好得很程序员自学网

<tfoot draggable='sEl'></tfoot>

SpringBoot 应用程序测试实现方案

Pre

本篇博文我们开始梳理下Spring 提供的测试解决方案。

对于 Web 应用程序而言, 一个应用程序中涉及数据层、服务层、Web 层,以及各种外部服务之间的交互关系时,我们除了对各层组件的单元测试之外,还需要充分引入集成测试保证服务的正确性和稳定性。

Spring Boot 中的测试解决方案

和 Spring Boot 1.x 版本一样,Spring Boot 2.x 也提供了一个用于测试的 spring-boot-starter-test 组件。

在 Spring Boot 中,集成该组件的方法是在 pom 文件中添加如下所示依赖:

?

1

2

3

4

5

6

7

8

9

10

11

< dependency >

     < groupId >org.springframework.boot</ groupId >

     < artifactId >spring-boot-starter-test</ artifactId >

     < scope >test</ scope >

</ dependency >

        

< dependency >

     < groupId >org.junit.platform</ groupId >

     < artifactId >junit-platform-launcher</ artifactId >

     < scope >test</ scope >

</ dependency >

其中,最后一个依赖用于导入与 JUnit 相关的功能组件。

然后,通过 Maven 查看 spring-boot-starter-test 组件的依赖关系,我们可以得到如下所示的组件依赖图:

从上图中可以看到,在代码工程的构建路径中,我们引入了一系列组件初始化测试环境。比如 JUnit、JSON Path、AssertJ、Mockito、Hamcrest 等

JUnit :JUnit 是一款非常流行的基于 Java 语言的单元测试框架 JSON Path :类似于 XPath 在 XML 文档中的定位,JSON Path 表达式通常用来检索路径或设置 JSON 文件中的数据。 AssertJ :AssertJ 是一款强大的流式断言工具,它需要遵守 3A 核心原则,即 Arrange(初始化测试对象或准备测试数据)——> Actor(调用被测方法)——>Assert(执行断言)。 Mockito :Mockito 是 Java 世界中一款流行的 Mock 测试框架,它主要使用简洁的 API 实现模拟操作。在实施集成测试时,我们将大量使用到这个框架。 Hamcrest :Hamcrest 提供了一套匹配器(Matcher),其中每个匹配器的设计用于执行特定的比较操作。 JSONassert :JSONassert 是一款专门针对 JSON 提供的断言框架。 Spring Test & Spring Boot Test :为 Spring 和 Spring Boot 框架提供的测试工具。

以上组件的依赖关系都是自动导入, 无须做任何变动。

测试 Spring Boot 应用程序

接下来,我们将初始化 Spring Boot 应用程序的测试环境,并介绍如何在单个服务内部完成单元测试的方法和技巧。

导入 spring-boot-starter-test 依赖后,我们就可以使用它提供的各项功能应对复杂的测试场景了。

初始化测试环境

对于 Spring Boot 应用程序而言,我们知道其 Bootstrap 类中的 main() 入口将通过 SpringApplication.run() 方法启动 Spring 容器.

?

1

2

3

4

5

6

7

8

import org.springframework.boot.SpringApplication;

import org.springframework.boot.autoconfigure.SpringBootApplication;

@SpringBootApplication

public class Application {

     public static void main(String[] args) {

         SpringApplication.run(Application. class , args);

     }

}

针对上述 Bootstrap 类,我们可以通过编写测试用例的方式,验证 Spring 容器能否正常启动。

基于 Maven 的默认风格,我们将在 src/test/java 和 src/test/resources 包下添加各种测试用例代码和配置文件。

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

import org.junit.Assert;

import org.junit.Test;

import org.junit.runner.RunWith;

import org.springframework.beans.factory.annotation.Autowired;

import org.springframework.boot.test.context.SpringBootTest;

import org.springframework.context.ApplicationContext;

import org.springframework.test.context.junit4.SpringRunner;

 

@SpringBootTest

@RunWith (SpringRunner. class )

public class ApplicationContextTests {

 

