This repository contains TestNG DataProvider wrapper (the latest version is based on TestNG 7.7.1) which helps to supply test data in a more flexible way.
Common DataProvider forces using quite old and ugly syntax which expects one of the following types to be returned from DP method's body:
That's weird as developers tend to use Stream
and Collection
API for data manipulation in the modern Java world.
Just imagine if you could use the following syntax to supply some filtered and sorted data into test method's signature:
@DataSupplier
public Stream<User> getData() {
return Stream
.of(
new User("Max", "password2"),
new User("Black Fox", "password3"),
new User("Mark", "password1")
)
.filter(u -> !u.getName().contains("Fox"))
.sorted(comparing(User::getPassword));
}
@Test(dataProvider = "getData")
public void shouldSupplyStreamData(final User user) {
// ...
}
Much better and flexible than two-dimensional arrays or iterators, isn't it?
And what if we don't want to iterate the same test N times depending on collection size? What if we want to inject it into test's signature like the following?
@DataSupplier(transpose = true)
public List<User> getTransposedData() {
return StreamEx
.of(
new User("username1", "password1"),
new User("username2", "password2")
)
.toList();
}
@Test(dataProvider = "getTransposedData")
public void shouldSupplyExtractedListData(final List<User> users) {
// ...
}
Or if you want to extract the values of your collection and inject into test's signature, you can combine transpose
with a flatMap
:
@DataSupplier(transpose = true, flatMap = true)
public Set<User> getExtractedData() {
return StreamEx.of("product1", "product2", "product1").toSet();
}
@Test(dataProvider = "getExtractedData")
public void shouldSupplyExtractedListData(final String... products) {
// ...
}
Java-like flatMap operation can be applied even to more complicated structures like Map
to extract values for each row:
@DataSupplier(flatMap = true)
public Map<Integer, String> getInternallyExtractedMapData() {
return EntryStream.of(asList("user3", "user4")).toMap();
}
@Test(dataProvider = "getInternallyExtractedMapData")
public void supplyInternallyExtractedMapData(final Integer key, final String value) {
// not implemented
}
Go top :point_up:
DataSupplier
(method name is used by default)DataSupplier
failure (skips by default)Go top :point_up:
Go top :point_up:
Add the following configuration into build.gradle:
repositories {
mavenCentral()
}
configurations {
agent
}
sourceCompatibility = JavaVersion.VERSION_1_8
ext {
aspectjVersion = '1.9.7'
}
[compileJava, compileTestJava]*.options*.compilerArgs = ['-parameters']
dependencies {
agent "org.aspectj:aspectjweaver:${aspectjVersion}"
implementation(
"org.aspectj:aspectjweaver:${aspectjVersion}",
'org.testng:testng:6.14.3',
'io.github.sskorol:test-data-supplier:1.7.0'
)
}
test {
doFirst {
jvmArgs("-javaagent:${configurations.agent.singleFile}")
}
useTestNG()
}
Check a separate project with usage examples.
Go top :point_up:
Add the following configuration into pom.xml:
<properties>
<aspectj.version>1.9.7</aspectj.version>
<java.version>1.8</java.version>
<compiler.plugin.version>3.8.0</compiler.plugin.version>
<surefire.plugin.version>2.20.1</surefire.plugin.version>
</properties>
<dependencies>
<dependency>
<groupId>org.aspectj</groupId>
<artifactId>aspectjweaver</artifactId>
<version>${aspectj.version}</version>
</dependency>
<dependency>
<groupId>org.testng</groupId>
<artifactId>testng</artifactId>
<version>6.14.3</version>
</dependency>
<dependency>
<groupId>io.github.sskorol</groupId>
<artifactId>test-data-supplier</artifactId>
<version>1.7.0</version>
</dependency>
</dependencies>
<build>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-compiler-plugin</artifactId>
<version>${compiler.plugin.version}</version>
<configuration>
<source>${java.version}</source>
<target>${java.version}</target>
<compilerArgs>
<arg>-parameters</arg>
</compilerArgs>
</configuration>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-surefire-plugin</artifactId>
<version>${surefire.plugin.version}</version>
<configuration>
<argLine>
-javaagent:"${settings.localRepository}/org/aspectj/aspectjweaver/${aspectj.version}/aspectjweaver-${aspectj.version}.jar"
</argLine>
</configuration>
</plugin>
</plugins>
</build>
Check a separate project with usage examples.
