Inheritance/Generics Basics. How to implement a code structure that's DRY - c#

In an attempt to DRY up my code today i'd like to do the following. (I don't know if its the best way, but it seems better than to have an ever increasing code base where I continually need to update multiple methods if i want to change something across the whole site)
What i know about Inheritance is scary. As Iv'e never questioned any of the code/libraries that I use, and Iv'e never really attempted writing anything like this before, but I want to learn... Hoping this will be my day of enlightenment :P
To my question:
Say Iv'e got an add method (in all my controllers) like this:
public ActionResult Add(VM_Down_Time_Capture viewModel)
{
using (Down_Time_CaptureRepository repository = new Down_Time_CaptureRepository())
{
if (!ModelState.IsValid)
return ReturnValidationFailure(ViewData.ModelState.Values);
Down_Time_Capture model = new Down_Time_Capture();
model.InjectFrom(viewModel);
string mserMsg = repository.Add(model, User.Identity.Name);
if (!string.IsNullOrEmpty(mserMsg))
return ReturnCustomValidationFailure(Server.HtmlEncode(mserMsg));
repository.Save();
return Json("Added successfully.", JsonRequestBehavior.AllowGet);
}
}
And at the moment I've got the following as well.
Generated by T4 Templates/EF.
ViewModels, Repositories, (Standard) EF Models
I'm thinking I need a ModelSpecfic base controller for each page (can be done using T4), that inherits from a custom ControllerBase class that contains the basic CRUD functionality. That way i can have custom code per controller, and my code base will be cleaner & smaller & that wont get affected should i need to regenerate the base files
I don't quite understand how to implement something in the lines of what i need. What i understand so far is that ill need to have my repositories, and view models inherit from a base as well and somehow specify in [B] which ones I'm using... but as to how to do that i don't know
For example (and this is my best attempt at it, not my actual code, extremely hacky as I'm amazingly confused :S)
public class Down_Time_CaptureController : Down_Time_CaptureBase
{
//[A]
}
//Generated by T4
public class Down_Time_CaptureBase: ControllerBase
{
//[B]
public override EntityObject CreateNewModel()
{
return new Down_Time_Capture();
}
public override Base_Repository CreateNewRepository()
{
return new Down_Time_CaptureRepository();
}
public override Base_ViewModel CreateNewViewModel()
{
return new VM_Down_Time_Capture();
}
//how would i go about specifying which repository & model & view model to use
//although i expect it to be something to what i did here above
//and how would i go about calling the new generic add method (but in context of this controller)?
}
//coded once
public abstract class ControllerBase: Controller
{
//[C]
//make abstract so i have to override it
public abstract Base_Controller CreateNewModel();
public abstract Base_Controller CreateNewRepository();
public abstract Base_Controller CreateNewViewModel();
//I'm assuming my generified add method would go in here
public virtual ActionResult Add(Base_ViewModel viewModel)
{
using (Base_Repository repository = CreateRepository())
{
if (!ModelState.IsValid)
return ReturnValidationFailure(ViewData.ModelState.Values);
EntityObject model = CreateNewModel();
model.InjectFrom(viewModel);
string mserMsg = repository.Add(model, User.Identity.Name);
if (!string.IsNullOrEmpty(mserMsg))
return ReturnCustomValidationFailure(Server.HtmlEncode(mserMsg));
repository.Save();
return Json("Added successfully.", JsonRequestBehavior.AllowGet);
}
}
}

Here's a simple generic interpretation of what you are asking for:
// concrete controller implementation
public class Down_Time_CaptureController: ControllerBase<Down_Time_Capture, VM_Down_Time_Capture, Down_Time_CaptureRepository>
{
}
// generic controller base
public abstract class ControllerBase<TModel, TViewModel, TRepository>: Controller
where TModel : Base_Model, new()
where TViewModel : Base_ViewModel, new()
where TRepository : Base_Repository, new()
{
protected virtual TModel CreateNewModel()
{
return (TModel)Activator.CreateInstance<TModel>();
}
protected virtual TRepository CreateNewRepository()
{
return (TRepository)Activator.CreateInstance<TRepository>();
}
protected virtual TViewModel CreateNewViewModel()
{
return (TViewModel)Activator.CreateInstance<TViewModel>();
}
//I'm assuming my generified add method would go in here
public virtual ActionResult Add(TViewModel viewModel)
{
using (var repository = CreateRepository())
{
if (!ModelState.IsValid)
return ReturnValidationFailure(ViewData.ModelState.Values);
var model = CreateNewModel();
model.InjectFrom(viewModel);
string mserMsg = repository.Add(model, User.Identity.Name);
if (!string.IsNullOrEmpty(mserMsg))
return ReturnCustomValidationFailure(Server.HtmlEncode(mserMsg));
repository.Save();
return Json("Added successfully.", JsonRequestBehavior.AllowGet);
}
}
}
A few notes:
You will probably want to create interfaces for the three types (Model, ViewModel, Repository) and use those as the generic constraints.
You will probably want a generic Repository interface and base implementation (so you don't have to code each repository independently, and copy similar logic from one to the other).
Consider using an Inversion of Control container and dependency injection. Rather than have the controller, for example, handle creating an instance of a repository, make it a property and set it from the constructor. You can then use an IoC of your choice (like Ninject or Autofac) and register concrete implementations, and it will manage creating and the lifetime of both the dependencies and the controller itself.

