Your own Validators

If you haven't already done so in the last chapter, import the svalidate classes:

scala> import svalidate._
import svalidate._

When it comes to validation, everyone has their own requirements:

• Your webshop accepts only m/f for gender, while a medical software might accept several other options
• A city zip code in Switzerland is 4 digits long, while it's 5 digits in most other countries
• Phone number layouts differ from country to country
• Usually you want to customize the error messages generated from your validation layer

Thus, svalidate does not impose any restrictions on what or how you validate.

The Validator trait

Let's have a closer look at the Validator trait:

scala> trait Validator[V] {
     | 
     |   def validate(value: V): Seq[String]
     | 
     | }
defined trait Validator

As we can see, it has a type parameter T: If we implement it, we always implement it for a certain type. It's type is the type of the property we validate.

There's a single method we implement: validate. It takes a value of our Type T and returns a Seq[String], which is the list of errors.

Lets look at an example, the NonEmptyString validator.

scala> object NonEmptyString extends Validator[String] {
     | 
     |   def validate(s: String): Seq[String] =
     |     if(s.isEmpty)
     |       Seq(
     |         "The field %s must not be empty"
     |       )
     |     else
     |       Seq.empty
     | 
     | }
defined module NonEmptyString

In the above example, we can see that the property type of the NonEmptyString validator is String. It's validate method thus takes a single string as it's argument. In the implementation, we just return a single error if the string is empty or no error if the string is not empty.