Go top :point_up:
plugins {
id 'java'
}
sourceCompatibility = JavaVersion.VERSION_11
repositories {
mavenCentral()
}
configurations {
agent
}
ext {
aspectjVersion = '1.9.7'
}
[compileJava, compileTestJava]*.options*.compilerArgs = ['-parameters']
dependencies {
agent "org.aspectj:aspectjweaver:${aspetjVersion}"
implementation(
"org.aspectj:aspectjweaver:${aspectjVersion}",
'org.testng:testng:7.4.0',
'io.github.sskorol:test-data-supplier:1.9.7'
)
}
test {
doFirst {
jvmArgs("-javaagent:${configurations.agent.singleFile}")
}
useTestNG()
}
Go top :point_up:
It's a bit tricky in terms of building and testing modular applications:
plugins {
id 'java-library'
id 'java'
}
ext {
moduleName = 'your.module.name'
}
sourceCompatibility = JavaVersion.VERSION_11
repositories {
mavenCentral()
}
configurations {
agent
}
ext {
aspectjVersion = '1.9.7'
}
dependencies {
agent "org.aspectj:aspectjweaver:${aspectjVersion}"
implementation(
"org.aspectj:aspectjweaver:${aspectjVersion}",
'org.testng:testng:7.4.0',
'io.github.sskorol:test-data-supplier:1.9.7'
)
}
compileJava {
inputs.property("moduleName", moduleName)
doFirst {
options.compilerArgs = [
'--module-path', classpath.asPath
]
classpath = files()
}
}
compileTestJava {
inputs.property("moduleName", moduleName)
doFirst {
options.compilerArgs = [
'--module-path', classpath.asPath,
'--patch-module', "$moduleName=" + files(sourceSets.test.java.srcDirs).asPath,
]
classpath = files()
}
}
test {
useTestNG()
inputs.property("moduleName", moduleName)
doFirst {
jvmArgs = [
"-javaagent:${configurations.agent.singleFile}",
'--module-path', classpath.asPath,
'--add-modules', 'ALL-MODULE-PATH',
'--add-opens', 'your.module.name/test.package.path=org.testng',
'--add-opens', 'your.module.name/test.package.path=org.jooq.joor',
'--patch-module', "$moduleName=" + files(sourceSets.test.java.outputDir).asPath
]
classpath = files()
}
}
Your module-info.java may look like the following:
module your.module.name {
requires io.github.sskorol.testdatasupplier;
requires org.testng;
// Optional
provides io.github.sskorol.core.IAnnotationTransformerInterceptor
with path.to.transformer.ImplementationClass;
provides io.github.sskorol.core.DataSupplierInterceptor
with path.to.interceptor.ImplementationClass;
}
Go top :point_up:
Note that test-data-supplier:2.0.0+
has been compiled with java 17. It means you must use the same language level in your build file.
plugins {
id 'java'
}
sourceCompatibility = JavaVersion.VERSION_17
repositories {
mavenCentral()
}
configurations {
agent
}
ext {
aspectjVersion = '1.9.19'
}
[compileJava, compileTestJava]*.options*.compilerArgs = ['-parameters']
dependencies {
agent "org.aspectj:aspectjweaver:${aspectjVersion}"
implementation(
"org.aspectj:aspectjweaver:${aspectjVersion}",
'org.testng:testng:7.8.0',
'io.github.sskorol:test-data-supplier:2.3.0'
)
}
test {
doFirst {
jvmArgs(
"-javaagent:${configurations.agent.singleFile}",
'--add-opens', 'java.base/java.lang=ALL-UNNAMED'
)
}
useTestNG()
}
Go top :point_up:
<properties>
<aspectj.version>1.9.19</aspectj.version>
<java.version>17</java.version>
<compiler.plugin.version>3.11.0</compiler.plugin.version>
<surefire.plugin.version>3.0.0-M9</surefire.plugin.version>
</properties>
<dependencies>
<dependency>
<groupId>org.aspectj</groupId>
<artifactId>aspectjweaver</artifactId>
<version>${aspectj.version}</version>
</dependency>
<dependency>
<groupId>org.testng</groupId>
<artifactId>testng</artifactId>
<version>7.8.0</version>
</dependency>
<dependency>
<groupId>io.github.sskorol</groupId>
<artifactId>test-data-supplier</artifactId>
<version>2.3.0</version>
</dependency>
</dependencies>
<build>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-compiler-plugin</artifactId>
<version>${compiler.plugin.version}</version>
<configuration>
<source>${java.version}</source>
<target>${java.version}</target>
<compilerArgs>
<arg>-parameters</arg>
</compilerArgs>
</configuration>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-surefire-plugin</artifactId>
<version>${surefire.plugin.version}</version>
<configuration>
<argLine>
-javaagent:"${settings.localRepository}/org/aspectj/aspectjweaver/${aspectj.version}/aspectjweaver-${aspectj.version}.jar"
</argLine>
</configuration>
</plugin>
</plugins>
</build>
Go top :point_up:
Instead of a common DataProvider annotation use the following:
@DataSupplier
public T getData() {
//...
