module JSON::Serializable
¶
The JSON::Serializable
module automatically generates methods for JSON serialization when included.
Example¶
require "json"
class Location
include JSON::Serializable
@[JSON::Field(key: "lat")]
property latitude : Float64
@[JSON::Field(key: "lng")]
property longitude : Float64
end
class House
include JSON::Serializable
property address : String
property location : Location?
end
house = House.from_json(%({"address": "Crystal Road 1234", "location": {"lat": 12.3, "lng": 34.5}}))
house.address # => "Crystal Road 1234"
house.location # => #<Location:0x10cd93d80 @latitude=12.3, @longitude=34.5>
house.to_json # => %({"address":"Crystal Road 1234","location":{"lat":12.3,"lng":34.5}})
houses = Array(House).from_json(%([{"address": "Crystal Road 1234", "location": {"lat": 12.3, "lng": 34.5}}]))
houses.size # => 1
houses.to_json # => %([{"address":"Crystal Road 1234","location":{"lat":12.3,"lng":34.5}}])
Usage¶
Including JSON::Serializable
will create #to_json
and self.from_json
methods on the current class,
and a constructor which takes a JSON::PullParser
. By default, these methods serialize into a json
object containing the value of every instance variable, the keys being the instance variable name.
Most primitives and collections supported as instance variable values (string, integer, array, hash, etc.),
along with objects which define to_json and a constructor taking a JSON::PullParser
.
Union types are also supported, including unions with nil. If multiple types in a union parse correctly,
it is undefined which one will be chosen.
To change how individual instance variables are parsed and serialized, the annotation JSON::Field
can be placed on the instance variable. Annotating property, getter and setter macros is also allowed.
require "json"
class A
include JSON::Serializable
@[JSON::Field(key: "my_key", emit_null: true)]
getter a : Int32?
end
JSON::Field
properties:
* ignore: if true
skip this field in serialization and deserialization (by default false)
* ignore_serialize: if true
skip this field in serialization (by default false)
* ignore_deserialize: if true
skip this field in deserialization (by default false)
* key: the value of the key in the json object (by default the name of the instance variable)
* root: assume the value is inside a JSON object with a given key (see Object.from_json(string_or_io, root)
)
* converter: specify an alternate type for parsing and generation. The converter must define from_json(JSON::PullParser)
and to_json(value, JSON::Builder)
. Examples of converters are a Time::Format
instance and Time::EpochConverter
for Time
.
* presence: if true
, a @{{key}}_present
instance variable will be generated when the key was present (even if it has a null
value), false
by default
* emit_null: if true
, emits a null
value for nilable property (by default nulls are not emitted)
Deserialization also respects default values of variables:
require "json"
struct A
include JSON::Serializable
@a : Int32
@b : Float64 = 1.0
end
A.from_json(%<{"a":1}>) # => A(@a=1, @b=1.0)
Extensions: JSON::Serializable::Strict
and JSON::Serializable::Unmapped
.¶
If the JSON::Serializable::Strict
module is included, unknown properties in the JSON
document will raise a parse exception. By default the unknown properties
are silently ignored.
If the JSON::Serializable::Unmapped
module is included, unknown properties in the JSON
document will be stored in a Hash(String, JSON::Any)
. On serialization, any keys inside json_unmapped
will be serialized and appended to the current json object.
require "json"
struct A
include JSON::Serializable
include JSON::Serializable::Unmapped
@a : Int32
end
a = A.from_json(%({"a":1,"b":2})) # => A(@json_unmapped={"b" => 2_i64}, @a=1)
a.to_json # => {"a":1,"b":2}
Class annotation JSON::Serializable::Options
¶
supported properties:
* emit_nulls: if true
, emits a null
value for all nilable properties (by default nulls are not emitted)
require "json"
@[JSON::Serializable::Options(emit_nulls: true)]
class A
include JSON::Serializable
@a : Int32?
end
Discriminator field¶
A very common JSON serialization strategy for handling different objects under a same hierarchy is to use a discriminator field. For example in GeoJSON each object has a "type" field, and the rest of the fields, and their meaning, depend on its value.
You can use JSON::Serializable.use_json_discriminator
for this use case.
Direct including types
Crystal::ContextResult
Crystal::ExpandResult
Crystal::ExpandResult::Expansion
Crystal::ImplementationResult
Crystal::ImplementationTrace
Crystal::Program::RecordedRequire
Crystal::Program::RequireWithTimestamp
Class methods¶
Methods¶
Macros¶
use_json_discriminator(field, mapping)
¶
(field, mapping)
Tells this class to decode JSON by using a field as a discriminator.
- field must be the field name to use as a discriminator
- mapping must be a hash or named tuple where each key-value pair maps a discriminator value to a class to deserialize
For example:
require "json"
abstract class Shape
include JSON::Serializable
use_json_discriminator "type", {point: Point, circle: Circle}
property type : String
end
class Point < Shape
property x : Int32
property y : Int32
end
class Circle < Shape
property x : Int32
property y : Int32
property radius : Int32
end
Shape.from_json(%({"type": "point", "x": 1, "y": 2})) # => #<Point:0x10373ae20 @type="point", @x=1, @y=2>
Shape.from_json(%({"type": "circle", "x": 1, "y": 2, "radius": 3})) # => #<Circle:0x106a4cea0 @type="circle", @x=1, @y=2, @radius=3>