Elmyr is a Kotlin library providing tools to generate “random” values, specifically useful for tests
Being an adept of testing code, I write a lot of tests. One thing I noticed is that in my tests, my fake / test data always look the same. My user names are always “Bob”
and “Alice”
, aged 42
or 69
, with userId 4816152342
or 24601
, and eating “spam”
, “eggs”
and “bacon”
.
The problem is, the more test I write, the less I'm confident in my fake values, because they're always the same.
This is where Elmyr kicks in, allowing you to create fake/fuzzy data based on a few constraints, making your test data random, and yet reproducible.
dependencies {
// Core library
testCompile("com.github.xgouchet.Elmyr:core:1.4.1")
// Testing Framework Integrations
testCompile("com.github.xgouchet.Elmyr:junit4:1.4.1")
testCompile("com.github.xgouchet.Elmyr:junit5:1.4.1")
testCompile("com.github.xgouchet.Elmyr:spek:1.4.1")
//
testCompile("com.github.xgouchet.Elmyr:jvm:1.4.1")
}
You can create an instance of the Forge
class, and from that generate:
- primitives, with basic constraints
- Strings matching simple predicates or even Regexes
- Any Kotlin
data class
- Your own custom data, by implementing the
ForgeryFactory
interface, then calling theForge::addFactory
method.
You can instantiate a ForgeRule
instance, which extends the Forge
class,
add factories to it, and then annotate fields on your test class with @Forgery
.
class FooTest {
@get:Rule
val forge = ForgeRule()
.withFactory(FooFactory())
.withFactory(BarFactory())
@Forgery
internal lateinit var fakeBar: Bar
@Forgery
lateinit var fakeFooList: List<Foo>
//…
}
You can add an extension and configure it. In addition to creating forgeries on fields/properties of your test class, you can inject parameters directly on your test methods.
@ExtendWith(ForgeExtension::class)
@ForgeConfiguration(KotlinAnnotationTest.Configurator::class)
internal class FooTest {
@Forgery
internal lateinit var fakeBar: Bar
@Forgery
lateinit var fakeFooList: List<Foo>
@Test
fun testSomething(@IntForgery i: Int, forge:Forge){
// …
}
}
You can create a custom Forge instance with spekForge
to be able to
add reproducibility to Spek tests.
class CalculatorSpek : Spek({
val forge = spekForge(
seeds = mapOf(
"CalculatorSpek/A calculator/addition/returns the sum of its arguments" to 0x1337L
)
)
describe("A calculator") {
val calculator by memoized { Calculator() }
describe("addition") {
it("returns the sum of its arguments") {
val a = forge.anInt()
val b = forge.anInt()
assertEquals(calculator.add(a, b), a + b)
}
}
}
})
The full documentation will be coming shortly. The Wiki contains a full reference of all the packages and classes distributed.
Contribution is always welcome, to know more, read our Contributing guide.
- Add enum forgery through reflexive factory
- Make reflexive factory available through injection (with
@Forgery
) - Make reflexive factory available via JUnit4 (with
@ForgeRule
) - Make reflexive factory available via JUnit5 (with
@ForgeExtension
) - Add set forgery (thanks @ambushwork)
- Add list forgery with indexed lambda (thanks @ambushwork)
This library is completely free to use and modify (as per the License). I try my best to make it as good as possible, but only do this on my free time. If you want to support my work, you can click the Donate button below.
Xavier F. Gouchet – @xgouchet
Distributed under the MIT license. See LICENSE.md for more information.