A powerful and simple Core Data wrapper framework written in Swift.
AlecrimCoreData provides a default NSManagedObjectContext
subclass that can be used as is, extended, or not used at all. If used it will help you with several Core Data related tasks.
It is possible to use the framework with a "vanilla" NSManagedObjectContext
too. In this case you will have the liberty to configure the Core Data stack as you want.
Mixing DataContext
and "vanilla" NSManagedObjextContext
instances is possible but is strongly discouraged.
// Initializes a `DataContext` instance with default options.
let dataContext = DataContext()
Note: The DataContext()
function provides a unique instance of a dataContext object on the main thread, so make sure to avoid core data concurrency violations.
The generic Table<T>
struct is the base for AlecrimCoreData functionality and is where the fun begins. T
, in this case, is a NSManagedObject
subclass type.
// Extends the `DataContext` class to include entity table properties.
// (It would be the `NSManagedObjectContext` class too.)
extension DataContext {
var people: Table<Person> { return Table<Person>(dataContext: self) }
var departments: Table<Department> { return Table<Department>(dataContext: self) }
}
Say you have an NSManagedObject
subclass type called Person
, related to a Department
. To get all of the Person
entities as an array, use the following methods:
for person in dataContext.people {
print(person.firstName)
}
You can also skip some results:
let people = dataContext.people.skip(3)
Or take some results only:
let people = dataContext.people.skip(3).take(7)
Or, to return the results sorted by a property:
let sortedPeople = dataContext.people.orderBy { $0.lastName }
Or, to return the results sorted by multiple properties:
let sortedPeople = dataContext.people
.orderBy { $0.lastName }
.thenBy { $0.firstName }
Or, to return the results sorted by multiple properties, ascending or descending:
let sortedPeople = dataContext.people
.orderByDescending { $0.lastName }
.thenByAscending { $0.firstName }
If you have a unique way of retrieving a single entity from your data store (such as via an identifier), you can use the first
method:
if let person = dataContext.people.first({ $0.identifier == 123 }) {
print(person.name)
}
You can filter the results using the filter
method:
let filteredPeople = dataContext.people.filter { $0.lastName == "Smith" }
You can combine multiple filters and other methods as well:
let filteredPeople = dataContext.people
.filter { $0.lastName == "Smith" }
.filter { $0.firstName.beginsWith("J") }
.orderBy { $0.lastName }
.thenBy { $0.firstName }
Or:
let filteredPeople = dataContext.people
.filter { $0.lastName == "Smith" && $0.firstName.beginsWith("J") }
.orderBy { $0.lastName }
.thenBy { $0.firstName }
You can count entities in your persistent store using the count
method.
let peopleCount = dataContext.people.count()
Or:
let filteredPeopleCount = dataContext.people.count { $0.lastName == "Smith" }
Or:
let filteredPeopleCount = dataContext.people
.filter { $0.lastName == "Smith" }
.count()
When you need to create a new instance of an Entity, use:
let person = dataContext.people.createEntity()
You can also create or get the first existing entity matching the criteria. If the entity does not exist, a new one is created and the specified attribute is assigned from the searched value automatically.
let person = dataContext.people.firstOrCreated { $ 0.identifier == 123 }
To delete a single entity:
if let person = dataContext.people.first({ $0.identifier == 123 }) {
dataContext.people.deleteEntity(person)
}
To delete many entities:
dataContext.departments.filter({ $0.people.count == 0 }).delete()
You can save the data context in the end, after all changes were made.
let person = dataContext.people.firstOrCreated { $0.identifier == 9 }
person.firstName = "Christopher"
person.lastName = "Eccleston"
person.additionalInfo = "The best Doctor ever!"
do {
dataContext.save()
}
catch let error {
// do a nice error handling here
}
Another important part of AlecrimCoreData is the use of strongly-typed query attributes. A lot of boilerplate code is required to support strongly typed queries. With this in mind, the ACDGen tool was created. All you have to do is point ACDGen
to your managed object model and the source code for the entities is automatically generated, including the AlecrimCoreData query attributes if you want.
Using the generated strongly-typed query attributes is completely optional, but with them the experience with AlecrimCoreData is greatly improved. The use of strongly-typed query attributes requires a project that has generated extensions of it's model classes using ACDGen
.
ACDGen binary and source code are avaible on "ACDGen/Bin" and "ACDGen/Source" folders respectively.