}
DataSupplier supports the following args: name, transpose, flatMap, runInParallel, indices and propagateTestFailure.
You can refer DataSupplier the same way as with TestNG DataProvider:
@Test(dataProvider = "getData")
public void supplyData(final T data) {
// ...
}
// or
@Test(dataProviderClass = ExternalDataProvider.class, dataProvider = "getData")
public void supplyExternalData(final T data) {
// ...
}
Check io.github.sskorol.testcases package for more examples.
Go top :point_up:
Test data supplier supports JSON, CSV, YML and XLSX data retrieval. Assuming you have the following resources:
username,password
"admin","admin"
"sskorol","password"
"guest","123"
[
{
"username": "admin",
"password": "admin"
},
{
"username": "sskorol",
"password": "password"
},
{
"username": "guest",
"password": "123"
}
]
---
username: admin
password: admin
---
username: sskorol
password: password
---
username: guest
password: '123'
USERNAME PASSWORD
admin admin
sskorol password
guest 123
You can now map Java entities to these data sources using @Source annotation, which accepts either local file name or URL:
@Data
@Source(path = "users.csv")
public class User {
@FieldName("username")
private final String name;
private final String password;
}
@Data
@Source(path = "users.json")
public class User {
@SerializedName("username")
private final String name;
private final String password;
}
@Data
@NoArgsConstructor
@Source(path = "users.yml")
public class User {
@JsonProperty("username")
private final String name;
private final String password;
}
@Data
@NoArgsConstructor
@Source(path = "users.xlsx")
@Sheet(name = "sheet_1")
public class User {
@Column(name = "USERNAME")
private String username;
@Column(name = "PASSWORD")
private String password;
}
In case if some Java field's name differs from its data source representation, you can assign a valid name via @FieldName for CSV, @SerializedName for JSON and @JsonProperty for YML data type.
Excel support is experimental. 2.0.0 version used ZeroCell library based on Apache POI to simplify corresponding files processing. Since 2.1.0, there's a custom implementation with similar approach, but minor improvements, e.g. there's no need to use column index anymore.
In terms of fields' mapping, you can use custom @Column
annotation (don't confuse with ZeroCell Column).
You should also make sure you provided a sheet name via corresponding @Sheet
annotation. Otherwise, the first one will be used.
In case if you have a similar structure on multiple sheets, you can use a repeatable @Sheets
annotation.
Dynamic sheets' specification is also possible via withAdditionalSources
builder method (see examples below).
Similarly to ZeroCell, you can use either default or custom fields' converters. Here's a list of defaults:
To use custom converter, you should specify its class via @Column
annotation.
@Column(name = "Tags", converter = StringToListConverter.class)
private List<String> data;
And the actual implementation may look like the following:
public class StringToListIConverter extends DefaultConverter<List<String>> {
@Override
public List<String> convert(final String value) {
return asList(value.split(","));
}
}
Custom converters must extend DefaultConverter
class.
Also note that by default test-data-supplier
uses an implicit conversion based on the field type.
So you don't have to explicitly specify a converter if it's among the defaults.
Local data sources must be located in a classpath. You usually use resources folder for that.
Then, within @DataSupplier
you can call a special TestDataReader builder to retrieve data from CSV, JSON, YML or XLSX data source.
See javadocs to get more details.
@DataSupplier
public StreamEx<User> getUsers() {
return use(CsvReader.class).withTarget(User.class).withSource("users.csv").read();
}
@DataSupplier
public StreamEx<User> getUsers() {
return use(JsonReader.class).withTarget(User.class).withSource("http://users.json").read();
}
@DataSupplier
public StreamEx<User> getUsers() {
return use(YamlReader.class).withTarget(User.class).read();
}
@DataSupplier
public StreamEx<User> getUsers() {
return use(XlsxReader.class).withTarget(User.class).read();
}
@DataSupplier
public StreamEx<User> getUsers() {
return use(XlsxReader.class).withTarget(User.class).withAdditionalSources("Sheet1", "Sheet2").read();
}
If you want to specify a custom source in runtime, you can remove @Source annotation and use withSource builder
method instead.
withAdditionalSources
builder method is experimental and implemented for XlsxReader
as a dynamic sheets provider.
