dart_json_mapper 1.5.2 dart_json_mapper: ^1.5.2 copied to clipboard
This package allows programmers to annotate Dart classes in order to serialize / deserialize them from / to JSON.
This package allows programmers to annotate Dart classes in order to Serialize / Deserialize them to / from JSON.
Why? #
- Compatible with all Dart platforms, including Flutter and Web platforms
- No need to extend your classes from any mixins/base/abstract classes to keep code leaner
- Clean and simple setup, transparent and straight-forward usage with no heavy maintenance
- No extra boilerplate, 100% generated code, which you'll never see.
- Custom converters per each class field, full control over the process
- NO dependency on
dart:mirrors
, one of the reasons is described here. - Because Serialization/Deserialization is NOT a responsibility of your Model classes.
Dart classes reflection mechanism is based on reflectable library. This means "extended types information" is auto-generated out of existing Dart program guided by the annotated classes only, as the result types information is accessible at runtime, at a reduced cost.
Typical Flutter.io project integration
sample can be found here
Basic setup #
Please add the following dependencies to your pubspec.yaml
:
dependencies:
dart_json_mapper:
dev_dependencies:
build_runner:
Say, you have a dart program main.dart having some classes intended to be traveling to JSON and back.
- First thing you should do is to put
@jsonSerializable
annotation on each of those classes - Next step is to auto generate main.reflectable.dart file. And afterwards import that file into main.dart
lib/main.dart
import 'package:dart_json_mapper/dart_json_mapper.dart';
import 'main.reflectable.dart'; // Import generated code.
@jsonSerializable // This annotation let instances of MyData traveling to/from JSON
class MyData {
int a = 123;
@JsonProperty(ignore: true)
bool b;
@JsonProperty(name: 'd')
String c;
MyData(this.a, this.b, this.c);
}
main() {
initializeReflectable(); // Imported from main.reflectable.dart
print(JsonMapper.serialize(MyData(456, true, "yes")));
}
output:
{
"a": 456,
"d": "yes"
}
Go ahead and create a build.yaml
file in your project root directory. Then add the
following content:
targets:
$default:
builders:
reflectable:
generate_for:
- lib/main.dart
options:
formatted: true
Now run the code generation step with the root of your package as the current directory:
> pub run build_runner build
You'll need to re-run code generation each time you are making changes to lib/main.dart
So for development time, use watch
like this
> pub run build_runner watch
Each time you modify your project code, all *.reflectable.dart files will be updated as well.
- Next step is to add "*.reflectable.dart" to your .gitignore
- And this is it, you are all set and ready to go. Happy coding!
Format DateTime / num types #
In order to format DateTime
or num
instance as a JSON string, it is possible to
provide intl based formatting patterns.
DateTime
@JsonProperty(converterParams: {'format': 'MM-dd-yyyy H:m:s'})
DateTime lastPromotionDate = DateTime(2008, 05, 13, 22, 33, 44);
@JsonProperty(converterParams: {'format': 'MM/dd/yyyy'})
DateTime hireDate = DateTime(2003, 02, 28);
output:
{
"lastPromotionDate": "05-13-2008 22:33:44",
"hireDate": "02/28/2003"
}
num
@JsonProperty(converterParams: {'format': '##.##'})
num salary = 1200000.246;
output:
{
"salary": "1200000.25"
}
As well, it is possible to utilize converterParams
map to provide custom
parameters to your custom converters.
Example with immutable class #
enum Color { Red, Blue, Green, Brown, Yellow, Black, White }
@jsonSerializable
class Car {
@JsonProperty(name: 'modelName')
String model;
@JsonProperty(enumValues: Color.values)
Color color;
@JsonProperty(ignore: true)
Car replacement;
Car(this.model, this.color);
}
@jsonSerializable
class Immutable {
final int id;
final String name;
final Car car;
const Immutable(this.id, this.name, this.car);
}
print(
JsonMapper.serialize(
Immutable(1, 'Bob', Car('Audi', Color.Green))
)
);
output:
{
"id": 1,
"name": "Bob",
"car": {
"modelName": "Audi",
"color": "Color.Green"
}
}
Constructor parameters #
Sometimes you don't really care or don't want to store some json property as a dedicated class field, but instead, you would like to use it's value in constructor to calculate other class properties. This way you don't have a convenience to annotate a class field, but you could utilize constructor parameter for that.
