Documentation

The Arrow instance allows for building non-cyclic directed graphs of churros.

The arr method allows for the creation of a that maps items with a pure function. This is equivalent to `fmap f id`.

>>> :set -XArrows
>>> :{
let sect  = process $ \x@(_x,_y,z) -> print x >> return z
    graph =
      proc i -> do
        j <- arr succ  -< i
        k <- arr show  -< j
        l <- arr succ  -< j
        m <- arr (> 5) -< j
        n <- sect      -< (k,l,m)
        o <- arr not   -< n
        p <- delay 0.1 -< o
        sinkPrint      -< p
in
runWaitChan $ sourceList [1,5,30] >>> graph
:}
("2",3,False)
("6",7,True)
("31",32,True)
True
False
False

The other Arrow methods are also usable:

>>> runWaitChan $ pure 1 >>> (arr show &&& arr succ) >>> sinkPrint
("1",2)

The Category instance allows for the creation of Churro pipelines.

All other examples of the form `a >>> b` use this instance.

The id method creates a passthrough arrow. There isn't usually a reason to use id directly as it has no effect:

>>> runWaitChan $ pure 1 >>> id >>> id >>> id >>> sinkPrint
1

Covariant functor instance for Churro - Maps over the output.

>>> let s = sourceList [1,2]
>>> runWaitChan $ s >>> sinkPrint
1
2

>>> runWaitChan $ fmap succ s >>> sinkPrint
2
3

The Applicative instance allows for pairwise composition of Churro pipelines. Once again this is covariat and the composition occurs on the output transports of the Churros.

The pure method allows for the creation of a Churro yielding a single item.