Note that in case of a data reading error or any kind of exception thrown in a @DataSupplier
body,
the corresponding test will be skipped. That's a default TestNG behaviour.
However, you can set propagateTestFailure
flag (introduced in TestNG 7.6.0) to explicitly mark the test as failed.
Go top :point_up:
Technically, there's no need to create an additional ORM wrapper to work with databases. But it's worth to show how to perform such integration.
Let's use ebean and Postgres as an example.
First, create an application-test.yaml
with db connection details in your test resources' folder:
ebean:
test:
useDocker: false
platform: postgres
ddlMode: none
dbName: your_db
dbSchema: your_schema
postgres:
username: your_username
password: your_password
url: jdbc:postgresql://localhost:5432/your_db
Next, create a mapping with your DB table:
@MappedSuperclass
public class BaseEntity extends Model {
@Id
long id;
}
@Entity
@Table(name = "testing.users")
public class UserEntity extends BaseEntity {
@NotNull
public String email;
@Column(name = "is_active")
public boolean isActive;
}
Now, you can supply users to your test the following way:
@DataSupplier
public List<UserEntity> usersData() {
return find(UserEntity.class)
.where()
.like("email", "%korol%@gmail.com")
.and()
.eq("is_active", true)
.findList();
}
Go top :point_up:
You can specify DataSupplier for Factory annotation as well as for common test methods.
@NoArgsConstructor
public class InternalFactoryTests {
@DataSupplier
public StreamEx getConstructorData() {
return IntStreamEx.rangeClosed(1, 3).boxed();
}
@DataSupplier
public String getTestData() {
return "data";
}
@Factory(dataProvider = "getConstructorData", dataProviderClass = InternalFactoryTests.class)
public InternalFactoryTests(final int index) {
// not implemented
}
@Test(dataProvider = "getTestData")
public void internalFactoryTest(final String data) {
// not implemented
}
}
Go top :point_up:
DataSupplierInterceptor interface allows tracking original DataProvider method calls for accessing additional meta-data. You can use the following snippet for getting required info:
public class DataSupplierInterceptorImpl implements DataSupplierInterceptor {
private static final Map<ITestNGMethod, DataSupplierMetaData> META_DATA = new ConcurrentHashMap<>();
@Override
public void beforeDataPreparation(final ITestContext context, final ITestNGMethod method) {
}
@Override
public void afterDataPreparation(final ITestContext context, final ITestNGMethod method) {
}
@Override
public void onDataPreparation(final DataSupplierMetaData testMetaData) {
META_DATA.putIfAbsent(testMetaData.getTestMethod(), testMetaData);
}
@Override
public Collection<DataSupplierMetaData> getMetaData() {
return META_DATA.values();
}
}
This class should be then loaded via SPI mechanism. Just create META-INF/services folder in resources root, and add a new file io.github.sskorol.core.DataSupplierInterceptor with a full path to implementation class.
Go top :point_up:
TestNG restricts users in a number of IAnnotationTransformer implementations. You may have not more than a single transformer within project's scope. As test-data-supplier uses this interface for its internal stuff, you won't be able to apply your own implementation.
In case if you still need to add a custom IAnnotationTransformer, you have to implement the following interface:
public class IAnnotationTransformerInterceptorImpl implements IAnnotationTransformerInterceptor {
@Override
public void transform(IFactoryAnnotation annotation, Method testMethod) {
}
@Override
public void transform(IConfigurationAnnotation annotation, Class testClass, Constructor testConstructor, Method testMethod) {
}
@Override
public void transform(IDataProviderAnnotation annotation, Method method) {
}
@Override
public void transform(IListenersAnnotation annotation, Class testClass) {
}
}
It's just an SPI wrapper for common TestNG feature. Use the same technique as for DataSupplierInterceptor to include it into your project.
Note that in case if you want to manage DataProviderTransformer manually, you have to use a special spi-off distribution:
dependencies {
implementation 'io.github.sskorol:test-data-supplier:2.3.0:spi-off'
}
Go top :point_up:
Test Data Supplier also has an IntelliJ IDEA plugin. Just install test-data-supplier-plugin from the official JetBrains repository.
More information about its features can be found on the related GitHub page.
Go top :point_up: