Pretty much every app nowadays requires you to connect to the internet to access some content. The majority of those apps use JSON to communicate that data from the backend systems.
There is high chance you will have some code in your app to download, parse and return objects for your app to use from an endpoint (unless you are using a network library such as Alamofire)
Building our codable objects
The first thing we need to do is build our codable objects. Objects that implement the Codable protocol allow Encoders and Decoders to encode or decode them to and from an external representation such as JSON.
Let’s take the response from the sample endpoint below as an example:
You can create codable classes yourself by hand. In simple examples this can be fairly straight forward, however if you have a response that has a more complex structure, doing so can be time consuming and error prone.
To create our codable objects we can use a generator, my weapon of choice is QuickType. We just paste in the JSON that is returned from the posts endpoint and it automatically generates the Codable structs for us. Easy!
If we paste in our post response, we should end up with some code looking like this:
How easy was that?! Obviously we will still need to check the structs, in the example above none of the fields are optional which means data must be passed in otherwise our decoding will fail. We don’t need to worry about that here, but worth remembering when checking the generated code in your examples.
URLSession extension and Generics
To solve our problem we are going to wrap the existing URLSession dataTask method. I’m sure if you have done any kind of request work in pure swift you will have used this method in some form so we aren’t going to go into the details of how it works.
So let’s step through this code sample step by step:
- First of we have defined a custom error for this extension, this is returned when no data has been returned from the request, covered in point 6. We also have an error case if we get an HTTPURLResponse with an incorrect status code, covered in point 5.
- Here we are making use of Generics to allow any type T being returned from this function as long as type T implements the Decodable protocol (which we need it to inorder to use the JSONDecoder)
- As discussed, here we are calling the existing dataTask method to run our request.
- First thing we do once the request has returned is check to see if there was a request error, if so we call the completion handler with the response and the error.
- The second check we perform is to check the status code if we have received an HTTPURLResponse. Note we aren’t stopping the code here if we don’t get a HTTPURLResponse as you could use this function to load a local JSON file for example, not just a remote URL. Any status code in the 200-299 range is considered a successful request, if we receive a status code outside this range we return an error along with the response for further processing by whoever passed the completion handler.
- The third check we perform is to unwrap data ready for decoding. If this fails (as in it’s nil) then we call the completionHandler with the response and our custom error defined in step 1.
- The final piece of the puzzle is to attempt to decode the data into type T we defined in the method signature as part of step 2. If this succeeds we can call our completion handler with our decoded type and response. If it throws an error we capture the error and return it using the catch block below.
See it in action
Now that we have put our function together, let’s take it for a test drive.
This shouldn’t look too scarey, infact if you have used the standard dataTask functions in your code previously this should look very familiar. The only different here being that our completion handler now returns our Codable User objects rather than just a blob of Data like before.
Hopefully that example makes sense and gives you a nice simple way to perform a request and have it decode some JSON into a struct / class. Now let’s have a look at some reactive programming using Combine.
Hopefully you have at least heard of Combine even if you haven’t had chance to use it yet in a production app. It is Apple’s own version of a reactive framework. Those of you who have already been using RxSwift will be right at home. We aren’t going to go into too much detail about what Combine is but here is a definition of what reactive programming is:
In computing, reactive programming is a declarative programming paradigm concerned with data streams and the propagation of change
In more simplistic terms, reactive programming uses a observer pattern to allow classes to monitor different streams of data or state. When this state changes it emits an event with the new value which can trigger other streams to perform work or update such as UI code. If you are familier with KVO you will understand the basic concept. However reactive programming is far less painful and a lot more powerful than KVO.
Now let’s take the previous pure Swift example and see how we can use it in Combine. The Combine framework adds new reactive functionality to the URLSession in the form of the dataTaskPublisher function.
Similar to our previous example we have extended URLSession to provide this functionality. Let’s step through it:
- As with the pure Swift example we are defining a custom error here to handle when we receive a status code that is not a success. The difference being here we are attaching the response to the error as we don’t have a completionHandler in Combine. That way whoever is handling the error can inspect the response and see why it failed.
- Here we are defining the method, again using generics to only accept a type T that has implemented the Decodable protocol. The function returns a publisher who returns our decoded object.
- As discussed previously, we are simply wrapping the existing dataTaskPublisher method.
- Now here is where things start to become reactive. The tryMap function is similar to the standard map function in that it attempts to convert / transfrom elements from one type to another. However the difference here being that it is almost wrapped in a try. In this case you can include code in the closure which throws errors and they will be pushed downstream and handled later instead of needing a do block. Similar to our pure Swift example, we are checking we have a valid status code, if not we throw our custom error. If not we map our data ready to be decoded.
- Here we are using the built in decode method to attempt to decode our custom type using the JSONDecoder. Similar to the tryMap function above, any errors are pushed downstream to be handler later.
- The final piece of the puzzle is to use type erasure. This removes the publisher class type and makes it AnyPublisher. For more info on type erasure see my previous post
Combine in action
Now that we have built our wrapper class let’s take a look at this in action:
- Here we have called our newly created dataTaskPublisher method which has returned our publisher. This is where reactive programming comes in. All of the code inside the dataTaskPublisher has not executed yet. We have simply returned a publisher who is waiting for a subscriber to come along and listen. A publisher will not execute unless a subscription has not been fulfilled. To subscribe to a stream we use the sink method. If you think of the chain of reactive methods flowing into a sink at the bottom, that is the best analogy here.
- The sink method has 2 parts. The first closure defines what happens once the stream is completed. Now this can come in the form of a finished state, which means the stream has completed what is doing and will no longer emit any more events. Or failure, which means some item further up the stream has raised an error which flows down into this sink where it can be handled.
- The second closure defines what we would like to do each time the event stream emits a change. In this case the publisher will send a users array once it has finished loading, here we are just printing out the user names.
What have we learnt:
- We have used QuickType to convert our JSON into codable structs for decoding.
- Wrapped the existing URLSession dataTask method with our own using Generics so we can using any Codable type to decode the response.
- Similarly, using reactive programming and Apple’s new Combine framework have created our own Generic wrapper for the existing dataTaskPublisher function.
Feel free to download the playground and play around with the examples yourself