     @Autowired

     private ApplicationContext applicationContext;

 

     @Test

     public void testContextLoads() throws Throwable {

         Assert.assertNotNull( this .applicationContext);

     }

}

该用例对 Spring 中的 ApplicationContext 作了简单非空验证。

执行该测试用例后,从输出的控制台信息中,我们可以看到 Spring Boot 应用程序被正常启动,同时测试用例本身也会给出执行成功的提示。

上述测试用例虽然简单,但是已经包含了测试 Spring Boot 应用程序的基本代码框架。其中,最重要的是 ApplicationContextTests 类上的 @SpringBootTest 和 @RunWith 注解,对于 Spring Boot 应用程序而言, 这两个注解构成了一套完成的测试方案。

接下来我们对这两个注解进行详细展开。

@SpringBootTest

因为 SpringBoot 程序的入口是 Bootstrap 类,所以 SpringBoot 专门提供了一个 @SpringBootTest 注解测试 Bootstrap 类。同时 @SpringBootTest 注解也可以引用 Bootstrap 类的配置,因为所有配置都会通过 Bootstrap 类去加载。

在上面的例子中,我们是通过直接使用 @SpringBootTest 注解提供的默认功能对作为 Bootstrap 类的 Application 类进行测试。

而更常见的做法是在 @SpringBootTest 注解中指定该 Bootstrap 类,并设置测试的 Web 环境,如下代码所示。

?

1

2

@SpringBootTest (classes = CustomerApplication. class ,

     webEnvironment = SpringBootTest.WebEnvironment.MOCK)

在以上代码中,@SpringBootTest 注解中的 webEnvironment 可以有四个选项,分别是 MOCK、RANDOM_PORT、DEFINED_PORT 和 NONE。

@SpringBootTest - webEnvironment

MOCK :加载 WebApplicationContext 并提供一个 Mock 的 Servlet 环境,此时内置的 Servlet 容器并没有正式启动。 RANDOM_PORT :加载 EmbeddedWebApplicationContext 并提供一个真实的 Servlet 环境,然后使用一个随机端口启动内置容器。 DEFINED_PORT :这个配置也是通过加载 EmbeddedWebApplicationContext 提供一个真实的 Servlet 环境,但使用的是默认端口,如果没有配置端口就使用 8080。 NONE :加载 ApplicationContext 但并不提供任何真实的 Servlet 环境。

在 Spring Boot 中,@SpringBootTest 注解主要用于测试基于自动配置的 ApplicationContext,它允许我们设置测试上下文中的 Servlet 环境。

在多数场景下,一个真实的 Servlet 环境对于测试而言过于重量级,通过 MOCK 环境则可以缓解这种环境约束所带来的困扰

@RunWith 注解与 SpringRunner

在上面的示例中,我们还看到一个由 JUnit 框架提供的 @RunWith 注解,它用于设置测试运行器。例如,我们可以通过 @RunWith(SpringJUnit4ClassRunner.class) 让测试运行于 Spring 测试环境。

虽然这我们指定的是 SpringRunner.class,实际上, SpringRunner 就是 SpringJUnit4ClassRunner 的简化,它允许 JUnit 和 Spring TestContext 整合运行,而 Spring TestContext 则提供了用于测试 Spring 应用程序的各项通用的支持功能。

执行测试用例

接下来我们将通过代码示例回顾如何使用 JUnit 框架执行单元测试的过程和实践,同时提供验证异常和验证正确性的测试方法。

单元测试的应用场景是一个独立的类,如下所示的 CustomerTicket 类就是一个非常典型的独立类:

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

public class CustomTicket {

     private Long id;

     private Long accountId;   

     private String orderNumber;

     private String description;

     private Date createTime;

     public CustomTicket (Long accountId, String orderNumber) {

         super ();

         Assert.notNull(accountId, "Account Id must not be null" );

         Assert.notNull(orderNumber, "Order Number must not be null" );

         Assert.isTrue(orderNumber.length() == 10 , "Order Number must be exactly 10 characters" );

         this .accountId = accountId;

         this .orderNumber = orderNumber;

     }

        …

}

我们可以看到,该类对CustomTicket 做了封装,并在其构造函数中添加了校验机制。

下面我们先来看看如何对正常场景进行测试。

例如 ArtisanTicket 中orderNumber 的长度问题,我们可以使用如下测试用例,通过在构造函数中传入字符串来验证规则的正确性:

?

