Skip to content
/ TRON Public
forked from MLSDev/TRON

Lightweight network abstraction layer, written on top of Alamofire

License

Notifications You must be signed in to change notification settings

vilimets/TRON

 
 

Repository files navigation

Build Status   codecov.io CocoaPod platform   CocoaPod version   Carthage compatible Packagist

TRON is a lightweight network abstraction layer, built on top of Alamofire. It can be used to dramatically simplify interacting with RESTful JSON web-services.

Features

  • Generic, protocol-based implementation
  • Built-in response and error parsing
  • Support for any custom mapper. Defaults to SwiftyJSON.
  • Support for upload tasks
  • Support for download tasks and resuming downloads
  • Robust plugin system
  • Stubbing of network requests
  • Modular architecture
  • Support for iOS/Mac OS X/tvOS/watchOS/Linux
  • Support for CocoaPods/Carthage/Swift Package Manager
  • RxSwift extension

Overview

We designed TRON to be simple to use and also very easy to customize. After initial setup, using TRON is very straightforward:

let request: APIRequest<User,MyAppError> = tron.request(path: "me")
request.perform(withSuccess: { user in
  print("Received User: \(user)")
}, failure: { error in
  print("User request failed, parsed error: \(error)")
})

Requirements

  • XCode 8
  • Swift 3
  • iOS 9 / macOS 10.11 / tvOS 9.0 / watchOS 2.0

Installation

CocoaPods

pod 'TRON', '~> 2.0.0-beta.1'

Only Core subspec, without SwiftyJSON dependency:

pod 'TRON/Core', '~> 2.0.0-beta.1'

RxSwift extension for TRON:

pod 'TRON/RxSwift', '~> 2.0.0-beta.1'

Carthage

github "MLSDev/TRON", ~> 2.0.0-beta.1

Migration Guides

Project status

TRON is under active development by MLSDev Inc. Pull requests and issues are welcome!

Request building

TRON object serves as initial configurator for APIRequest, setting all base values and configuring to use with baseURL.

let tron = TRON(baseURL: "https://api.myapp.com/")

You need to keep strong reference to TRON object, because it holds Alamofire.Manager, that is running all requests.

URLBuildable

URLBuildable protocol is used to convert relative path to URL, that will be used by request.

public protocol URLBuildable {
    func url(forPath path: String) -> URL
}

By default, TRON uses URLBuilder class, that simply appends relative path to base URL, which is sufficient in most cases. You can customize url building process globally by changing urlBuilder property on TRON or locally, for a single request by modifying urlBuilder property on APIRequest.

HeaderBuildable

HeaderBuildable protocol is used to configure HTTP headers on your request.

public protocol HeaderBuildable {
    func headers(forAuthorizationRequirement requirement: AuthorizationRequirement, including headers: [String:String]) -> [String: String]
}

AuthorizationRequirement is an enum with three values:

public enum AuthorizationRequirement {
    case none, allowed, required
}

It represents scenarios where user is not authorized, user is authorized, but authorization is not required, and a case, where request requires authorization.

By default, TRON uses HeaderBuilder class, which adds "Accept":"application/json" header to your requests.

Sending requests

To send APIRequest, call perform(withSuccess:failure:) method on APIRequest:

let alamofireRequest = request.perform(withSuccess: { result in }, failure: { error in})

Notice that alamofireRequest variable returned from this method is an Alamofire.Request?, that will be nil if request is stubbed.

Alternatively, you can use performCollectingTimeline(withCompletion:) method that contains Alamofire.Response inside completion closure:

request.performCollectingTimeline(withCompletion: { response in
    print(response.timeline)
    print(response.result)
})

In both cases, you can additionally chain Alamofire.Request methods, if you need:

request.perform(withSuccess: { result in }, failure: { error in })?.progress { bytesWritten, totalBytesWritten, totalBytesExpectedToWrite in
    print(bytesWritten, totalBytesWritten, totalBytesExpectedToWrite)
}

Response parsing

Generic APIRequest implementation allows us to define expected response type before request is even sent. We use Alamofire DataResponseSerializerProtocol, and are adding to it ErrorHandlingDataResponseSerializerProtocol, which basically allows us to have two generics for both success and error values.

