This play module provides a simpler alternative to play Actions using for-comprehensions.

NOTE: We just started work on this repository, expect many changes.

	Scala 2.10	Scala 2.11	Scala 2.12	Scala 2.13
Play 2.3
Play 2.4
Play 2.5
Play 2.6
Play 2.7
Play 2.8

To add to your project, include this depedency in the build.sbt file:

Write simple and reusable Play Action's using for-comprehensions.

def saveAction = MonadicAction {
  for {
    auth <- header("Authorization")
    user <- authService.authorize(auth)
    jsonBody <- body(BodyParsers.parse.tolerantJson)
  } yield {
    // your business logic here
  }
}

Each step in the for-comprehension might fail. Just as a Future carries the concept of a hidden Throwable, this Monad carries a hidde failure Result response from Play. If the header is missing a BadRequest will be returned, if the authService doesn't authorize the request, a Forbidden will be returned and errors processing the body would return other error codes as well.

The original spark for the idea came to be when using Play Framework for a project where requests included huge file uploads. These uploads could be in the hundreds of MB or even a few GBs. Uploading took a long time. It was possible headers were invalid, user was not authenticated and other validations could go wrong. This could lead to a lot of wasted time when headers where invalid. HTTP supports the 100-continue feature. It allows users to send the headers first, wait for confirmation these are accepted and that it may continue to send the body.

Play supports the 100-continue feature in principle but when using Play's Actions with body parsers, the body is processed fully before the headers. To leverage this feature a different way of constructing Actions is needed.

Finally, to make sure every endpoint supports this, it is imperative all header processing happens before using the body. To enable that it was needed to make the Monad stateful. Headers cannot be accessed after processing the body and this is enforced by the compiler.

The idea for this library was conceived by @rcano around 2012-2014 and was implemented in a closed project. Since then, this pattern has been re-implemented in other places by Tomás Lázaro (@tlazaro) in a different shape but same spirit. After many years seeing Play Actions has not changed and no implementation exists in the wild we decided to do this work in open-source and share it.

The play-monadic-actions shares the same core idea and has a variety of conversions that are very useful. However, it does not include the concept of a stateful Monad and does not enable usage of the 100-continue HTTP feature.