OK. You can write code like this:
// No data access is made here.
let peopleInDepartments = dataContext.people
.filter { $0.department << [dept1, dept2] }
.orderBy { $0.firstName }
.thenBy { $0.lastName }
let itemsPerPage = 10
for pageNumber in 0..<5 {
print("Page: \(pageNumber)")
// No data access is made here either.
let peopleInCurrentPage = peopleInDepartments
.skip(pageNumber * itemsPerPage)
.take(itemsPerPage)
// Now is when the data is read from persistent store.
for person in peopleInCurrentPage {
print("\(person.firstName) \(person.lastName) - \(person.department.name)")
}
}
But you can do even more with AlecrimCoreData.
There is a implementation of NSFetchedResultsController
(for OS X) and FetchRequestController
that is a strongly-typed wrapper for NSFetchedResultsController
.
You are invited to read the code and discover more possibilities (and to help us to improve them and create new ones).
There are methods for aggregating, asynchronous fetching in background and many others. You can read the AlecrimCoreData documentation at http://cocoadocs.org/docsets/AlecrimCoreData for more information.
You can order and filter entities not using the AlecrimCoreData query attributes at all. In this case you lose the strongly-typed attributes, but gain in flexibility. You can even mix the two approaches without any problem.
You can order the entities using NSSortDescriptor
instances:
let sortDescriptor: NSSortDescriptor = ...
let orderedPeople = dataContext.sortUsingSortDescriptor(sortDescriptor)
Or:
let sd1: NSSortDescriptor = ...
let sd2: NSSortDescriptor = ...
let orderedPeople = dataContext.sortUsingSortDescriptors([sd1, sd2])
You can also use the sortByAttributeName:::
method:
let orderedPeople = dataContext.sortByAttributeName("lastName", ascending: true)
You can filter entities using NSPredicate
instances:
let predicate: NSPredicate = ...
let filteredPeople = dataContext.people.filterUsingPredicate(predicate)
If you want to use the DataContext
class and initialize it's instance without parameters AlecrimCoreData will try to infer the managed object model and the persistent store locations, based on most common cases.
You can however create and configure a DataContextOptions
struct and pass it as parameter to DataContext
initializer.
There are helper initializers on DataContextOptions
, but if you know the URLs or know how to construct them it may be better to pass these locations yourself to the DataContextOptions
initializer directly.
Other options can be configured using DataContextOptions
(the defaultBatchSize
and defaultComparisonPredicateOptions
properties). Unlike the managed object model and persistent store locations, these options are global and static for the entire framework (and your project).
Since AlecrimCoreData version 4 the configuration of your Core Data managed object contexts to include iCloud integration are not made by the framework anymore. This type of integration is very trick and we think it is better a manually approach, case by case, including your own observers and handlers. You can, however, use the configureUbiquityWithContainerIdentifier:::
method and the ubiquityEnabled
property from the DataContextOptions
struct to help you to configure iCloud integration if you want.
Since AlecrimCoreData version 4 the DataContext
is an NSManagedObjectContext
subclass and the framework can work with "vanilla" NSManagedObjectContext
instances as well. So you can integrate and use other frameworks as you are using only NSManagedObjectContext
instances and there should be no side effects.
- Swift 2
- Xcode 7.0
- OS X 10.9 / iOS 8.0 / watchOS 2.0
CocoaPods is a dependency manager for Cocoa projects.
CocoaPods 0.36 adds supports for Swift and embedded frameworks. You can install it with the following command:
$ gem install cocoapods
To integrate AlecrimCoreData into your Xcode project using CocoaPods, specify it in your Podfile
:
source 'https://github.com/CocoaPods/Specs.git'
platform :ios, '8.0'
use_frameworks!
pod 'AlecrimCoreData', '~> 4.0'
Then, run the following command:
$ pod install
You can add AlecrimCoreData as a git submodule, drag the AlecrimCoreData.xcodeproj
file into your Xcode project and add the framework product as an embedded binary in your application target.
- master - The production branch. Clone or fork this repository for the latest copy.
- develop - The active development branch. Pull requests should be directed to this branch.
If you want to contribute, please feel free to fork the repository and send pull requests with your fixes, suggestions and additions. :-)
The main areas the framework needs improvement:
- Correct the README, code and examples for English mistakes;
- Write more and better code documentation;
- Write unit tests;
- Replace some pieces of code with more "elegant" ones.
AlecrimCoreData is released under an MIT license. See LICENSE for more information.