// Alamofire 4:

public protocol DataResponseSerializerProtocol {
associatedtype SerializedObject

var serializeResponse: (URLRequest?, HTTPURLResponse?, Data?, Error?) -> Result<SerializedObject> { get }
}

// TRON 2:

public protocol ErrorHandlingDataResponseSerializerProtocol : DataResponseSerializerProtocol {
associatedtype SerializedError

var serializeError: (Alamofire.Result<SerializedObject>?,URLRequest?, HTTPURLResponse?, Data?, Error?) -> APIError<SerializedError> { get }
}

TRON provides JSONDecodable protocol, that allows us to parse models using SwiftyJSON:

public protocol JSONDecodable {
    init(json: JSON) throws
}

To parse your response from the server using SwiftyJSON, all you need to do is to create JSONDecodable conforming type, for example:

class User: JSONDecodable {
  let name : String
  let id: Int

  required init(json: JSON) {
    name = json["name"].stringValue
    id = json["id"].intValue
  }
}

And send a request:

let request: APIRequest<User,MyAppError> = tron.request(path: "me")
request.perform(withSuccess: { user in
  print("Received user: \(user.name) with id: \(user.id)")
})

There are also default implementations of JSONDecodable protocol for Swift built-in types like String, Int, Float, Double and Bool, so you can easily do something like this:

let request : APIRequest<String,MyAppError> = tron.request(path: "status")
request.perform(withSuccess: { status in
    print("Server status: \(status)") //
})

You can also use EmptyResponse struct in cases where you don't care about actual response.

There is also an array extension for JSONDecodable, however it's commented out due to this.

RxSwift

let request : APIRequest<Foo, MyError> = tron.request(path: "foo")
_ = request.rxResult().subscribeNext { result in
    print(result)
}
let multipartRequest : APIRequest<Foo,MyError> = tron.upload(path: "foo", formData: { _ in })
multipartRequest.rxMultipartResult().subscribeNext { result in
    print(result)
}

Error handling

TRON includes built-in parsing for errors. APIError is a generic class, that includes several default properties, that can be fetched from unsuccessful request:

struct APIError<T> {
    public let request : URLRequest?
    public let response : NSHTTPURLResponse?
    public let data : Data?
    public let error : Error?

    public var errorModel: T?
}

When APIRequest fails, you receive concrete APIError instance, for example, let's define MyAppError we have been talking about:

class MyAppError : JSONDecodable {
  var errors: [String:[String]] = [:]

  required init(json: JSON) {
    if let dictionary = json["errors"].dictionary {
      for (key,value) in dictionary {
          errors[key] = value.arrayValue.map( { return $0.stringValue } )
      }
    }
  }
}

This way, you only need to define how your errors are parsed, and not worry about other failure details like response code, because they are already included:

request.perform(withSuccess: { response in }, failure: { error in
    print(error.request) // Original URLRequest
    print(error.response) // NSHTTPURLResponse
    print(error.data) // Data of response
    print(error.error) // Error from Foundation Loading system
    print(error.errorModel.errors) // MyAppError parsed property
  })

Using Alamofire custom response serializers

Any custom response serializer for Alamofire can be used with TRON, you just need to specify error type, that will be used, for example, if CustomError is JSONDecodable:

extension Alamofire.DataResponseSerializer : ErrorHandlingDataResponseSerializerProtocol {
    public typealias SerializedError = CustomError

    public var serializeError: (Result<SerializedObject>?, URLRequest?, HTTPURLResponse?, Data?, Error?) -> APIError<SerializedError> {
        return { erroredResponse, request, response, data, error in
            let serializationError : Error? = erroredResponse?.error ?? error
            var error = APIError<ErrorModel>(request: request, response: response,data: data, error: serializationError)

            // Here you can define, how error needs to be parsed
            error.errorModel =  try? ErrorModel.init(json: JSON(data: data ?? Data()))
            return error
        }
    }
}

CRUD

struct Users
{
    static let tron = TRON(baseURL: "https://api.myapp.com")

    static func create() -> APIRequest<User,MyAppError> {
        let request: APIRequest<User,MyAppError> = tron.request(path: "users")
        request.method = .post
        return request
    }

    static func read(id: Int) -> APIRequest<User, MyAppError> {
        return tron.request(path: "users/\(id)")
    }

    static func update(id: Int, parameters: [String:Any]) -> APIRequest<User, MyAppError> {
        let request: APIRequest<User,MyAppError> = tron.request(path: "users/\(id)")
        request.method = .put
        request.parameters = parameters
        return request
    }

    static func delete(id: Int) -> APIRequest<User,MyAppError> {
        let request: APIRequest<User,MyAppError> = tron.request(path: "users/\(id)")
        request.method = .delete
        return request
    }
}

