Scaberry fields forever (WIP)

Scaberry is a library to reify fields on case classes. This allows for the creation of filter or update objects without any boilerplate. It also allows for easy access to the annotations on your fields, in case you want to do something with that.

The base element is the simple @berry annotation. Just put it on your case class and voilà:

@berry
case class Dog(name: String, owner: Option[String], good: Boolean = true)

You now have access on the companion object to a meta value with all the fields of the class. Those fields can be applied on instance, combined to form filters or updates…

val inouk = Dog("Inouk", Some("Gael"))
// inouk: scaberry.tests.models.Dog = Dog(Inouk,Some(Gael),true)

val dogName = Dog.meta.fields.name
// dogName: scaberry.tests.models.Dog.meta.fields.name.type = ...

dogName(inouk).get
// res0: String = Inouk

dogName(inouk).copy("Rex")
// res1: scaberry.tests.models.Dog = Dog(Rex,Some(Gael),true)

val filterName = dogName.filterEq("Inouk")
// filterName: scaberry.core.Filter.Field[scaberry.tests.models.Dog,String] = <function1>

filterName(inouk)
//res2: Boolean = true

val filterMultiple = filterName |@| Dog.meta.fields.owner.filterWith(_.isEmpty)
// filterMultiple: scaberry.core.Filter[scaberry.tests.models.Dog] = <function1>

filterMultiple(inouk)
//res3: Boolean = false

val updateName = dogName.updateVal("Rex")
// updateName: scaberry.core.Update.Field[scaberry.tests.models.Dog,String] = <function1>

updateName(inouk)
// res4: scaberry.tests.models.Dog = Dog(Rex,Some(Gael),true)

val updateMultiple = updateName |@| Dog.meta.fields.good.updateWith(!_)
// updateMultiple: scaberry.core.Update[scaberry.tests.models.Dog] = <function1>

updateMultiple(inouk)
// res5: scaberry.tests.models.Dog = Dog(Rex,Some(Gael),false)

Filter

Let's say you have a case class:

Let's say you want a filter for it, to pass as an argument to your search function. A few common approaches are:

• Using a different argument in the search method for each field, all of them optional. This is highly boilerplatey, scales badly and is hard to maintain, especially when you realize all functions calling search are going to need the exact same arguments. Next!
• Using the class Dog itself as your filter. This is way better, but you run into another problem: how not to filter on some field you want to ignore ?
• Using an anonymous function as your filter. This works and is very flexible, but how do you serialize that to JSON or whatever exchange format you use ?
• Using a specific Filter class. It works, is a lot of boilerplate and is what Scaberry does for you !

JSON ser/deserialization in progress…

Known problems

IntelliJ integration

A few things are not well handled by IntelliJ and appear as compile errors, although they do actually compile:

• renaming the meta value on the companion object;
• access to the BerryProps declared on fields;
• secondary constructors on the class annotated with @berry.