Related

Inject object to the constructor of controller using unity container

I couldn’t find any similar issue so I’m writing this post. There is sample controller with private field IBaseClass. Sample code looks like follows:
public class TmpController : Controller
{
private IBaseClass _baseClass;
public TmpController()
{
_baseClass = new BaseClass(this);
}
}
public interface IBaseClass
{
//...
}
public class BaseClass : IBaseClass
{
protected TmpController TmpController;
public BaseClass(TmpController tmpController)
{
TmpController = tmpController;
}
//IBaseClass implementation
}
My question is; how to inject BaseClass object to the constructor of TmpController using Unity framework?
I want to make my controller "slimmer". I want to put the logic about validation and preparing dataSource of my controls like comboBox etc. to different class. I try to make some kind of SOC in my .Web project in that very specific case, which will make my controller easier to read and maintain. I'm using approach one controller per one view but I met the case with very complex form. Currently I have controller with more than 3000 lines of code and it's hard to maintain so I want to do something with it.
And yes I'm using Services and Repositories but the problem is about validation of ViewModels, mapping ViewModel objects into DTOs and backwards, preparing data source of given components etc.
#Razem, what you guess from my comment is correct. And the minus point you described is also valid.
What you are asking "Service depending on the controller" can surely be achieved but that would be a bad design.
Currently BaseClass is only dependent on TempController. How would you handle the scenario when you need the BaseClass in some other controller also? The code will start breaking and you will end up adding new dependency to BaseClass.
Also as per the design recommendations Top Layers should be dependent on the Bottom Layers not the vice versa.
Being said that, you can still achieve the feature you are looking for that too by making controller dependent on the IBaseClass.
I am not sure the specific reasons you need to access controller inside BaseClass. I have made certain assumptions while creating following suggestions. One of such assumption is BaseClass, IBaseClass and Controller classes are part of the same assembly.
//Have a BaseController Class with the properties and/or method which you will be using in the `BaseClass` class and make them virtual so that derived controller classes can override them to have specific implementation.
public class BaseController : Controller
{
public virtual string ControllerMethod()
{
return "Controller Method from Base Controller";
}
public virtual string RandomValue
{
get
{
return "Random value from Base Controller";
}
}
}
Create a method in IBaseClass which will Set the Controller for it.
public interface IBaseClass
{
void SetController(BaseController controller);
void Method1();
}
public class BaseClass : IBaseClass
{
private BaseController controller;
public void SetController(BaseController controller)
{
this.controller = controller;
}
public void Method1()
{
var str = this.controller.RandomValue;
}
}
And derive the TempController from the BaseController and make it dependent on IBaseClass. And in the constructor of TempController call SetController method of IBaseClass by passing this argument to it. You also can override method/properties of BaseController here.
After this you can call any method of IBaseClass without passing controller instance to it.
public class TempController : BaseController
{
private IBaseClass baseClass;
public HomeController(IBaseClass productService)
{
this.baseClass = productService;
this.baseClass.SetController(this);
}
public override string RandomValue
{
get
{
return "Random value from Derived Class.";
}
}
public ActionResult Index()
{
this.baseClass.Method1();
ViewBag.Title = "Home Page";
return View();
}
}
Install nuget package Unit.Mvc in your web project. Open file Unity.Config located under App_Start folder and change method RegisterTypes as following.
public static void RegisterTypes(IUnityContainer container)
{
container.RegisterType<IBaseClass, BaseClass>(new PerRequestLifetimeManager());
}
I am sure I don't need to explain how this is going to work.
P.S. : You need to make sure that you calls IBaseClass.SetController method in controller constructor to avoid NullReferenceException when you use controller in BaseClass. This is small overhead you need to take to achieve good and maintainable design.

What are better alternatives in asp.net MVC to directly accessing the database in the Controller?

