ObjectMapperAdditions

• Adds simple calls to include NULL values in output JSON.
• Adds ability to simply type cast JSON values to specified type.
• Adds ability to map Swift base type arrays into Realm arrays.
• Adds TimestampTransform to simply transform to/from UNIX timestamps.
• Adds ISO8601JustDateTransform to simplty transform to/from ISO8601 date string. Because ObjectMapper's ISO8601DateTransform actually is date and time transform.

Example

To run the example project, clone the repo, and run pod install from the Example directory first.

Installation

Swift Package Manager

• In Xcode select File > Add Packages...
• Copy and paste the following into the search: https://github.com/APUtils/ObjectMapperAdditions
• ‼️Make sure Up to Next Major Version is selected and put 13.0.0 into the lower bound if needed. There is a bug in 14.2 Xcode, it does not select versions higher than 9.0.0 by default‼️
• Tap Add Package
• Select ObjectMapperAdditions to add core functionality
• Optionally, select ObjectMapperAdditionsRealm to add Realm related functionality
• Tap Add Package

CocoaPods

ObjectMapperAdditions is available through CocoaPods.

To install Core features, simply add the following line to your Podfile:

pod 'ObjectMapperAdditions/Core', '~> 13.1'

To add Realm transform to your project add the following line to your Podfile:

pod 'ObjectMapperAdditions/Realm', '~> 13.1'

Carthage DEPRECATED

Please check official guide

Cartfile:

github "APUtils/ObjectMapperAdditions" ~> 13.1

If you do not need Realm part, add those frameworks: ObjectMapperAdditions, ObjectMapper, RoutableLogger.

If you are going to use Realm part, add those frameworks: ObjectMapperAdditions, ObjectMapperAdditionsRealm, ObjectMapper, Realm, RealmSwift, RoutableLogger.

Usage

Core Features

It's a common case when app gets Int in JSON instead of String even if backend guy said you it'll be String. Worst of all sometimes it could be String and sometimes something else so it'll look like you released broken app even if you tested it well.

After several projects I made a rule for myself: Never trust a backend!. I always make optional fields and cast values to type I'll use. Right now I'm using a great framework ObjectMapper to map my objects but it doesn't have transforms I need so I wrote them as this separate pod.

Example model:

import Foundation
import ObjectMapper
import ObjectMapperAdditions


struct MyModel: Mappable {
    var string: String?
    var stringsArray: [String]?
    var double: Double?
    var myOtherModel: MyOtherModel?
    var myOtherModelsArray: [MyOtherModel]?
    
    init?(map: Map) {}
    
    mutating func mapping(map: Map) {
        // You could specify proper type transform directly
        string <- (map["string"], StringTransform.shared)
        
        // Or you could just use TypeCastTransform
        string <- (map["string"], TypeCastTransform())
        
        // No doubt it also works with Double
        double <- (map["double"], TypeCastTransform())
        
        // Works with arrays too but for TypeCastTransform you must specify type
        stringsArray <- (map["stringsArray"], TypeCastTransform<String>())
        
        // Or just use StringTransform directly
        stringsArray <- (map["stringsArray"], StringTransform.shared)
        
        // No need to transform your types. They should specify transforms by themselfs.
        myOtherModel <- map["myOtherModel"]
        myOtherModelsArray <- map["myOtherModelsArray"]
    }
}

Right now there are 4 base type transforms you could use: BoolTransform, DoubleTransform, IntTransform and StringTransform. But for basic types it's easier to just use TypeCastTransform which will type cast to proper type automatically.

Typecasting for Bool, Double, Int and String raw representable enums are also supported with EnumTypeCastTransform.

Moreover this pod has extension to simplify creation of JSON with NULL values included from objects. Just call .toJSON(shouldIncludeNilValues: true) on BaseMappable object or array/set.

Date transformers example usage:

// If date in timestamp format (1506423767)
date <- (map["date"], TimestampTransform.shared)

// If date in ISO8601 full-date format (yyyy-MM-dd)
date <- (map["date"], ISO8601JustDateTransform.shared)

See example and tests projects for more details.

Realm Features

This part of ObjectMapperAdditions solves issues that prevent simply using ObjectMapper and Realm in one model. RealmListTransform to transform custom types into realm lists was taken from ObjectMapper-Realm but it can't transform simple type arrays nor optional values.

import Foundation
import ObjectMapper
import ObjectMapperAdditions
import RealmSwift

class MyRealmModel: Object, Mappable {
    @objc dynamic var id: Int = 0
    @objc dynamic var double: Double = 0
    let optionalDouble = RealmProperty<Double?>()
    @objc dynamic var string: String?
    @objc dynamic var myOtherRealmModel: MyOtherRealmModel?
    let myOtherRealmModels = List<MyOtherRealmModel>()
    var strings: List<String> = List<String>()
    
    override class func primaryKey() -> String? { "id" }
    
    override init() {
        super.init()
    }

    required init?(map: ObjectMapper.Map) {
        super.init()
        
        // Primary kay should not be reassigned after object is added to the Realm so we make sure it is assigned during init only
        id <- (map["id"], IntTransform.shared)
    }

    func mapping(map: ObjectMapper.Map) {
        performMapping {
            // Read-only primary key
            id >>> map["id"]
            
            // Same as for ordinary model
            double <- (map["double"], DoubleTransform.shared)
            
            // Using ObjectMapperAdditions's RealmPropertyTypeCastTransform
            optionalDouble <- map["optionalDouble"]
            
            // Custom transform support
//            optionalDouble <- (map["optionalDouble"], DoubleTransform.shared)
            
            // You could also use RealmPropertyTransform if you don't like type cast but you need to declare `optionalDouble` as a `var` then
//            optionalDouble <- (map["optionalDouble"], RealmPropertyTransform<Double>())
            
            string <- (map["string"], StringTransform.shared)
            myOtherRealmModel <- map["myOtherRealmModel"]
            
            // Using ObjectMapper+Realm's RealmListTransform to transform custom types
            myOtherRealmModels <- map["myOtherRealmModels"]
            
            // Using ObjectMapperAdditions's RealmTypeCastTransform
            strings <- map["strings"]
            
            // // Custom transform support
//            strings <- (map["strings"], StringTransform.shared)
            
            // You could also use RealmTransform if you don't like type cast but you need to declare `optionalDouble` as a `var` then
//            strings <- (map["strings"], RealmTransform())
        }
    }
}

Swift optionals cast to realm optionals this way: Int? -> RealmProperty<Int>>, Double? -> RealmProperty<Double?>, Bool? -> RealmProperty<Bool?>, etc.

Swift arrays cast to realm arrays this way: [String] -> List<String>, [Int] -> List<String>, [Double] -> List<Double>, [Bool] -> List<Bool>, etc.

Be sure to check that properties of type RealmProperty and List are not dynamic nor optional. Also despite of they defined as var they should be handled as constants if model is added to Realm. Use .value to change RealmOptional value or use .removeAll() and append(objectsIn:) methods to change List content

See example and tests projects for more details.

Contributions

Any contribution is more than welcome! You can contribute through pull requests and issues on GitHub.

Author

Anton Plebanovich, [email protected]

License

ObjectMapperAdditions is available under the MIT license. See the LICENSE file for more info.