使用 testcontainers 测试 kafka 和 spark

Question

我正在尝试与 testcontainers 一起检查流式管道作为集成测试，但我不知道如何获取 bootstrapServers，至少在最后一个 testcontainers 版本中并在那里创建特定主题。如何使用 'containerDef' 提取引导服务器并添加主题？

import com.dimafeng.testcontainers.{ContainerDef, KafkaContainer}
import com.dimafeng.testcontainers.scalatest.TestContainerForAll
import munit.FunSuite
import org.apache.spark.sql.SparkSession

class Mykafkatest extends FunSuite with TestContainerForAll {
  //val kafkaContainer: KafkaContainer      = KafkaContainer("confluentinc/cp-kafka:5.4.3")
  override val containerDef: ContainerDef = KafkaContainer.Def()

  test("do something")(withContainers { container =>
    val sparkSession: SparkSession = SparkSession
      .builder()
      .master("local[*]")
      .appName("Unit testing")
      .getOrCreate()

    // How add a topic in that container?

    // This is not posible:
    val servers=container.bootstrapServers

    val df = sparkSession.readStream
      .format("kafka")
      .option("kafka.bootstrap.servers", servers)
      .option("subscribe", "topic1")
      .load()


    df.show(false)

  })

}

我的 sbt 配置：

lazy val root = project
  .in(file("./pipeline"))
  .settings(
    organization := "org.example",
    name := "spark-stream",
    version := "0.1",
    scalaVersion := "2.12.10",
    libraryDependencies := Seq(
      "org.apache.spark" %% "spark-sql-kafka-0-10"       % "3.0.3"  % Compile,
      "org.apache.spark" %% "spark-sql"                  % "3.0.3"  % Compile,
      "com.dimafeng"     %% "testcontainers-scala-munit" % "0.39.5" % Test,
      "org.dimafeng"     %% "testcontainers-scala-kafka" % "0.39.5" % Test,
      "org.scalameta"    %% "munit"                      % "0.7.28" % Test
    ),
    testFrameworks += new TestFramework("munit.Framework"),
    Test / fork := true
  )

文档未显示完整示例：https://www.testcontainers.org/modules/kafka/

Answer 1

这里唯一的问题是您将 KafkaContainer.Def 显式转换为 ContainerDef。

withContianers、Containter提供的容器类型由path dependent type在提供的ContainerDef决定，

trait TestContainerForAll extends TestContainersForAll { self: Suite =>

  val containerDef: ContainerDef

  final override type Containers = containerDef.Container

  override def startContainers(): containerDef.Container = {
    containerDef.start()
  }

  // inherited from TestContainersSuite
  def withContainers[A](runTest: Containers => A): A = {
    val c = startedContainers.getOrElse(throw IllegalWithContainersCall())
    runTest(c)
  }

}

trait ContainerDef {

  type Container <: Startable with Stoppable

  protected def createContainer(): Container

  def start(): Container = {
    val container = createContainer()
    container.start()
    container
  }
}

当您在 override val containerDef: ContainerDef = KafkaContainer.Def() 中明确指定类型 ContainerDef 时，这会破坏整个“类型诡计”，因此 Scala 编译器会留下 type Container <: Startable with Stoppable 而不是 KafkaContainer。

因此，只需删除转换为 ContainerDef 的显式类型，val servers = container.bootstrapServers 就会按预期工作。

import com.dimafeng.testcontainers.KafkaContainer
import com.dimafeng.testcontainers.munit.TestContainerForAll
import munit.FunSuite

class Mykafkatest extends FunSuite with TestContainerForAll {
  override val containerDef = KafkaContainer.Def()

  test("do something")(withContainers { container =>
    //...

    val servers = container.bootstrapServers

    println(servers)

    //...
  })
}