if micro-services?

When we are coding for some micro-services, one case is the cache is already exists on the distributed cache server (e.g. Redis), however, we would also like to cache the data within the service itself, in order to avoid high traffic to the cache server when we have a million instances of the services running.

In this case, there is a IStorageManager which allows developers to consume more than 1 cache sources and control its behavior.

using MemoryStorage = CodingCat.Cache.Memory.Storage;
using RedisStorage = CodingCat.Cache.Redis.Storage;

....

public IStorageManager StorageManager { get; }

public Constructor() {
  var expiry = TimeSpan.FromSeconds(60);
  var memoryStorage = new MemoryStorage(expiry);
  var redisStorage = new RedisStorage(
    this.redisConnection.GetDatabase(),
    expiry
  );
  
  // -- Default: if the cache is found in fallback storages but not the default storage, only return the value but not saving into the default storage
  // -- SaveFromFallback: if the cache is found in the fallback storages but not the default storage, the manager will save the value from fallback to the default storage
  this.StorageManager = new StorageManager(FallbackPolicy.SaveFromFallback)
    .SetDefault(memoryStorage)
    .AddFallback(redisStorage);
}

...
var product = this.StorageManager.Get(usingKey); // -- Serialized JSON
var featuredProducts = this.StorageManager
  .Get(otherUsingKey, FallbackPolicy.Default); // -- Support override the behavior!
  
var comments = this.StorageManager
  .Get(
    commentsUsingKey,
    () => this.SocialNetworkProvider.GetCommentsOf(id) // -- Need to deserialize from JSON before retun
  );

Target Frameworks

• .Net 4.6.1
• .Net Standard 2.0