Using these requests is really simple:

Users.read(56).perform(withSuccess: { user in
  print("received user id 56 with name: \(user.name)")
})

It can be also nice to introduce namespacing to your API:

enum API {}
extension API {
  enum Users {
    // ...
  }
}

This way you can call your API methods like so:

API.Users.delete(56).perform(withSuccess: { user in
  print("user \(user) deleted")
})

Stubbing

Stubbing is built right into APIRequest itself. All you need to stub a successful request is to set apiStub property and turn stubbingEnabled on:

let request = API.Users.get(56)
request.stubbingEnabled = true
request.apiStub.model = User.fixture()

request.perform(withSuccess: { stubbedUser in
  print("received stubbed User model: \(stubbedUser)")
})

Stubbing can be enabled globally on TRON object or locally for a single APIRequest. Stubbing unsuccessful requests is easy as well:

let request = API.Users.get(56)
request.stubbingEnabled = true
request.apiStub.error = APIError<MyAppError>.fixtureError()
request.perform(withSuccess: { _ in }, failure: { error in
  print("received stubbed api error")
})

You can also optionally delay stubbing time or explicitly set that api stub should fail:

request.apiStub.stubDelay = 1.5
request.apiStub.successful = false

Upload

  • From file:
let request = tron.upload(path: "photo", fromFileAt: fileUrl)
  • NSData:
let request = tron.upload(path: "photo", data: data)
  • Stream:
let request = tron.upload(path: "photo", fromStream: stream)
  • Multipart-form data:
let request: MultipartAPIRequest<EmptyResponse,MyAppError> = tron.uploadMultipart(path: "form") { formData in
    formData.appendBodyPart(data: data,name: "cat", mimeType: "image/jpeg")
}
request.performMultipart(withSuccess: { result in
    print("form sent successfully")
})

Note Multipart form data uploads use MultipartAPIRequest class instead of APIRequest and have different perform method.

Download

let request = tron.download(path: "file", to: destination)

Resume downloads:

let request = tron.download(path: "file", to: destination, resumingFrom: data)

Plugins

TRON includes simple plugin system, that allows reacting to some request events.

public protocol Plugin {
    func willSendRequest(request: URLRequest?)    
    func requestDidReceiveResponse(response: (URLRequest?, NSHTTPURLResponse?, Data?, Error?))
}

Plugins can be used globally, on TRON instance itself, or locally, on concrete APIRequest. Keep in mind, that plugins that are added to TRON instance, will be called for each request. There are some really cool use-cases for global and local plugins.

By default, no plugins are used, however two plugins are implemented as a part of TRON framework.

NetworkActivityPlugin

NetworkActivityPlugin serves to monitor requests and control network activity indicator in iPhone status bar. This plugin assumes you have only one TRON instance in your application.

let tron = TRON(baseURL: "https://api.myapp.com", plugins: [NetworkActivityPlugin()])

NetworkLoggerPlugin

NetworkLoggerPlugin is used to log responses to console in readable format. By default, it prints only failed requests, skipping requests that were successful.

Local plugins

There are some very cool concepts for local plugins, some of them are described in dedicated PluginConcepts page.

Alternatives

We are dedicated to building best possible tool for interacting with RESTful web-services. However, we understand, that every tool has it's purpose, and therefore it's always useful to know, what other tools can be used to achieve the same goal.

TRON was heavily inspired by Moya framework and LevelUPSDK

License

TRON is released under the MIT license. See LICENSE for details.

About MLSDev

MLSDev.com

TRON is maintained by MLSDev, Inc. We specialize in providing all-in-one solution in mobile and web development. Our team follows Lean principles and works according to agile methodologies to deliver the best results reducing the budget for development and its timeline.

Find out more here and don't hesitate to contact us!

About

Lightweight network abstraction layer, written on top of Alamofire

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Swift 98.0%
  • Other 2.0%