For building DSLs that allow the composition of pure and fully type-safe expressions such that any sub-expression may generate or receive errors as input, short-circuiting evaluation.

resolvers ++= Seq(
    Resolver.sonatypeRepo("releases"),
    Resolver.sonatypeRepo("snapshots")
)

libraryDependencies ++= Seq(
    "io.nary" %% "espresso" % "0.0.1"
)

espresso makes extensive use of the Typelevel cats and shapeless libraries as well as the kind-projector compiler plugin.

Take a look at https://github.com/ryanonsrc/espresso/tree/master/src/test/scala/io/nary/espresso/sample

• Term[E, A] - Represents a term within an expression, evaluating to an error of type E or result of type A
• Expr[E, A, B] - For any expression taking a value of type A and evaluating to either a result of type B or an error of type E
• In[E, A] - Input that can be "read" into an expression as either value of type A or an error of type E
• opN - Lift a function f: (HList of N terms ⇒ Term[E, B]) into a Expr[E, A1 :: ... :: AN :: HNil, B]
• joinN - Combine an Hlist of N expressions: Expr[E, A, Bi] into a single Expr[E, A, B1 :: ... :: BN :: HNil]
• evalN - Given N expressions: Expr[E, A, Bi] and operation: Expr[E, B1 :: ... :: BN :: HNil, C], generate Expr[E, A, C]