1

2

3

4

5

6

7

8

9

@RunWith (SpringRunner. class )

public class CustomerTicketTests {

     private static final String ORDER_NUMBER = "Order00001" ;

     @Test

     public void testOrderNumberIsExactly10Chars() throws Exception {

         CustomerTicket customerTicket = new CustomerTicket(100L, ORDER_NUMBER);

         assertThat(customerTicket.getOrderNumber().toString()).isEqualTo(ORDER_NUMBER);

     }

}

使用 @DataJpaTest 注解测试数据访问组件

数据需要持久化,接下来我们将从数据持久化的角度出发,讨论如何对 Repository 层进行测试的方法。

首先,我们讨论一下使用关系型数据库的场景,并引入 针对 JPA 数据访问技术的 @DataJpaTest 注解 。

@DataJpaTest 注解会自动注入各种 Repository 类,并初始化一个内存数据库和及访问该数据库的数据源。在测试场景下,一般我们可以使用 H2 作为内存数据库,并通过 MySQL 实现数据持久化 ,因此我们需要引入以下所示的 Maven 依赖:

?

1

2

3

4

5

6

7

8

9

< dependency >

        < groupId >com.h2database</ groupId >

        < artifactId >h2</ artifactId >

</ dependency >

< dependency >

        < groupId >mysql</ groupId >

        < artifactId >mysql-connector-java</ artifactId >

        < scope >runtime</ scope >

</ dependency >

另一方面,我们需要准备数据库 DDL 用于初始化数据库表,并提供 DML 脚本完成数据初始化。其中,schema-mysql.sql 和 data-h2.sql 脚本分别充当了 DDL 和 DML 的作用。

在 customer-service 的 schema-mysql.sql 中包含了 CUSTOMER 表的创建语句,如下代码所示:

?

1

2

3

4

5

6

7

8

9

DROP TABLE IF EXISTS `customerticket`;

create table `customerticket` (

        `id` bigint (20) NOT NULL AUTO_INCREMENT,

        `account_id` bigint (20) not null ,

        `order_number` varchar (50) not null ,

        `description` varchar (100) not null ,

         `create_time` timestamp not null DEFAULT CURRENT_TIMESTAMP ,

        PRIMARY KEY (`id`)

);

而在 data-h2.sql 中,我们插入了一条测试需要使用的数据,具体的初始化数据过程如下代码所示:

?

1

INSERT INTO customerticket (`account_id`, `order_number`,`description`) values (1, 'Order00001' , ' DemoCustomerTicket1' );

接下来是提供具体的 Repository 接口,我们先通过如下所示代码回顾一下 CustomerRepository 接口的定义。

?

1

2

3

public interface CustomerTicketRepository extends JpaRepository<CustomerTicket, Long> {

     List<CustomerTicket> getCustomerTicketByOrderNumber(String orderNumber);

}

这里存在一个方法名衍生查询 getCustomerTicketByOrderNumber,它会根据 OrderNumber 获取 CustomerTicket。

基于上述 CustomerRepository,我们可以编写如下所示的测试用例:

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

@RunWith (SpringRunner. class )

@DataJpaTest

public class CustomerRepositoryTest {

     @Autowired

     private TestEntityManager entityManager;

 

     @Autowired

     private CustomerTicketRepository customerTicketRepository;

 

     @Test

     public void testFindCustomerTicketById() throws Exception {            

         this .entityManager.persist( new CustomerTicket(1L, "Order00001" , "DemoCustomerTicket1" , new Date()));

