HPNetwork

HPNetwork is a lightweight but customizable network stack.

Posting Request

To submit a request, you can use the singleton:

Network.shared.schedule(request) { result in
    switch result {
    case .success(let output):
        // handle result
    case .failure(let error):
        // handle error
    }
}

The result is Result<DataRequest.Output, Error> where DataRequest.Output is inferred from the request object. Network.shared will do its networking on “com.henrikpanhans.Network” (which is a concurrent queue). If you want to use a custom queue, you can pass it in the initialiser:

let customQueue = DispatchQueue(label: "com.henrikpanhans.CustomQueue", qos: .userInitiated, attributes: .concurrent)
let network = Network(queue: customQueue)

You can limit the maximum number of concurrent requests to be executed by settings Network.shared.maximumConcurrentRequests = 5 for example

Creating Requests

Basics

HPNetwork is following a rather protocol based approach, so to schedule a request, it has to conform to NetworkRequest. In the most simple terms, that means you supply a URL and a request method.

Example 1:

struct BasicDataRequest: NetworkRequest {

    typealias Output = Data
    
    var url: URL? {
        // construct your URL here
    }
    
    var requestMethod: NetworkRequestMethod {
        .get
    }

}

Example 2:

struct BasicDataRequest: NetworkRequest {

    typealias Output = Data

    let url: URL?
    let requestMethod: NetworkRequestMethod

}

let basicRequest = BasicDataRequest(
    url: URL(string: "https://panhans.dev/index.html"),
    requestMethod: .get
)

JSON

If you're working with JSON, you can also use DecodableRequest which requires a JSONDecoder to be supplied. The request will use that decoder to automatically convert the received data to the specified output type

Example 3:

struct BasicDecodableRequest<Output: Decodable>: DecodableRequest {

    let url: URL?
    let requestMethod: NetworkRequestMethod
    
    var decoder: JSONDecoder {
        JSONDecoder() // use default or custom decoder
    }

}

Combine

You can also call dataTaskPublisher() on any NetworkRequest instance to get an instance of AnyPublisher<NetworkRequest.Output, Error. The publisher will walk through the same validation and error handling process as the regular Network.

Intercepting Errors

By default, instances of NetworkRequest will simply forward any encountered errors to the completion block. If you want to do some custom error conversion based on the raw Data that was received, you can implement func convertError(_ error: Error, data: Data?, response: URLResponse?) -> Error in your request model.

Everything else

Things like httpBody: Data?, authentication: NetworkRequestAuthentication? and headerFields: [NetworkRequestHeaderField] should be pretty self-explanatory so I'm gonna let you figure those out on your own.

Progress Callback

You can pass a progress handler block to Network like this:

Network.shared.schedule(request: request) { progress in
    print(progress.fractionComplete)
} completion: { result in
    // Result handling as usual
}

Cancelling Requests

Any call to schedule(request) { result in ... } returns an instance of NetworkTask that you can cancel by calling task.cancel()

WIP

• ☑ Cancellation support
• ☑ Progress callback
• ☑ Improving the documentation
• ☐ Cookie support