DataSeed

– is a little tool for generating test data for scala.

Why?

Because relevant test data is too often lost in the clutter of all the setup code.

Example of bad practice:

test("Full name"):
  // Given clutter
  val p: Person = Person(
    firstName = "Long John",
    surname = "Silver",
    address = Address(
      street = "My street",
      zipCode = ZipCode(12345),
      area = "Tiny Town"
    ),
    phone = "555 9999"
  )

  //Then
  assert(p.fullName == "Long John Silver")

There is a lot of setup data. Distinguishing the important setup data from the need clutter is difficult at first glans. Only firstName = "Long John" and surname = "Silver" is important. The rest is just needed to be able to create p: Person.

What?

... does it do? It focuses tests on the important parts of the setup data. The rest of the data is randomly generated stuff that nobody should care about.

Example with good practice:

test("Full name"):
  // Given
  val p: Person = Gen: // generate
    Person(
      firstName = "Long John", // Important
      surname = "Silver", // Important
      address = Address( // Don't care
        street = string(between(2, 30)), // Just a random string
        zipCode = ZipCode(between(10000, 99999)), // A random number
        area = string(between(2, 30)) // Another random string
      ),
      phone = string(8).? // An optional random string 
    )

  //Then
  assert(p.fullName == "Long John Silver")

There is still a lot of setup. The difference is that only the important data is stated explicitly. Unimportant data is just randomly generated from a seed.

Example with splittig up the generated data

// generated address
val address: Gen[Address] = Address( 
  street = string(between(2, 30)), 
  zipCode = ZipCode(between(10000, 99999)), 
  area = string(between(2, 30)) 
)

val p: Person = Gen.fromSeed(seed = 12345): // generate
  Person(
    firstName = "Long John",
    surname = "Silver", 
    address = address, // using the generated address
    phone = string(8).?
  )

Seed can be set explicitly.

Example where setting the seed

val p: Person = Gen.fromSeed(seed = 12345): // generate
  Person(
    firstName = "Long John", 
    surname = "Silver", 
    address = address, 
    phone = string(8).?
  )

RELEASE

There is no release at the moment as all the important stuff is in core/src/main/scala/dataseed/generate.scala. It is less than 100 lines of code, and should stay that way.

So if you want to use it, copy the file and make your own adjustments.

The names subproject is just a playroom for generating names.