Normally I do my data access by instanciating my DbContext globally in my Controller and then I use that manipulate my data.
See below:
public class UserController : Controller
{
private OrtundEntities db = new OrtundEntities();
public ActionResult Create(CreateUserViewModel model)
{
try
{
UserDataModel user = new UserDataModel
{
// map view model fields to data model ones
};
db.Users.Add(user);
db.SaveChanges();
}
catch (Exception ex)
{
// some or other error handling goes here
}
}
}
It occurs to me that this might not be the ideal way to do it in all applications but aside from implementing a web service for every project I do, I can't think of any alternatives to the above.
So what's a better way to handle the data access on larger projects where the above wouldn't be ideal?
I'm just looking for so-called "best practice" for this or that particular situation. Many opinions will differ on what's the best way so what do you think it is and why?
To help keep your controllers concise and free of direct access to your database, you can implement the repository and dependency injection patterns. For even more concise code, you can also use the unit of work pattern.
Say you had this model:
public class Person {
public int Id { get; set; }
public string Name { get; set; }
}
With the help of generics, you can create an interface to provide a CRUD blueprint:
public interface IRepository<T> {
IEnumerable<T> Get();
T Get(int? i);
void Create(T t);
void Update(T t);
void Delete(int? i);
}
Then create a Repository class that implements the IRepository. This is where all your CRUD will take place:
public class PersonRepository : IRepository<Person> {
private OrtundEntities db = new OrtundEntities();
public IEnumerable<Person> Get() {
return db.Persons.ToList();
}
//invoke the rest of the interface's methods
(...)
}
Then in your controller you can invoke the dependency injection pattern:
private IRepository<Person> repo;
public PersonController() : this(new PersonRepository()) { }
public PersonController(IRepository<Person> repo) {
this.repo = repo;
}
And your controller method for, say, Index() could look like this:
public ActionResult Index() {
return View(repo.Get());
}
As you can see this has some useful benefits, including structure to your project, and keeping your controllers easy to maintain.
I think you need to read this
http://chsakell.com/2015/02/15/asp-net-mvc-solution-architecture-best-practices/
Larger proyets ?
Maybe https://msdn.microsoft.com/es-es/library/system.data.sqlclient.sqlcommand(v=vs.110).aspx
I use this in some big requests.

Resolve controller: "Make sure that the controller has a parameterless public constructor"

I have several Web API controllers in my project. After a lot of redundant code, I've refactored them into the code below, which seemed to be highly reusable. However, I'm suddenly getting the error Make sure that the controller has a parameterless public constructor, which seems to be caused by Ninject not being able to resolve the controller bindings. I'm not sure how to bind them.
My code:
public interface IController<T, TK>
{
DataSourceResult Get(DataSourceRequest request);
T Get(TK id);
HttpResponseMessage Post(T model);
T Put(T model);
TK Delete(TK id);
}
public abstract class BaseController<T, TK> : ApiController, IController<T, TK>
{
private readonly IRepository<T, TK> repository;
public BaseController(IRepository<T, TK> repository)
{
this.repository = repository;
}
/* methods here */
}
public class ReceiptsController : BaseController<ReceiptViewModel, long>
{
public ReceiptsController(IRepository<ReceiptViewModel, long> repository) :
base(repository)
{
}
}
In the ninject RegisterServices method, I've tried the following:
kernel.Bind<IController<OntvangstViewModel, long>>().To<OntvangstenController>();
kernel.Bind<BaseController<OntvangstViewModel, long>>().To<OntvangstenController>();
But neither seem to work. Is my implementation or inheritance wrong? Or should I bind them differently?
There is so much discussion around Repositories that you could read for days. One thing that I could point out to make your code better is this:
public class ReceiptsController : ApiController
{
public ReceiptsController()
{
}
public List<Receipt> Get()
{
List<Receipt> receipts = new List<Receipt>();
using (var context = new DbContext())
{
receipts = context.Receipts.ToList();
}
return View(receipts);
}
}
You don't need a repository. They don't really give you anything in benefits. In fact they remove a lot of the goodness from the DbContext. In my example, you don't have to worry about any injection at all.
Take a close look at DbContext.
It is wrapped in a using. That means, when you are done using the database connection, or your database transactions throw an error then your connection gets disposed of properly. Something that doesn't seem to be happening in your scenario - AFAIK.
Secondly, my example takes less time to write because I haven't written; a controller class, its interface, a generic implementation for a repository, a concrete implementation for a repository. So that's 4 classes that I have circumvented.
IMHO - my way is tonnes easier, less code to write. Safer.

