Scala with Cats 学习笔记1-Type Class 的思考

Abstract: Scala 中有一个新的概念就是 implicit ,在阅读 Scala with Cats 这本书的第一章着重在介绍这个概念。本文记录在学习过程中个人的一些理解,从而能够帮助更好的在实际项目中应用。

应用场景

首先我们介绍一下这个 implicitType Class 的一个最典型应用场景:对象序列化

在编程中 ObjectJSON 之间的转化是一项基本的需求,在 Java 中最常用的是使用 Jackson 通过注解在类上定义其序列化的要求,比如字段名称等等,然后在应用时通过注解自动的实现了对象和 JSON 的转化。

Scala 中,我们通向希望能够达到这个目标,提前定义好序列化的方法,在需要的时候自动调用。

下面介绍一下我们的需求:

定义一个类,可以直接对实例对象转化为 JSON,比如 Person 对象,可以通过 person.toJson 或者是 XXX.toJson(person)

对象转 JSON 实现

序列化的过程无非是把不同类型的字段转化为字符串的过程。

1. 创建抽象语法树

1
2
3
4
5
6
// Define a very simple JSON AST
sealed trait Json
final case class JsObject(get: Map[String, Json]) extends Json
final case class JsString(get: String) extends Json
final case class JsNumber(get: Double) extends Json
case object JsNull extends Json

2. 创建 Type Class

A type class is an interface or API that represents some functionality we want to implement. In Cats a type class is represented by a trait with at least one type parameter.

1
2
3
4
// The "serialize to JSON" behaviour is encoded in this trait
trait JsonWriter[A] {
def write(value: A): Json
}

这个包含有一个方法的接口被称为 Type Class,为什么叫 type,参考这篇文章scala类型系统:26) type classes模式

这个模式被称为type classes,源于haskell,我对haskell不熟悉,为什么用这么奇怪的一个名字?从这儿看到:在haskell里没有类似于java/scala里的class的概念,所以class这个术语可以随意使用,它的意思相当于”class of types”,可以理解为对某一系列类型的抽象(即高阶类型)。

scala里的type classes模式主要通过隐式参数来实现,但需要注意的是,并不是所有的隐式参数都可以理解为type classes模式,隐式参数的类型必须是泛型的(高阶),它表示某种抽象行为,这种行为的具体实现要由它的具体类型参数决定。

3. 实现 Type Class 的接口

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
final case class Person(name: String, email: String)

object JsonWriterInstances {
implicit val stringWriter: JsonWriter[String] =
new JsonWriter[String] {
def write(value: String): Json =
JsString(value)
}

implicit val personWriter: JsonWriter[Person] =
new JsonWriter[Person] {
def write(value: Person): Json =
JsObject(Map(
"name" -> JsString(value.name),
"email" -> JsString(value.email)
))
}

// etc...
}

我们可以看到,我们定义 trait 的时候,类型是一个泛型,这里实现了两种类型的 write 方法,一种是基础的 String,一种是我们业务的 Person 类。

这里我们使用 implicit,说明这个方法并不会直接被调用。

4. 使用 Type Class 的接口

这里有两种方法,

使用静态方法 Interface Objects

1
2
3
4
object Json {
def toJson[A](value: A)(implicit w: JsonWriter[A]): Json =
w.write(value)
}

使用接口文法 Interface Syntax

1
2
3
4
5
6
object JsonSyntax {
implicit class JsonWriterOps[A](value: A) {
def toJson(implicit w: JsonWriter[A]): Json =
w.write(value)
}
}

5. 应用

Interface Objects

1
2
3
4
5
import JsonWriterInstances._

Json.toJson(Person("Dave", "[email protected]"))
// 等价于
//Json.toJson(Person("Dave", "[email protected]"))(personWriter)

Interface Syntax

1
2
3
4
import JsonWriterInstances._
import JsonSyntax._

Person("Dave", "[email protected]").toJson

原理

当我们最后调用 Json.toJson 或者是 .toJson 的时候,因为其定义带有 implicit w: JsonWriter[A] 这个隐含参数,所以 Scala 就会尝试去寻找对应匹配类型的 JsonWriter,如果是 String 就会找到 stringWriter, 如果是 Person 就会找到 personWriter。然后在使用这个函数传入到 Json.toJson 或者是 .toJson 方法中,最后调用其 write 方法实现转换。

总的来说,就是其他定义 implicit 方法,然后传入参数指定其为 implicit,最后 Scala 自动寻找匹配类型的隐含方法并应用。

Cats Show 做了什么?

我们阅读源码 org.typelevel/cats-core_2.12/srcs/cats-core_2.12-1.4.0-sources.jar!/cats/Show.scala

1
2
3
4
5
6
7
8
9
10
11
12
/**
* A type class to provide textual representation. It is meant to be a
* better "toString". Whereas toString exists for any Object,
* regardless of whether or not the creator of the class explicitly
* made a toString method, a Show instance will only exist if someone
* explicitly provided one.
*/
trait Show[T] extends Show.ContravariantShow[T]

trait ContravariantShow[-T] extends Serializable {
def show(t: T): String
}

这里的 Show 和上面的 JsonWriter 同样的角色,定义了 show 方法。

对于 Show 的实现并不在这个文件中,但是 Cats 包含了一个 package object instances 这里面有很多基础类型的 instance, 比如:

1
2
3
4
trait StringInstances extends cats.kernel.instances.StringInstances {
implicit val catsStdShowForString: Show[String] =
Show.fromToString[String]
}

接口定义:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
trait Ops[A] {
def typeClassInstance: Show[A]
def self: A
def show: String = typeClassInstance.show(self)
}

trait ToShowOps {
implicit def toShow[A](target: A)(implicit tc: Show[A]): Ops[A] = new Ops[A] {
val self = target
val typeClassInstance = tc
}
}

trait ShowSyntax extends Show.ToShowOps {
implicit final def showInterpolator(sc: StringContext): Show.ShowInterpolator = Show.ShowInterpolator(sc)
}

如何使用呢?只需要定义一个 我们需要类型的 instance,就可以直接使用了。

1
2
3
4
5
6
7
8
9
implicit val catShow = Show.show[Cat] { cat =>
import cats.instances.int._ // for Show
import cats.instances.string._ // for Show
val name = cat.name.show
val age = cat.age.show
val color = cat.color.show
s"$name is a $age year-old $color cat."
}
println(Cat("Garfield", 38, "ginger and black").show)

总结

Scala 的 Type Class 模式三部曲:

  1. Type Class Definition
  2. Type Class Instance
  3. Type Class Interface (Syntax/Objects)

另外这段话说的很好:

简单总结,type classes模式通过泛型来描述某种通用行为,对每个想要用到这种行为的具体类型,在现实时行为部分并不放在类型自身中,而是通过实现一个type class实例(对泛型具化),最后在调用时(通过隐式参数)让编译器自动寻找行为的实现。

我们可以那这 Show 思考一下,传统做法是什么呢? 应该是定义一个统一的接口,接口中有一个 show 的方法,然后哪个类想实现 show,就实现这个接口。如果出现类型依赖,就要依赖其 show 的实现。(其实这就是 Java Object 类的 toString)

虽然这样做并没有问题,但是 type class 模式进一步解耦合,类型天生并不绑定的其能力,能力是单独定义 (Type Class),一个类型如果需要这种能力,只需要单独创建其能力的实现(Instance),并由编译器自动绑定。

有时间我们可以阅读以下 Play Framework 关于请求 Json 序列化过程的代码 (或者 Lagom Framework)里面大量使用了这种模式。

Gist