        

         CustomerTicket customerTicket = this .customerTicketRepository.getOne(1L);

         assertThat(customerTicket).isNotNull();

         assertThat(customerTicket.getId()).isEqualTo(1L);

     }

        

     @Test

     public void testFindCustomerTicketByOrderNumber() throws Exception {   

         String orderNumber = "Order00001" ;

        

         this .entityManager.persist( new CustomerTicket(1L, orderNumber, "DemoCustomerTicket1" , new Date()));

         this .entityManager.persist( new CustomerTicket(2L, orderNumber, "DemoCustomerTicket2" , new Date()));

        

         List<CustomerTicket> customerTickets = this .customerTicketRepository.getCustomerTicketByOrderNumber(orderNumber);

         assertThat(customerTickets).size().isEqualTo( 2 );

         CustomerTicket actual = customerTickets.get( 0 );

         assertThat(actual.getOrderNumber()).isEqualTo(orderNumber);

     }

 

     @Test

     public void testFindCustomerTicketByNonExistedOrderNumber() throws Exception {             

         this .entityManager.persist( new CustomerTicket(1L, "Order00001" , "DemoCustomerTicket1" , new Date()));

         this .entityManager.persist( new CustomerTicket(2L, "Order00002" , "DemoCustomerTicket2" , new Date()));

        

         List<CustomerTicket> customerTickets = this .customerTicketRepository.getCustomerTicketByOrderNumber( "Order00003" );

         assertThat(customerTickets).size().isEqualTo( 0 );

     }

}

这里可以看到,我们使用了 @DataJpaTest 实现 CustomerRepository 的注入。同时,我们还注意到另一个核心测试组件 TestEntityManager,它的效果相当于不使用真正的 CustomerRepository 完成数据的持久化,从而提供了一种数据与环境之间的隔离机制。

执行这些测试用例后,我们需要关注它们的控制台日志输入,其中核心日志如下所示(为了显示做了简化处理):

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

Hibernate: drop table customer_ticket if exists

Hibernate: drop sequence if exists hibernate_sequence

Hibernate: create sequence hibernate_sequence start with 1 increment by 1

Hibernate: create table customer_ticket (id bigint not null , account_id bigint, create_time timestamp, description varchar( 255 ), order_number varchar( 255 ), primary key (id))

Hibernate: create table localaccount (id bigint not null , account_code varchar( 255 ), account_name varchar( 255 ), primary key (id))

Hibernate: call next value for hibernate_sequence

Hibernate: call next value for hibernate_sequence

Hibernate: insert into customer_ticket (account_id, create_time, description, order_number, id) values (?, ?, ?, ?, ?)

Hibernate: insert into customer_ticket (account_id, create_time, description, order_number, id) values (?, ?, ?, ?, ?)

Hibernate: select customerti0_.id as id1_0_, customerti0_.account_id as account_2_0_, customerti0_.create_time as create_t3_0_, customerti0_.description as descript4_0_, customerti0_.order_number as order_nu5_0_ from customer_ticket customerti0_ where customerti0_.order_number=?

Hibernate: drop table customer_ticket if exists

Hibernate: drop sequence if exists hibernate_sequence

从以上日志中,我们不难看出执行各种 SQL 语句的效果。

Service层和Controller的测试

与位于底层的数据访问层不同,这两层的组件都依赖于它的下一层组件,即 Service 层依赖于数据访问层,而 Controller 层依赖于 Service 层。因此,对这两层进行测试时,我们将使用不同的方案和技术。

使用 Environment 测试配置信息

在 Spring Boot 应用程序中,Service 层通常依赖于配置文件,所以我们也需要对配置信息进行测试。

配置信息的测试方案分为两种,第一种依赖于物理配置文件,第二种则是在测试时动态注入配置信息。

第一种测试方案比较简单,在 src/test/resources 目录下添加配置文件时,Spring Boot 能读取这些配置文件中的配置项并应用于测试案例中。

在介绍具体的实现过程之前,我们有必要先来了解一下 Environment 接口,该接口定义如下:

?