With the input JSON like this:
{"LogistikTeileInOrdnung":"true"}
You could potentially have a class like this:
@jsonSerializable
class BusinessObject {
final bool logisticsChecked;
final bool logisticsOK;
BusinessObject()
: logisticsChecked = false,
logisticsOK = true;
@jsonConstructor
BusinessObject.fromJson(
@JsonProperty(name: 'LogistikTeileInOrdnung') String processed)
: logisticsChecked = processed != null && processed != 'null',
logisticsOK = processed == 'true';
}
Unmapped properties #
If you are looking for an alternative to Java Jackson @JsonAnySetter / @JsonAnyGetter
It is possible to configure the same scenario as follows:
@jsonSerializable
class UnmappedProperties {
String name;
Map<String, dynamic> _extraPropsMap = {};
@jsonProperty
void unmappedSet(String name, dynamic value) {
_extraPropsMap[name] = value;
}
@jsonProperty
Map<String, dynamic> unmappedGet() {
return _extraPropsMap;
}
}
// given
final json = '''{"name":"Bob","extra1":1,"extra2":"xxx"}''';
// when
final instance = JsonMapper.deserialize<UnmappedProperties>(json);
// then
expect(instance.name, 'Bob');
expect(instance._extraPropsMap['name'], null);
expect(instance._extraPropsMap['extra1'], 1);
expect(instance._extraPropsMap['extra2'], 'xxx');
Iterable types #
Since Dart language has no possibility to create typed iterables dynamically, it's a bit of a challenge to create exact typed lists/sets/etc via reflection approach. Those types has to be declared explicitly.
For example List() will produce List<dynamic>
type which can't be directly set to the concrete
target field List<Car>
for instance. So obvious workaround will be to cast
List<dynamic> => List<Car>
, which can be performed as List<dynamic>().cast<Car>()
.
In order to do so, we'll use Value Decorator Function inspired by Decorator pattern.
final String json = '[{"modelName": "Audi", "color": "Color.Green"}]';
JsonMapper().useAdapter(JsonMapperAdapter(
valueDecorators: {
typeOf<List<Car>>(): (value) => value.cast<Car>(),
typeOf<Set<Car>>(): (value) => value.cast<Car>()
})
);
List<Car> myCarsList = JsonMapper.deserialize(json);
Set<Car> myCarsSet = JsonMapper.deserialize(json);
Basic iterable based generics using Dart built-in types like List<num>, List<Sring>, List<bool>, List<DateTime>, Set<num>, Set<Sring>, Set<bool>, Set<DateTime>, etc.
supported out of the box.
For custom iterable types like List<Car> / Set<Car>
you have to provide value decorator function
as showed in a code snippet above before using deserialization. This function will have explicit
cast to concrete iterable type.
OR an easy case #
When you are able to pre-initialize your Iterables with an empty instance, like on example below, you don't need to mess around with value decorators.
@jsonSerializable
class ListItem {}
@jsonSerializable
class IterablesContainer {
List<ListItem> list = [];
Set<ListItem> set = {};
}
// given
final json = '''{"list":[{}, {}],"set":[{}, {}]}''';
// when
final target = JsonMapper.deserialize<IterablesContainer>(json);
// then
expect(target.list, TypeMatcher<List<ListItem>>());
expect(target.list.first, TypeMatcher<ListItem>());
expect(target.list.length, 2);
expect(target.set, TypeMatcher<Set<ListItem>>());
expect(target.set.first, TypeMatcher<ListItem>());
expect(target.set.length, 2);
List of Lists of Lists ... #
Using value decorators, it's possible to configure nested lists of virtually any depth.
@jsonSerializable
class Item {}
@jsonSerializable
class ListOfLists {
List<List<Item>> lists;
}
// given
final json = '''{
"lists": [
[{}, {}],
[{}, {}, {}]
]
}''';
// when
JsonMapper().useAdapter(JsonMapperAdapter(
valueDecorators: {
typeOf<List<List<Item>>>(): (value) => value.cast<List<Item>>(),
typeOf<List<Item>>(): (value) => value.cast<Item>()
})
);
final target = JsonMapper.deserialize<ListOfLists>(json);
// then
expect(target.lists.length, 2);
expect(target.lists.first.length, 2);
expect(target.lists.last.length, 3);
expect(target.lists.first.first, TypeMatcher<Item>());
expect(target.lists.last.first, TypeMatcher<Item>());
Enum types #
Enum construction in Dart has a specific meaning, and has to be treated accordingly.
Enum declarations should not be annotated with @jsonSerializable
, since they are not a classes
technically, but a special built in types.
enum Color { Red, Blue, Green, Brown, Yellow, Black, White }
...
@JsonProperty(enumValues: Color.values)
Color color;
Each enum based class field has to be annotated as showed in a snippet above.
Enum.values
refers to a list of all possible enum values, it's a handy built in capability of all
enum based types. Without providing all values it's not possible to traverse it's values properly.
Inherited classes derived from abstract / base class #
Please use complementary @Json(typeNameProperty: 'typeName')
annotation for subclasses
derived from abstract or base class. This way dart-json-mapper
will dump the concrete object type to the JSON output during serialization process.