Programming to Interfaces and using loose coupling - studing TDD, unit testing and DI

Greetins,
Lets suppose I have a class called Formater:IFormater, and a method LoadData(), like this:
public interface IFormater { void LoadData(string strSomeData); }
public class Formater : IFormater { public void LoadData(string strSomeData) {/*do some magic*/} }
Suppose also I have a class called View:IView and a method LoadRawData(), like this:
public interface IView { void LoadRawData(string strSomeData); }
public class View : IView {
private IFormater _formater;
public View(IFormater formater) { _formater = formater; }
public void LoadRawData(string strSomeData) { _formater.LoadData(strSomeData); }
}
Now, in my prog I create an object of the View class and call LoadRawData(), something like this:
static void Main(string[] args) {
kernel = new StandardKernel(new Modules()); //ninject
formater = kernel.Get<IFormater>(); //ninject
IView view = new View(formater);
view.LoadRawData(args[0]);
}
I use ninject as DI mechanism and actually everything works fine now.
The questions:
Is it a correct to have the IView interface to call Formater::LoadDtata() internally? (For myself I don't see any other way).
Should I have Different names for Formater::LoadData() and View::LoadRawData(), or they should have the same name, or maybe there is a practice how to name them?
I'm sorry. It looks like I made several serius inaccuracies. I have corrected them above.
Actually I don't use MVC pattern. I only thought so.
The point is that I want to isolate display logic from parsing and formating it.
My console app should open a file, read text contents and pass it to formater class, that will do some formating and bring it back (but before that formater will pass it to parser to extract necessary contents from a raw text), like this
view <-> formater <-> parser
This is simple prog and I use it to better understand some best methodologies and practicies like TDD, unit testing and DI.
I'm sorry for previous inconsistiencis.
To answer question 1), as you are using the MVC pattern, your concrete view (not IView which is an interface) should not call IController::LoadData(). The relationship between the Model, View, Controller looks like:
(Image taken from this article which gives a basic definition of MVC)
Instead the controller should place the LoadData data into a model, and then the controller should return a view which is associated with (i.e. can access) this model.
e.g in ASP.NET MVC you could do something like:
public class YourController : Controller
{
IDataRepository db;
public YourController(IDataRepository db)
{
this.db = db;
}
public ViewResult Index()
{
YourModel model = db.LoadData();
return View(model);
}
}
What is not helping you is the fact that you are trying to use the MVC pattern from a console application. I'd suggest working through an MVC tutorial (e.g. this is pretty good) to get a better understanding.

EF Model Validation against database

I want to use EF 5 model validation to avoid duplicate values in the database, so I'm using a model class like this:
[Table("MeasureUnits")]
public class MeasureUnit : IValidatableObject
{
public int MeasureUnitId { get; set; }
public string Symbol { get; set; }
public string Name { get; set; }
public IEnumerable<ValidationResult> Validate(ValidationContext validationContext)
{
using (MeasureUnitRepository rep = new MeasureUnitRepository())
{
MeasureUnit measureUnit = rep.FindDuplicateBySymbol(this);
if (measureUnit != null)
yield return new ValidationResult(
"There is another unit with this symbol, you can't duplicate it",
new[] { "Symbol" });
}
}
The repository class creates the DbContext, implements IDisposable, has the logic to find the duplicate and it all works just as intended.
However, using the debugger I realized the validation is performed twice for every insert or update, so the repository (and DbContext) gets instantiated and disposed twice also.
Besides that, there is another DbContext living in the controller but just don't find the way to use it inside the model class, other than including the DbContext in the model's constructor, but I feel it's not the right solution.
Is there a better o "right" way to achieve this validation?
Thanks in advance.
When you have to go to the database then you need to use DbContext and DbContext has an Overridable method called ValidateEntity. See this article: Entity Framework Validation.
I put the code I use in another answer here
And more about how I've structured the validation in MVC here.
Also, instantiating a context inside your repository is likely to cause you grief. The repositories will need to share a context. You could treat the context as your unit of work and pass it into the repository in the constructor, or you could wrap the context in your own unit of work and pass that in.
You can use any IOC container available out there like Unity, Ninject, Autofac or StructureMap to inject your repository as a dependency.
This way you would be able to access the same object in the controller, your Validate method or wherever you need to use it.
Some of these IOC(Ninject for sure - look for 'request scope') containers are capable of integrating with ASP.NET MVC so that the dependency(your repository in that case) is created once per request and disposed when the request ends.
Example using Ninject:
You create a globally accessible(the design is up to you) ninject kernel
public static class NinjectKernel
{
public static IKernel Kernel = new StandardKernel();
static NinjectKernel()
{
Kernel.Bind<IMyRepository>().To<MyRepositoryImpl>().InRequestScope();
}
}
and a controller factory for MVC controllers
public class NinjectControllerFactory : DefaultControllerFactory
{
protected override IController GetControllerInstance(RequestContext requestContext,
Type controllerType)
{
return controllerType == null ? null : (IController)NinjectKernel.Kernel.Get(controllerType);
}
}
You can then set your controller factory in Global.asax like this
ControllerBuilder.Current.SetControllerFactory(new NinjectControllerFactory());
and get the repository in your Validate method in a similar way it's done in the Controller factory.

Categories