1

2

3

4

5

public interface Environment extends PropertyResolver {

     String[] getActiveProfiles();

     String[] getDefaultProfiles();

     boolean acceptsProfiles(String... profiles);

}

在上述代码中我们可以看到,Environment 接口的主要作用是处理 Profile,而它的父接口 PropertyResolver 定义如下代码所示:

?

1

2

3

4

5

6

7

8

9

10

11

public interface PropertyResolver {

     boolean containsProperty(String key);

     String getProperty(String key);

     String getProperty(String key, String defaultValue);

     <T> T getProperty(String key, Class<T> targetType);

     <T> T getProperty(String key, Class<T> targetType, T defaultValue);

     String getRequiredProperty(String key) throws IllegalStateException;

     <T> T getRequiredProperty(String key, Class<T> targetType) throws IllegalStateException;

     String resolvePlaceholders(String text);

     String resolveRequiredPlaceholders(String text) throws IllegalArgumentException;

}

显然,PropertyResolver 的作用是根据各种配置项的 Key 获取配置属性值。

现在,假设 src/test/resources 目录中的 application.properties 存在如下配置项:

?

1

springcss.order.point = 10

那么,我们就可以设计如下所示的测试用例了。

?

1

2

3

4

5

6

7

8

9

10

@RunWith (SpringRunner. class )

@SpringBootTest

public class EnvironmentTests{

     @Autowired

     public Environment environment;

     @Test

     public void testEnvValue(){

         Assert.assertEquals( 10 , Integer.parseInt(environment.getProperty( "springcss.order.point" )));

     }

}

这里我们注入了一个 Environment 接口,并调用了它的 getProperty 方法来获取测试环境中的配置信息。

除了在配置文件中设置属性,我们也可以使用 @SpringBootTest 注解指定用于测试的属性值,示例代码如下:

?

1

2

3

4

5

6

7

8

9

10

@RunWith (SpringRunner. class )

@SpringBootTest (properties = { " springcss.order.point = 10" })

public class EnvironmentTests{

     @Autowired

     public Environment environment;

     @Test

     public void testEnvValue(){

         Assert.assertEquals( 10 , Integer.parseInt(environment.getProperty( "springcss.order.point" )));

     }

}

使用 Mock 测试 Service 层

Service 层依赖于数据访问层。因此,对 Service 层进行测试时,我们还需要引入新的技术体系,也就是应用非常广泛的 Mock 机制。

接下来,我们先看一下 Mock 机制的基本概念。

Mock 机制

Mock 的意思是模拟,它可以用来对系统、组件或类进行隔离。

在测试过程中,我们通常关注测试对象本身的功能和行为,而对测试对象涉及的一些依赖,仅仅关注它们与测试对象之间的交互(比如是否调用、何时调用、调用的参数、调用的次数和顺序,以及返回的结果或发生的异常等),并不关注这些被依赖对象如何执行这次调用的具体细节。

因此,Mock 机制就是使用 Mock 对象替代真实的依赖对象,并模拟真实场景来开展测试工作。

使用 Mock 对象完成依赖关系测试的示意图如下所示:

可以看出,在形式上,Mock 是在测试代码中直接 Mock 类和定义 Mock 方法的行为,通常测试代码和 Mock 代码放一起。因此,测试代码的逻辑从测试用例的代码上能很容易地体现出来。

下面我们一起看一下如何使用 Mock 测试 Service 层。

使用 Mock

@SpringBootTest 注解中的 SpringBootTest.WebEnvironment.MOCK 选项,该选项用于加载 WebApplicationContext 并提供一个 Mock 的 Servlet 环境,内置的 Servlet 容器并没有真实启动。接下来,我们针对 Service 层演示一下这种测试方式。

首先,我们来看一种简单场景,在 customer-service 中存在如下 CustomerTicketService 类:

?

