Home

Awesome

Official repo for ficus. Adopted from ceedubs

Ficus

Ficus is a lightweight companion to Typesafe config that makes it more Scala-friendly.

Ficus adds an as[A] method to a normal Typesafe Config so you can do things like config.as[Option[Int]], config.as[List[String]], or even config.as[MyClass]. It is implemented with type classes so that it is easily extensible and many silly mistakes can be caught by the compiler.

Build Status Join the chat at https://gitter.im/iheartradio/ficus Coverage Status Latest version

Examples

import net.ceedubs.ficus.Ficus._
import com.typesafe.config.ConfigFactory

object Country extends Enumeration {
  val DE = Value("DE")
  val IT = Value("IT")
  val NL = Value("NL")
  val US = Value("US")
  val GB = Value("GB")
}

case class SomeCaseClass(foo: String, bar: Int, baz: Option[FiniteDuration])

class Examples {
  val config: Config = ConfigFactory.load() // standard Typesafe Config

  // Note: explicit typing isn't necessary. It's just in these examples to make it clear what the return types are.
  // This line could instead be: val appName = config.as[String]("app.name")
  val appName: String = config.as[String]("app.name") // equivalent to config.getString("app.name")

  // config.as[Option[Boolean]]("preloadCache") will return None if preloadCache isn't defined in the config
  val preloadCache: Boolean = config.as[Option[Boolean]]("preloadCache").getOrElse(false)

  val adminUserIds: Set[Long] = config.as[Set[Long]]("adminIds")

  // something such as "15 minutes" can be converted to a FiniteDuration
  val retryInterval: FiniteDuration = config.as[FiniteDuration]("retryInterval")

  // can extract arbitrary Enumeration types
  // Note: it throws an exception at runtime, if the enumeration type cannot be instantiated or 
  // if a config value cannot be mapped to the enumeration value
  import net.ceedubs.ficus.readers.EnumerationReader._
  val someEnumerationType: Seq[Country.Value] = config.as[Seq[Country.Value]]("countries")

  // can hydrate most arbitrary types
  // it first tries to use an apply method on the companion object and falls back to the primary constructor
  // if values are not in the config, they will fall back to the default value on the class/apply method
  import net.ceedubs.ficus.readers.ArbitraryTypeReader._
  val someCaseClass: SomeCaseClass = config.as[SomeCaseClass]("someCaseClass")
}

For more detailed examples and how they match up with what's defined in a config file, see the example spec.

Adding the dependency

Now add the Ficus dependency to your build SBT file as well:

// for Scala 2.10.x
libraryDependencies += "com.iheart" %% "ficus" % "1.0.2"

// for Scala 2.11.x and Java 7
libraryDependencies += "com.iheart" %% "ficus" % "1.1.3"

// for Scala 2.11.x, 2.12.x, 2.13.x and Java 8
// See the latest version in the download badge below.
libraryDependencies += "com.iheart" %% "ficus" % <latest version> //see latest version in the badge below

Latest version

If you want to take advantage of Ficus's ability to automatically hydrate arbitrary traits and classes from configuration, you need to be on Scala version 2.10.2 or higer, because this functionality depends on implicit macros.

Built-in readers

Out of the box, Ficus can read most types from config:

In this context, A means any type for which a ValueReader is already defined. For example, Option[String] is supported out of the box because String is. If you want to be able to extract an Option[Foo[A]] for some type Foo that doesn't meet the supported type requirements (for example, this Foo has a type parameter), the option part is taken care of, but you will need to provide the implementation for extracting a Foo[A] from config. See Custom extraction.

Imports

The easiest way to start using Ficus config is to just import net.ceedubs.ficus.Ficus._ as was done in the Examples section. This will import all of the implicit values you need to start easily grabbing most basic types out of config using the as method that will become available on Typesafe Config objects.