This ensures, that dart-json-mapper will be able to reconstruct the object with
the proper type during deserialization process.
@jsonSerializable
@Json(typeNameProperty: 'typeName')
abstract class Business {
String name;
}
@jsonSerializable
class Hotel extends Business {
int stars;
Hotel(this.stars);
}
@jsonSerializable
class Startup extends Business {
int userCount;
Startup(this.userCount);
}
@jsonSerializable
class Stakeholder {
String fullName;
List<Business> businesses;
Stakeholder(this.fullName, this.businesses);
}
// given
final jack = Stakeholder("Jack", [Startup(10), Hotel(4)]);
// when
JsonMapper().useAdapter(JsonMapperAdapter(
valueDecorators: {
typeOf<List<Business>>(): (value) => value.cast<Business>()
})
);
final String json = JsonMapper.serialize(jack);
final Stakeholder target = JsonMapper.deserialize(json);
// then
expect(target.businesses[0], TypeMatcher<Startup>());
expect(target.businesses[1], TypeMatcher<Hotel>());
Name casing styles [Pascal, Kebab, Snake, SnakeAllCaps] #
Assuming your Dart code is following Camel case style, but that is not
always true
for JSON models, they could follow
one of those popular - Pascal, Kebab, Snake, SnakeAllCaps styles, right?
That's why we need a way to manage that in an organized way, instead of
hand coding each property using @JsonProperty(name: ...)
it is possible to pass
CaseStyle
parameter to serialization / deserialization methods.
@jsonSerializable
class NameCaseObject {
String mainTitle;
String description;
bool hasMainProperty;
NameCaseObject({this.mainTitle, this.description, this.hasMainProperty});
}
/// Serialization
// given
final instance = NameCaseObject(
mainTitle: 'title', description: 'desc', hasMainProperty: true);
// when
final json = JsonMapper.serialize(instance,
SerializationOptions(indent: '', caseStyle: CaseStyle.Kebab));
// then
expect(json, '''{"main-title":"title","description":"desc","has-main-property":true}''');
/// Deserialization
// given
final json = '''{"main-title":"title","description":"desc","has-main-property":true}''';
// when
final instance = JsonMapper.deserialize<NameCaseObject>(
json, DeserializationOptions(caseStyle: CaseStyle.Kebab));
// then
expect(instance.mainTitle, 'title');
expect(instance.description, 'desc');
expect(instance.hasMainProperty, true);
Nesting configuration #
In case if you need to operate on particular portions of huge JSON object and you don't have a true desire to reconstruct the same deep nested JSON objects hierarchy with corresponding Dart classes. This section is for you!
Say, you have a json similar to this one
{
"root": {
"foo": {
"bar": {
"baz": {
"items": [
"a",
"b",
"c"
]
}
}
}
}
}
And with code similar to this one
@jsonSerializable
@Json(name: 'root/foo/bar')
class RootObject {
@JsonProperty(name: 'baz/items')
List<String> items;
RootObject({this.items});
}
// when
final RootObject instance = JsonMapper.deserialize(json);
// then
expect(instance.items.length, 3);
expect(instance.items, ['a', 'b', 'c']);
you'll have it done nice and quick.
@Json(name: 'root/foo/bar')
provides a root nesting for the entire annotated class,
this means all class fields will be nested under this 'root/foo/bar' path in Json.
@JsonProperty(name: 'baz/items')
provides a field nesting relative to the class root nesting
name
is compliant with RFC 6901 JSON pointer
Schemes #
Scheme - is a set of annotations associated with common scheme id. This enables the possibility to map a single Dart class to many different JSON structures.
This approach usually useful for distinguishing [DEV, PROD, TEST, ...] environments, w/o producing separate Dart classes for each environment.
enum Scheme { A, B }
@jsonSerializable
@Json(name: 'default')
@Json(name: '_', scheme: Scheme.B)
@Json(name: 'root', scheme: Scheme.A)
class Object {
@JsonProperty(name: 'title_test', scheme: Scheme.B)
String title;
Object(this.title);
}
// given
final instance = Object('Scheme A');
// when
final json = JsonMapper.serialize(instance, SerializationOptions(indent: '', scheme: Scheme.A));
// then
expect(json, '''{"root":{"title":"Scheme A"}}''');
// given
final instance = Object('Scheme B');
// when
final json = JsonMapper.serialize(instance, SerializationOptions(indent: '', scheme: Scheme.B));
// then
expect(json, '''{"_":{"title_test":"Scheme B"}}''');
// given
final instance = Object('No Scheme');
// when
final json = JsonMapper.serialize(instance, SerializationOptions(indent: ''));
// then
expect(json, '''{"default":{"title":"No Scheme"}}''');
Custom types #
For the very custom types, specific ones, or doesn't currently supported by this library, you can provide your own custom Converter class per each custom runtimeType.