1

2

3

4

5

6

7

8

9

@Service

public class CustomerTicketService {

     @Autowired

     private CustomerTicketRepository customerTicketRepository;

     public CustomerTicket getCustomerTicketById(Long id) {

         return customerTicketRepository.getOne(id);

     }

     …

}

这里我们可以看到,以上方法只是简单地通过 CustomerTicketRepository 完成了数据查询操作。

显然,对以上 CustomerTicketService 进行集成测试时,还需要我们提供一个 CustomerTicketRepository 依赖。

下面,我们通过以下代码 演示一下如何使用 Mock 机制完成对 CustomerTicketRepository 的隔离 。

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

@RunWith (SpringRunner. class )

@SpringBootTest (webEnvironment = SpringBootTest.WebEnvironment.MOCK)

public class CustomerServiceTests {

     @MockBean

     private CustomerTicketRepository customerTicketRepository;

     @Test

     public void testGetCustomerTicketById() throws Exception {

         Long id = 1L;

        Mockito.when(customerTicketRepository.getOne(id)).thenReturn( new CustomerTicket(1L, 1L, "Order00001" , "DemoCustomerTicket1" , new Date()));

         CustomerTicket actual = customerTicketService.getCustomerTicketById(id);

         assertThat(actual).isNotNull();

         assertThat(actual.getOrderNumber()).isEqualTo( "Order00001" );

     }

}

首先,我们通过 @MockBean 注解注入了 CustomerTicketRepository;然后,基于第三方 Mock 框架 Mockito 提供的 when/thenReturn 机制完成了对 CustomerTicketRepository 中 getCustomerTicketById() 方法的 Mock。

当然, 如果你希望在测试用例中直接注入真实的CustomerTicketRepository,这时就可以使用@SpringBootTest 注解中的 SpringBootTest.WebEnvironment.RANDOM_PORT 选项,示例代码如下:

?

1

2

3

4

5

6

7

8

9

10

11

12

13

@RunWith (SpringRunner. class )

@SpringBootTest (webEnvironment = SpringBootTest.WebEnvironment.RANDOM_PORT)

public class CustomerServiceTests {

     @Autowired

     private CustomerTicketRepository customerTicketRepository;

     @Test

     public void testGetCustomerTicketById() throws Exception {

         Long id = 1L;

         CustomerTicket actual = customerTicketService.getCustomerTicketById(id);

         assertThat(actual).isNotNull();

         assertThat(actual.getOrderNumber()).isEqualTo( "Order00001" );

     }

}

运行上述代码后就会以一个随机的端口启动整个 Spring Boot 工程,并从数据库中真实获取目标数据进行验证。

以上集成测试的示例中只包含了对 Repository 层的依赖,而有时候一个 Service 中可能同时包含 Repository 和其他 Service 类或组件,下面回到如下所示的 CustomerTicketService 类:

?

1

2

3

4

5

6

7

8

9

@Service

public class CustomerTicketService {

     @Autowired

     private OrderClient orderClient;

     private OrderMapper getRemoteOrderByOrderNumber(String orderNumber) {

         return orderClient.getOrderByOrderNumber(orderNumber);

     }

     …

}

这里我们可以看到,在该代码中,除了依赖 CustomerTicketRepository 之外,还同时依赖了 OrderClient。

请注意:以上代码中的 OrderClient 是在 customer-service 中通过 RestTemplate 访问 order-service 的远程实现类,其代码如下所示:

?

1

2

3

4

5

6

7

8

9

10

11

12

@Component

public class OrderClient {

     @Autowired

     RestTemplate restTemplate;

     public OrderMapper getOrderByOrderNumber(String orderNumber) {

         ResponseEntity<OrderMapper> restExchange = restTemplate.exchange(

                 "http://localhost:8083/orders/{orderNumber}" , HttpMethod.GET, null ,

                 OrderMapper. class , orderNumber);

          OrderMapper result = restExchange.getBody();

         return result;

     }

}

CustomerTicketService 类实际上并不关注 OrderClient 中如何实现远程访问的具体过程。因为对于集成测试而言,它只关注方法调用返回的结果,所以我们将同样采用 Mock 机制完成对 OrderClient 的隔离。

对 CustomerTicketService 这部分功能的测试用例代码如下所示,可以看到,我们采用的是同样的测试方式。

?