To enable Ficus's reading of Enumeration types, you can also import net.ceedubs.ficus.readers.EnumerationReader._. See Enumeration support

To enable Ficus's macro-based reading of case classes and other types, you can also import net.ceedubs.ficus.readers.ArbitraryTypeReader._. See Arbitrary type support

If you would like to be more judicial about what you import (either to prevent namespace pollution or to potentially speed up compile times), you are free to specify which imports you need.

You will probably want to import net.ceedubs.ficus.Ficus.toFicusConfig, which will provide an implicit conversion from Typesafe Config to FicusConfig, giving you the as method.

You will then need a ValueReader for each type that you want to grab using as. You can choose whether you would like to get the reader via an import or a mixin Trait. For example, if you want to be able to call as[String], you can either import net.ceedubs.ficus.FicusConfig.stringValueReader or you can add with net.ceedubs.ficus.readers.StringReader to your class definition.

If instead you want to be able to call as[Option[String]], you would need to bring an implicit ValueReader for Option into scope (via import net.ceedubs.ficus.FicusConfig.optionValueReader for example), but then you would also need to bring the String value reader into scope as described above, since the Option value reader delegates through to the relevant value reader after checking that a config value exists at the given path.

Don't worry. It will be obvious if you forgot to bring the right value reader into scope, because the compiler will give you an error.

Enumeration support

Ficus has the ability to parse config values to Scala's Enumeration type.

If you have the following enum:

object Country extends Enumeration {
  val DE = Value("DE")
  val IT = Value("IT")
  val NL = Value("NL")
  val US = Value("US")
  val GB = Value("GB")
}

You can define the config like:

countries = [DE, US, GB]

To get an Enumeration type from your config you must import the EnumerationReader into your code. Then you can fetch it with the as method that Ficus provides on Typesafe Config objects.

import net.ceedubs.ficus.readers.EnumerationReader._
val countries: Seq[Country.Value] = config.as[Seq[Country.Value]]("countries")

Arbitrary type support

Supported types

If the apply method or constructor used has default arguments, Ficus will fall back to those for values not in the configuration.

If it exists, a valid apply method will be used instead of a constructor.

If Ficus doesn't know how to read an arbitrary type, it will provide a helpful compile-time error message explaining why. It won't risk guessing incorrectly.

Arbitrary type support requires Scala 2.10.2 or higher, because it takes advantage of implicit macros. To enable it, import net.ceedubs.ficus.readers.ArbitraryTypeReader._. Note that having the arbitrary type reader in scope can cause some implicit shadowing that you might not expect. If you define MyClass and define an implicit val myClassReader: ValueReader[MyClass] in the MyClass companion object, the arbitrary type reader will still win the implicit prioritization battle unless you specifically import MyClass.myClassReader.

By default the config keys has to match exactly the field name in the class, which by java convention is camel cased. To enable hyphen cased mapping, i.e. hyphen cased config keys, you can import a hyphen cased name mapper into the scope, such as:

import net.ceedubs.ficus.readers.namemappers.implicits.hyphenCase

Custom extraction

When you call as[String]("somePath"), Ficus config knows how to extract a String because there is an implicit ValueReader[String] in scope. If you would like, you can even teach it how to extract a Foo from the config using as[Foo]("fooPath") if you create your own ValueReader[Foo]. You could pass this Foo extractor explicitly to the as method, but most likely you just want to make it implicit. For an example of a custom value reader, see the ValueReader[ServiceConfig] defined in ExampleSpec.

Contributions

Many thanks to all of those who have contributed to Ficus.

Would you like to contribute to Ficus? Pull requests are welcome and encouraged! Please note that contributions will be under the MIT license. Please provide unit tests along with code contributions.

Binary Compatibility

MiMa can be used to check the binary compatibility between two versions of a library.T To check for binary incompatibilities, run mimaReportBinaryIssues in the sbt repl. The build is configured to compare the current version against the last released version (It does this naïvely at the moment by merely decrementing bugfix version). If any binary compatibility issues are detected, you may wish to adjust your code to maintain binary compatibility, if that is the goal, or modify the minor version to indicate to consumers that the new version should not be considered binary compatible.