/// Abstract class for custom converters implementations
abstract class ICustomConverter<T> {
dynamic toJSON(T object, [JsonProperty jsonProperty]);
T fromJSON(dynamic jsonValue, [JsonProperty jsonProperty]);
}
All you need to get going with this, is to implement this abstract class
class CustomStringConverter implements ICustomConverter<String> {
const CustomStringConverter() : super();
@override
String fromJSON(dynamic jsonValue, [JsonProperty jsonProperty]) {
return jsonValue;
}
@override
dynamic toJSON(String object, [JsonProperty jsonProperty]) {
return '_${object}_';
}
}
And register it afterwards, if you want to have it applied for all occurrences of specified type
JsonMapper().useAdapter(JsonMapperAdapter(
converters: {
String: CustomStringConverter()
})
);
OR use it individually on selected class fields, via @JsonProperty
annotation
@JsonProperty(converter: CustomStringConverter())
String title;
Annotations #
@JsonSerializable()
or@jsonSerializable
for short, It's a required class only marker annotation. Use it to mark all the Dart classes you'd like to be traveling to / from JSON- Has NO params
@JsonConstructor()
or@jsonConstructor
for short, It's an optional constructor only marker annotation. Use it to mark specific Dart class constructor you'd like to be used during deserialization.- scheme dynamic Scheme marker to associate this meta information with particular mapping scheme
@Json(...)
It's an optional class only annotation, describes a Dart class to JSON Object mapping. Why it's not a@JsonObject()
? just for you to type less characters 😄- name Defines RFC 6901 JSON pointer, denotes the json Object root name/path to be used for mapping.
Example:
'foo', 'bar', 'foo/bar/baz'
- typeNameProperty declares the necessity for annotated class and all it's subclasses to dump their own type name to the property named as this param value
- ignoreNullMembers If set to
true
Null class members will be excluded from serialization process - allowCircularReferences As of
int
type. Allows certain number of circular object references during serialization. - scheme dynamic Scheme marker to associate this meta information with particular mapping scheme
- name Defines RFC 6901 JSON pointer, denotes the json Object root name/path to be used for mapping.
Example:
@JsonProperty(...)
It's an optional class member annotation, describes JSON Object property mapping.- name Defines RFC 6901 JSON pointer, denotes the name/path to be used for property mapping relative to the class root nesting
Example:
'foo', 'bar', 'foo/bar/baz'
- scheme dynamic Scheme marker to associate this meta information with particular mapping scheme
- converter Declares custom converter instance, to be used for annotated field serialization / deserialization
- converterParams A
Map<String, dynamic>
of named parameters to be passed to the converter instance - ignore A bool declares annotated field as ignored so it will be excluded from serialization / deserialization process
- ignoreIfNull A bool declares annotated field as ignored if it's value is null so it will be excluded from serialization / deserialization process
- enumValues Provides a way to specify enum values, via Dart built in capability for all Enum instances.
Enum.values
- defaultValue Defines field default value
- name Defines RFC 6901 JSON pointer, denotes the name/path to be used for property mapping relative to the class root nesting
Example:
Complementary adapter libraries #
If you want a seamless integration with popular use cases, feel free to pick an existing adapter or create one for your use case and make a PR to this repo.
Adapter - is a library which contains a bundle of pre-configured:
- custom converters
- custom value decorators
- custom typeInfo decorators
For example, you would like to use Int32
type provided by Fixnum library
in your app.
-
Make sure you have following dependencies in your
pubspec.yaml
:dependencies: fixnum: dart_json_mapper: dart_json_mapper_fixnum: dev_dependencies: build_runner:
-
Usually, adapter library exposes
final
adapter definition instance, to be provided as a parameter toJsonMapper().useAdapter(adapter)
import 'package:fixnum/fixnum.dart' show Int32; import 'package:dart_json_mapper/dart_json_mapper.dart'; import 'package:dart_json_mapper_fixnum/dart_json_mapper_fixnum.dart'; import 'main.reflectable.dart'; // Import generated code. @jsonSerializable class FixnumExample { Int32 integer32; FixnumExample(this.integer32); } main() { initializeReflectable(); // Imported from main.reflectable.dart JsonMapper().useAdapter(fixnumAdapter); // fixnumAdapter imported from dart_json_mapper_fixnum print(JsonMapper.serialize( FixnumExample(Int32(1234567890)) )); }
output:
{ "integer32": 1234567890 }
You can easily mix and combine several adapters using following one-liner: #
JsonMapper()
.useAdapter(fixnumAdapter)
.useAdapter(mobXAdapter)
.info(); // print out a list of used adapters to console