1

2

3

4

5

6

7

8

9

10

11

12

@Test

public void testGenerateCustomerTicket() throws Exception {

         Long accountId = 100L;

         String orderNumber = "Order00001" ;

         Mockito.when( this .orderClient.getOrderByOrderNumber( "Order00001" ))

             .thenReturn( new OrderMapper(1L, orderNumber, "deliveryAddress" ));

         Mockito.when( this .localAccountRepository.getOne(accountId))

             .thenReturn( new LocalAccount(100L, "accountCode" , "accountName" ));

 

         CustomerTicket actual = customerTicketService.generateCustomerTicket(accountId, orderNumber);

         assertThat(actual.getOrderNumber()).isEqualTo(orderNumber);

}

这里提供的测试用例演示了 Service 层中进行集成测试的各种手段,它们已经能够满足一般场景的需要。

测试 Controller 层

对 Controller 层进行测试之前,我们先来提供一个典型的 Controller 类,它来自 customer-service,如下代码所示:

?

1

2

3

4

5

6

7

8

9

10

11

12

@RestController

@RequestMapping (value= "customers" )

public class CustomerController {

     @Autowired

     private CustomerTicketService customerTicketService;

     @PostMapping (value = "/{accountId}/{orderNumber}" )

     public CustomerTicket generateCustomerTicket( @PathVariable ( "accountId" ) Long accountId,

             @PathVariable ( "orderNumber" ) String orderNumber) {

         CustomerTicket customerTicket = customerTicketService.generateCustomerTicket(accountId, orderNumber);

         return customerTicket;

     }

}

关于上述 Controller 类的测试方法,相对来说比较丰富,比如有 TestRestTemplate、@WebMvcTest 注解和 MockMvc 这三种,下面我们逐一进行讲解。

使用 TestRestTemplate

Spring Boot 提供的 TestRestTemplate 与 RestTemplate 非常类似,只不过它专门用在测试环境中。

如果我们想在测试环境中使用 @SpringBootTest,则可以直接使用 TestRestTemplate 来测试远程访问过程,示例代码如下:

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

@RunWith (SpringRunner. class )

@SpringBootTest (webEnvironment = SpringBootTest.WebEnvironment.RANDOM_PORT)

public class CustomerController2Tests {

     @Autowired

     private TestRestTemplate testRestTemplate;

 

     @MockBean

     private CustomerTicketService customerTicketService;

 

     @Test

     public void testGenerateCustomerTicket() throws Exception {

         Long accountId = 100L;

         String orderNumber = "Order00001" ;

         given( this .customerTicketService.generateCustomerTicket(accountId, orderNumber))

                 .willReturn( new CustomerTicket(1L, accountId, orderNumber, "DemoCustomerTicket1" , new Date()));

 

         CustomerTicket actual = testRestTemplate.postForObject( "/customers/" + accountId+ "/" + orderNumber, null , CustomerTicket. class );

         assertThat(actual.getOrderNumber()).isEqualTo(orderNumber);

     }

}

上述测试代码中,首先,我们注意到 @SpringBootTest 注解通过使用 SpringBootTest.WebEnvironment.RANDOM_PORT 指定了随机端口的 Web 运行环境。然后,我们基于 TestRestTemplate 发起了 HTTP 请求并验证了结果。

特别说明:这里使用 TestRestTemplate 发起请求的方式与 RestTemplate 完全一致

使用 @WebMvcTest 注解

接下来测试方法中,我们将引入一个新的注解 @WebMvcTest,该注解将初始化测试 Controller 所必需的 Spring MVC 基础设施,CustomerController 类的测试用例如下所示:

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

@RunWith (SpringRunner. class )

@WebMvcTest (CustomerController. class )

public class CustomerControllerTestsWithMockMvc {

     @Autowired

     private MockMvc mvc;

 

     @MockBean

     private CustomerTicketService customerTicketService;

 

     @Test

     public void testGenerateCustomerTicket() throws Exception {

         Long accountId = 100L;

         String orderNumber = "Order00001" ;

         given( this .customerTicketService.generateCustomerTicket(accountId, orderNumber))

                 .willReturn( new CustomerTicket(1L, 100L, "Order00001" , "DemoCustomerTicket1" , new Date()));

 

         this .mvc.perform(post( "/customers/" + accountId+ "/" + orderNumber).accept(MediaType.APPLICATION_JSON)).andExpect(status().isOk());

     }

}

MockMvc 类提供的基础方法分为以下 6 种,下面一一对应来看下。

Perform :执行一个 RequestBuilder 请求,会自动执行 SpringMVC 流程并映射到相应的 Controller 进行处理。 get/post/put/delete :声明发送一个 HTTP 请求的方式,根据 URI 模板和 URI 变量值得到一个 HTTP 请求,支持 GET、POST、PUT、DELETE 等 HTTP 方法。 param :添加请求参数,发送 JSON 数据时将不能使用这种方式,而应该采用 @ResponseBody 注解。 andExpect :添加 ResultMatcher 验证规则,通过对返回的数据进行判断来验证 Controller 执行结果是否正确。 andDo :添加 ResultHandler 结果处理器,比如调试时打印结果到控制台。 andReturn :最后返回相应的 MvcResult,然后执行自定义验证或做异步处理。

执行该测试用例后,从输出的控制台日志中我们不难发现,整个流程相当于启动了 CustomerController 并执行远程访问,而 CustomerController 中使用的 CustomerTicketService 则做了 Mock。

显然,测试 CustomerController 的目的在于验证其返回数据的格式和内容。在上述代码中,我们先定义了 CustomerController 将会返回的 JSON 结果,然后通过 perform、accept 和 andExpect 方法模拟了 HTTP 请求的整个过程,最终验证了结果的正确性。

请注意 @SpringBootTest 注解不能和 @WebMvcTest 注解同时使用。

使用 @AutoConfigureMockMvc 注解

在使用 @SpringBootTest 注解的场景下,如果我们想使用 MockMvc 对象,那么可以引入 @AutoConfigureMockMvc 注解。

通过将 @SpringBootTest 注解与 @AutoConfigureMockMvc 注解相结合,@AutoConfigureMockMvc 注解将通过 @SpringBootTest 加载的 Spring 上下文环境中自动配置 MockMvc 这个类。

使用 @AutoConfigureMockMvc 注解的测试代码如下所示:

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

@RunWith (SpringRunner. class )

@SpringBootTest

@AutoConfigureMockMvc

public class CustomerControllerTestsWithAutoConfigureMockMvc {

 

     @Autowired

     private MockMvc mvc;

 

     @MockBean

     private CustomerTicketService customerTicketService;

 

     @Test

     public void testGenerateCustomerTicket() throws Exception {

         Long accountId = 100L;

         String orderNumber = "Order00001" ;

         given( this .customerTicketService.generateCustomerTicket(accountId, orderNumber))

                 .willReturn( new CustomerTicket(1L, 100L, "Order00001" , "DemoCustomerTicket1" , new Date()));

 

         this .mvc.perform(post( "/customers/" + accountId+ "/" + orderNumber).accept(MediaType.APPLICATION_JSON)).andExpect(status().isOk());

     }

}

在上述代码中,我们使用了 MockMvc 工具类完成了对 HTTP 请求的模拟,并基于返回状态验证了 Controller 层组件的正确性。

小结

以上为个人经验,希望能给大家一个参考,也希望大家多多支持。

原文链接:https://blog.csdn.net/yangshangwei/article/details/117231037

查看更多关于SpringBoot 应用程序测试实现方案的详细内容...

  阅读:14次